[House Hearing, 111 Congress]
[From the U.S. Government Publishing Office]


 
                 EXAMINATION OF THE U.S. DEPARTMENT OF 
                    VETERANS AFFAIRS REGIONAL OFFICE 
                DISABILITY CLAIMS QUALITY REVIEW METHODS 

=======================================================================

                                HEARING

                               before the

       SUBCOMMITTEE ON DISABILITY ASSISTANCE AND MEMORIAL AFFAIRS

                                 of the

                     COMMITTEE ON VETERANS' AFFAIRS
                     U.S. HOUSE OF REPRESENTATIVES

                     ONE HUNDRED ELEVENTH CONGRESS

                             SECOND SESSION

                               __________

                             March 24, 2010

                               __________

                           Serial No. 111-68

                               __________

       Printed for the use of the Committee on Veterans' Affairs

                               ----------
                         U.S. GOVERNMENT PRINTING OFFICE 

57-012 PDF                       WASHINGTON : 2010 

For sale by the Superintendent of Documents, U.S. Government Printing 
Office Internet: bookstore.gpo.gov Phone: toll free (866) 512-1800; 
DC area (202) 512-1800 Fax: (202) 512-2104 Mail: Stop IDCC, 
Washington, DC 20402-0001 























                     COMMITTEE ON VETERANS' AFFAIRS

                    BOB FILNER, California, Chairman

CORRINE BROWN, Florida               STEVE BUYER, Indiana, Ranking
VIC SNYDER, Arkansas                 CLIFF STEARNS, Florida
MICHAEL H. MICHAUD, Maine            JERRY MORAN, Kansas
STEPHANIE HERSETH SANDLIN, South     HENRY E. BROWN, Jr., South 
Dakota                               Carolina
HARRY E. MITCHELL, Arizona           JEFF MILLER, Florida
JOHN J. HALL, New York               JOHN BOOZMAN, Arkansas
DEBORAH L. HALVORSON, Illinois       BRIAN P. BILBRAY, California
THOMAS S.P. PERRIELLO, Virginia      DOUG LAMBORN, Colorado
HARRY TEAGUE, New Mexico             GUS M. BILIRAKIS, Florida
CIRO D. RODRIGUEZ, Texas             VERN BUCHANAN, Florida
JOE DONNELLY, Indiana                DAVID P. ROE, Tennessee
JERRY McNERNEY, California
ZACHARY T. SPACE, Ohio
TIMOTHY J. WALZ, Minnesota
JOHN H. ADLER, New Jersey
ANN KIRKPATRICK, Arizona
GLENN C. NYE, Virginia

                   Malcom A. Shorter, Staff Director

                                 ______

       SUBCOMMITTEE ON DISABILITY ASSISTANCE AND MEMORIAL AFFAIRS

                    JOHN J. HALL, New York, Chairman

DEBORAH L. HALVORSON, Illinois       DOUG LAMBORN, Colorado, Ranking
JOE DONNELLY, Indiana                JEFF MILLER, Florida
CIRO D. RODRIGUEZ, Texas             BRIAN P. BILBRAY, California
ANN KIRKPATRICK, Arizona

Pursuant to clause 2(e)(4) of Rule XI of the Rules of the House, public 
hearing records of the Committee on Veterans' Affairs are also 
published in electronic form. The printed hearing record remains the 
official version. Because electronic submissions are used to prepare 
both printed and electronic versions of the hearing record, the process 
of converting between various electronic formats may introduce 
unintentional errors or omissions. Such occurrences are inherent in the 
current publication process and should diminish as the process is 
further refined.


















                            C O N T E N T S

                               __________

                             March 24, 2010

                                                                   Page
Examination of the U.S. Department of Veterans Affairs Regional 
  Office Disability Claims Quality Review Methods................     1

                           OPENING STATEMENTS

Chairman John J. Hall............................................     1
    Prepared statement of Chairman Hall..........................    30
Hon. Doug Lamborn, Ranking Republican Member.....................     3
    Prepared statement of Congressman Lamborn....................    31

                               WITNESSES

U.S. Department of Veterans Affairs:

  Belinda J. Finn, Assistant Inspector General for Audits and 
    Evaluations, Office of Inspector General.....................     4
      Prepared statement of Ms. Finn.............................    32
  Bradley G. Mayes, Director, Compensation and Pension Service, 
    Veterans Benefits Administration.............................    24
      Prepared statement of Mr. Mayes............................    51
  U.S. Government Accountability Office, Daniel Bertoni, 
    Director, Education Workforce, and Income Security...........     6
      Prepared statement of Mr. Bertoni..........................    36

                                 ______

American Legion, Ian C. de Planque, Assistant Director, Veterans 
  Affairs and Rehabilitation Commission..........................    16
    Prepared statement of Mr. de Planque.........................    48
American Veterans (AMVETS), Raymond C. Kelley, National 
  Legislative Director...........................................    15
    Prepared statement of Mr. Kelley.............................    48
Disabled American Veterans, John L. Wilson, Assistant National 
  Legislative Director...........................................    14
    Prepared statement of Mr. Wilson.............................    45
National Veterans Legal Services Program, Ronald B. Abrams, Joint 
  Executive Director.............................................    13
    Prepared statement of Mr. Abrams.............................    42

                   MATERIAL SUBMITTED FOR THE RECORD

Follow-up Letter:

      Hon. George J. Opfer, Inspector General, U.S. Department of 
        Veterans Affairs, to Hon. John Hall, Chairman, 
        Subcommittee on Disability Assistance and Memorial 
        Affairs, Committee on Veterans' Affairs, letter dated 
        April 29, 2010...........................................    55

Post-Hearing Questions and Responses for the Record:

      Hon. John J. Hall, Chairman, Subcommittee on Disability 
        Assistance and Memorial Affairs, Committee on Veterans' 
        Affairs, to Daniel Bertoni, Director, Education, 
        Workforce, and Income Security, U.S. Government 
        Accountability Office, letter dated April 5, 2010, and 
        response letter dated May 4, 2010........................    55
      Hon. John J. Hall, Chairman, Subcommittee on Disability 
        Assistance and Memorial Affairs, Committee on Veterans' 
        Affairs, to Bradley Mayes, Director, Compensation and 
        Pension Service, Veterans Benefits Administration, U.S. 
        Department of Veterans Affairs, letter dated April 5, 
        2010, and VA responses...................................    57


                 EXAMINATION OF THE U.S. DEPARTMENT OF
                    VETERANS AFFAIRS REGIONAL OFFICE
                DISABILITY CLAIMS QUALITY REVIEW METHODS

                              ----------                              


                       WEDNESDAY, MARCH 24, 2010

             U.S. House of Representatives,
                    Committee on Veterans' Affairs,
Subcommittee on Disability Assistance and Memorial Affairs,
                                                    Washington, DC.

    The Subcommittee met, pursuant to notice, at 2:28 p.m., in 
Room 334, Cannon House Office Building, Hon. John Hall 
[Chairman of the Subcommittee] presiding.
    Present: Representatives Hall, Donnelly, and Lamborn.
    Mr. Hall. Good afternoon, everybody. Thank you for your 
patience.
    Would we all please rise for the Pledge of Allegiance.
    [Pledge of Allegiance.]

               OPENING STATEMENT OF CHAIRMAN HALL

    Mr. Hall. Welcome to the Subcommittee on Disability 
Assistance and Memorial Affairs' hearing entitled, Examination 
of the U.S. Department of Veterans Affairs (VA) Regional Office 
Disability Claims Quality Review Methods--Is the Veterans 
Benefits Administration's (VBA's) Systematic Technical Accuracy 
Review or STAR Making the Grade?
    We are going to try to make abbreviated opening statements 
by myself and Ranking Member Lamborn as we understand votes 
will be called any time.
    That said, I welcome you all here in what has been a 
profoundly historic and important week for the Nation and for 
our veterans. Over the last 7 days, the full Committee convened 
a successful Summit, which brought many of you and dozens of 
other top veteran stakeholders together to aid us in fixing the 
VA compensation and pension claims process.
    From the Summit came a lot of very useful information which 
we welcome and look forward to using to solve the problems that 
VA faces in processing disability claims of our Nation's 
veterans.
    Next in a rare Sunday session, Congress passed and the 
President signed a sweeping health care reform package. And I 
am pleased that Secretary Shinseki as well as the Chairman of 
the full VA Committee and the Armed Services Committee have 
signed a letter and sent it to the veterans service 
organizations (VSOs) stating unequivocally that TRICARE and VA 
care will not be adversely affected by the national health care 
reforms.
    Also in the past few days, we passed the End Veteran 
Homelessness Act of 2010 to provide funding to help Secretary 
Shinseki's goal of ending homelessness for America's warriors, 
the Help Heroes Keep Their Homes Act, the COLA, the cost of 
living increase for veterans, and the National Guard Employment 
Protection Act.
    This particular hearing will be about the Systematic 
Technical Accuracy Review (STAR) technical review system and we 
will look at the accuracy of assessing disability compensation 
and pension claims rating and the disparity between accuracy in 
the different regional offices (ROs).
    Using this quality control tool, VBA should be able to 
focus attention on poorly-performing ROs and help the agency 
direct additional staff and training to the problem offices and 
at the same time look at those who are the highest-performing 
ROs and find out what they are doing right.
    The STAR system was implemented in October of 1998. Since 
fiscal year 2007, VBA has set for itself a goal of completing 
compensation claims ratings without error 90 percent of the 
time.
    Its long-term strategy goal is 98 percent. Unfortunately, 
we are still far from achieving that goal. And until the STAR 
system provides an accurate accounting of the error rate at 
VBA, it is difficult to envision a path to meeting this goal.
    So we are honored to have today the Office of the Inspector 
General (OIG), the U.S. Government Accountability Office (GAO), 
both of which have produced studies revealing issues that may 
be those impeding the efficiency and consistency of the STAR 
system. We are looking forward to hearing their testimony.
    I am personally troubled by GAO's finding that the VBA 
claims processing accuracy rate is particularly low in cases 
pertaining to post-traumatic stress disorder (PTSD) and 
traumatic brain injury (TBI). Today we hope to analyze these 
studies and hear testimony regarding those issues from OIG and 
GAO.
    There have been improvements made by the VBA to the claims 
rating system and the STAR system. We look forward to hearing 
about these improvements and how we, the Subcommittee, the full 
Committee, and Congress, can help.
    To fully understand STAR, it is important to review the 
compensation and pension (C&P) rating system itself. To assess 
a claim, the application for disability assistance must be 
developed, a process that involves obtaining all necessary 
evidence to support the veteran's claim. After development, the 
claims go to a rating veterans service representative (RVSR).
    The RVSRs determine if that disability is service-connected 
and assigns a percentage rating that is intended to represent 
the average earning reduction a veteran with that condition 
would experience in a civilian occupation. The veteran is then 
notified of that decision.
    For reopened claims, the assigned diagnostic codes affect 
the veteran's overall combined percentage of disability. The 
ROs, regional offices personnel applying the appropriate 
diagnostic code percentages to determine the combined level of 
disability.
    Once that claim is completed, it is declared an end product 
and the result of that claim is cleared and a work credit is 
given to the regional office. So a completed claim and 
corresponding cleared end product is then subject to review by 
STAR reviewers.
    In the 110th Congress, I introduced H.R. 5892, which sought 
to improve VBA's quality control measures. This bill was 
incorporated into an omnibus veterans' package, which was 
signed into law as Public Law 110-389. The Veterans Disability 
Benefits Claims Modernization Act of 2008, H.R. 5892, was part 
of that. So our hearing should also provide a chance to gauge 
how well these quality control measures are working.
    I thank you for being here, in advance for your testimony, 
and now I would like to recognize Ranking Member Lamborn for 
his opening statement.
    [The prepared statement of Chairman Hall appears on p. 30.]

             OPENING STATEMENT OF HON. DOUG LAMBORN

    Mr. Lamborn. Well, thank you, Mr. Chairman.
    And thank you for waiting, although I would have been happy 
for you to go ahead and start. But I know as a courtesy, you in 
a spirit of bipartisanship, you wanted to wait. So I thank you 
for that.
    And I do apologize. There was a misunderstanding between my 
staff and I. I thought that we were going to start after this 
first series of votes because I specifically asked that. And 
once they found the mistake, they notified me immediately, but 
I was two buildings away. So I got here as quickly as I could.
    But I know everyone's time here is extremely valuable and 
so I apologize. It is totally the fault of me and my office.
    Mr. Chairman, I want to, like you, welcome everyone to this 
hearing on the Department of Veterans Affairs' STAR Program. 
Throughout my tenure on this Committee, my fellow Members and I 
have called for stronger accountability within the VA claims 
system. For too long, the primary focus has been on production 
and this has led to an error rate that is unacceptable.
    I believe that the VA's greatest challenge, the claims 
backlog, is largely attributable to hasty decisions made 
without proper regard for accuracy. The ramifications of this 
approach can be seen throughout the entire system.
    Therefore, VA employee performance awards cannot be based 
entirely on production. There must also be a valid measure of 
quality.
    Under the STAR Program, a statistically valid sample of 
rating decisions from various regional offices is reviewed for 
accuracy. While this method may be useful from a macro 
perspective, it is not sufficient for ensuring individual 
accountability.
    VA must be able to identify employees in need of 
individualized remedial training. Without this essential 
component of the quality assurance process, VA will have 
perpetual problems in its claims system.
    In the 110th Congress, this Committee passed a provision 
that was included in the Veterans Benefits Improvement Act of 
2008 that required VA to conduct a study on the effectiveness 
of the current employee work credit system.
    I believe the upcoming report, along with the testimony 
that we will hear today, will provide valuable feedback for the 
Department to improve its Quality Assurance and Accountability 
Program.
    I look forward to hearing from our witnesses today, and I 
thank you all for your participation.
    And, Mr. Chairman, I thank you also and I yield back.
    [The prepared statement of Congressman Lamborn appears on 
p. 31.]
    Mr. Hall. Thank you, Mr. Lamborn.
    We are moving at lightning speed now. We would like to ask 
our first panel to join us at the witness table. We have 
Belinda J. Finn, Assistant Inspector General for Audits and 
Evaluations, Office of Inspector General, U.S. Department of 
Veterans Affairs, and Daniel Bertoni, Director of Education, 
Workforce, and Income Security with the Government 
Accountability Office.
    Welcome to you both, and your full written statements are 
entered in the record. So we will recognize you for 
5 minutes each for however much of that you would like to give 
to us directly.
    Ms. Finn, welcome, and you are now recognized for 5 
minutes.

STATEMENTS OF BELINDA J. FINN, ASSISTANT INSPECTOR GENERAL FOR 
   AUDITS AND EVALUATIONS, OFFICE OF INSPECTOR GENERAL, U.S. 
     DEPARTMENT OF VETERANS AFFAIRS; ACCOMPANIED BY LARRY 
 REINKEMEYER, DIRECTOR, KANSAS CITY AUDIT OPERATIONS DIVISION, 
   OFFICE OF INSPECTOR GENERAL, U.S. DEPARTMENT OF VETERANS 
 AFFAIRS; AND DANIEL BERTONI, DIRECTOR, EDUCATION, WORKFORCE, 
   AND INCOME SECURITY, U.S. GOVERNMENT ACCOUNTABILITY OFFICE

                  STATEMENT OF BELINDA J. FINN

    Ms. Finn. Thank you, Chairman Hall.
    I am fighting some allergies that sometimes cause coughing 
fits, so I am trying to avoid that.
    Mr. Hall. You are excused.
    Ms. Finn. Thank you for having us here today. I am pleased 
to be here to discuss our review of VBA's Quality Assurance 
Program and how VBA can improve the programs to better serve 
our veterans.
    I am joined today by Larry Reinkemeyer, who is the Director 
of the Kansas City Audit Operations Division.
    The OIG is committed to proactively reviewing the 
Department's key internal controls to identify weaknesses 
before they escalate into significant problems.
    Over the past 2 years, my office has audited three 
components of VBA's Quality Assurance Program, rating accuracy, 
rating consistency, and VBA's Site Visit Program.
    Although our written statement covers our work in all three 
areas, my comments today are only addressing our audit of the 
Systematic Technical Accuracy Review or STAR Program.
    In March 2009, we issued the audit of VBA's compensation 
rating accuracy and consistency reviews. We concluded that 
VBA's STAR process did not effectively identify and report 
errors in compensation claim rating decisions. We projected 
that about 77 percent of claims were accurate as opposed to 
VBA's reported accuracy rate of 87 percent.
    This equated to approximately 88,000 additional claims 
where veterans' monthly benefits may be incorrect.
    We identified five areas where VBA needed to improve the 
STAR Program. VBA agreed with all of our recommendations and 
just recently reported that all corrective actions were 
complete.
    First, the STAR reviewers did not identify some errors 
because they either did not thoroughly review available 
evidence or they inappropriately misclassified benefit 
entitlement errors as comments that did not count against the 
accuracy rate.
    Second, STAR management required regional offices to report 
quarterly on actions taken to correct errors but did not 
follow-up to ensure that these offices actually took the 
corrective actions on comments.
    Third, VBA excluded brokered claims from STAR reviews. The 
officials told us that the STAR Program's primary focus was to 
assess and report rating accuracy for each of the individual 
regional offices. Since two regional offices are involved in 
brokered claims, these officials believed it would be difficult 
to assign responsibility for rating accuracy on a brokered 
claim.
    Fourth, the STAR reviewers did not ensure regional offices 
submitted all of the selected compensation claim ratings for 
review. In fact, the offices did not submit about seven percent 
of the requested ratings for the 12-month period ending 
February 2008. We reviewed a sample of these unsubmitted claims 
and identified a benefit entitlement error rate of 
approximately 22 percent.
    Last, the STAR reviewers were not required to complete 
formal training on an annual basis. In contrast, the regional 
office staff that prepare and complete ratings are required to 
complete 80 hours of training per year to stay current on laws, 
policies, and processes that affect rating claims.
    The STAR management relies on the regional offices to take 
corrective actions on the issues identified by the STAR team. 
Since April 2009, the OIG Benefits Inspection Division has 
issued eight reports where we looked at regional office 
procedures to ensure the accurate and timely correction of 
errors identified by STAR.
    Our analysis of 148 errors found that regional office staff 
had not corrected 27 percent of these and in some cases had 
erroneously reported to STAR that the errors had been 
corrected.
    In closing, VBA is under tremendous pressure to process 
claims and reduce the growing backlog. Without an effective and 
reliable Quality Assurance Program, VBA leadership cannot 
adequately monitor performance to make necessary program 
improvements and ensure veterans receive accurate and 
consistent ratings.
    Mr. Chairman, thank you again for the opportunity to be 
here today. Mr. Reinkemeyer and I would be pleased to answer 
any questions you may have.
    [The prepared statement of Ms. Finn appears on p. 32.]
    Mr. Hall. Thank you, Ms. Finn. You had 4 seconds remaining. 
Good job.
    Ms. Finn. Thank you.
    Mr. Hall. Mr. Bertoni, welcome. You are now recognized for 
5 minutes.

                  STATEMENT OF DANIEL BERTONI

    Mr. Bertoni. Mr. Chairman, Members of the Subcommittee, 
good afternoon. I am pleased to be here to discuss the 
Department of Veterans Affairs' efforts to improve the quality 
of disability decisions.
    For years, we have noted that VA's claims processing 
challenges not only include making quicker decisions in 
reducing its claims backlog but also improving accuracy and 
consistency.
    My statement today focuses on steps VA has taken in 
response to recommendations from us and others to enhance its 
quality assurance tools, namely the Systematic Technical 
Accuracy Review or STAR Program, as well as other programs 
designed to address decisional consistency.
    Since STAR was first implemented, we have made numerous 
recommendations for improvement. For example, very early on, we 
noted that STAR reviewers lacked organizational independence 
because they also had claims processing duties and reported 
directly to regional managers whose claims they may review.
    Per our recommendation, VA moved to require 
organizationally independent STAR reviewers who are precluded 
from making claims decisions.
    In subsequent work, we found that STAR sampling was 
insufficient to ensure the accuracy of disability pension 
decisions.
    VA addressed our findings by consolidating pension 
processing into three locations and establishing a separate 
STAR review for pension claims.
    More recently we reported that VA is not using STAR to 
separately assess the accuracy of the benefits delivery at 
discharge and quick-start claims, alternative processes for 
fast tracking VA disability compensation claims for active-duty 
servicemembers.
    To date, the Agency has opted not to evaluate the extent to 
which staff are accurately developing and rating these claims, 
although such information could better inform training and 
focus program monitoring efforts.
    VA's Office of Inspector General has also recommended 
changes to the STAR Program which VA has begun to address, 
including establishing minimum annual training requirements for 
reviewers, mandating an additional supervisory review of STAR 
reports, sampling brokered claims for accuracy, and 
implementing more stringent procedures for conducting STAR 
reviews.
    Finally, the Agency has begun to take steps to address 
deficiencies that both we and VA's Inspector General have 
identified with its consistency review programs, which assess 
the extent to which regional offices and individual raters make 
consistent decisions on the same claim.
    We recommended that VA conduct systematic studies of 
impairments identified as having potentially inconsistent 
decisions. And in fiscal year 2008, VA did initiate this 
effort.
    However, last year, the Inspector General reported that VA 
had not followed through on its plans to conduct additional 
reviews, which was attributed in part to insufficient STAR 
staffing resources.
    The Agency has since developed a consistency review 
strategy and is in the process of conducting fiscal year 2010 
reviews. However, these efforts have only recently begun and it 
is too early to assess their impact.
    And despite various recommendations for improvement and 
actions to address them, VA has struggled over the years to 
improve the accuracy rate for disability compensation decisions 
which was 84 percent in fiscal year 2009 and well short of VA's 
stated goal of 90 percent.
    VA has attributed its inability to meet accuracy goals in 
part to the large numbers of newly-hired personnel conducting 
claims development work and their general lack of training and 
experience. We have also noted that human capital challenges 
associated with providing training to help new staff become 
more proficient will likely continue into the near future and 
could impact quality.
    Thus, it is important that VA continue to improve and 
maintain a robust quality assurance framework that not only 
supports staff in their understanding of the very complex 
business of making timely, accurate, and consistent disability 
decisions but also ensures that all veterans receive the 
benefits they are legally entitled to.
    Over time, VA's newly-implemented quality assurance 
initiatives have the potential to improve decisional accuracy 
and consistency if compliance with enhanced protocols, 
procedures, and standards is sustained.
    However, it is imperative that VA remain proactive in its 
quality assurance efforts going forward, especially as aging 
veterans and more veterans from current conflicts add to VA's 
already substantial claims workloads.
    And, Mr. Chairman, this does conclude my statement. I am 
happy to answer any questions you may have. Thank you.
    [The prepared statement of Mr. Bertoni appears on p. 36.]
    Mr. Hall. Thank you, Mr. Bertoni.
    I am just going to yield to Mr. Lamborn to make a comment 
about his questions.
    Mr. Lamborn. Mr. Chairman, I am going to waive my questions 
for the sake of time and just follow-up in writing to the 
degree that we need further explanations.
    So thank you.
    Mr. Hall. Thank you, Mr. Lamborn.
    And I will try to squeeze a few questions in here before we 
have to go across the street to vote.
    Ms. Finn, your OIG report indicates that there are about 
203,000 claims where a veteran is receiving incorrect monthly 
benefits.
    Does this number reflect only veterans who are being 
underpaid? I ask because it is important to know whether there 
is a systematic bias toward underpaying rather than overpaying 
veterans. What does your data suggest.
    Ms. Finn. No. We found errors are reported both for 
overpayments and underpayments.
    Mr. Hall. Can you give us a ratio or is that roughly equal.
    Mr. Reinkemeyer. No. I think most of the errors are 
underpayments where the claims examiner did not identify all 
the issues. A veteran may have filed his claim and had eight or 
nine issues and a couple of them were omitted in the rating 
decision. So I would say most, but we have no numbers on that. 
And the 203,000 is the projected number based on our sample.
    Mr. Hall. Okay. Well, that would be a good thing if you can 
cull that from your data. That would be good for us to know if 
it is tilted toward underpayment and how much.
    Mr. Reinkemeyer. Sure.
    [The VA Inspector General George Opfer, subsequently 
followed up in a letter dated April 29, 2010, which appears on 
p. 55.]
    Mr. Hall. Ms. Finn, your testimony also indicates that VA 
reported that it has completed actions to implement the 
recommendations of the OIG report of March 2009.
    Are there plans at OIG to follow-up on this?
    Ms. Finn. We do not have immediate plans for follow-up 
right now. We do over time, though, follow-up on selected 
reports to actually assess the effectiveness of corrective 
actions.
    We always follow-up with management to ensure that they can 
report to us what they have done.
    Mr. Hall. Thank you.
    Your study observed that STAR reviewers do not identify 
some errors because they either do not thoroughly review 
available evidence, they fail to identify the absence of 
necessary medical information, and sometimes they misclassify 
errors in a way that resulted in errors not being counted 
against the accuracy rate.
    What do you believe can cure these deficiencies? Is it a 
training issue alone?
    Ms. Finn. I believe our recommendation in the report 
related to increased supervisory reviews on those issues. 
Certainly to ensure that comments that should be corrected 
should be counted as errors were not counted.
    Mr. Hall. Thank you.
    And you noted that in your testimony that STAR management 
required that regional offices report quarterly on actions 
taken to correct benefit entitlement errors, but they did not 
require or follow-up to ensure regional offices were actually 
taking corrective actions on the comments made by the STAR 
reviewers.
    Do you have any indication that VA has remedied this 
concern and what assurances the corrective action at VA as 
purported has actually been implemented in a satisfactory 
manner?
    Ms. Finn. The only assurance we have is VBA's response to 
our recommendation. And as I alluded to in my oral and written 
testimony, we look at this issue when we go on our regional 
office inspections and we do not always find that STAR errors 
have been actually corrected.
    Mr. Hall. And lastly for you, Ms. Finn, your review 
indicates that brokered claims are experiencing a 69 percent 
accuracy rate.
    What impact do you believe brokered claims will have on the 
overall VBA claims accuracy rate in the short term and long 
term or can you tell yet?
    Ms. Finn. No, I cannot tell you. We are conducting a 
national audit looking at the brokering or redistribution of 
claims program. And we are assessing the impact on the 
timeliness of claims processing and also how this could 
possibly impact accuracy or how the brokering program is being 
implemented in respect to accuracy.
    So I do not at this point have any real solid projections. 
Our sampling that we did a year ago was based on a sample of 
the brokered claims to determine our error rate at that time.
    Mr. Hall. I would ask you, Ms. Finn, and then Mr. Bertoni 
also, the following question.
    In the list of the 2009 STAR accuracy ratings for all 
regional offices, that we were given, the most accurate RO was 
Des Moines with an accuracy rating of 92.34 percent, ranging 
down to Baltimore with a 69.34 percent accuracy rate. The 
Washington Regional Office is not rated but we understand its 
quality level may be lower than Baltimore. Have you already or 
do you have plans to try to identify what it is that the Des 
Moines office is doing that others maybe should be emulating or 
what it is that Baltimore is doing that others should avoid or 
is it not that simple? Are there other factor?
    Ms. Finn. I do not know that it is that simple. We have not 
yet visited the Des Moines office on our benefit inspection. We 
have, however, gone to Baltimore. And we do a selected review 
of claims. We are looking at specifically claims for PTSD, TBI. 
We have looked at diabetes and also we look at brokered claims.
    So our review of claims is not consistent necessarily with 
the STAR methodology and, therefore, the numbers are not 
directly comparable.
    We did find when we looked at those selected claims in 
Baltimore that they had about 38 percent inaccuracy on those 
types of selected claims. We also found a number of management 
issues in Baltimore in that they had had leadership vacancies 
and they also had had staff removed from the regular regional 
office operations to work on a disability pilot.
    Mr. Hall. Right.
    Ms. Finn. So we felt that impacted their ability to do 
quality work.
    Mr. Hall. So the virtual RO pilot project and DES pilot may 
have been hurting, at least temporarily, the overall accuracy 
rating?
    Ms. Finn. We felt it was an issue that needed to be 
addressed. We could not directly quantify the impact.
    Mr. Hall. Okay. Well, thank you very much.
    Mr. Bertoni, would you care to address that same question.
    Mr. Bertoni. Yes. In terms of the variation in quality, I 
think the OIG testimony bears out some interesting issues with 
the accuracy rate itself.
    I think one question might be variation might be due 
perhaps to quality of the review across the regions. I mean, we 
see lack of training, folks who may not know completely the 
rules of the game who may be reviewing these claims. And in 
some regions, they may be doing a more thorough job. In other 
regions, they may not.
    The other part is that there is some blame to lay on the 
regions in terms of quality. I think you cannot underrate 
regional management. There are good managers out there who 
really embrace quality and it is ingrained in the staff. They 
are doing innovative things in terms of training.
    And we have a report we are about to issue to this 
Committee in regard to training and some actually best practice 
type things that are going on out there.
    And also the ability to hire staff and keep staff. And I 
think it is more difficult in certain areas than others, 
especially some of the larger urban areas, and supervisors.
    So I think it is a range of things. It is a combination of 
things that sort of lead to this variance.
    But I would also, again, hearken back to the quality of the 
reviews. Is it true that these two regions are entirely 
different? It may be reviewers are making mistakes and it might 
lead to some of this disparity.
    Mr. Hall. Mr. Bertoni, in our last hearing examining the 
benefit delivery at discharge (BDD) and quick-start programs, 
you noted that VA lacks sufficient and specific performance 
measures for assessing the accuracy of decisions on BDD claims. 
And you recommended that VA consider options for separately 
estimating the accuracy of such decisions.
    However, VA has asserted that the cost of sampling pre-
discharge claims as part of STAR would outweigh the benefits.
    If VA subjected BDD and quick-start claims to STAR review, 
could we make these programs better, even better than they 
apparently are? If so, what staffing, training, or other 
resources do you think would be needed?
    Mr. Bertoni. I think the answer is yes. There were 250,000 
initial claims that year and BDD represented a little over 
50,000. So that is about 20 percent of the new claims coming 
into the pipeline.
    So if you can make a difference there, you can make a 
difference. And the goal is to make BDD claims an even larger 
part of the initial claims.
    The idea that it is not cost effective, I do not know if I 
have seen enough to agree to that. VA did do some analysis for 
us, but it was only on the cost side, full-time equivalents, 
space, modifications to STAR. They never pulled a sample of BDD 
claims to say what did we get from this additional review. So 
what is the return on investment on the other side is not 
clear.
    And I continue to believe that VA continues to process more 
claims outside the traditional process. They are reengineering 
processes. They are reengineering, you know, how they do work. 
And I think they are using in many cases sort of STAR as more 
of a blunt instrument rather than using it to really drill down 
into these specialized new business processes to see how they 
are doing.
    And I think at the end of the day, they need to do more of 
that to find out where they are getting the return on 
investment and where they need to walk away from things.
    Mr. Hall. You asserted in your testimony that the accuracy 
rate was 86 percent in fiscal year 2008 and 84 percent for 
fiscal year 2009, well short of VBA's goal of 90 percent.
    So at this point, what additional recommendations do you 
have for improving the STAR Program.
    Mr. Bertoni. I think there is a mixture here again. There 
are program design issues that they have to address, but there 
is also management and oversight. And I think the OIG really 
hit on many of those.
    I was extremely surprised to hear that you have such an 
important program and had no training requirements. I was very 
surprised. The fact that they have not done that yet is 
something that they really need to move on.
    A supervisory review, you know, having another look at 
these cases before they are blessed or deemed to be correct is 
very important. It is part of their quality assurance process. 
They need to build in some management controls and certainly 
training.
    We have a lot of new staff and integrating those staff and 
putting them in a position to be able to accurately assess 
claims is going to take good management, good supervisory 
review to sort of teach people as to, you know, what needs to 
be done. So I think that is very important.
    Mr. Hall. Thank you, Mr. Bertoni.
    You also note that due to increased Congressional support, 
VBA has increased its compensation and pension staffing. 
Specifically you point out that VBA more than doubled the size 
of the quality assurance staff, allowing it to increase the 
scope of quality assurance reviews.
    Do you believe this additional staff has made a difference 
or are you of the opinion that staffing alone is insufficient 
to address deficiencies? I think you started to answer this 
question, but----
    Mr. Bertoni. Staffing never hurts. And if you staff up a 
program and you give folks the tools that they need to do the 
job they need to do, it could make a difference. And that means 
again training them up and putting them in a position to make a 
difference.
    So the numbers will help. But, again, it is going to take 
time. There is a learning curve. But over time, I think it will 
be effective, but you also have to build in the management 
controls to make sure that the reviews do have the authority 
and the oversight teeth that they should have.
    Mr. Hall. Thank you.
    Some VSOs point to the high rate of remand ordered by the 
Board of Veterans' Appeals (BVA) or the U.S. Court of Appeals 
for Veterans Claims (CAVC) as indications that the accuracy 
rate is a lot lower than reported by VBA.
    What is your assessment of how the errors found by the 
Board and the Court impact VBA's reported claims processing 
accuracy rate?
    And maybe, Ms. Finn, if you could comment on that as well.
    Ms. Finn. Mr. Chairman, we have not done any specific work 
on the appeals process, so I would prefer not to comment.
    I would note, however, that remands involve other issues 
other than just whether the rating decision was accurate. There 
could be procedural reasons for a remand. But other than that, 
I will leave that one to Mr. Bertoni.
    Mr. Bertoni. We had worked in that area not very recently, 
but I would say again there are a range of reasons why a case 
would be remanded back. Sometimes it is a matter of it has been 
so long that the impairments change, new impairments are 
introduced, other evidence has not been considered. So there 
are a lot of reasons.
    But I think overall in hearing what I have heard today with 
the OIG report, what we know and work that we have done in the 
past in terms of missing claims and how that could affect error 
rates, I would say the error rate--I do not have a whole lot of 
faith in the error rate or, I am sorry, the accuracy rate of 84 
percent. It is probably some other lower figure, but I do not 
know what that is.
    Mr. Hall. And, lastly, Mr. Bertoni, before the changes in 
the STAR Program were mandated by Public Law 110-389, only ten 
cases per month for each RO were being reviewed by the STAR 
system. Today VA staff advises us that twice as many cases are 
being reviewed, 240 cases per RO per year.
    How many total cases were decided by the regional offices 
last year and is a sample of 20 cases per month per RO 
sufficient or should that sample be larger, particularly for 
ROs to process a larger number of claims?
    Mr. Bertoni. Unfortunately, I do not have the number of 
total cases. But any time you can ramp up the numbers is going 
to make the projectability of what you do more rigorous.
    Our concern again with the current system is STAR tends to 
be used more as a blunt instrument to account for accuracy of 
all claims regardless of varied business processes.
    More and more, we see these claims being processed through 
these alternative or re-engineered processes in order to 
expedite things. And the Agency could be using STAR more 
effectively to assess how they are doing with these alternative 
processes. They are not. And BDD again is a good example of 
that.
    Every re-engineered process should include some kind of 
quality assurance whether it is continuous or a snapshot or a 
limited investment. Without doing that, you waste a lot of 
time. You do not know what you are getting for your money and 
you might end up where you do not want to be at the end of the 
day.
    I think failure to build any regular or limited investment 
quality assurance into any of these alternative processes is a 
failure of management. And we have noted along the way these 
varied again re-engineered or alternative means by which cases 
are being decided where the quality assurance, the STAR review 
was not sufficient in our view to get a good sense of how 
accurate and consistent many claims are.
    Mr. Hall. Well, thank you, Mr. Bertoni and Ms. Finn and Mr. 
Reinkemeyer. I am grateful for your testimony. And we may have 
further questions for you that we will send you in writing.
    We now have a series of votes called on the floor, 
approximately 30 minutes we are told, so we will ask our next 
panels to be patient, please, as you are used to, I am sure, by 
now.
    And the first panel is excused. Thank you again.
    The Subcommittee is recessed for votes.
    [Recess.]
    Mr. Hall. Welcome again, and thank you for your patience. 
The Subcommittee on Disability Assistance and Memorial Affairs 
will now resume its hearing on the STAR Review, Making the 
Grade. And our second panel, thank you for joining us again. 
Ronald B. Abrams, Joint Executive Director, National Veterans 
Legal Services Program (NVLSP); John L. Wilson, Assistant 
National Legislative Director, Disabled American Veterans 
(DAV); Raymond C. Kelley, National Legislative Director of 
AMVETS; and Ian C. de Planque, Assistant Director, Veterans 
Affairs and Rehabilitation Commission of the American Legion.
    Gentlemen, thank you so much for being here today. And your 
full statement, written statement, of course is entered into 
the record. So we will give you each 5 minutes to expound upon 
it, and starting with Mr. Abrams. You are now recognized.

   STATEMENTS OF RONALD B. ABRAMS, JOINT EXECUTIVE DIRECTOR, 
   NATIONAL VETERANS LEGAL SERVICES PROGRAM; JOHN L. WILSON, 
  ASSISTANT NATIONAL LEGISLATIVE DIRECTOR, DISABLED AMERICAN 
  VETERANS; RAYMOND C. KELLEY, NATIONAL LEGISLATIVE DIRECTOR, 
 AMERICAN VETERANS (AMVETS); AND IAN C. DE PLANQUE, ASSISTANT 
   DIRECTOR, VETERANS AFFAIRS AND REHABILITATION COMMISSION, 
                        AMERICAN LEGION

                 STATEMENT OF RONALD B. ABRAMS

    Mr. Abrams. Thank you, Mr. Chairman. I have some good news. 
We have just met with hundreds of service officers for Legion 
and Purple Heart. And I am happy to tell you that when they 
talked to me they said STAR does catch errors, they do order 
the VA to fix them, and there is some training based on the 
errors caught by STAR. However, based on the same comments of 
these service officers, and our own personal experience, we 
find that the training based on the errors caught by STAR is 
not generally doing the job. The errors still continue and the 
emphasis is still on production in the regional offices.
    NVLSP believes that the quality of the regional office 
adjudications is much worse than what is reported by the VA. 
And all you have to do if you want to look at it in an 
independent check is look at the statistics produced by the 
Board and the Court of Appeals for Veterans Claims. And I can 
tell you as far as consistency goes, I defy almost anyone to 
look at 20 evaluations of mental conditions and make rhyme or 
reason out of them. The rating schedule is not applied on a 
consistent basis.
    I am not going to over the BVA statistics. You all have 
them. I will say that the results of what we do with the 
American Legion, and we have done over forty quality checks in 
regional offices, we find that the quality is much worse. In 
many instances claims are denied in a premature manner or 
benefits are not paid at the proper rate because the regional 
office was more concerned about claiming work credit and 
reducing the backlog than taking the time to develop and 
analyze the claim properly. STAR shows about a 16 percent error 
rate. We are finding over a 30 percent error rate, in general.
    Here are some of the errors that we have found that you 
should know about. Assignment of erroneous low ratings for 
service-connected mental conditions. Erroneous denial of claims 
for service-connection for mental conditions. Failure to 
consider 38 USC 1154(b), the Combat Veteran Statute. Erroneous 
denials of claims for individual unemployability. Failure to 
consider presumptive service-connection. And inadequate 
requests for medical opinions. Sometimes doctors are not asked 
the right questions and that leads to a litany of problems.
    Based on that we ask that the VA work measurement system be 
altered so that quality, production, and timeliness are 
concepts that really drive the system. That is the most 
important thing. You need to change that to make any effective 
change. We would also like a quality control with teeth and an 
independent quality check.
    I will leave it at that except to say that what the VA is 
doing now is obviously not working. We need to have things 
changed. Thank you.
    [The prepared statement of Mr. Abrams appears on p. 42.]
    Mr. Hall. Thank you, Mr. Abrams. Mr. Wilson, you are now 
recognized.

                  STATEMENT OF JOHN L. WILSON

    Mr. Wilson. Thank you, sir. Mr. Chairman and Members of the 
Subcommittee, I am pleased to appear before you on behalf of 
Disabled American Veterans to address VBA's Systematic 
Technical Accuracy Review program.
    Successfully reforming the veterans benefits claims process 
will require training and accountability, two elements central 
to producing quality results for veterans. VBA continues to 
struggle with the quantity of claims, with quality, training, 
and accountability taking a backseat. VBA's primary quality 
assurance program is the STAR program. The STAR program can 
identify three types of errors: benefit entitlement, decision 
documentation and notification, and administrative.
    The STAR program was evaluated by the VA OIG in the March 
2009 report, which determined the program did not provide a 
complete assessment of rating accuracy. The OIG found STAR 
processors did not effectively identify and report all errors 
in compensation claim rating decisions. While VBA stated STAR 
reviewers achieved an 87 percent technical accuracy rate, the 
OIG projected an accuracy rate of only 78 percent based on its 
review of STAR reviewed cases. The OIG determined that this 
equates to approximately 203,000 claims in that 1 year alone 
with veterans monthly benefits may be therefore incorrect. The 
OIG determined the STAR reviewers did not identify some of the 
missed errors because they either did not thoroughly review 
available medical and non-medical evidence, did not identify 
the absence of necessary medical information, or 
inappropriately misclassified benefit entitlement errors in the 
comments section. These findings point to the need for greater 
management oversight and an effective formal training program 
for the STAR reviewers.
    We had heard reports that while STAR reviewers could 
benefit from formal training current workload requirements do 
not allow for the necessary training time. This common theme 
for VBA underlies even STAR reviews; quantity over quality. 
Even in the area that is supposed to ensure quality of at least 
a technical nature.
    The need for a quality control program as an adjunct to the 
STAR quality assurance program can also be seen when considered 
through a review of the Board of Veterans' Appeals summary of 
remands. Nineteen thousand one hundred cases, or 34 percent of 
appeals reaching the BVA in fiscal year 2009, were due to 
remands for notice not being provided to claimants, failed 
requests for service medical records and personnel records, or 
ignored travel Board requests. These elementary errors were 
either undetected or ignored. A 34 percent error rate on such 
basic elements in the claims process is simply unacceptable.
    With no incentive to prevent such errors, and a constant 
focus on production, quality will continue to decline. DAV 
agrees with the VA OIG recommendations to improve the STAR 
program. In addition, we recommend VBA establish a quality 
control program that looks at claims in process in order to 
determine not just whether a proper decision was made but how 
it was arrived at in order to identify ways to improve the 
system. Combining results from such quality control reviews 
with STAR's quality assurance results, and the data from 
remands from the Board of Veterans' Appeals, and the Court of 
Appeals for Veterans Claims, could yield valuable information 
on trends and causes of errors. If this data could be 
incorporated into a robust IT system, proper analysis of such 
data would provide management and employees insights into 
processes and decisions. With a modern IT system, VBA would be 
able to do quality control in real time, not just after the 
fact. This in turn would lead to quicker and more accurate 
decisions on benefits claims and more importantly to the 
delivery of all benefits earned by the veteran, particularly 
disabled veterans, in a timely manner.
    That concludes my testimony. I would be happy to answer any 
questions.
    [The prepared statement of Mr. Wilson appears on p. 45.]
    Mr. Hall. Thank you, Mr. Wilson. Mr. Kelley. You are 
recognized.

                 STATEMENT OF RAYMOND C. KELLEY

    Mr. Kelley. Thank you for giving AMVETS the opportunity to 
present our views on the STAR program. AMVETS agrees with VA's 
Office of Inspector General's March, 2009 report that 
identified eight issues that will improve the process of 
reviewing claims for errors, and AMVETS is please to see VBA is 
taking action to correct these issues. AMVETS is concerned, 
however, with what is done with the information that is gleaned 
from STAR.
    For the STAR program to truly be effective AMVETS believes 
three things must be done. First, STAR must be enhanced so 
trends and errors can be easily identified by regional offices. 
With this information, VBA must hold ROs accountable for 
failures in accuracy and insist that the ROs develop 
improvement strategies and include training for these accuracy 
issues.
    Second, VBA must change its culture of timeliness and 
strive for a culture of accuracy. Whether or not STAR is 
completely accurate in its review is important but not nearly 
as important as what is done to ensure the same mistakes are 
not made again.
    Third, OIG must conduct periodic reviews of the STAR 
program to ensure that its accuracy ratings are within a 3 
percent error margin.
    Even though AMVETS believes the STAR program is effective, 
we believe it could be expanded to ensure that specific 
programs such as BDD and specific conditions can be tracked for 
anomalies that occur so improvement strategies and specific 
training can be implemented. After the Veterans Claims 
Assistance Act (VCAA) was introduced in the first quarter of 
fiscal year 2002, nearly half of all errors were VCAA related. 
VBA had the ROs retrain their claims processors and in the last 
8 months of the fiscal year these types of errors were reduced 
by one-third. This type of accountability needs to be the rule 
and not the exception.
    As you know, the STAR program identifies errors in claims 
decisions each month through random sampling of each regional 
office. The results of the reviews are sent to the ROs. Errors 
that are found are to be corrected by the RO who made the 
decision. The corrections are made, but all too often the ROs 
do not implement strategies to ensure that claims processors do 
not continually repeat the same mistakes.
    AMVETS believes the reason these strategies are not 
developed is that the culture within the regional offices is 
one of timeliness and not one of accuracy. STAR has 
consistently found that nearly 20 percent of claims are in 
error over the past decade, but VBA has done little to ensure 
that mistakes that are made in the regional offices are 
understood and that strategies for improvements are put in 
place. On the other hand, VBA does require ROs to develop 
corrective action plans if they do not reach strategic goals 
for production, inventory, and timeliness. This paradigm must 
be flipped.
    The March, 2009 OIG report clearly defines the gaps in the 
STAR program that have caused the 10 percent disparity in 
compensation and pension rating accuracy. AMVETS believes VBA 
is taking action to close these gaps, however we believe it is 
important to have this accuracy fall within a 3 percent margin 
of error. Therefore, AMVETS requests that OIG conduct a follow 
up to the 2009 report to ensure VBA's gap solutions are 
productive and that OIG continue to conduct periodic reviews of 
STAR to be sure the program reaches and maintains that 3 
percent margin of error.
    Mr. Chairman, this concludes my remarks and I will be happy 
to answer any questions that you have.
    [The prepared statement of Mr. Kelley appears on p. 48.]
    Mr. Hall. Thank you, Mr. Kelley. Mr. de Planque.

                 STATEMENT OF IAN C. DE PLANQUE

    Mr. De Planque. Thank you Mr. Chairman and Members of the 
Committee. On behalf of the American Legion I would like to say 
that the STAR program is a potentially effective tool that if 
used in a more effective manner could really help VA deal with 
the accuracy issues. It is refreshing to see that in the OIG 
report from March 2009, that the issues that OIG identified VA 
concurred with and has recently reported that they are making 
the changes to correct those.
    The American Legion would recommend three points which 
could help VA enhance the efficiency of STAR and the ability 
for it to be effective. The first point is to create an 
aggregate record of all of the errors that are reported through 
STAR so that it can be analyzed. The second point would be to 
use the collected data to develop a targeted training program. 
And the third point would be to have regular, outside, 
impartial oversight of the process.
    Going back to elaborate on the first part. The errors are 
being reported back to the ROs. However, to the best of our 
knowledge there is no one consolidated effort to aggregate 
these mistakes, common errors, and deficiencies that are being 
noted. If you combine this with common errors and deficiencies 
noted from the Board of Veterans' Appeals, the Appeals 
Management Center (AMC), and the Court of Appeals for Veterans 
Claims, you can develop an overall picture both on a regional 
office level, perhaps even on a team level, but more 
importantly on a national level of what the common errors 
within VA are. And you can use that to determine where to 
devote the resources to improve accuracy within VA.
    That leads directly into point number two, targeted 
training. We have reports and when we have spoken to VA 
employees on the American Legion quality review visits to 
regional offices, it is noted and VA notes that when STAR 
identifies problems they are reported back to the teams and the 
coaches for correction involved. However, if you can notice a 
national trend, if you can notice that consistently VA is 
experiencing problems with, say, rating mental health 
disorders, or improperly asking for exams, then this can be set 
into a targeted training program so that VA is getting the most 
use possible out of their 80 to 85 hours of mandated training, 
that it is going to correct the areas that you need corrected. 
If you take a math test and you realize that you are having 
problems with binomial equations, then you need to go back and 
do some extra work on binomial equations so that the next time 
you take that math test you get it right. This is exactly the 
same sort of thing that VA could use this aggregate data to 
accomplish.
    And the third point that I want to make is about outside 
oversight, third party oversight. In the recent report this 
morning that was published in the Federal Register on the OIG 
investigation of Togus, Maine, one of the things that was 
pointed out, 25 percent of the STAR errors were not corrected 
in accordance with VBA policy. And two examples that are listed 
here, STAR instructed the regional office to inform a widow 
that her child could be entitled to Dependency and Indemnity 
Compensation (DIC) benefits. There is no evidence in the claims 
folder showing staff informed the widow of this potential 
entitlement. Furthermore, the RO staff erroneously informed 
STAR that they corrected the error. Secondly, STAR instructed 
the RO to send a notification letter for a burial claim to the 
proper claimant. While the staff informed STAR that they 
corrected the error they did not send the revised error to the 
proper claimant.
    So clearly, and this is not new that we have seen in the 
reports of the OIG of the various specific regional offices, 
even if STAR is capturing the errors they are not necessarily 
being followed up on, which is why you need third party 
oversight. You need somebody to go in and double check that 
they are crossing the I's, dotting the T's, and correcting 
these errors.
    VA has potentially a very effective tool here and we want 
them to be able to use it effectively. And with the three 
points that we believe will be effective to create an aggregate 
of that information that can be searched for trends, to use 
those trends to target the training, and to add outside 
oversight, we believe that this can be an effective tool for 
VA.
    I thank you very much, Mr. Chairman and the Committee, and 
we would be happy to answer any questions.
    [The prepared statement of Mr. de Planque appears on p. 
48.]
    Mr. Hall. Thank you, is it Mr. de Planque or de Planque, 
just----
    Mr. De Planque. It is de Planque, technically.
    Mr. Hall. De Planque, okay, thank you. Just like to try to 
get the correct pronunciation. You mentioned that training 
should be more specialized. Who do you think should oversee 
this training? Should it be standardized at either the RO or 
national levels, or individualized for each local VA office?
    Mr. De Planque. What I am proposing here, what the American 
Legion is proposing here, is to have some national training 
that is targeted to national areas of deficiency. However, 
within the individual ROs you also, there are different 
training staff within the individual ROs. And so when they 
notice something that is within their particular office 
obviously they have some initiative to direct the training on 
their own.
    However, what we are talking about here is, when you are 
identifying a national problem that a consistent number, or 
even a regional problem, that ROs in a certain area are 
deficient in performance in a certain aspect, such as applying 
the mental health disability schedule, or properly sending out 
VCAA notice, then it is important to have the national training 
specifically craft training that will address the issues that 
are noted. And that is why it is important to actually have a 
method for capturing that that takes into account not only STAR 
but also, as I said, there needs to be a method for capturing 
the errors that are being seen at the Appeals Management 
Center, that are being seen at the Board of Veterans' Appeals, 
and that are being seen at the courts. And if you combine those 
four elements, STAR, the AMC, the BVA, and the CAVC, then you 
actually have a big picture, the picture from, you know, 3,000 
feet of what is going on. And you can create the targeted 
training that is going to be more effective.
    Mr. Hall. Thank you. And just one more question. Sir, you 
suggested incentives for bonuses in the management of the ROs 
are driven by quantity rather than quality. How would you 
change the incentive structure to help encourage the accuracy 
and quality of claims?
    Mr. De Planque. There is technically one point that 
addresses accuracy. However, it does not substantially impact 
the bonus structure. The vast majority of it is driven by 
timeliness or quantity of claims. And what the American Legion 
has maintained for a long time is that the work credit system 
needs to be addressed. We have proposed, although we are open 
to other proposals, the idea that work credit not be counted 
until a claim is finally adjudicated so that you can determine 
that it was correctly done every step of the way. That may be a 
more complicated system. However, some system that takes into 
account the accuracy of the claim in counting the credit for 
that work rather than simply that the claim was passed along to 
another desk needs to be implemented. As the system stands 
right now where you are counting people on three and a half 
claims a day rather than three and a half correctly done claims 
a day, that is where you are running into the problem. If you 
start counting them on accurately and effectively completed 
claims, then that is when VA is going to be in a situation 
where the incentive is to get the claims done right and not to 
get a large number of claims done.
    Mr. Hall. Thank you. Mr. Kelley, in your testimony AMVETS 
asserts that the STAR program should be expanded to ensure that 
specific programs such as BDD and specific conditions can be 
tracked for anomalies that occur so that improvement strategies 
and specific training can be implemented. VA has claimed that 
the benefits of such changes do not surpass the cost of these 
measures. Do you concur with that position, and if not why?
    Mr. Kelley. I do not agree with the, that it is not worth 
the cost. These are veterans who are waiting for a benefit, an 
accurate benefit. And I do not think there should be a cost 
analysis against what veterans are waiting and deserve to have 
done accurately.
    Mr. Hall. You point out that not only is training for VBA 
raters needed but sometimes retraining is needed to reteach key 
lessons based on updated procedures. In the 10 years since 
STAR's establishment, the VA has yet to surpass a 90 percent 
rating accuracy. With that in mind, do you think that the STAR 
review team needs to be retrained to make sure that they are 
correctly employing the STAR quality control system?
    Mr. Kelley. I have not really put any thought into that. It 
is a good concept. I think just having enough people in place 
within the STAR, I think they had 24 billets and they were 
using eight to process these, you are putting a lot of pressure 
on individuals again to do a lot of work in a little bit of 
time. I think making sure that they have the right resources 
and that they are trained properly, and retrained. If you start 
seeing deficiencies it is absolutely time to go back and 
retrain.
    Mr. Hall. Thank you, Mr. Kelley. I just wanted to comment 
that we had three veterans who were from, members of AMVETS, 
with their service dogs in our office last week. And it was a 
remarkable experience. My dogs were very interested when I got 
home. But it is a great program and I want to thank you for the 
organization's work with service dogs, and trying to make them 
more available to all veterans who need them.
    Mr. Abrams, you suggested VBA should increase its claim 
processing accuracy by employing a system of awards and 
disincentives. What would the system look like in your mind? 
Mr. ABRAMS. Well, if you adjudicate claims from the point of 
view of the veteran as opposed to VA, and the disincentive is 
if you do not do it right and the appellate period extends, the 
managers in the regional office cannot claim work credit, that 
is a disincentive. If you give the managers the incentive to do 
it right because you are going to reward them with work credit, 
they are going to encourage their workers to do it right. VA 
workers will work to what their managers want them to do, as do 
postal workers, and patent office workers. And you have read 
about the problems with those groups. Set the system up so that 
what the managers want is what we want, quality adjudications.
    Then we could also reward people who do good work and 
plenty of it. But instead of just plenty of it work, that is 
not enough. That is a good way to make people do the right 
thing, and get benefits promptly to our veterans.
    Mr. Hall. Which was what this Committee intended in the 
legislation we passed in 2008.
    Mr. Abrams. We certainly support that. We were happy that 
we did it, and we anxiously await the response.
    Mr. Hall. As do we. You suggest in your testimony that VBA 
staff, or some VBA staff, take an adversarial attitude toward 
claims filed by veterans seeking benefits for mental conditions 
such as PTSD and TBI. What examples do you have of such bias 
that you can come up with off the top of your head?
    Mr. Abrams. If you want, we have plenty of cases where not 
only the regional office but the Board of Veterans' Appeals 
denied claims even though the veteran met the requirements in 
the diagnostic code to get the higher evaluation. In fact we 
had one, I just trained on it yesterday, where we screened it. 
It was a BVA denial, and the VA General Counsel did not even 
let us brief the case. They said let us award the benefit. They 
completely conceded. There is a reason. There is something 
going on where the VA does not want to evaluate properly. And I 
am not saying every case. We are talking about a certain 
percentage of cases that need to be corrected. But we are 
finding a high level there. And we have other instances, but I 
would have to go back through files and go over them, and if 
they are public I can share them.
    Mr. Hall. I would appreciate that and I am sure the 
Committee would. Mr. Wilson, in your testimony you indicate 
that the primary problem with the current accountability system 
is that VBA employees are not usually held responsible for 
their errors. In what specific way would you propose that VBA 
hold its employees accountable? Are Mr. Abrams' suggestions 
similar to what you would suggest?
    Mr. Wilson. We view the STAR program as a quality assurance 
program. It does what it does rather well generally speaking, 
although it certainly has an error rate that is nonetheless 
unacceptable. But it does measure the process itself. It tells 
you what the results of that process will provide. It does 
that. What it does not provide, however, is quality control. 
You do not have a quality control program where you can 
determine if rating decisions are correct based on the merits 
of the case. There is no effective program in place beyond each 
review to determine whether this process is giving the correct 
rating decisions. What we need is a program that will provide 
quality inspections and deliverable quality peer reviews of the 
of the work done at the end of the day. This work, currently by 
coaches, is done only on an occasional basis, unfortunately. 
What is needed is a program that would aggregate this quality 
data, roll it up and provide VBA a national level, trend 
analysis providing opportunities for recognition and training, 
focused on various regional offices or individual employees. 
That data could also be tied into the kinds of compensation 
packages and appraisals that would be provided to not only the 
senior supervisors, but to all individuals in the VBA.
    Mr. Hall. Your testimony alludes to the benefits of 
creating a new, more robust IT system that would incorporate 
many different measurements of accuracy. We heard from our IT 
witness previously that at least temporarily in Baltimore there 
may be a lengthening of the time or an increase in inaccuracy, 
a decrease in accuracy, due to the fact that they have siphoned 
people off to work on the virtual RO who otherwise would be 
working on the normal, old-fashioned way of processing claims. 
So I am just curious what your thoughts are about this move 
that is being attempted into the virtual world and increased 
use of IT? Are there less costly alternatives? Or should the VA 
pursue this and keep going down the road that they are 
currently on.
    Mr. Wilson. It is an important initiative and I would like 
to respond for the record to the question.
    [Mr. Wilson subsequently provided the following 
information:]

    VA must continue moving forward on the path towards a fully 
electronic IT system--the Veterans Benefits Management System (VBMS)--
which will become the backbone of the claims processing system. 
Although there may be temporary losses in production, and potentially 
accuracy as well, while new methods and technology are developed and 
implemented, VA must remain focused on long term reform. Over the next 
few years, VA should seek to minimize such transitory losses of 
production and quality through increased resources and personnel.

    If VA were simply focused on reducing the backlog quickly they 
could probably achieve some short term reductions through non-
technological means. However that would neither be sustainable nor 
conducive to creating a claims processing system that ensures veterans 
receive accurate and timely benefits.

    We also continue to have concerns about whether VA's focus on 
production could come at the expense of quality and accuracy, and how 
that is influencing the development of the VBMS. Given the enormous 
pressure on VA to reduce the backlog, we believe that an independent, 
outside expert review of the VBMS development program would provide 
Congress with greater certainty that VA is on the right path, as well 
as provide early warnings if changes to the VBMS program are required.

    Mr. Hall. Okay. I have more questions for all of you that 
we are going to submit in writing because of the possibility of 
votes being called, and we still have another panel to hear 
testimony from. We will send you some questions in writing in 
addition to the ones that have been asked already. But one last 
one maybe each of you could respond to. According to VA's 
assessment of accuracy we range from a high in Des Moines of 
92.34 percent accuracy to a low in Baltimore of 69.33 percent, 
and Washington lower than that, although we do not have a 
number. I am sorry to say that New York, my home State, is 
close to the bottom at 76 percent, according to VA's estimate 
which the OIG suggests may be actually optimistic by as much as 
ten percentage points. But I, maybe starting with Mr. Abrams 
you could tell what you suggest that the VA do to take, to 
learn from those who are getting the best results? Mr. Abrams, 
do you want to----
    Mr. Abrams. We have been to Des Moines to do a quality 
check, and outside of the fact that we could not get back to DC 
from there, which is a long story, and an ice storm somewhere 
involved, too, we did not find such a high quality of work. It 
was better than some offices, but not at 92 percent. In my 
experience, I do not really trust VA statistics except to note 
that they are generally worse than what are reported. I say 
that because I have seen that over 36 years.
    However, some offices in rural areas attract a higher level 
of worker. And we do not talk about that a lot, but working for 
the VA in New York with the kicker to get a little more money 
does not give you the same quality of life as working in Des 
Moines. So one of the problems is attracting a higher level of 
worker in the regional office. Also, a smaller office tends to 
be on top of things a little more. The worst office we found 
was St. Petersburg, the largest one. It was a disaster, in my 
view.
    Mr. Hall. Right. And they are about two-thirds of the way 
down the list here, but not all the way at the bottom.
    Mr. Abrams. Well, I am only looking at what we actually 
found. And since I did a lot of those, you know, cases I can 
talk to it personally.
    Mr. Hall. Sure. Mr. Wilson.
    Mr. Wilson. This is a response for the record on that as 
well, sir.
    [Mr. Wilson subsequently provided the following 
information:]

    Rather than focusing on the disparities in performance at various 
Regional Offices, we would emphasize that quality and accuracy must be 
prioritized in every location, as well as at level of leadership and 
management. While VA's STAR program provides a standard assessment of 
technical accuracy of claims work, it fails to provide true quality 
control at every stage of the process. Furthermore, the IG's review of 
the STAR program found that it failed to properly measure even 
technical accuracy, with about an additional 10 percent error rate 
going undetected.

    The only solution is for VA to foster a management culture that 
measures and rewards quality of results, not just quantity, and 
provides sufficient training of VA's management and workforce in order 
to achieve this outcome. Employees at all levels of the organization 
must be encouraged to provide constructive feedback at each step in the 
claims process to ensure that optimum performance is achieved. Creating 
and maintaining such an atmosphere will benefit employees, management 
and veterans. In the long run, the only way that VA will ever reduce 
and eliminate the backlog of overdue claims is by building a quality 
system that `gets it right the first time.'

    Mr. Hall. Thank you. Mr. Kelley.
    Mr. Kelley. I think if you look at the ROs there is a 
personality that comes out in each one of these ROs. And in 
that there is a culture. And we need to figure out how can we 
get everybody to look at things the same way. Localized 
training on certain issues that have been identified as a true 
problem for that one area need to be identified, but there 
needs to be national, standardized training throughout to 
ensure that everybody is thinking from the same book and 
everybody is moving forward with the same thought in mind.
    Mr. Hall. Thank you. Mr. de Planque.
    Mr. De Planque. Well the simple answer that you would want 
to do in a case like that is figure out a way to clone Iowa, if 
that is your top performing thing, and replicate it everywhere 
else. But as Mr. Abrams pointed out, you do not necessarily 
have the same base, the same resources, the same workload at 
every regional office. And so there are going to be 
fluctuations.
    We agree completely on what you have to address in the pool 
of workers that you can attract with the salaries that are 
there for VA. Obviously, any employment is great in a difficult 
economy. But people want to be competitive, and they want to be 
in a position to do that. And in some areas where the cost of 
living is much higher it is less attractive and so that is 
difficult. And we do not talk about that a lot, but it is 
something that is there, and it is there to consider.
    The nationally standardizing training and having good, 
solid means of national analysis to see, ``Hey, how do we 
replicate the Iowa results in other offices?'' Is there a way 
to condense that, put it into a training package and get that 
to everyplace else? That is what you are dealing with. But also 
in each regional office there is different staff, there is 
different management. They are all trying to do the same task, 
but you have a different climate in each one. The accuracy 
rates that we have seen have been fairly consistently, when the 
American Legion does quality reviews, have been fairly 
consistently in the 25 percent to 33 percent range of error, 
which is, and that is an across the board sort of average. 
Which is below obviously what the STAR averages are finding, 
and if you look across the board, then something like 92 
percent may be optimistic, although you should still consider 
that that is better than what a number of others are doing.
    So in addition to finding out where you are seeing the 
deficiencies are, also look for where you are seeing the 
strengths are and use that in your training programs to, again, 
like we said, create targeted training. Create training that is 
targeting what you know is both positive and negative about 
your organization and to correct those errors. Play to your 
strengths, and beef up your weaknesses.
    Mr. Hall. Thank you, Mr. de Planque, and Mr. Kelley, Mr. 
Wilson, Mr. Abrams, for your written and oral testimony, and 
for your patience while we wait for votes to happen today. It 
is always good to hear from you and to see you. So I thank you 
for your service to our veterans and to our country. And the 
second panel may now be excused.
    And we will ask our third panel to join us. Bradley G. 
Mayes, the Director of Compensation and Pension Service for the 
Veterans Benefits Administration, U.S. Department of Veterans 
Affairs; accompanied by Edna MacDonald, Acting Deputy Director 
of Compensation and Pension Service of VBA; and Terence Meehan, 
Director of Employee Development and Training of the VBA. Thank 
you all for being here. Thank you for your patience. As you 
know, your full written statement is already made a part of the 
hearing record so feel free to abridge it or add to it, or 
whatever you would like to do in 5 minutes.
    Mr. Mayes.

   STATEMENT OF BRADLEY G. MAYES, DIRECTOR, COMPENSATION AND 
    PENSION SERVICE, VETERANS BENEFITS ADMINISTRATION, U.S. 
DEPARTMENT OF VETERANS AFFAIRS; ACCOMPANIED BY EDNA MACDONALD, 
 ACTING DEPUTY DIRECTOR, COMPENSATION AND PENSION SERVICE FOR 
 POLICY AND PROCEDURES, VETERANS BENEFITS ADMINISTRATION, U.S. 
DEPARTMENT OF VETERANS AFFAIRS; AND TERENCE MEEHAN, DIRECTOR OF 
     EMPLOYEE DEVELOPMENT AND TRAINING, VETERANS BENEFITS 
      ADMINISTRATION, U.S. DEPARTMENT OF VETERANS AFFAIRS

    Mr. Mayes. Mr. Chairman, thank you for providing me this 
opportunity to discuss the Veterans Benefits Administration's 
quality assurance program and the positive effect it has on the 
processing of veterans disability claims. And joining me today, 
as you said, are Edna MacDonald to my right, Acting Deputy 
Director of Compensation and Pension Service for Policy and 
Procedures; and to my left, Terence Meehan, Director of VBA 
Employee Development and Training.
    Mr. Chairman, I have prepared remarks. But let me just say 
before I get into that, that I agree completely with virtually 
every panelist that has been up here, that quality must be at 
the core of our philosophy for serving veterans. There is no 
question about that.
    The Subcommittee has indicated a special interest in our 
Systematic Technical Accuracy Review program. However, I want 
to emphasize that STAR is only one of four tiers of our 
multifaceted national quality assurance program. The STAR 
component focuses on claims processing accuracy, while the 
other three components address regional office oversight, 
rating consistency, and special focus reviews. Along with the 
STAR program, these components collaborate to ensure high 
quality and consistent decisions for veterans.
    STAR is the quality assurance component that focuses on 
claims processing accuracy. STAR reviews are focused on 
outcomes for veterans rather than specific processes by which 
the outcomes are reached. STAR reviews evaluate the quality of 
the rating decision product that VBA provides for veterans. It 
is from the veteran's perspective that there is an expectation 
that we understand the claim, evaluate it accurately and 
fairly, and provide proper compensation under the law.
    The purpose of STAR reviews is to ensure that rating 
decision outcomes meet these expectations. The STAR program 
incorporates review of three types of work, claims that usually 
require a rating decision, authorization work that does not 
usually require a rating decision, and fiduciary work. A 
determination of benefit entitlement accuracy for rating and 
authorization work utilizes a structured checklist to ensure 
all issues were addressed, claims assistance was provided, and 
the decision was correct. Accuracy results are calculated on 
the results of the benefit entitlement reviews. STAR findings 
provide statistically valid accuracy results at both the 
regional office and the national level.
    VBA continues to focus on expanding and improving the STAR 
program. In 2008, the STAR staff was consolidated to Nashville, 
which provided space for expansion and allowed aggressive 
recruitment for more staff. Since then, STAR has completed an 
extensive expansion effort, more than doubling the staff and 
increasing the sample size to obtain a statistically valid 
sample at the regional office level. In 2010, quality review of 
fiduciary cases was also transferred to Nashville.
    During fiscal year 2009, a little over 20,000 cases were 
reviewed under the program. The targeted number of cases for 
review in 2010 is 37,932. The STAR sample was expanded in 2009 
to include a review of brokered work completed by VBA's 
resource centers and Tiger Team. Sampling was increased for the 
pension management centers to allow measurement of pension 
entitlement decisions. Ongoing reviews of the disability 
evaluation system cases, and Appeals Management Center cases 
became part of the monthly compensation quality sample in 
fiscal year 2009.
    The VBA quality assurance program has undergone significant 
change over the past several years, and has become more 
comprehensive by expanding the type and breadth of cases 
reviewed. But we acknowledge this is a journey that we are on, 
and we look forward to working with our stakeholders to 
continuously improve our quality assurance program.
    Mr. Chairman, thank you for the opportunity to provide you 
an update on our accomplishments at this hearing. And with 
that, I would be glad to answer any questions that you may 
have.
    [The prepared statement of Mr. Mayes appears on p. 51.]
    Mr. Hall. Thank you, Mr. Mayes. I appreciate the challenge 
that you face, and the need to evaluate things even as numbers 
are growing, numbers of staff are growing, numbers of claims 
being brought in by the new diseases that are service-connected 
automatically, not to mention the returning Operation Iraqi 
Freedom/Operation Enduring Freedom (OIF/OEF) soldiers and 
veterans, and the aging of our older veterans population. But I 
am curious just, you know, for instance what action you have, 
or have taken, or are considering taking regarding for instance 
Baltimore or Washington given the disparity between, you know, 
their accuracy rate under your evaluation and the more 
successful. I mean, I would just guess that maybe Nashville is 
rated near the top because STAR is based there? Is that a 
correct assumption? And, but what, is there something underway 
already to try to bring Baltimore and Washington up.
    Mr. Mayes. Well first, I will tackle that last one. I do 
not think it hurts that we have our quality assurance operation 
in Nashville, although it is not actually in the regional 
office. But in Nashville you had a director who I understand is 
going to be retiring soon, Brian Corley, who came up through 
the system understanding the C&P program and the adjudication 
process, and was very dedicated, I know personally, in training 
and focusing on developing his workforce. He knew the value of 
developing that workforce because that in the end would reap a 
long-term reward. While you may have to pay the price early on 
to train these employees, down the road they would become more 
effective. He preached that, and he enforced that.
    Now, I think one of the earlier panelists mentioned that 
this has to be a philosophy, and I echoed that. I am not 
suggesting that in Baltimore they do not have that philosophy 
or they do. What I am saying, though, is what the service has 
to do, is help shine the light on areas where there are 
deficiencies. STAR helps do that. With STAR, you have that 
array of the quality by regional offices. But we also do site 
surveys. We have gone into Baltimore on a site visit, and we 
have highlighted some areas that they need to focus their 
attention on. We work closely with our operations element 
through our Office of Field Operations. We share this 
information with them. And then the expectation is that the 
management team will execute on these recommendations. That is 
how this process works. We have provided that information, and 
I believe that they have set about executing a plan for 
addressing these deficiencies that we have identified. It has 
resulted, I believe, in a higher quality rating.
    Mr. Hall. Does the VA keep records of how performances 
changed or improved with new leadership at the ROs? For 
instance, in New York, I know, as it serves my constituents in 
the Hudson Valley and other New York regional veterans, 
recently had a new director appointed. And I am curious if, 
well currently in this ranking has the fifth lowest accuracy 
rating. Is that changing? Have you had time yet to ascertain 
whether the quality, the accuracy rate has started to improve 
since the new director has taken over there.
    Mr. Mayes. Mr. Chairman, I want to be clear. I am beginning 
to get out of my lane a little bit. The way we are organized, 
with me being the head of the policy component in VBA, our 
program goes in and takes a look and identifies these gaps, and 
then we work closely with that operation's component to then 
come back and rectify those deficiencies.
    My understanding is the Director just reported there. So I 
think it might be too early for us to tell whether some of the 
countermeasures that she has put in place are beginning to reap 
some benefit.
    Mr. Hall. Do you have accuracy ratings for the previous 
year, 2008?
    Mr. Mayes. We do. It is a 12-month cumulative accuracy 
score, so it is looking back at the 12-month cumulative number. 
We have those numbers going back historically, yes.
    Mr. Hall. Thank you, Mr. Mayes. Could you provide the 
Committee, please, with any changes in the directors, the 
different ROs within the last 2 years? And any correlation that 
you see in terms of the accuracy before and after the new 
management took their office? That is something, just homework 
for you.
    [The VA subsequently provided the information in the answer 
to Question 5 of the Post-Hearing Questions and Responses for 
the Record, which appears on p. 58.]
    Mr. Mayes. All right, we can provide that.
    Mr. Hall. I think it is worth looking at. Would you address 
the question that the previous panel was talking about, in 
terms of incentives and disincentives, to try to encourage the 
higher accuracy performers and maybe disincentivize those who 
are not producing the same rate of accuracy?
    Mr. Mayes. Well one thing I would say is that accuracy is 
in the performance plan for every individual that is involved 
in the claims process, from the director right down to the 
claims examiner who is developing the claim, to the RVSR who is 
rating the claim. Every one of those decision makers, and the 
managers in that process, every one of those employees has a 
critical component related to accuracy. So the assertion that 
we somehow do not value accuracy is not borne out by how we 
measure individuals' performance. It is production as well, but 
there is equal weight in the performance plan for all of those 
employees for accuracy.
    Mr. Hall. I appreciate that. And I note that it is a 
complicated thing that is not explained by any one factor. For 
instance, you do have areas where it may be easier to attract, 
due to the cost of living and the standard of living one can 
achieve on a given salary. Just as with air traffic 
controllers, or with doctors in the private medical field and, 
you know, other areas. There is obviously in Des Moines, as was 
pointed out before, a chance to do better on a given salary 
than in New York or Washington.
    But there has been conversation in the veterans community, 
in the public press, and here in Congress and in this Committee 
about the evaluations, year end evaluations and bonuses that 
are given to people at different, in different positions, at 
VBA. And, you know, it seems to me that that is something that 
I would look at if I were going to try to get some teeth into 
this quickly. We need to try to motivate people to try to get 
up to where Des Moines is. In fact, we need to get more Des 
Moines here. So are you looking at or have you already made any 
changes in the way bonuses are set?
    Mr. Mayes. Well Mr. Chairman, I am making the assumption 
that you are talking about the senior leadership level. I am 
not involved in the process by which the bonuses are 
determined.
    Mr. Hall. But I believe it goes down below that, also. Are 
not bonuses given out to career VBA employees?
    Mr. Mayes. Yes. We can certainly give performance awards to 
employees, up and down the line, across the spectrum of 
employees. I was saying, quality, even for the journey-level 
employee that is processing this work, is a core component of 
their performance plan. I know when I was in Cleveland as the 
Director, or previously in Milwaukee, that if somebody was 
failing to achieve their performance plan target for quality, 
then that counted against them in the mix with respect to who 
was going to get performance awards, because we valued quality.
    Mr. Hall. So if we looked at offices that are not 
performing as well we would see lower year end bonuses or 
performance awards?
    Mr. Mayes. Mr. Chairman, because I have not been involved 
in that process. That is one I have to take back for the 
record.
    [The VA subsequently provided the information in the answer 
to Question 6 of the Post-Hearing Questions and Responses for 
the Record, which appears on p. 58.]
    Mr. Hall. Okay. We will try to find out, and if you could 
find out for us that would be good. Is it safe to conclude 
since Atlanta, Chicago, New York, L.A., and Baltimore are all 
near the bottom of the list that accuracy levels are driven 
down by volume?
    Mr. Mayes. I think that Mr. Abrams made some points about 
smaller stations. It is easier to control your workload. I 
think that is a valid observation. But there is evidence that 
larger stations can successfully manage that workload. It can 
be more challenging because you have a lot more work coming 
through there; you have a lot more employees that you have to 
manage. Even though the span of control at the first line 
supervisor level may be the same, there certainly are some 
complexities when you have a large operation. Edna, do you want 
to add to that?
    Ms. MacDonald. I believe that some of the other panels 
already talked about the challenges in hiring and retaining a 
good workforce. As you point out, most of those are 
significantly urban cities. We do recognize that is a challenge 
for us to compete and to retain a workforce. Not that we cannot 
do it, but it is sometimes easier in your smaller stations like 
Muskogee, which is a very large office for us, but we are able 
to retain a very strong workforce.
    Mr. Mayes. That is precisely why we built out our quality 
assurance operation in Nashville because we were having a 
challenge in recruiting and retaining employees at that grade 
level here in the Washington, DC, area. I know the OIG report 
mentioned our site survey staff and the difficulty that we have 
had in getting employees to come in and stay and do that work. 
We have made the decision we are going to move that part of our 
operation to Nashville as well because we are achieving more 
success there.
    Mr. Hall. And when is that going to happen?
    Mr. Mayes. We have an announcement on the street right now 
for employees for part of that operation in Nashville. So as 
soon as we can through the hiring process, we will begin that 
transition, just like we did for the reviewers. We moved that 
from Washington to Nashville.
    Mr. Hall. Could you please provide us with, if you have it, 
the incentives or inducements that the Department is offering 
to hire people in the cities that you are having a harder time 
retaining or hiring staff in? Just so we have some idea of what 
the range is. As I said, there are other branches of government 
that are doing the same thing for the same reasons. And it 
would be helpful for us to know that.
    [The VA subsequently provided the following information:]

    VBA offers relocation incentive bonuses on a case-by-case basis for 
management positions in geographic areas where recruitment and 
retention of a skilled and qualified workforce are challenging. These 
challenges historically occur in areas where:

          Employers outside of the Federal Government pay 
        higher salaries for similar positions;
          Local candidates do not possess the unique 
        competencies required for VBA positions; or,
          The geographic location, position, or duties are 
        undesirable to qualified candidates.

    All relocation incentive bonus requests are pre-approved by the 
Associate Deputy Under Secretary for Field Operations and the Under 
Secretary for Benefits prior to announcing the position. Relocation 
incentive amounts typically range from $5,000 to $15,000 depending on 
the position, location, and past recruitment efforts.

    OIG reports that the STAR quality control system was 
hindered by VBA's exclusion of brokered claims from STAR 
reviews. Specifically, based on OIG's 2009 review it found a 69 
percent accuracy rate. While VBA initially, as I understand it, 
resisted the idea of reviewing brokered claims there is an 
indication that recently you have embraced this recommendation 
and begun to review those claims. What impact do you believe a 
review of brokered claims will have on the overall accuracy 
rate?
    Mr. Mayes. Well, we are reporting it now. We started with 
cases worked in April. Right now the national accuracy rate for 
brokered work is 77 percent. For the Tiger Team it is 78 
percent. When you factor that in to the national numbers, I 
think it stays the same at 83 percent. Is that correct, Edna? 
Right now given the sample sizes, it is not impacting the 
overall nationwide accuracy rate.
    But I think the fact that we are looking at this work now 
is important, because it means that folks that are promulgating 
decisions out of our resource centers know that there is the 
possibility that you are going to have your case selected for 
review at that national level.
    Mr. Hall. How do you explain the difference between STAR 
comments on a claim and issues identified as errors? What can 
you do, or what are you doing, to ensure that all issues are 
correctly identified as errors or comments.
    Mr. Mayes. Well, when we select our sample, we are sampling 
claims that have been completed over a 1-month period previous 
to the month that we are doing the reviews. The sample is 
driven by the end product control that is associated with the 
work that was done. The way we set our system up, we are doing 
the review for the claim that was at issue associated with that 
control. But sometimes during the course of a review we might 
find where we missed something, but it is not associated with 
the claim that is subject to the review. So we will annotate 
that as a comment, rather than call an error on the claim that 
is subject to the review in this particular review. So what the 
OIG commented on was that for error calls where we are saying, 
``You made a benefit entitlement error on that particular 
claim,'' we were requiring a second level review. But in those 
reviews where there was no error associated with that claim, 
but we noted a comment, we were not requiring a second level 
review. We concurred with the OIG said that we need to have the 
second level review not just on those cases where we call a 
benefit entitlement error but also in those situations where we 
found something in the record that warranted a comment. We have 
implemented that. So that was something that we agreed with the 
Inspector General on.
    Mr. Hall. Thank you, Mr. Mayes. Thank you, Ms. MacDonald 
and Mr. Meehan. I am grateful to you for the work you are doing 
and for coming here again, yet again, to tell us about it. I 
will also ask you to put yourselves in the shoes of those 
veterans in Washington and Baltimore, New York, and in areas 
served by the other RO regions which according to the OIG, have 
quality levels maybe as much as 10 percent below what your 
numbers show. And I realize you are dealing with a very 
complicated situation. But each veteran is also dealing with a 
very complicated situation that they may not have the resources 
to deal with. So I know you understand the seriousness of this, 
and the Committee does also.
    You and all the witnesses will have, and Members of the 
Committee will have, 5 days to extend or revise remarks. And 
thank you to all of our panels. The hearing is now adjourned.
    [Whereupon, at 4:55 p.m., the Subcommittee was adjourned.]
















                            A P P E N D I X

                              ----------                              


    Prepared Statement Hon. John J. Hall, Chairman, Subcommittee on
               Disability Assistance and Memorial Affairs
    Good Afternoon.
    Would everyone please rise for the Pledge of Allegiance?
    Flags are located at the front and back of the room.
    Ladies and gentlemen, I welcome you all to the halls of Congress in 
what has been a profoundly historic and important week for the nation 
and for our veterans. Over the past seven days the Full Committee 
convened a successful Claims Summit which brought together dozens of 
the nation's top veterans' stakeholders including VA officials, GAO 
representatives, Veterans Service Organizations (VSOs), veterans' 
advocates, and academia and even industry leaders in this area. The big 
news from the Summit is that VA admitted what many of us have known for 
a while, the system for processing compensation and pension (C&P) 
claims is broken beyond repair, and must be replaced. We welcome this 
acknowledgement and look forward to working with VA in developing this 
new system.
    Of course, on a rare Sunday session, Congress passed, and 
yesterday, the President signed part one of the sweeping package to 
reform the nation's health care system. I for one look forward to the 
second part of health reform reaching the President's desk.
    Finally, this week, we have passed a number of bills in the full 
House that will significantly help our veterans. These include the End 
Veteran Homelessness Act of 2010 to provide $200 million in support of 
Sec. Shinseki's goal of ending homelessness for America's warriors. We 
also passed the Helping Heroes Keep Their Homes Act of 2009, sponsored 
by our colleague Tom Perriello, which would protect home mortgages of 
veterans and servicemembers. We also passed the disability compensation 
COLA bill, that would allow a Cost of living increase by the end of the 
year, also sponsored by Mr. Perriello. The National Guard Employment 
Protection Act of 2010 which aims to help preserve jobs of soldiers 
ordered to full time deployments with the National Guard is set to be 
voted on today. And I remain committed to helping Chairman Filner push 
legislation to help our veterans who were harmed by Agent Orange 
exposure.
    In this afternoon's hearing entitled: Examination of VA Regional 
Office Disability Quality Review Methods: Is VBA's Systematic Technical 
Review System (STAR) Making the Grade?, we will examine the primary 
quality review tool employed by the Department of Veteran Affairs (VA) 
to assess the accuracy of disability compensation and pension claims 
processing. The STAR system can help VBA monitor the quality and 
accuracy of ratings at its regional offices (ROs). Through this quality 
control tool, VBA can focus attention on poorly performing ROs and help 
the agency direct additional staff and training to problem offices. At 
the end of the day the goal is for VA to get the claim right the first 
time so that veterans are not improperly denied the benefits they 
deserve or faced with lengthy appeals cycles.
    The STAR system was implemented by VBA in October 1998 to improve 
the measurement of the accuracy of claims processing. Since FY 2007, 
VBA has set for itself a performance goal of completing compensation 
claim ratings without error, 90% of the time. Its long term strategic 
accuracy goal is 98%. Unfortunately, the VA is far from achieving this 
goal. Until the STAR system provides an accurate accounting of the 
error rate at the VBA, it is difficult to envision a path for meeting 
this goal.
    VA's Office of Inspector General (OIG) and the Government 
Accountability Office (GAO) produced studies which revealed several 
issues that impede the efficiency and consistency of the STAR system. 
Specifically, OIG suggests that any claims ratings accuracy numbers 
reported by VA should be discounted by as much as 10% to yield an 
accurate performance measurement. I'm also personally troubled by GAO's 
finding that VBA claim processing accuracy rate is particularly low in 
cases filed by veterans seeking benefits based upon Post Traumatic 
Stress Disorder (PTSD) and/or Traumatic Brain Injury (TBI). Today, we 
intend to analyze these studies through hearing testimony from 
representatives of VA's OIG and GAO.
    VBA has made some improvements to its claims rating system since 
the implementation of STAR system. We look forward to hearing from them 
about these improvements and how we can help them with any areas of 
concern.
    To fully understand the STAR quality assurance program, it's 
important to review the C&P rating system itself. Through its 
disability compensation program, VBA pays monthly benefits to veterans 
for injuries or diseases incurred or aggravated while on active 
military duty. To access a claim, the application for disability 
benefits must be ``developed,'' a process that involves obtaining all 
necessary evidence of the veteran's service and disability to support 
the claim. After development, claims go to a Rating Veterans Service 
Representative (RVSR) for a decision. RVSRs determine if a veteran's 
disability is service connected and assigns a percentage rating 
(ranging from 0 to 100 percent) that is intended to represent the 
average earning reduction a veteran with that condition would 
experience in civilian occupations. The veteran is then notified of the 
decision.
    For reopened claims, the veteran's previously assigned diagnostic 
codes with evaluations also affect the veteran's overall combined 
percentage of disability. Regional offices use previously assigned 
diagnostic codes, along with their percentages, in combination with 
current assigned diagnostic code percentages to determine the combined 
level of disability. Once a claim is completed, the result of that 
claim or ``end product'' is cleared, and work credit is given to the 
regional office.
    A completed claim and corresponding cleared end product is then 
subject to review by STAR reviewers based on a statistical sample of 
all completed rating end products.
    In the 110th Congress, I introduced H.R. 5892, which outlined a 
series of steps to improve the quality of VA's claims processing 
system. These recommendations formed the core of directives which were 
codified by P.L. 110-389, the Veterans Disability Benefits Claims 
Modernization Act of 2008. Today's hearing also provide us a chance to 
gauge how these quality control measures have been implemented by VA.
    With that, I look forward to the informed testimony of our 
witnesses and insightful comments and questions from my colleagues on 
the Subcommittee.
    I now recognize Ranking Member Lamborn for his Opening Statement.

                                 
  Prepared Statement of Hon. Doug Lamborn, Ranking Republican Member,
       Subcommittee on Disability Assistance and Memorial Affairs

    Thank you Mr. Chairman,
    And welcome everyone, to this hearing on the Department of Veterans 
Affairs STAR program.
    STAR is an acronym for Systematic Technical Accuracy Review, which 
is VA's program for quality assurance.
    Throughout my tenure on this Committee, my fellow members and I 
have called for stronger accountability within the VA claims system.
    For too long, the primary focus has been on production and this has 
led to an error rate that is unacceptable.
    I believe that the VA's greatest challenge, the claims backlog, is 
largely attributable to hasty decisions made without proper regard for 
accuracy.
    The ramifications of this approach can be seen throughout the 
entire system.
    Therefore, VA employee performance awards cannot be based entirely 
on production, there must also be a valid measure of quality.
    Under the STAR program, a statistically valid sample of rating 
decisions from various regional offices is reviewed for accuracy.
    While this method may be useful from a macro perspective, it is not 
sufficient for ensuring individual accountability.
    VA must be able to identify employees in need of individualized 
remedial training.
    Without this essential component of the quality assurance process, 
VA will have perpetual problems in its claims system.
    I also have other concerns based on the written statements of 
today's participants.
    There seems to be some contradiction as to whether brokered claims 
are excluded from STAR review, or whether STAR requires follow-up to 
ensure that VA Regional Offices provide corrective action or remedial 
training.
    Perhaps these were former deficiencies that have since been 
corrected.
    I am encouraged by GAO's report that VA has made a number of 
improvements, but I also agree that challenges remain.
    In the 110th Congress this committee passed a provision that was 
included in the Veterans' Benefits Improvement Act of 2008 that 
required VA to conduct a study on the effectiveness of the current 
employee work-credit system.
    I believe the upcoming report, along with the testimony that we 
will hear today, will provide valuable feedback for the Department to 
improve its quality assurance and accountability program.
    I look forward to hearing from our witnesses today, and I thank you 
all for your participation
    Thank you, I yield back.

                                 
       Prepared Statement of Belinda J. Finn, Assistant Inspector
    General for Audits and Evaluations, Office of Inspector General,
                  U.S. Department of Veterans Affairs

INTRODUCTION

    Mr. Chairman and Members of the Subcommittee, thank you for the 
opportunity to discuss the Office of Inspector General's (OIG) work 
regarding the Veterans Benefits Administration's Systematic Technical 
Accuracy Review (STAR) program. I am accompanied by Mr. Larry 
Reinkemeyer, Director of the OIG's Kansas City Audit Operations 
Division.
    The OIG is committed to proactively reviewing the effectiveness of 
key management controls to assure the accomplishment of mission 
critical service responsibilities to veterans, such as the delivery of 
accurate and timely disability benefits. Further, we strive to focus 
our efforts on identifying control weaknesses before they escalate into 
significant problems. Over the past 2 years, we issued audit reports 
covering aspects of VBA's quality assurance process--the STAR, Rating 
Consistency Review, and Site Visit programs. In March 2009, we issued 
the Audit of Veterans Benefits Administration Compensation Rating 
Accuracy and Consistency Reviews (Report No. 08-02073-96). In May 2009, 
we issued the Audit of Veterans Benefits Administration Compensation 
and Pension Site Visit Program (Report No. 08-02436-126).
    Also in fiscal year (FY) 2009, we established a Benefits Inspection 
Division to provide recurring oversight of regional offices by focusing 
on disability compensation claims processing and performance of 
Veterans Service Center (VSC) operations.

STAR PROGRAM

    Improving the quality of rating decisions, which includes accuracy 
and consistency of disability compensation claim rating decisions, is 
among VBA's highest priorities. The STAR program is a key mechanism for 
evaluating regional office performance in processing accurate benefit 
claims for veterans and beneficiaries. The STAR program provides a 
comprehensive review and analysis of compensation rating processing 
associated with specific claims or issues. VBA's FY 2008 and FY 2009 
goal for compensation claim ratings was a 90 percent accuracy rate. 
STAR reviewers select a number of claims to review from VA regional 
offices nationwide and use a checklist designed to facilitate a 
consistent, structured review and classify errors into three 
categories: benefit entitlement errors, decision documentation/
notification errors, and administrative errors.

Results

    During the 12-month period ending February 2008, STAR reviewers 
found that VBA staff accurately rated about 87 percent of the claims 
(approximately 767,000 claims of 882,000 claims reviewed). We reviewed 
a random sample of STAR reviewed claims and found additional errors 
affecting veterans' benefits that STAR reviewers did not identify. As a 
result, we projected VBA's accuracy rate for claims reviewed was only 
78 percent.
    The STAR program does not provide a complete assessment of 
compensation claim rating accuracy. We found that VBA officials 
excluded brokered claims from STAR reviews. Brokered claims are those 
assigned to one regional office but sent to another office to be rated. 
When we combined the results of our review of brokered claims with our 
review of STAR reviewed claims, we projected that about 77 percent 
(approximately 679,000 claims) of compensation claims were accurate and 
VBA's reported error rate was understated by approximately 10 
percentage points. This equates to approximately 88,000 additional, or 
about 203,000 total, claims where veterans' monthly benefits may be 
incorrect. This difference occurred because STAR reviewers did not 
identify all errors, and VBA officials excluded a significant number of 
compensation claim ratings from review. Additionally, VBA officials had 
not implemented an effective formal training program for the STAR 
reviewers.
    We identified five areas in our review of the STAR program where 
VBA needed to take action to improve its quality assurance and 
oversight efforts.

          STAR reviewers did not identify some errors because 
        they either did not thoroughly review available medical and 
        non-medical evidence or identify the absence of necessary 
        medical information. STAR reviewers also misclassified some 
        errors in a way that resulted in the error not being counted 
        against the regional office's accuracy rate. In these cases, 
        they recorded a ``comment'' instead of an ``error'' although 
        the errors clearly affected veterans' benefits entitlements and 
        should have been counted as errors.
          STAR management required regional offices to report 
        quarterly on actions taken to correct benefit entitlement 
        errors but they did not require or follow up to ensure regional 
        offices took corrective actions on comments made by STAR 
        reviewers. Ultimately, they relied on the regional office to 
        take corrective action on all issues identified whether the 
        STAR reviewer identified an error or a comment. From our 
        sample, we identified 33 compensation claim ratings where STAR 
        reviewers made comments instead of reporting issues as errors. 
        At least six of the comments related to issues that could 
        affect the veterans' benefits. We found that regional office 
        staff had not corrected any of the six comments potentially 
        affecting the veteran's benefits.
          VBA officials excluded brokered claims from STAR 
        reviews. VBA officials told us STAR reviewers do not review 
        brokered claims because the STAR program's primary focus is on 
        assessing and reporting rating accuracy for each of the 
        regional offices. Since two or more regional offices are 
        involved in brokered work, VBA officials stated it would be 
        difficult to assign responsibility for the rating accuracy to 
        one specific regional office. Thus we found that STAR 
        management was replacing brokered claims selected for review 
        with non-brokered claims. We reviewed a sample of brokered 
        claims that were not evaluated and found a 69 percent accuracy 
        rate.
          The STAR reviewers did not ensure regional offices 
        submitted all of the selected compensation claim ratings to the 
        STAR program for review. Regional offices did not submit about 
        600 (7 percent) of the approximately 9,000 requested claim 
        ratings for the 12-month period ending February 2008. We 
        reviewed 54 of the 600 pending requests and identified 12 (22 
        percent) benefit entitlement errors. STAR management relies on 
        regional office staff to submit the requested claims and only 
        follows up with the regional offices that do not submit any 
        requested claims for a given month. A STAR manager stated they 
        did not have sufficient resources to follow up on individual 
        claims that regional office staff do not submit. Therefore, 
        regional office staff can cherry pick claims because STAR 
        reviewers do not reconcile the claim requests. This control 
        weakness provided opportunities for regional office staff to 
        withhold claims if they suspect the claims to have errors.
          STAR reviewers are not required to complete formal 
        training on an annual basis. The reviewers met infrequently to 
        discuss issues, and had no set formal training schedule or 
        requirements. Regional office staffs that prepare and complete 
        ratings and awards for compensation claims are required to 
        achieve 80 hours of training per year to stay competent on 
        laws, policies, and processes on rating-related issues. 
        However, the STAR program manager stated their program workload 
        requirements currently do not allow for the amount of training 
        time necessary, yet agreed the program staff could benefit from 
        formal training.

    We recommended that VBA:

          Ensure STAR reviewers evaluate all documentation 
        related to the claim selected for review.
          Establish a requirement that all STAR reviewer 
        comments receive a second review to make sure the comment was 
        not a benefit entitlement error.
          Establish procedures to review brokered claims as 
        part of the STAR program.
          Enforce procedures requiring regional offices to 
        submit all requested claims to the STAR program office for 
        their review or submit written justification to the STAR 
        program's office requesting to exclude the claim from the 
        review.
          Establish minimum annual training requirements for 
        STAR reviewers that are comparable to regional office rating 
        staff training requirements.

    VBA agreed with all five recommendations and reported that it had 
completed actions to implement the recommendations in our March 2009 
report.

RATING CONSISTENCY PROGRAM

    To address variances in compensation rating decisions within 
individual VA regional offices and across the Nation, VBA developed a 
rating consistency review plan that included metrics to monitor rating 
consistency and a method to identify variances in compensation claim 
ratings. VBA's plan would identify unusual patterns of variance in 
claims, and then review selected variances to assess the level of 
decision consistency within and between regional offices. However, as 
of March 2009, VBA had not fully implemented its rating consistency 
review plan.

Results

    In FY 2008, VBA officials identified 61 diagnostic codes where 
offices appeared to be making inconsistent decisions in the evaluation 
of a granted claim or whether offices granted or denied the claim. VBA 
officials planned to conduct 22 of the 61 reviews in FY 2008 consisting 
of 20 grant/denial rate and 2 evaluation reviews. However, they only 
initiated two grant/denial rate reviews and did not complete either 
review until December 2008. Additionally, VBA did not initiate either 
of the evaluation reviews designed to reduce variances in compensation 
claim ratings. In March 2010, VBA informed us that insufficient 
staffing prevented them from completing any consistency reviews in FY 
2009. However, they have now hired the necessary employees. The first 
FY 2010 review was completed in January 2010; a second review started 
in February 2010.
    We identified three areas that impaired VBA's ability to fully 
implement its rating consistency review plan.

          VBA did not have an accurate universe of claims to 
        review. From April through May 2008, VBA officials encountered 
        a delay in completing planned consistency reviews because the 
        universe of claims included completed claims with diagnostic 
        codes outside the scope for the requested period. VBA officials 
        notified the Office of Performance, Accountability, and 
        Integrity of the data integrity issue and worked with 
        Compensation and Pension (C&P) Service staff to identify an 
        appropriate universe by June 2008.
          Data captured by STAR reviewers contained veterans' 
        personally identifiable information (PII). In July 2008, VBA 
        officials stopped all consistency reviews until they could take 
        action to secure PII in STAR's electronic records used to 
        capture and review data and analyze the results. VBA officials 
        took the necessary actions to correct this condition and 
        secured the database in December 2008.
          VBA officials did not assign a sufficient number of 
        staff to accomplish consistency reviews. The STAR program 
        office was authorized 26 reviewers for rating accuracy and 
        consistency reviews, and had 18 reviewers on board. However, 
        only 8 of the 18 reviewers conducted consistency reviews. The 
        eight reviewers tasked with completing reviews of consistency 
        were not sufficient to complete the assigned work. A STAR 
        manager estimated that approximately 12 reviewers were needed 
        to complete the 22 planned reviews. However, in addition to the 
        consistency reviews, STAR management also assigned these same 
        eight reviewers to conduct at least seven special focus reviews 
        involving thousands of compensation claim ratings. We concluded 
        that the assigned staffing was insufficient to complete this 
        planned work.

    As part of future rating consistency review plans, inter-rater 
reliability reviews (IRRR) should be included, along with the annual 
rating consistency reviews. In July 2008, C&P Service staff conducted 
an IRRR and found that 76 (31 percent) of the 246 participants 
incorrectly rated a relatively simple back strain claim. In August 
2008, C&P Service staff conducted another IRRR and found that 30 (13 
percent) of the 247 participants incorrectly rated a Post Traumatic 
Stress Disorder compensation claim. According to VBA officials, the 
most common errors were incorrect evaluations due to misinterpretation 
of the appropriate facts and criteria. C&P Service managers used the 
results of the IRRRs to plan focused training efforts, and they plan to 
conduct follow-up IRRRs to evaluate the effectiveness of the training. 
The IRRRs allow VBA officials to target a single rating issue to ensure 
the consistent application of policies and procedures nationally.
    We recommended VBA develop an annual rating consistency review 
schedule and complete all planned reviews as scheduled, dedicate 
sufficient staff to conduct consistency reviews in order to complete 
planned workload and reviews, and include inter-rater reliability 
reviews as a permanent component of their consistency review program. 
VBA agreed with these recommendations and reported that it had 
completed actions to implement the recommendations.

Site Visit Program

    The C&P Service's Site Visit program was established to ensure 
centralized oversight and provide technical assistance to VBA's 57 
regional offices. Site Visit teams monitor compliance with policies and 
procedures and identify best practices to assist in achieving high 
performance. This includes determining if regional offices are 
correcting errors identified by STAR teams.

Results

    The Site Visit program lacks an adequate infrastructure and 
management strategy to meet its mission and goals. We identified three 
areas that VBA needed to address to leverage the benefits of their Site 
Visit program.

          The Site Visit team experienced significant turnover 
        during FY 2006 and FY 2008 and has never been fully staffed to 
        the allotted eight full-time equivalent positions. Program 
        officials stated that they have been unable to maintain 
        adequate staffing of the Site Visit program because of 
        difficulties in recruiting from field offices qualified 
        candidates who are willing to relocate to the Washington, DC, 
        area. In addition, we found that C&P Service cannot ensure that 
        onsite evaluations are performed in compliance with generally 
        applicable governmental standards for independence or that 
        sufficient independence exists between the Site Visit program's 
        employees and VSCs reviewed.
          C&P Service did not review all 57 VSCs in any 3-year 
        period, and 7 (12 percent) of 57 VSCs were only visited once 
        from FY 2001 to FY 2008. Because the planned 3-year cycle of 
        review coverage has not been met, potentially low-performing 
        VSCs who could most benefit from a Site Visit program 
        evaluation may not be visited frequently enough. In addition, 
        C&P Service does not have formal policies and procedures to 
        ensure site survey protocols are modified to reflect emerging 
        C&P issues and systemic deficiencies identified during site 
        visits.
          C&P Service has not established procedures and 
        guidelines to identify and disseminate best practices. Also, 
        C&P Service has not developed reports that adequately develop 
        the causes of errors identified, and a follow-up process to 
        ensure that action items are resolved. In addition, C&P Service 
        does not adequately identify and report system-wide trends to 
        senior VBA managers, thus missing out on opportunities to 
        proactively address issues and concerns found during individual 
        site visits nationwide.

    We recommended C&P Service:

          Develop a staffing plan to ensure that sufficient 
        resources are made available to complete VSC reviews on a 3-
        year cycle.
          Comply with generally applicable government standards 
        for independence when performing site visits.
          Develop a procedure to continuously monitor and 
        update protocols to address systemic issues identified during 
        Site Visit reviews, management concerns and priorities, and 
        changes in program operations.
          Develop a process for the identification of best 
        practices and resolution of inconsistencies in the application 
        of policies and procedures.
          Develop and implement policies, procedures, and 
        performance measures to strengthen follow-up on corrective 
        action plans developed by regional offices on issues identified 
        during onsite evaluations.
          Ensure Site Visit reports issued for VSC operations 
        to more fully identify the root cause of issues affecting VSC 
        performance. VBA agreed with all six recommendations and 
        reported that it had completed actions to implement the 
        recommendations.

    VBA agreed with these recommendations and reported that it had 
completed actions to implement the recommendations.

OIG REGIONAL OFFICE INSPECTIONS RESULTS

    STAR management relies on the regional office managers to take 
corrective action on all issues identified by the STAR team. Since 
April 2009, the OIG's Benefits Inspection Division has issued eight 
reports that include a review of regional office procedures to ensure 
the accurate and timely correction of errors identified by the VBA's 
STAR program. We found that regional offices did not have formal 
procedures established to ensure employees took corrective actions on 
the identified errors and as a result, five of the eight regional 
offices had not corrected all of the errors identified by the STAR 
team. Our analysis of 145 errors identified by STAR found that regional 
office staff did not correct 40 (28 percent) of the errors. Further, 
regional office staff erroneously reported to STAR that 21 (53 percent) 
of those 40 errors were corrected, although no corrections were made.
    We will continue to review and report on regional offices' 
performance in correcting errors identified during STAR reviews in 
future OIG benefit inspections and to report on other issues affecting 
accuracy and timeliness of claims processing. We will also confirm 
during these inspections whether the actions taken by VBA to implement 
our recommendations to improve the C&P Site Visit and Rating 
Consistency programs were effective. Currently, we plan on conducting 
up to 18 inspections in FY 2010, which means that each regional office 
will be inspected on a 3-year basis. Once we have sufficient data, we 
plan to issue roll-up reports that identify trends affecting accuracy 
and timeliness and include recommendations for improvement across the 
VBA system.

CONCLUSION

    VBA is under tremendous pressure to process claims and reduce the 
growing backlog. Without an effective and reliable quality assurance 
program, VBA leadership cannot adequately monitor performance to make 
necessary program improvements and ensure veterans receive accurate and 
consistent ratings.
    Mr. Chairman, thank you for the opportunity to discuss these 
important issues. We would be pleased to answer any questions that you 
or other members of the Subcommittee may have.

                                 
            Prepared Statement of Daniel Bertoni, Director,
               Education, Workforce, and Income Security,
                 U.S. Government Accountability Office
    VETERANS' DISABILITY BENEFITS: VA Has Improved Its Programs for 
       Measuring Accuracy and Consistency, but Challenges Remain
                             GAO Highlights
Why GAO Did This Study

    For years, in addition to experiencing challenges in making 
disability claims decisions more quickly and reducing its claims 
backlog, the Department of Veterans Affairs (VA) has faced challenges 
in improving the accuracy and consistency of its decisions.
    GAO was asked to discuss issues surrounding VA's Systematic 
Technical Accuracy Review (STAR) program, a disability compensation and 
pension quality assurance program, and possible ways, if any, this 
program could be improved.
    This statement focuses on actions VA has taken; including those in 
response to past GAO recommendations, to (1) address identified 
weaknesses with STAR and (2) improve efforts to monitor the consistency 
of claims decisions. This statement is based on GAO's prior work, which 
examined several aspects of STAR, as well as VA's consistency review 
activities, and on updated information GAO obtained from VA on quality 
assurance issues that GAO and VA's Office of Inspector General (OIG) 
have identified. GAO also reviewed VA's OIG March 2009 report on STAR.
    GAO is not making any new recommendations.

What GAO Found

    Over the past several years, GAO has identified several 
deficiencies with the Veterans Benefit Administration's (VBA) STAR 
program, and although VBA has taken actions to address these issues, it 
continues to face challenges in improving claims accuracy. For example, 
GAO found that STAR reviewers lacked organizational independence, a 
basic internal control principle. In response to our finding, VA began 
utilizing organizationally independent reviewers that do not make 
claims decisions. GAO also found that sample sizes for pension claims 
were insufficient to provide assurance about decision accuracy. In 
response to GAO's recommendation, in fiscal year 2009, VA began 
increasing the number of pension claims decisions it reviews annually 
at each of its offices that process pension decisions. VA has also 
taken a number of other steps to address weaknesses that VA's OIG found 
in the STAR program, including (1) establishing minimum annual training 
requirements for reviewers and (2) requiring additional supervisory 
review of STAR reviewers' work. Although it has made or has started 
making these improvements, VBA remains challenged to improve its 
decision accuracy for disability compensation decisions, and it has not 
met its stated accuracy goal of 90 percent. VBA's performance has 
remained about the same over the past several fiscal years.
    In addition, VA has taken steps to address deficiencies that GAO 
and the VA's OIG have identified with consistency reviews--assessments 
of the extent to which individual raters make consistent decisions on 
the same claims. For example, in prior work, GAO reported that VA did 
not conduct systematic studies of impairments that it had identified as 
having potentially inconsistent decisions. In response to GAO's 
recommendation, in fiscal year 2008, VBA's quality assurance staff 
began conducting studies to monitor the extent to which veterans with 
similar disabilities receive consistent ratings across regional offices 
and individual raters. However, last year, VA's OIG reported that VA 
had not followed through on its plans to conduct such reviews. In 
response to this and other OIG findings and recommendations, VA took a 
number of actions, including developing an annual consistency review 
schedule and hiring additional quality assurance staff. However, VBA 
has only recently begun these programs to improve consistency, and it 
is too early to assess the effectiveness of their actions.

                               __________

    Mr. Chairman and Members of the Subcommittee:
    I am pleased to have the opportunity to comment on the Department 
of Veterans Affairs' (VA) efforts to improve the accuracy and 
consistency of its disability compensation and pension benefit 
decisions. As we and other organizations have reported over the last 
decade, VA's claims processing challenges are not limited to making 
decisions more quickly and reducing its claims backlog; but also 
includes improving the accuracy and consistency of its decisions. The 
number of veterans awaiting decisions could grow as servicemembers 
returning from ongoing conflicts and aging veterans submit claims. 
According to VA, about 35 percent of veterans from ongoing hostilities 
file claims. It is important not only that decisions be timely, but 
also accurate. Accurate initial claims decisions can help ensure that 
VA is paying cash disability benefits to those entitled to such 
benefits and also help prevent lengthy appeals. Meanwhile, consistent 
decisions help ensure that comparable medical conditions of veterans 
are rated the same, regardless of which VA regional benefits office 
processes the claim.
    You asked us to discuss issues surrounding VA's disability 
compensation and pension quality assurance programs; particularly, the 
Systematic Technical Accuracy Review (STAR) program. My statement 
focuses on STAR, which deals with accuracy, and two other VA quality 
assurance activities that focus on consistency.\1\ More specifically, 
my remarks will focus on actions VA has taken to (1) address 
deficiencies identified with STAR and (2) improve efforts to monitor 
the consistency of claim decisions. This statement is based on our 
prior work, which examined several aspects of STAR, as well as VA's 
consistency review programs, and on updated information we obtained 
from VA on quality assurance vulnerabilities that we and VA's Office of 
Inspector General (OIG) have identified. We also reviewed VA OIG's 
March 2009 report on STAR and consistency reviews.\2\ Our work was 
conducted in accordance with generally accepted government auditing 
standards. Those standards require that we plan and perform the audit 
to obtain sufficient, appropriate evidence to provide a reasonable 
basis for our findings and conclusions based on our audit objectives. 
We believe the evidence obtained provides a reasonable basis for our 
findings and conclusions based on our audit objectives.
---------------------------------------------------------------------------
    \1\ These are (1) reviews of consistency of claims decisions across 
VA's Veterans Benefits Administration, which is responsible for 
administering VA's disability compensation and pension programs, by 
type of disabling condition; and (2) inter-rater reliability reviews, 
which examine the consistency of raters when evaluating the same 
condition based on a comparable body of evidence.
    \2\ Office of Inspector General, Department of Veterans Affairs, 
Audit of Veterans Benefits Administration Compensation Accuracy and 
Consistency Reviews (Washington, D.C.: Mar. 12, 2009).
---------------------------------------------------------------------------
Background
    Through its disability compensation program, VA pays monthly 
benefits to veterans with service-connected disabilities.\3\ Under its 
disability pension program, VA pays monthly benefits to low-income 
veterans who have disabilities not related to their military service or 
are age 65 or older. VA also pays compensation to the survivors of 
certain veterans who had service-connected disabilities and of 
servicemembers who died while on active duty.
---------------------------------------------------------------------------
    \3\ The amount of disability compensation depends largely on the 
severity of the disability, which VA measures in 10 percent increments 
on a scale of 0 percent to 100 percent. In 2010, basic monthly payments 
for veterans range from $123 for 10 percent disability to $2,673 for 
100 percent disability.
---------------------------------------------------------------------------
    Veterans and their survivors claim benefits at one of the Veterans 
Benefits Administration's (VBA) 57 regional offices. Once the claim is 
received, a service representative assists the veteran in gathering the 
relevant evidence to evaluate the claim. Such evidence includes the 
veteran's military service records, medical examinations, and treatment 
records from VA medical facilities and private medical service 
providers. Also, if necessary for reaching a decision on a claim, the 
regional office arranges for the veteran to receive a medical 
examination. Once all necessary evidence has been collected, a rating 
specialist evaluates the claim and determines whether the claimant is 
eligible for benefits. If so, the rating specialist assigns a 
percentage rating. Veterans with multiple disabilities receive a single 
composite rating. Since 2001, VBA has created 15 resource centers that 
are staffed exclusively to process claims or appeals from backlogged 
regional offices. Most of these centers focus either on making rating 
decisions, or on developing the information needed to evaluate claims.
    In addition to the traditional claims process, any member of the 
armed forces who has seen active duty--including those in the National 
Guard or Reserves--is eligible to apply for VA disability benefits 
prior to leaving military service through VA's Benefits Delivery at 
Discharge (BDD) program or the related Quick Start program.\4\ In 2006, 
VA completed its consolidation of BDD rating activity into its Salt 
Lake City, Utah, and Winston-Salem, North Carolina, regional offices, 
to increase the consistency of BDD claims decisions. Also, under the 
Department of Defense (DoD)--VA disability evaluation system pilot 
program, servicemembers undergoing disability evaluations, if found 
medically unfit for duty, receive VA disability ratings. This rating 
covers both the unfitting conditions identified by the military service 
and conditions identified by the servicemember during the process. The 
rating is used by both DoD and VA to determine entitlement for 
disability benefits.\5\
---------------------------------------------------------------------------
    \4\ In order to be eligible for the BDD program, servicemembers 
must meet several requirements, which include filing a VA claim 60 to 
180 days prior to an honorable discharge and completing a medical 
examination. Under BDD, the examination also serves as Department of 
Defense's separation physical examination. Quick Start is for those 
servicemembers--primarily members of the National Guard and Reserve--
who cannot meet the BDD time frame.
    \5\ For our review of the DoD--VA disability evaluation system 
pilot program, see GAO, Military Disability System: Increased Supports 
for Servicemembers and Better Pilot Planning Could Improve the 
Disability Evaluation Process, GAO-08-1137 (Washington, D.C.: Sept. 24, 
2008).
---------------------------------------------------------------------------
    Enacted in October 2008, the Veterans' Benefits Improvement Act of 
2008 required VA to contract for an independent, 3-year review of VBA's 
quality assurance program.\6\ This review is to include, among other 
items, assessments of the accuracy of disability ratings and their 
consistency across VA regional offices. VA contracted with the 
Institute for Defense Analyses (IDA) to conduct this study. According 
to VA, IDA will provide preliminary findings in the Summer of 2010, and 
VA is scheduled to report to the Congress in October 2011.
---------------------------------------------------------------------------
    \6\ Pub. L. No. 110-389, Sec. 224; 38 U.S.C. Sec. 7731(c).

---------------------------------------------------------------------------
STAR Program

    Under the STAR program, which was implemented in fiscal year 1999, 
VBA selects a random sample of completed claims decisions each month 
from each of its regional offices to review for accuracy. STAR 
reviewers assess decision accuracy using a standard checklist. For 
decisions affecting benefit entitlement, this review includes an 
assessment of whether (1) all issues in the claim were addressed; (2) 
assistance was provided to the claimant, as required by the Veterans 
Claims Assistance Act of 2000; and (3) the benefit entitlement decision 
was correct. If a claim has any error, VBA counts the entire claim as 
incorrect for accuracy rate computation purposes. The STAR reviewer 
then returns the case file and the results of the review to the 
regional office that made the decision. If an error was found, the 
regional office is required to either correct it or request 
reconsideration of the error determination. VBA uses the national 
accuracy rate from STAR reviews of compensation entitlement decisions 
as one of its key claims processing performance measures. VA also uses 
STAR data to estimate improper compensation and pension benefit 
payments.

Consistency Review Activities

    One VA consistency review activity involves conducting studies of 
regional offices' decisions on specific conditions such as post-
traumatic stress disorder where VBA found differences, such as in 
benefit grant rates, across regional offices through comparative 
statistical analysis. VBA uses the results of these reviews to identify 
root causes of inconsistencies and to target training. Under another VA 
consistency review activity, called inter-rater reliability reviews, 
VBA provides rating specialists a sample case file to assess how well 
raters from various regional offices agree on an eligibility 
determination when reviewing the same body of evidence. These reviews 
allow VBA officials to target a single rating issue and take remedial 
action to ensure the consistent application of policies and procedures 
nationally.

VA Has Implemented Procedures to Address Deficiencies Identified with 
        the STAR Program, but Continues to Face Challenges in Improving 
        Accuracy

    Over the past decade, VBA has taken several actions to improve its 
STAR program and to address deficiencies identified by both GAO and 
VA's OIG. For example, in March 1999, we found that STAR review staff 
lacked sufficient organizational independence because they were also 
responsible for making claims decisions and reported to regional office 
managers responsible for claims processing.\7\ In response to our 
findings, VBA took steps to address this by utilizing reviewers who do 
not process claims and who do not report to managers responsible for 
claims processing. More recently, in February 2008, we found that STAR 
was not sampling enough initial pension claims to ensure the accuracy 
of pension claims decisions.\8\ Because initial pension claims 
constituted only about 11 percent of the combined compensation and 
pension caseload subject to accuracy review, few were likely to be 
included in the STAR review sample. We recommended that VBA take steps 
to improve its quality assurance review of initial claims, which could 
include reviewing a larger sample of pension claims. According to VBA, 
it has addressed this issue by consolidating pension claims processing 
in its three Pension Management Centers \9\ and establishing a separate 
STAR sample for pension claims. During fiscal year 2009, VBA began 
reviewing more pension claim decisions and reported that, for fiscal 
year 2009, its pension entitlement accuracy was 95 percent, exceeding 
its goal.
---------------------------------------------------------------------------
    \7\ GAO, Veterans' Benefits Claims: Further Improvements Needed in 
Claims-Processing Accuracy, GAO/HEHS-99-35 (Washington, D.C.: Mar. 1, 
1999).
    \8\ GAO, Veterans' Benefits: Improved Management Would Enhance VA's 
Pension Program, GAO-08-112 (Washington, D.C.: Feb. 14, 2008).
    \9\ The Pension Management Centers are located in St. Paul, 
Minnesota; Philadelphia, Pennsylvania; and Milwaukee, Wisconsin.
---------------------------------------------------------------------------
    In a September 2008 report, we noted that VA lacked sufficient and 
specific performance measures for assessing the accuracy of decisions 
on BDD claims and recommended that VA consider options for separately 
estimating the accuracy of such claims decisions.\10\ VA conducted an 
analysis of the costs of sampling pre-discharge claims as part of STAR 
and concluded that the costs would outweigh possible, unquantifiable 
benefits. VA also noted that the two sites that rate BDD claims 
surpassed the national average in accuracy for claims overall.\11\ 
While generally responsive to our recommendation, VA's analysis did not 
specifically review the accuracy of BDD claims relative to traditional 
claims. Moreover, because BDD claims do not comprise all claims 
reviewed at the two rating sites, we continue to believe VA's analysis 
was not sufficient to estimate the relative accuracy of BDD claims at 
these sites. While we agree that the benefits of reviewing accuracy are 
difficult to measure, if VA had better information on the accuracy of 
BDD claims, VA could use such information to inform training and focus 
its monitoring efforts. In contrast, VA currently performs STAR reviews 
that target rating decisions made by its Baltimore and Seattle offices 
under the DoD--VA disability evaluation system pilot program. Such a 
targeted review could also be conducted for BDD claims.
---------------------------------------------------------------------------
    \10\ GAO, Veterans' Disability Benefits: Better Accountability and 
Access Would Improve the Benefits Delivery at Discharge Program, GAO-
08-901 (Washington, D.C.: Sept. 9, 2008).
    \11\ BDD claims are rated at the regional offices in Winston-Salem, 
North Carolina, and Salt Lake City, Utah.
---------------------------------------------------------------------------
    In its March 2009 report, VA's OIG also identified several 
deficiencies in the STAR program and recommended corrective actions. 
The OIG found that (1) regional offices did not always submit all 
requested sample cases for review, (2) reviewers did not evaluate all 
documentation in sample files, and (3) reviewers were not properly 
recording some errors. The OIG also found that VBA was not conducting 
STAR reviews of redistributed cases (for example, claims assigned to 
resource centers for rating). The OIG reviewed a sample of 
redistributed claims and found that 69 percent had accurate entitlement 
decisions, well below VBA's reported rate of 87 percent for the 12-
month period ending in February 2008. Further, the OIG found that VBA 
did not have minimum training requirements for STAR reviewers.
    As of March 2010, VBA had taken actions to respond to all of the 
OIG's recommendations related to STAR, including (1) implementing 
procedures to follow up on cases not submitted by regional offices; (2) 
adding a mechanism to the STAR database to remind reviewers of key 
decision points; (3) requiring a second-level review of STAR reviewers' 
work; and (4) establishing a requirement that STAR reviewers receive 80 
hours of training per year. In addition, during fiscal year 2009, based 
in part on the OIG's recommendation, VBA also began monitoring the 
accuracy of claims decided by rating resource centers as it does for 
regional offices. As we noted in our January 2010 report, VBA has 
significantly expanded its practice of redistributing regional offices' 
disability claims workloads in recent years,\12\ and gathering 
timeliness and accuracy data on redistributed claims could help VBA 
assess the effectiveness of workload redistribution.
---------------------------------------------------------------------------
    \12\ VBA refers to the practice of redistributing claims as 
``brokering.''
---------------------------------------------------------------------------
    In addition, as the Congress has provided more resources to VBA to 
increase compensation and pension staffing, VBA has devoted more 
resources to quality review. In fiscal year 2008, VBA more than doubled 
the size of the quality assurance staff, allowing it to increase the 
scope of quality assurance reviews. VA states that in the 12-month 
period ending in May 2009, STAR staff reviewed over 14,000 compensation 
and pension benefit entitlement decisions.
    Although VBA has taken steps to address deficiencies in the STAR 
program, the accuracy of its benefit entitlement decisions has not 
improved. The accuracy rate was 86 percent in fiscal year 2008 and 84 
percent in fiscal year 2009, well short of VBA's fiscal year 2009 goal 
of 90 percent.\13\ VA attributed this performance to the relatively 
large number of newly hired personnel conducting claims development 
work and a general lack of training and experience. Human capital 
challenges associated with providing the needed training and acquiring 
the experience these new claims processors need to become proficient at 
their jobs will likely continue in the near future. According to VBA 
officials, it can take 3 to 5 years for rating specialists to become 
proficient.
---------------------------------------------------------------------------
    \13\ This rating-related accuracy measure includes original and 
reopened claims for disability compensation and dependency and 
indemnity (survivor) compensation benefits. Reopened claims include 
cases where a veteran seeks a higher rating for a disability or seeks 
compensation for an additional condition.

VA Has Taken Actions to Strengthen Efforts to Monitor Consistency of 
---------------------------------------------------------------------------
        Claims Decisions

    VA has taken actions to address deficiencies identified with its 
consistency review programs, but it is still too early to determine 
whether these actions will be effective. In prior work, we reported 
that VBA did not systematically assess the consistency of decision-
making for any specific impairments included in veterans' disability 
claims. We noted that if rating data identified indications of decision 
inconsistency, VA should systematically study and determine the extent 
and causes of such inconsistencies and identify ways to reduce 
unacceptable levels of variations among regional offices. Based on our 
recommendation, VBA's quality assurance staff began conducting studies 
to monitor the extent to which veterans with similar disabilities 
receive consistent ratings across regional offices and individual 
raters.\14\ VBA began these studies in fiscal year 2008. VBA identified 
61 types of impairments for consistency review and conducted at least 
two inter-rater reliability reviews, which found significant error 
rates.
---------------------------------------------------------------------------
    \14\ GAO, Veterans' Benefits: Quality Assurance for Disability 
Claims and Appeals Processing Can Be Further Improved, GAO-02-806 
(Washington, D.C.: Aug. 16, 2002).
---------------------------------------------------------------------------
    In its March 2009 report, the OIG noted that, while VBA had 
developed an adequate rating consistency review plan, including metrics 
to monitor rating consistency and a method to identify variances in 
compensation claim ratings, it had not performed these reviews as 
scheduled. In fact, VBA had initiated only 2 of 22 planned consistency 
reviews in fiscal year 2008. The OIG reported that VBA had not 
conducted these reviews because STAR staffing resources were not 
sufficient to perform all of their assigned responsibilities and noted 
that VBA's quality review office had not staffed all of its authorized 
positions. In addition, the OIG found that inter-rater reliability 
reviews were not included in VBA's quality assurance plan. The OIG 
recommended that VBA (1) develop an annual rating consistency review 
schedule and complete all planned reviews as scheduled; (2) dedicate 
sufficient staff to conduct consistency reviews in order to complete 
planned workload and reviews; and (3) include inter-rater reliability 
reviews as a permanent component of its consistency review program.
    VBA reported that it has developed an annual consistency review 
schedule and is in the process of conducting scheduled fiscal year 2010 
reviews. As of January 2010, VBA also added six staff members to 
perform quality assurance reviews. Further, VBA incorporated inter-
rater reliability reviews into its fiscal year 2009 quality assurance 
plan. Because VBA has only recently implemented these initiatives, it 
is too early to determine their impact on the consistency of claims 
decisions.

Conclusion

    Over the years, VA has been challenged in its efforts to ensure 
that veterans get the correct decisions on disability claims the first 
time they apply for them, regardless of where the claims are decided. 
Making accurate, consistent, and timely disability decisions is not 
easy, but it is important. Our veterans deserve timely service and 
accurate decisions regardless of where their claims for disability 
benefits are processed. To fulfill its commitment to quality service, 
it is imperative that VA continue to be vigilant in its quality 
assurance efforts, as this challenge will likely become even more 
difficult as aging veterans and veterans returning from ongoing 
conflicts add to VA's workload.
    Mr. Chairman, this concludes my prepared statement. I would be 
pleased to respond to any questions you or Members of the Subcommittee 
may have at this time.

GAO Contact and Staff Acknowledgments

    For further information about this testimony, please contact Daniel 
Bertoni at (202) 512-7215 or [email protected]. Contact points for our 
Offices of Congressional Relations and Public Affairs may be found on 
the last page of this testimony. In addition to the contact named 
above, key contributors to this statement include Shelia Drake, Jessica 
Orr, Martin Scire, and Greg Whitney.

Related GAO Products

    Veterans' Disability Benefits: Further Evaluation of Ongoing 
Initiatives Could Help Identify Effective Approaches for Improving 
Claims Processing. GAO-10-213. Washington, D.C.: January 29, 2010.
    Veterans' Disability Benefits: Preliminary Findings on Claims 
Processing Trends and Improvement Efforts. GAO-09-910T. Washington, 
D.C.: July 29, 2009.
    Military Disability System: Increased Supports for Servicemembers 
and Better Pilot Planning Could Improve the Disability Evaluation 
Process. GAO-08-1137. Washington, D.C.: September 24, 2008.
    Veterans' Disability Benefits: Better Accountability and Access 
Would Improve the Benefits Delivery at Discharge Program. GAO-08-901. 
Washington, D.C.: September 9, 2008.
    Veterans' Benefits: Improved Management Would Enhance VA's Pension 
Program. GAO-08-112. Washington, D.C.: February 14, 2008.
    Veterans' Benefits: Further Changes in VBA's Field Office Structure 
Could Help Improve Disability Claims Processing. GAO-06-149. 
Washington, D.C.: December 9, 2005.
    Veterans Benefits: VA Needs Plan for Assessing Consistency of 
Decisions. GAO-05-99. Washington, D.C.: November 19, 2004.
    VA Disability Benefits: Routine Monitoring of Disability Decisions 
Could Improve Consistency. GAO-06-120T. Washington, D.C.: October 20, 
2005.
    Veterans' Benefits: Improvements Needed in the Reporting and Use of 
Data on the Accuracy of Disability Claims Decisions. GAO-03-1045. 
Washington, D.C.: September 30, 2003.
    Veterans' Benefits: Quality Assurance for Disability Claims and 
Appeals Processing Can Be Further Improved. GAO-02-806. Washington, 
D.C.: August 16, 2002.
    Veterans' Benefits: Quality Assurance for Disability Claims 
Processing. GAO-01-930R. Washington, D.C.: August 23, 2001.
    Veterans' Benefits Claims: Further Improvements Needed in Claims-
Processing Accuracy. GAO/HEHS-99-35. Washington, D.C.: March 1, 1999.

                                 
        Prepared Statement of Ronald B. Abrams, Joint Executive
           Director, National Veterans Legal Services Program
    Mr. Chairman and Members of the Committee:
    I am pleased to have the opportunity to submit this testimony on 
behalf of the National Veterans Legal Services Program (NVLSP). NVLSP 
is a nonprofit veterans service organization founded in 1980 that has 
been assisting veterans and their advocates for thirty years. We 
publish numerous advocacy materials, recruit and train volunteer 
attorneys, train service officers from such veterans service 
organizations as The American Legion and Military Order of the Purple 
Heart in veterans benefits law, and conduct quality reviews of the VA 
regional offices on behalf of The American Legion. NVLSP also 
represents veterans and their families on claims for veterans benefits 
before VA, the U.S. Court of Appeals for Veterans Claims (CAVC), and 
other federal courts. Since its founding, NVLSP has represented over 
2,000 claimants before the Board of Veterans' Appeals and the Court of 
Appeals for Veterans Claims (CAVC). NVLSP is one of the four veterans 
service organizations that comprise the Veterans Consortium Pro Bono 
Program, which recruits and trains volunteer lawyers to represent 
veterans who have appealed a Board of Veterans' Appeals decision to the 
CAVC without a representative. In addition to its activities with the 
Pro Bono Program, NVLSP has trained thousands of veterans service 
officers and lawyers in veterans benefits law, and has written 
educational publications that thousands of veterans advocates regularly 
use as practice tools to assist them in their representation of VA 
claimants.
    It is clear that the quality of VA adjudications is not 
satisfactory and is a major contributor to the size of the backlog. 
Because many claims are improperly denied, because many VA adjudicators 
are inadequately trained, because many VA regional offices are 
improperly managed, because many VA regional offices are inadequately 
staffed, and because VA Central Office management has not acted to fix 
these problems in any meaningful way, many veterans and other claimants 
for VA benefits have to file unnecessary appeals, wait several years 
for a BVA remand, and wait for the VA to obtain evidence that should 
have been requested during the original adjudication of the claim. 
These appeals clog the system and create unneeded work for the VA. Of 
course, it would have been better for the VARO to do the work correctly 
the first time.
    NVLSP believes that the quality of VARO adjudications is much worse 
than what is reported by the VA. Valid indicators of the poor quality 
of adjudications performed by the VAROs are the remand and reversal 
statistics produced by decisions issued by the Board of Veterans' 
Appeals (BVA or Board) and decisions issued by the U.S. Court of 
Appeals for Veterans Claims (CAVC).
    BVA statistics provided by the Board for FY 2009 reveal that Board 
decided over 48,000 appeals. The Board granted additional benefits in 
24 percent of the appeals and remanded 37.3 percent of these appeals 
back to the VAROs. Therefore, 61.3 percent of the VARO decisions that 
were appealed and decided by the BVA were either reversed or remanded. 
These statistics are about 4 percent worse than the statistics I 
reported to you in 2008.
    The news gets worse. The BVA, in its rush to make final decisions 
and to avoid remands quite often prematurely denies claims that should 
have been remanded. The CAVC in a large number of appeals finds fault 
with how the VA processed the claim and remands the case back to the VA 
to correct claims processing errors. The CAVC reports that of the 3,270 
merits decisions it issued in fiscal year 2009, it affirmed the BVA 
decision in less than 20 percent of the cases. In over 80 percent of 
all the BVA decisions appealed to the CAVC and decided on the merits, 
the CAVC either reversed or remanded the BVA decision
    Even Chief Justice Roberts was startled when he learned that in 
litigating with veterans before the CAVC, the government more often 
than not takes a position that is substantially unjustified. These 
statistics are shocking and revealing. The VA is required to conduct 
ex-parte, non-adversarial claims adjudications in a veteran friendly 
environment. (38 U.S.C. 5107(b), 38 C.F.R. 3.102) and provide veteran 
claimant's with a non-adversarial adjudication system.
    The results of the Legion/NVLSP quality reviews continue to be 
discomforting. The American Legion/NVLSP team usually spends a week in 
a VARO reviewing the quality of recently adjudicated claims where The 
American Legion represented the veteran. The results of these quality 
reviews reveal that in many instances claims are improperly denied or 
benefits are not paid at the proper rate because the RO was more 
concerned about claiming work credit and reducing the VARO backlog than 
taking the time to develop and analyze the claim properly. The Legion 
team generally finds a much higher error rate than the 12 percent 
generally reported by STAR.
    The good news is that most of the VA service center: managers, 
coaches, decision review officers, and raters that we have interviewed 
on these many quality reviews are sincere when they mention the need 
for quality adjudications. We have, however, met many VA regional 
office officials who are free to admit that their main focus is on 
production, not quality.

INDEPENDENT QUALITY REVIEWS

    During the past several years the National Veterans Legal Services 
Program (NVLSP) has performed, on behalf of The American Legion and 
several state departments of veterans' affairs, quality reviews of 
decisions issued by several VA Regional Offices (ROs). Our conclusion, 
based on these reviews and on information received at our service 
officer trainings, is that although the VA is to be commended for 
initiatives to stop blatant work measurement (end-product) cheating and 
to emphasize quality, the most needed change--full and fair 
adjudication of veterans' claims--has not become a reality.
    Essentially, while NVLSP commends VBA for its quality initiatives, 
we are forced to conclude that these initiatives combined with the STAR 
program have not achieved the desired result.

Premature Adjudications Resulting in Adverse Decisions

    The most important and pervasive problem facing veterans seeking VA 
disability benefits is the eagerness of some ROs to adjudicate claims 
before all necessary evidence has been obtained. For example, some ROs 
prematurely deny claims based on inadequate VA examinations. In some 
cases, even where the VA examiner clearly fails to respond to a 
specific question asked by the RO, the examination report is not 
returned as inadequate. Instead, the claim is adjudicated and denied on 
the basis of the inadequate report. In other instances, claims are 
denied before all service medical records are received. Other claims 
are sometimes denied before the veteran has a fair opportunity to 
submit independent medical evidence. These all-too-frequent cases of 
premature denial result from an over-emphasis on timeliness and a lack 
of accountability.
    We certainly believe that claims for VA disability benefits should 
be accurately adjudicated in a timely manner. However, because of a 
management emphasis on timeliness, or a perceived emphasis on 
timeliness, some VA adjudicators appear to believe that they are 
pressured to make premature final decisions. In most instances, we have 
discovered that a decision made prematurely is likely to take the form 
of a denial of benefits rather than an award of benefits.
    Let us make something very clear: The timeliness of VA adjudication 
is but one factor in the overall assessment of the VA disability claims 
adjudication system. We realize that the overall timeliness statistics 
provided by the VBA show that VBA has not met its goal to reduce the 
time it takes to adjudicate claims for disability benefits. Even though 
the VA has not met its goal in this respect, we urge that you not 
overemphasize timeliness to the detriment of quality. It does veterans 
little good to have their claims promptly, but inaccurately, denied. 
The errors found by STAR and the subsequent trainings based on STAR 
findings have not significantly improved this situation.
    One may wonder why VA adjudicators would want to prematurely deny 
claims. The answer lies in the VA work measurement system. When a claim 
for VA benefits is prematurely and inaccurately denied, many veterans 
submit new evidence to reopen their claim. The VA considers the new 
evidence a second claim and the employee earns double work credit. 
Adjudication officers, now called service center managers, have 
informed us off-the-record that they feel pressured to prematurely 
adjudicate claims because they expect other ROs will do the same, and 
they want to show that their productivity and timeliness is as good as 
other ROs.
    The VA work measurement system should encourage a timely and 
accurate adjudication, not just a timely adjudication. Section 2 of 
H.R. 3047 would change when VA regional offices (VAROs) can claim work 
credit, and was a good bill that would have helped to accomplish this 
goal. NVLSP looks forward to reviewing the overdue VA report that was 
mandated by PL 110-389.

Adversarial Attitude

    Our quality review has identified a systemic attitude problem in 
some ROs, which may take one of several forms. One example is that 
despite the general tendency to deny prematurely, some ROs ``develop to 
deny.'' That is, these ROs consistently seek to develop negative 
evidence in cases where all the evidence of record before the RO, 
without further development, would reasonably support the grant of 
benefits.
    Another attitude problem is that some ROs have biases against 
certain types of VA claims for benefits. For example, veterans seeking 
service connection for mental conditions, entitlement to individual 
unemployability benefits, or entitlement to compensation based upon 
secondary service connection, in some instances, have to jump over a 
higher bar than those who file other types of claims.
    In addition, some ROs either refuse to consider or are unaware of 
beneficial statutes in Title 38, United States Code. For example our 
quality reviews have found that 38 U.S.C. Sec. 1154(b), which provides 
in most cases that the statement of a combat veteran about an injury 
that occurred during combat will be accepted as true even though there 
is no official record of the injury, is sometimes conspicuously 
disregarded.

Communication Problems

    In many cases, the VA's communication with its veteran-claimants 
causes real problems. For example, VA notifications often fail to 
provide an adequate explanation of the reasons and bases for the 
adverse VA determination. Other communication problems noted by NVLSP 
are:

          Inadequate development letters (development letters 
        are sent by the VA to the veteran and his or her 
        representative, asking for further information or evidence) 
        that do not comply with VA's guidance that letters should 
        clearly tell the claimant what evidence is needed and what 
        exactly has to be done to establish entitlement to the benefit 
        sought (see M21-1, Part III, para. 1.04a.); and
          Telephone communication with the veteran that is not 
        monitored or sanctioned by the veteran's representative (the VA 
        does not even inform the representative that it is about to 
        contact the representative's client).

Widespread Errors

    The following is a list of a systemic pattern of errors that we 
have noticed during our quality review checks. These errors are:

          Assignment of erroneously low disability ratings for 
        service-connected mental conditions;
          Erroneous denial of claims for service connection for 
        mental conditions;
          Failure to consider 38 U.S.C. Sec. 1154(b);
          Erroneous denial of claims of individual 
        unemployability;
          Inadequate requests for medical opinions (for 
        example, the standard of proof in the VA claims process is 
        rarely explained to VA doctors, and in many instances 
        conclusions regarding critical facts are not communicated to 
        doctors who are asked to provide medical opinions); and
          Non-responsive VA examination reports (for example, 
        some VA examiners do not comply with the AMIE protocol, and 
        other examiners fail to respond to specific questions), coupled 
        with the acceptance of these inadequate examination reports by 
        ROs.

    In general, there is a lack of coordinated local (RO) quality 
control and a subsequent failure to act on recognized patterns of 
errors.

NVLSP Recommendations

    Based on the foregoing observations, NVLSP makes the following 
suggestions:

          VA's work measurement system should be altered so 
        that quality as well as timeliness are twin concepts that 
        together drive the system.
          To provide VA quality control with ``teeth'' and 
        prevent end-product and work measurement abuses, an aggressive 
        independent quality control should be performed.
          VBA should conduct regular meetings with its 
        stakeholders to inform them of any actions VBA has taken to 
        correct systemic adjudication problems. The stakeholders should 
        be informed about the patterns of errors identified nationally, 
        the ROs where there are significant problems, VBA's plans to 
        correct these problems, changes in management, progress reports 
        on previous initiatives, and an invitation for the stakeholders 
        to participate and coordinate in the correction of problems.
          VA should institute a system of awards and 
        disincentives for managers and adjudicators. VA managers and 
        adjudicators who perform accurate and timely work should be 
        rewarded. Managers who do not perform adequately should be 
        appropriately chastised.
          VA employees who do a good job should be paid a 
        reasonable salary, receive bonuses and be promoted.
          VA management should more clearly communicate with 
        its employees what it wants from them. If management focuses on 
        quality as well as efficient work, veterans will be better off.

    NVLSP acknowledges that the adjudication of claims for VA benefits 
is very complicated. However, we believe the stakeholders want to help 
correct adjudication problems. We would be happy to meet regularly with 
the VA to talk about the problems we have identified and suggested 
solutions.
    We would like to commend VBA managers for initiatives in reducing 
outright end-product and work measurement dishonesty and efforts to 
emphasize quality. While these efforts are commendable, it is time to 
see results.
    Our experience has taught us that VA managers are reasonable people 
who want to do the right thing. These managers care about veterans and 
know that the claims adjudication system is not working properly. To 
help these managers we ask you to encourage the VA to make at least the 
most necessary changes, alter VA's work measurement system, institute 
an aggressive quality control program, and support its efforts to 
coordinate with its stakeholders.
    We appreciate the opportunity to provide the subcommittee with this 
testimony.
    Thank you.

                                 
            Prepared Statement of John L. Wilson, Assistant
       National Legislative Director, Disabled American Veterans
    Mr. Chairman and Members of the Committee:
    The Disabled American Veterans (DAV) has consistently stated that 
the keys to successfully reforming the veterans' benefits claims 
process are training and accountability, two elements central to 
producing quality results for veterans. However, today the Veterans 
Benefits Administration (VBA) remains production driven from the Monday 
morning workload reports to personnel awards.
    The DAV strongly believes that quality should be rewarded at least 
on parity with production. However, in order for this to occur, VBA 
must first implement and then inculcate a comprehensive quality control 
program to complement its quality assurance program.
    VBA's primary quality assurance program is the Systematic Technical 
Accuracy Review (STAR) program. The STAR program can identify three 
types of errors--benefit entitlement, decision documentation/
notification, and administrative. STAR looks at whether a proper VCAA 
pre-decision ``notice'' was provided and whether the rating decision 
was merited based on the available evidence.
    Under the STAR program, VA reviews a sampling of decisions from 
regional offices and bases its national accuracy measures on the 
percentage with errors that affect entitlement, benefit amount, and 
effective date. This is a technical review to ensure a variety of 
transactions are properly carried out. Inconsistency in technical 
accuracy may signal several types of processing problems, including: 
uneven or insufficient understanding of governing criteria, operating 
rules that are too vague and allow an overly broad interpretation 
which, in turn, leads to significant variance, or outright 
arbitrariness in decision-making. Obviously, VA must detect 
inconsistencies before the cause or causes can be determined and 
remedied.
    The STAR program was implemented in the late 1990s. It was intended 
to be a national quality assurance program that would assist VBA in 
identifying processing vulnerabilities and error trends. It was 
designed to draw statistically significant case samples from all 
regional offices, using stratified random sampling algorithms. Using 
this tool, VBA could review a statistically valid case sample each 
year, and calculate its national error rates. Using the STAR program 
was also intended to identify major national error trends, so the 
Compensation and Pension (C&P) program could initiate corrective 
measures. Such corrective measures could include training, improved 
procedural guidance, or automated system improvements.
    The STAR program was not designed to provide evaluative data at the 
working unit or individual level. There are two major reasons for this. 
First, the sample sizes used in STAR are small. While totals may be 
significant at the national level, breaking out the data to Regional 
Offices may not provide numbers significant enough to ascertain a trend 
with an individual employee. Second, the STAR program essentially 
assesses the outcome, not the process of getting there. So, a claim 
that took two years to process because of piece-meal development would 
not have an error called if the resulting decision was correct and all 
pre- and post-notification requirements were met. Such processing 
delays would fall under the purview of quality control in-process 
reviews and not quality assurance programs such as STAR.
    Quality control findings from local in-process reviews would assist 
a station in assessing overall performance towards achieving the goal 
of timely and accurate decisions on veterans' claims. VBA recognized 
the importance of such quality control when the STAR program was 
created to replace the ineffective Statistical Quality Control program. 
At that time, VBA requested funding to implement a local C&P quality 
control program. That program--called Systematic Individual Performance 
Assessment (SIPA)--was announced in 2000 as a new initiative to monitor 
individual performance. Under this program, the VA would review an 
annual sample of 100 decisions for each adjudicator to identify 
individual deficiencies, ensure maintenance of skills, promote accuracy 
and consistency of claims adjudication, and restore credibility to the 
system. The reviewers would have performed related administrative 
functions, such as providing feedback on reviews, maintaining reports, 
and playing a role in employee development and ongoing training. 
Unfortunately, the VA abandoned this initiative during 2002, and 
proficiency is now apparently subjectively assessed by supervisors 
based on their day-to-day perceptions of employee performance. The SIPA 
program may have been abandoned due to inadequate resources. Without 
any quality assurance review on the individual level, the VA is 
unlikely to impose effective accountability down to the individual 
adjudicator level, where it must go if optimum quality is expected.
    VBA elected instead to install a much reduced quality control 
procedure, where coaches' review several cases per month for Veterans 
Service Representatives (VSR) and Rating Veterans Service 
Representatives (RVSR), as their quality control mechanism. This hybrid 
process has not provided adequate individual accountability, or 
sufficiently robust data to identify local process improvements. 
Coaches typically do not have the time to manage day-to-day operations 
and pull case files for ad hoc reviews of employees.
    With significant attention on the C&P claims backlog, it is 
understandable that VBA wants to maximize its personnel and resources 
to direct claim decision-making. However, technical accuracy is 
arguably the most important component of the C&P claims process. 
Pushing cases faster will not help if significant numbers of cases are 
done wrong.
    VBA needs to elevate quality to the highest priority. This means 
they should dedicate adequate resources to both quality assurance and 
quality control programs. The VSRs, RVSRs and local management teams 
need to understand that high quality work will be recognized and 
rewarded. Further, they need to understand that there will be clear 
thresholds for individual quality, repeated errors will be identified 
and associated with their processor, and that there will be appropriate 
consequences for insufficient accuracy rates.
    The STAR program was evaluated by the VA Office of Inspector 
General as part of its review of compensation rating accuracy in March 
2009, in the report titled ``Audit of Veterans Benefits Administration 
Compensation Rating Accuracy and Consistency Reviews.'' The OIG 
determined that VBA's STAR program does not provide a complete 
assessment of rating accuracy.
    During the 12-month period ending in February 2008, VBA's STAR 
process did not effectively identify and report all errors in 
compensation claim rating decisions. VBA identified a national 
compensation claim rating accuracy of 87 percent. Of the approximately 
882,000 compensation claims measured by STAR reviewers, VBA estimated 
that about 87 percent were technically accurate. The OIG, on the other 
hand, reviewed a random sampling of cases that had also been reviewed 
by STAR reviewers and found additional errors. They projected an 
accuracy rate of only 78 percent. They also audited brokered cases. Of 
that sampling, they found an accuracy rate of 69 percent. Combining the 
audit of brokered claims with those STAR reviewed claims, results in a 
projected accuracy rate of about 77 percent of claims. The OIG 
determined that this equates to approximately 203,000 claims in that 
one year alone where veterans' monthly benefits may be incorrect.
    The OIG found that STAR reviewers did not identify some of the 
missed errors because they either did not thoroughly review available 
medical and non-medical evidence, did not identify the absence of 
necessary medical information, or inappropriately misclassified benefit 
entitlement errors as comments.
    These findings are, on their surface, a result of STAR reviewers 
not finding all the errors in the cases they reviewed. They also point 
to the need for greater management oversight and an effective formal 
training program for the STAR reviewers. STAR reviewers could benefit 
from formal training; however STAR managers have said that current 
workload requirements do not allow for the amount of training time 
necessary. This is a common theme for VBA that underlies even STAR 
reviews--quantity over quality--even in the area that is supposed to 
ensure quality of at least a technical nature.
    The need for a quality control program as an adjunct to the STAR 
program can also be seen when considered through a review of the Board 
of Appeals for Veterans Claims' Summary of Remands. The summary 
represents a statistically large and reliable sample of certain 
measurable trends. The examples must be viewed in the context of the VA 
(1) deciding over 880,000 cases per year; (2) receiving over 133,000 
Notice of Disagreements; and (3) over 49,000 appeals to the Board. The 
examples below are from fiscal year (FY) 2009:

        1.  Remands resulted in 801 cases because no ``notice'' under 
        section 5103 was ever provided to the claimant. In addition, 
        there were 4,048 remanded for inadequate or incorrect notice, 
        some of which may result from the current generic notice 
        letters sent VBA. The DAV continues to call for changes to 
        these letters to include more specific information, which could 
        help lower the incidence of this error.
        2.  VA failed to request for Service Medical Records in 1,739 
        cases and failed to request for personnel records in 1,511 
        cases. These numbers are disturbing because initially 
        requesting a veteran's service records is the foundation to 
        every compensation claim.
        3.  The Board remanded 7,814 cases for failure to request VA 
        medical records. The disturbing factor here is that a VA 
        employee can usually obtain VA medical records without ever 
        leaving the confines of one's computer screen.
        4.  Another 3,187 cases were remanded because the claimant had 
        requested a travel board hearing or video-conference hearing. 
        Again, there is a disturbing factor here. A checklist is 
        utilized prior to sending an appeal to the Board that contains 
        a section that specifically asked whether the claimant has 
        asked for such a hearing.

    The examples above totaled 19,100 cases or 34 percent of appeals 
reaching the Board, all of which cleared the local rating board and the 
local appeals board with errors that are elementary in nature. Yet they 
were either not detected or they were ignored. Many more cases were 
returned for more complex errors. Regardless, a 34 percent error rate 
on such basic elements in the claims process involving VBA's most 
senior rating specialists is simply unacceptable.
    The problem with the current accountability system is that VBA 
employees who commit such errors are usually not held responsible. With 
no incentive to prevent such errors and a constant focus on production, 
quality will continue to decline.
    DAV agrees with the VA OIG that VBA could improve the STAR program 
by establishing: a mechanism to ensure STAR reviewers evaluate all 
documentation related to the claim selected for review; a requirement 
that all STAR reviewer comments receive a second review to make sure 
the reviewer appropriately recorded the comment instead of a benefit 
entitlement error; procedures to review brokered claims as part of the 
STAR program; and minimum annual training requirements for each STAR 
reviewer that are comparable to regional office rating staff training 
requirements.
    In addition, DAV recommends that VBA establish a quality control 
program that looks at claims in-process in order to determine not just 
whether a proper decision was made, but how it was arrived at in order 
to identify ways to improve the system. Combining results from such 
quality control reviews with STAR's quality assurance results and the 
data from remands from the Board of Veterans' Appeals and the Court of 
Appeals for Veterans Claims could yield valuable information on trends 
and cause of errors. If the data from all such reviews could be 
incorporated into a robust IT system, proper analysis of such data 
would provide management and employees important insights into 
processes and decisions. This in turn would lead to quicker and more 
accurate decisions on benefits claims, and most importantly, to the 
delivery of all earned benefits to veterans, particularly disabled 
veterans, in a timely manner
    That concludes my testimony. I would be pleased to answer any 
questions the Committee may have.
                                 
           Prepared Statement of Raymond C. Kelley, National
            Legislative Director, American Veterans (AMVETS)
    Chairman Hall, Ranking Member Lamborn, and members of the 
subcommittee, thank you for the opportunity to appear before you today 
to provide AMVETS' views regarding VBA's Systematic Technical Accuracy 
Review (STAR) Program.
    The Systematic Technical Accuracy Review (STAR) Program is VBA's 
compensation and pension accuracy measurement system. AMVETS agrees 
with the VA's Office of Inspector General's March 12, 2009 report that 
identified eight issues that will improve the process of reviewing 
claims for errors. And AMVETS is pleased to see VBA is taking action to 
correct these issues. AMVETS is concerned though, with what is done 
with the information that is gleaned from STAR.
    For the STAR program to truly be effective, AMVETS believes three 
things must be done. First, STAR must be enhanced so trends in errors 
can be easily identified by the regional offices. With this 
information, VBA must hold ROs accountable for failures in accuracy and 
insist that ROs develop improvement strategies that include training 
for all accuracy issues. Second, VBA must change its culture of 
timeliness and strive for a culture of accuracy. Wither or not STAR is 
completely accurate in its reviews is important but not nearly as 
important as what is done to ensure the same mistakes are not repeated. 
Third, OIG must conduct periodic review of the STAR program to ensure 
that STAR's accuracy ratings are within a 3 percent margin of error.
    Even though AMVETS believes the STAR program is effective, we 
believe it should be expanded to ensure that specific programs, such as 
the Benefits Delivery at Discharge (BDD) program and specific 
conditions can be tracked for anomalies that occur, so improvement 
strategies and specific training can be implemented. When the Veterans 
Claims Assistance Act (VCAA) was introduced in the first quarter of FY 
2002, nearly half of all errors were VCAA related. VBA had the ROs re-
train their claims processors and in the last 8 months of the FY, these 
types of errors were reduced by one-third. This type of accountability 
needs to be the rule and not the exception.
    As you know, the STAR program identifies errors in claims decisions 
each month through random sampling of each Regional Office (RO). The 
results of the reviews are sent to the ROs. Errors that are found are 
to be corrected by the RO that made the decision. The corrections are 
made, but all too often, the ROs don't implement strategies to ensure 
that claims processors don't continually repeat the same mistakes. 
AMVETS believes the reason these strategies are not developed is that 
the culture within the regional offices is one of timeliness and not 
one of accuracy. STAR has consistently found that 20 to 25 percent of 
claims are in error over the past decade, but VBA has done little to 
ensure that mistakes that are made in the regional offices are 
understood and that strategies for improvements are put in place. On 
the other hand, VBA does require ROs to develop corrective action plans 
if they do not reach strategic goals for production, inventory and 
timeliness. This paradigm must be flipped.
    The March, 2009 OIG Report clearly defined the gaps in the STAR 
program that have caused the 10 percent disparity in compensation and 
pension rating accuracy. AMVETS believes VBA is taking action to close 
these gaps. However, AMVETS believes it is important to have this 
accuracy fall within a 3 percent margin of error. Therefore, AMVETS 
requests that OIG conduct a follow-up to the 2009 report to ensure 
VBA's gap solutions are productive, and that OIG continue to conduct 
periodic reviews of STAR to ensure the program reaches and maintains a 
3 percent margin of error.
    Mr. Chairman, thank you again for providing AMVETS the opportunity 
to present our views on the STAR program. This concludes my testimony 
and I will be happy to answer any questions you may have.

                                 
      Prepared Statement of Ian C. de Planque, Assistant Director,
    Veterans Affairs and Rehabilitation Commission, American Legion
    Mr. Chairman and Members of the Subcommittee:
    I appreciate this opportunity to express the views of the nearly 3 
million members of The American Legion on the Department of Veterans 
Affairs' (VA's) Systemic Technical Accuracy Review (STAR). VA 
maintains, as stated by Acting Under Secretary for Benefits Michael 
Walcoff in a January 3, 2010 appearance on 60 Minutes, that: ``We 
stress over and over again to our employees that quality is our number 
one indicator, that's absolutely a requirement for successful 
performance.'' STAR has the potential to be an effective tool to 
achieve this end; however, VA is currently falling short in their 
effective implementation of this tool. This failure can, and should be 
corrected. The American Legion recommends VA consider taking focused 
actions to reverse this current situation.
    On March 12, 2009, VA's Office of the Inspector General (VAOIG) 
issued Report No. 08-02073-96, Audit of Veterans Benefits 
Administration Compensation Rating Accuracy and Consistency Reviews. 
This report detailed numerous flaws in the current implementation of 
the STAR system. Although VA acknowledged some of those criticisms, The 
American Legion is unaware of any significant corrective actions by VA 
leadership.
    The VAOIG audit was conducted to evaluate the Veterans Benefits 
Administration's (VBA's) quality assurance program, of which STAR is a 
component. Some of the more troubling findings included:

         VBA's STAR process did not effectively identify and report 
        errors in compensation claim rating decisions. VBA identified a 
        national compensation claim rating accuracy of 87 percent for 
        the 12-month period ending February 2008. We projected that VBA 
        officials understated the error rate by about 10 percent, which 
        equates to approximately 88,000 additional claims where 
        veterans' monthly benefits may be incorrect. In total, we 
        projected about 77 percent (approximately 679,000) of the 
        almost 882,000 claims completed were accurate for the 12-month 
        period ending February 2008. The 87 percent rate is not 
        comparable to the accuracy rate VBA reports in the Performance 
        and Accountability Report because that rate includes pension 
        claims. Further, this accuracy rate only included errors that 
        affected a veteran's benefits. STAR identifies, but does not 
        report, other errors related to the rating decision's 
        documentation, notification, and administration.

    That 88,000 claims are potentially incorrect is certainly troubling 
and clearly unacceptable to The American Legion. This possible trend 
will undoubtedly add to the current claims backlog challenge being 
aggressively addressed by VA Secretary Eric Shinseki. Further 
problematic is the inaccuracy of VA's system of reporting its own 
errors and isolating problem areas and/or responsible individuals in 
the claims rating process. Without external observation, such as this 
audit by VAOIG, such errors may never have come to light and an 
inaccurate picture of the overall flaws in the disability system may 
never have been recorded. More importantly, corrective actions cannot 
be taken in a timely manner.
    Furthermore, the VAOIG audit reports:

         VBA officials planned to conduct 22 reviews in FY 2008 
        consisting of 20 grant/denial rate and 2 evaluation reviews. 
        However, they only initiated two grant/denial rate reviews and 
        these were not completed until December 2008. Furthermore, VBA 
        officials did not initiate either of the two planned evaluation 
        reviews to analyze and improve the consistency of disability 
        compensation ratings and to reduce the variances between 
        states.

    Even where problem areas or potential problem areas are identified, 
VA is not conducting followup or analysis on their projected plans. 
Effective as STAR may be, if it is not implemented as intended, it 
cannot help correct problems. Aggressive congressional oversight would 
seem essential to assuring timely application of these procedures.

    VAOIG concluded in this instance:

         Without an effective and reliable quality assurance program, 
        VBA leadership cannot adequately monitor performance to make 
        necessary program improvements and ensure veterans receive 
        accurate ratings. Further, without implementing an effective 
        rating consistency program, VBA officials cannot successfully 
        assess or prioritize the improvements needed for claim rating 
        consistency.

    If this was the only problem discovered, it would be enough to call 
into question the effectiveness of the system as currently configured. 
However, with further investigation it becomes clear that the current 
use of STAR does not accurately assess the claims benefits operational 
picture.
    The report states:

         In addition, VBA officials excluded brokered claims from STAR 
        reviews. We reviewed a sample of brokered claims and found an 
        accuracy rate of 69 percent.

    The brokering of claims is increasingly becoming an integral part 
of the way VBA conducts operations today. Brokering is a system 
utilized by VA to shift claims from Regional Offices with a larger 
workload to Regional Offices with less substantial workloads to 
increase the ability to process a greater number of claims overall. 
Brokering claims also raises other serious issues that merit further 
investigation. Accountability is one of the major concerns. The 
American Legion believes that without STAR analysis of the brokered 
claims there is a lack of accountability for these claims, which is 
deeply troubling.
    How effective is the actual STAR analysis? This quote from the 
report raises additional unsettling issues for The American Legion:

         STAR reviewers did not identify some of the missed errors 
        because they either did not thoroughly review available medical 
        and non-medical evidence (or identify the absence of necessary 
        medical information), or they inappropriately misclassified 
        benefit entitlement errors as comments.

    The American Legion asks how can a system intended to assess the 
quality and accuracy of actions by VBA be effective, if evaluators do 
not have access to the full file information required to make the 
decision.
    Even with these errors and flaws in the current system, such a 
potentially powerful tool for self correction should not be abandoned 
completely. Perhaps a better solution is to look for ways, both 
internal and external, in which the present system could be adapted to 
effectively improve the use of the existing components.
    In order to rectify existing problems within STAR, The American 
Legion suggests VA could make improvements by implementing a three-step 
process for change.

        1.  Compile results nationwide of rating errors identified by 
        STAR evaluations. Currently there appears to be no systemic 
        method to track errors. This data could be critical to identify 
        patterns, whether in an individual sense, on a Regional Office 
        (RO) level, or nationally across the scope of VA operations. If 
        this information is currently gathered, it does not appear to 
        be used for analysis to detect trends which could indicate 
        systemic problems within VA. This data, coupled with data 
        gathered on errors at lower levels from the Board of Veterans 
        Appeals (BVA) and the Appeals Management Center (AMC), would be 
        an excellent tool to assess VA's quality overall by supplying 
        details that could indicate problem areas.

            It is not enough to know what VA's accuracy rate across the 
        system is. VA must also know where are their greatest areas of 
        concern in order to determine areas that could be addressed and 
        provide the most efficient and effective use of resources.

        2.  Utilize data and patterns gathered from STAR to plan and 
        implement a program of training. Adequate and effective 
        training is a key concern noted often in the adjudication of 
        claims. This data could show specific topics in which employees 
        need additional training to improve accuracy of ratings. Such 
        information could help VA leadership craft an effective 
        training schedule to maximize the training resources. Future 
        training would not be generalized, but rather targeted to fix 
        specifically identified problems. This focused approach would 
        assure that training resources would be used to the greatest 
        possible impact.
        3.  Augment STAR for accuracy with outside oversight to ensure 
        the effectiveness of the program. One obvious complaint about 
        the current implementation of STAR is a lack of adequate 
        followup. The American Legion believes third-party oversight 
        offers the opportunity to provide impartial and critical 
        followup to assure compliance. The American Legion strongly 
        advocates the use of external oversight for validation.

    The American Legion recommends VA should closely monitor and record 
common errors from previous BVA and AMC remands and grants. Such status 
monitoring and documentation could help VA identify errors consistently 
made in a Regional Office or by specific individuals that are 
eventually recognized and corrected later in the process. The American 
Legion believes this would help isolate trends needing immediate 
corrective action. Unless there is a mechanism for identifying, 
compiling, reporting and correcting those errors, the office(s) or 
individual(s) making repeated mistakes continue. This concept also 
applies to the systemic review nationwide of claims by STAR. The 
American Legion believes if the error reporting of all three entities 
is combined, it would constitute an even more effective pool of data to 
enhance VA's analysis of internal problems.
    As Acting Under Secretary Walcoff stated, that quality of 
processing and not quantity of processing is the primary concern of VA, 
then an overarching and thorough assessment of every aspect of claims 
rating in which quality falls short is essential to rectifying the 
system. The American Legion believes there is plenty of data available 
to be harvested that would reveal exactly where VA is making its most 
common mistakes. There is no question that there are issues where VA 
struggles more than others. The American Legion recommends VA should 
effectively use this data to focus available resources where VA could 
improve accuracy. VBA should analyze where the process is weakest and 
use the analysis to develop meaningful training to improve performance.
    The most obvious solution to improve accuracy is meaningful and 
timely training of employees. Utilizing the wealth of information on 
weak points that can be generated by compiling data on common errors, 
VA can create a targeted training plan focused on weakness most in need 
of improvement. If focused and meaningful, VA's continuing education 
and training can maximize effectiveness of performance.
    When The American Legion conducts its quality assessment visits to 
Regional Offices around the country, one of the most common complaints 
from VA personnel, who actually process the claims, is the lack of 
useful training. An all too often made complaint from VA personnel is 
that the training is repetitive, unclear and not in areas that they 
struggle with most. Employees identify segments of the rating schedule 
that they find:

          most confusing,
          raise questions about compliance with notification 
        due to the Veterans' Claims Assistance Act (VCAA), or
          questions on how to better communicate with examining 
        physicians to achieve exams that provide enough information to 
        adequately rate a claim.

    STAR review, in conjunction with data gathered from the BVA, AMC 
and perhaps employee review can find where VA employees are struggling 
most and get them the training they need to get the job done right the 
first time. The American Legion is confident that VA employees want to 
get the claims right, but struggle under the dual burdens of inadequate 
training in a complex system and time pressures that place a greater 
value on production rather than on quality. VA recently increased the 
required training time annually from 80 hours to 85 hours. More of this 
training will be standardized. By reforming training with regard to 
common errors identified through internal review processes such as 
STAR, VA can build on its commitment to more effective and meaningful 
training and provide its employees all the skill sets and knowledge 
they need to do their jobs right--the first time.
    The fact that VA has been inconsistent following up on corrections 
identified by STAR highlights the need for outside oversight. Even when 
problems have been identified, VA has fallen short at the goal of 
rectifying those problems. VA states a target of 90 percent accuracy, 
yet VAOIG has pointed out that VA's internal numbers are artificially 
inflated by inaccurate self-reporting and still fall short of those 
numbers. Incentives for bonuses in the management of Regional Offices 
are all numbers-based and driven by workload measurement (quantity not 
quality). Consequently, there appears to be little to no incentive 
within VBA to accurately report numbers that would work counter to 
achieving these bonus goals. The American Legion recommends that 
external audits and congressional oversight would greatly help redirect 
emphasis on quality over quantity.
    Under the provisions of Public Law 110-389, the Veterans' Benefits 
Improvement Act of 2008, the VA Secretary is to contract with an 
independent entity to conduct, over a three-year period, an assessment 
of the VA quality assurance program. In addition, the VA Secretary is 
to report the entity's findings and conclusions to Congress. The 
American Legion looks forward to this report, but would encourage VA to 
aggressively strive for marked improvements in accuracy before, during 
and after this mandated assessment. The American Legion is confident 
that VA is capable of internal improvement. There is sufficient 
information now to begin making STAR work the way it should for VA 
employees and veterans.
    The American Legion stands ready to answer any questions of this 
Subcommittee and thank you again for this opportunity to provide 
testimony.

                                 
           Prepared Statement of Bradley G. Mayes, Director,
          Compensation and Pension Service, Veterans Benefits
          Administration, U.S. Department of Veterans Affairs
    Mr. Chairman and Members of the Subcommittee, thank you for 
providing me with this opportunity to discuss the Veterans Benefits 
Administration's (VBA) Quality Assurance Program and the positive 
effect it has on the processing of Veterans' disability claims. Joining 
me today are Edna MacDonald, Acting Deputy Director of Compensation & 
Pension Service for Policy and Procedures, and Terence Meehan, Director 
of VBA Employee Development and Training.
    The Subcommittee has indicated a special interest in our Systemic 
Technical Accuracy Review (STAR) Program. However, I want to emphasize 
that STAR is only one of four tiers of our multi-faceted national 
Quality Assurance Program. The STAR component focuses on claims 
processing accuracy, while the other three components address regional 
office oversight, rating consistency, and special focus reviews. Along 
with the STAR Program, these components collaborate to ensure high 
quality and consistent decisions for Veterans. Before discussing the 
STAR program, I will briefly discuss the three other facets of our 
Quality Assurance Program, which endeavors to ensure that compensation 
and pension benefits are provided in a timely, accurate, and consistent 
manner.

Quality Assurance Program

Oversight

    The oversight component involves compliance oversight visits to 
regional offices by members of Central Office site survey teams. Each 
regional office is visited on a three-year rotating basis. All 
operations of the visited regional office are reviewed and evaluated, 
with recommendations provided for improving claims processing 
efficiency, accuracy, and timeliness. Additionally, the site visit team 
assesses whether the regional office is in compliance with VBA policy 
and procedures and consistent with national standards. A Web site has 
recently been created to share the ``best practices'' identified during 
site visits.

Consistency

    The consistency component is based on mining of source data from 
the VBA corporate database that houses information from all regional 
office rating decisions. Given the possibility for variation in 
disability decisions, it is incumbent on VA to ensure program integrity 
by having a credible system for identifying indications of 
inconsistency among its regional offices and then remedying any 
inconsistencies found to be unreasonable. The two key pieces of 
information obtained are the grant rate and the evaluation 
distribution.
    Data analysis of recently completed rating decisions identifies the 
most frequently rated diagnostic codes, and plots both the grant/denial 
rate and evaluation assigned across all regional offices. This 
information focuses on rating decisions for specific disabilities and 
provides a method to evaluate consistency and accuracy on a regional 
office level. A focused review of files from the regional offices that 
are above or below the national average is conducted with the goal of 
identifying causes for the statistical anomaly. Once root causes are 
identified, the regional offices are notified of any recommendations, 
which may include training to correct problems or improper procedures 
identified during the review.

Special Focus Reviews

    Special focus reviews address topics of special interest to VBA or 
other stakeholders where accuracy and consistency are an issue and are 
conducted as needed in support of VA's mission and needs. They address 
a specified purpose or type of claim and may involve a nationwide 
review or a review of work assigned to a specific regional office. The 
ad hoc reviews can be one-time or recurring in nature. For example, 
consolidation of the processing of radiation claims began on October 
16, 2006. In 2008 the STAR staff conducted a special focused review of 
radiation claims completed between October 2006 and October 2007. The 
findings from this review provided a means for assessing the 
consolidation effort. The review found the overall accuracy and 
processing timeliness of radiation claims improved following 
consolidation.

STAR

    STAR is the quality assurance component that focuses on claims 
processing accuracy. STAR reviews are focused on outcomes for Veterans 
rather than specific processes by which the outcomes are reached. STAR 
reviews evaluate the quality of the rating decision product that VBA 
provides for Veterans. From the Veteran's perspective, there is an 
expectation that we understand the claim, evaluate it accurately and 
fairly, and provide proper compensation under the law. The purpose of 
STAR reviews is to ensure that rating decision outcomes meet these 
expectations.
    The STAR system includes review of three types of work: claims that 
usually require a rating decision; authorization work that does not 
generally require a rating decision; and fiduciary work. The focus on 
rating-related decisions and authorization actions is a benefit 
entitlement review using a structured checklist to ensure all issues 
were addressed, claims assistance was provided, and the decision was 
correct (to include a correct effective date). Accuracy results are 
calculated on the results of the benefit entitlement review. STAR 
findings provide statistically valid accuracy results at both the 
regional office and national level. In addition, quality reviews of the 
accuracy of VBA examination requests and VHA examination reports are 
conducted in collaboration with the Compensation and Pension 
Examination Program (CPEP) office.
    VBA continues to focus on expanding and improving the STAR Program. 
In 2008 the STAR staff was consolidated to Nashville, which provided 
space for expansion and allowed aggressive recruitment for more STAR 
staff Since then, STAR has completed an extensive expansion effort, 
more than doubling the staff and increasing the sample size to obtain a 
statistically valid sample at the regional office level. In 2010, 
quality review of fiduciary cases was transferred to Nashville. During 
fiscal year (FY) 2009, 24,747 cases were reviewed for rating and 
authorization accuracy, and 3,671 cases were reviewed for fiduciary 
accuracy. The targeted number of cases for STAR review in FY 2011 is 
37,932. The STAR sample was expanded in 2009 to include a review of 
brokered work completed by VBA's 12 Resource Centers and one Tiger Team 
(a Cleveland-based team focus on processing a unique subset of claims), 
and sampling was increased for the Pension Management Centers to allow 
measurement of pension entitlement decisions. Ongoing reviews of the 
Disability Evaluation System cases and Appeals Management Center cases 
became part of the monthly compensation quality sample in FY 2009.
    STAR error trends are identified and used as training topics. 
Training continues to be a priority and is conducted using a variety of 
methods, including a monthly national Quality Call, where Compensation 
& Pension Service's training, policy, and procedures staffs collaborate 
with the STAR staff to address national error trends identified in STAR 
assessments.
    To assure accuracy of STAR findings, a second-level peer review of 
all comments was implemented in FY 2009. Regional offices are provided 
explanations on all error calls, and they are required to take 
corrective action. On a quarterly basis, regional offices are required 
to certify to VBA headquarters the corrective action taken for all 
errors identified by STAR. The reported actions are validated during 
the oversight visits conducted by the site survey teams.
    Section 224 of the Veterans' Benefits Improvement Act of 2008 (PL 
110-389) required VA to contract with a third-party entity to conduct 
an assessment of the quality assurance program, evaluate a sample of 
employees' work, measure the performance of VA regional offices and 
accuracy of rating, assess employee and manager performance, and 
produce data to help identify trends. VBA contracted with the Institute 
for Defense Analyses (IDA) to conduct this assessment. IDA furnished 
preliminary findings concerning its evaluation of the national accuracy 
of work, the regional office accuracy of work, the accuracy of 
disability ratings, and the consistency of disability ratings. Its 
preliminary findings include an assessment that the current STAR 
accuracy review program is adequately designed to estimate, within a 5 
percent margin of error, the percentage of claims completed that 
contain errors both nationally and by regional office. IDA is 
continuing to determine options for identifying the accuracy and 
consistency of disability rating decisions.
    VBA anticipates that further improvements in the STAR Program, as 
well as other components of the Quality Assurance Program, will result 
from recommendations from the IDA study. We are looking forward to the 
recommendations from IDA and working collaboratively with them to 
further improve our Quality Assurance Program.

Quality Assurance and Training

Training Improvements

    VBA is committed to using the error trends and accuracy findings to 
improve overall quality. VBA uses nationwide error patterns identified 
by STAR reviews, as well as information from other components of the 
Quality Assurance Program, to adjust and develop the employee training 
curricula.
    All employees, regardless of training level, must receive 80 hours 
of instruction annually. Instructional methods may include Training 
Performance Support System (TPSS) modules, lectures, or practical 
application exercises. For intermediate and journey-level employees, 
the 85 hours must include 40 Core Technical Training Requirement (CTTR) 
hours. These involve standardized training curricula of essential 
topics and information. Employees must complete an additional 20 hours 
of training from a list of standardized topics provided by VBA. The 
final 20 hours may be used by regional offices to train on local issues 
and areas of concern. This approach ensures that new and experienced 
employees are grounded in standardized claims processing fundamentals.
    Data from STAR reviews, consistency reviews, special focus reviews, 
and regional office site visits, are used to develop training for our 
new hires, as well as our intermediate and journey-level employees. 
Claims processing personnel are timely informed of errors and 
inconsistency trends and provided with constructive feedback, including 
instructions on how to avoid such errors in the future. The error 
trends identified in STAR reviews provide us the information we need to 
assess the effectiveness of our training programs and make necessary 
adjustments. This promotes our goal of providing accurate, fair, and 
consistent claims processing.

Office of Employee Development and Training

    To ensure that our training is up to date and incorporates the 
lessons learned from the Quality Assurance Program, VBA has a robust 
training evaluation program conducted by the Technical Training and 
Evaluation section of the Office of Employee Development and Training. 
With the assistance of professionally qualified outside evaluators, 
this office has undertaken two formal evaluations of the centralized 
Challenge program for newly hired claims personnel. The first formal 
evaluation, conducted during 2007-2008, included visits to 16 regional 
offices and surveys with 1,405 respondent trainees. This led to the 
following three recommendations, which were adopted in reformulating 
the Challenge curricula:

          Re-sequencing the content to eliminate redundancy and 
        make better use of resources,
          Providing standardized hands-on claim processing 
        during the centralized training portion, and
          Creating more formal requirements for delivery and 
        compliance.

    The second evaluation of Challenge training began at the end of 
2009. Preliminary findings show that the changes made to Challenge 
significantly improved its effectiveness, as demonstrated by new 
trainees now being able to correctly process simple cases immediately 
after completing the centralized portion of the training. TPSS training 
was separately evaluated during 2005-2006 and 2006-2007, and then 
subsumed into the Challenge evaluations because of its high usage 
during the initial training phase.
    In total, the Office of Employee Development and Training visited 
34 regional offices during the programmatic evaluations of Challenge 
and collected more than 3,200 surveys. This is in addition to the 
hundreds of individual participant Challenge surveys done in the course 
of the three phases of Challenge classes.
    Performance support training tools allied to TPSS modules continue 
to show high and increasing usage, reflecting their utility to the 
field. For example, the Medical Electronic Performance Support System 
provides computerized visual images of the various human body systems. 
It was developed with STAR review input to assist with identifying the 
appropriate rating codes associated with different body systems and to 
facilitate medical examination requests. This tool had 633,585 unique 
user sessions in FY 2009, a 32 percent increase from FY 2008. Another 
training tool, the Veterans Service Representative Assistant, was 
designed to assist with claims development and had 34,696 unique 
sessions in FY 2009, a 62 percent increase from FY 2008. The increased 
use of these performance support tools may be attributable to a growing 
population of new claims processors who have learned about them during 
Challenge training, where their use is integrated into practical 
applications.
    VBA continues to improve its assessment of training program quality 
and effectiveness by collecting more timely feedback from participants. 
Improvements in the last two years, for instance, include adding 
surveys during the home-station segments of Challenge training and 
promulgating an official requirement to complete surveys for both 
Challenge and CTTR training.

Conclusion

    The VBA Quality Assurance program has undergone significant change 
over the past several years and has become more comprehensive by 
expanding the type and breadth of cases reviewed. This is a journey 
that we are on, and we look forward to learning from the IDA review and 
working with our stakeholders to continuously improve our Quality 
Assurance Program.
    Thank you for the opportunity to provide you an update on our 
accomplishments.
                   MATERIAL SUBMITTED FOR THE RECORD

                                U.S. Department of Veterans Affairs
                                        Office of Inspector General
                                                    Washington, DC.
                                                     April 29, 2010
The Honorable John Hall
Chairman, Subcommittee on
Disability Assistance and Memorial Affairs
Committee on Veterans' Affairs
United States House of Representatives
Washington, DC 20515

Dear Mr. Chairman:

    At the March 24, 2010, hearing before the Subcommittee on 
Examination of VA Regional Office Disability Claims Quality Review 
Methods--Is VBA's Systematic Technical Accuracy Review (STAR) Making 
the Grade?, you requested additional information on the results of file 
reviews associated with our March 2009 report, Audit of Veterans 
Benefits Administration Compensation Rating Accuracy and Consistency 
Reviews. Specifically, you asked for the ratio of underpayments and 
overpayments in files that contained errors. We did not collect data 
from the VA regional offices during our audit to determine the ratio. 
The Veterans Benefits Administration may be able to determine the ratio 
from other data sources.
    Thank you for your interest in the Department of Veterans Affairs.
            Sincerely,

                                           /S/Joanne M. Moffett for
                                                    GEORGE J. OPFER
                                                  Inspector General

                                 

                                     Committee on Veterans' Affairs
         Subcommittee on Disability Assistance and Memorial Affairs
                                                    Washington, DC.
                                                      April 5, 2010
Daniel Bertoni
Director, Education, Workforce, and Security
U.S. Government Accountability Office
441 G Street, NW
Washington, DC 20548

Dear Mr. Bertoni:

    Thank you for testifying at the House Committee on Veterans' 
Affairs' Subcommittee on Disability Assistance and Memorial Affairs' 
oversight hearing on the ``Examination of VA Regional Office Disability 
Claims Quality Review Methods--Is VBA's Systematic Technical Accuracy 
Review (STAR) Making the Grade?,'' held on March 24, 2010. I would 
greatly appreciate if you would provide answers to the enclosed follow-
up hearing questions by Thursday, May 6, 2010.
    In an effort to reduce printing costs, the Committee on Veterans' 
Affairs, in cooperation with the Joint Committee on Printing, is 
implementing some formatting changes for material for all full 
committee and subcommittee hearings. Therefore, it would be appreciated 
if you could provide your answers consecutively on letter size paper, 
single-spaced. In addition, please restate the question in its entirety 
before the answer.
    Due to the delay in receiving mail, please provide your responses 
to Ms. Megan Williams by fax at (202) 225-2034. If you have any 
questions, please call (202) 225-3608.
            Sincerely,

                                                       John J. Hall
                                                           Chairman

                               __________

                     United States Government Accountability Office
                                                    Washington, DC.
                                                        May 4, 2010
The Honorable John J. Hall
Chairman
Subcommittee on Disability Assistance and Memorial Affairs
Committee on Veterans' Affairs
House of Representatives

Subject: VA Has Improved Its Programs for Measuring Accuracy and 
Consistency, but Challenges Remain

    This letter responds to your April 5, 2010, request that we answer 
questions related to our testimony on March 24, 2010.\1\ During that 
hearing, we discussed the progress the Department of Veterans Affairs 
(VA) has made in improving its ability to measure the accuracy and 
consistency of its decisions on disability benefit claims; and 
remaining challenges to improving accuracy and consistency. Your 
questions, along with our responses, follow.
---------------------------------------------------------------------------
    \1\ Veterans' Disability Benefits: VA Has Improved Its Programs for 
Measuring Accuracy and Consistency, but Challenges Remain, GAO-10-530T 
(Washington, D.C.: March 24, 2010).

        1.  In your testimony, you observed that while the STAR system 
        was used to review the accuracy of some Veterans Benefits 
        Administration (VBA) claims programs, Benefits Delivery at 
        Discharge (BDD) and Quick Start pre-discharge claims were not 
        assessed for quality by the STAR program. Are there other 
        programs that are not being assessed by the STAR system, which 
        should be, in your opinion? If there are such programs, what 
        are the benefits of using STAR to examine the quality of these 
        VBA programs? Could STAR reviews for all VBA ratings programs 
        help improve the agency's ability to more timely deliver 
---------------------------------------------------------------------------
        benefits to deserving veterans.

    In our September 2008 report on BDD,\2\ we noted that VBA lacked 
measures that would allow it to assess whether decisions on BDD claims 
were more or less accurate than decisions under its traditional claims 
process, or whether BDD locations faced challenges in processing 
claims. As noted in my statement, improved accuracy would improve VBA's 
ability to provide veterans and their survivors the benefits to which 
they are entitled. Also, improved accuracy helps to ensure that 
decisions are correct the first time, which can prevent delays in VA's 
lengthy appeals process. In addition, VBA's consistency reviews have 
the potential to improve VBA's ability to provide comparable benefits 
to comparable veterans across its 57 regional offices. We have not 
identified any other specific types of claims where we have recommended 
targeted reviews under the Systematic Technical Accuracy Review (STAR) 
program.
---------------------------------------------------------------------------
    \2\ GAO, Veterans' Disability Benefits: Better Accountability and 
Access Would Improve the Benefits Delivery at Discharge Program, GAO-
08-901 (Washington, D.C.: Sept. 9, 2008).

        2.  Detail all the VBA programs that are currently being 
        assessed by the STAR system, and provide the annual accuracy 
---------------------------------------------------------------------------
        rates for those programs for 2009.

    VA provided information on several types of claims where it was 
conducting STAR reviews, with fiscal year 2009 accuracy rates.

  Table 1: Accuracy rates of VBA programs assessed by STAR, Fiscal Year
                                  2009
------------------------------------------------------------------------
                                               Accuracy rate (percent)
 
------------------------------------------------------------------------
Compensation entitlement (key             84
 performance measure)
------------------------------------------------------------------------
Pension entitlement                       97
------------------------------------------------------------------------
Redistributed (brokered) cases            77 \a\
------------------------------------------------------------------------
Disability Evaluation System pilot cases  88
 \b\
------------------------------------------------------------------------
Appeals Management Center \c\             83
------------------------------------------------------------------------
\a\ Benefit entitlement decisions made by VBA's 9 rating resource
  centers. VBA also reported a 78 percent accuracy rate for ratings by
  the Tiger Team at the Cleveland Regional Office. According to VBA, it
  began STAR reviews of these cases in July 2009.
\b\ According to VBA, it began STAR reviews of Disability Evaluation
  System pilot ratings by the Baltimore and Seattle regional offices in
  June 2009.
\c\ According to VBA, it began reviewing Appeals Management Center
  decisions in July 2009.
Source: VBA.


        3.  You both observed that STAR review staff lacked the 
        manpower and training to fully utilize the quality control 
        mechanisms offered by the STAR program. With these issues in 
        mind, how many Full Time Equivalent (FTE) employees would be 
        ideal, and what training for those employees would you 
        recommend?

    In its March 2009 report, VA's Office of Inspector General noted 
that VBA lacked sufficient staff to conduct planned consistency 
reviews. Since then, VBA has hired additional quality assurance staff. 
We have not done the work needed to estimate the number of Full Time 
Equivalent (FTE) staff the quality assurance programs need to conduct 
its planned quality assurance reviews.
    In responding to these questions, we relied on information we 
collected in preparing our March 24, 2010, written statement. For 
further information, please contact Daniel Bertoni at (202) 512-7215 or 
[email protected].

                                                     Daniel Bertoni
                Director, Education, Workforce, and Income Security

                                 

                                     Committee on Veterans' Affairs
         Subcommittee on Disability Assistance and Memorial Affairs
                                                    Washington, DC.
                                                      April 5, 2010
Bradley Mayes
Director, Compensation and Pension Service
Veterans Benefits Administration
U.S. Department of Veterans Affairs
810 Vermont Avenue, NW
Washington, DC 20240

Dear Mr. Mayes:

    Thank you for testifying at the House Committee on Veterans' 
Affairs' Subcommittee on Disability Assistance and Memorial Affairs' 
oversight hearing on the ``Examination of VA Regional Office Disability 
Claims Quality Review Methods--Is VBA's Systematic Technical Accuracy 
Review (STAR) Making the Grade?,'' held on March 24, 2010. I would 
greatly appreciate if you would provide answers to the enclosed follow-
up hearing questions by Thursday, May 6, 2010.
    In an effort to reduce printing costs, the Committee on Veterans' 
Affairs, in cooperation with the Joint Committee on Printing, is 
implementing some formatting changes for material for all full 
committee and subcommittee hearings. Therefore, it would be appreciated 
if you could provide your answers consecutively on letter size paper, 
single-spaced. In addition, please restate the question in its entirety 
before the answer.
    Due to the delay in receiving mail, please provide your responses 
to Ms. Megan Williams by fax at (202) 225-2034. If you have any 
questions, please call (202) 225-3608.
            Sincerely,

                                                       John J. Hall
                                                           Chairman

                               __________

                   Questions for the Record from the
                   HVAC DAMA Subcommittee Hearing on
                   Examination of VA Regional Office
                Disability Claims Quality Review Methods
 Is VBA's Systematic Technical Accuracy Review (STAR) Making the Grade?
                             March 24, 2010
    Question 1. Does VA know why the Des Moines Regional Office (RO) is 
the highest performing RO with an accuracy rate reported by STAR of 92 
percent?

    Response: Regional office performance varies as a result of a 
number of factors including workforce experience, local economic and 
employment environment, and staff turnover. Regional offices that 
consistently perform well are in areas where VA is an employer of 
choice. In these locations, VA is able to recruit and retain high-
performing employees. The Des Moines RO has a stable and well-trained 
workforce, allowing the RO to achieve a high accuracy rate.

    Question 2. Does the VA know why the Des Moines RO is performing at 
such a higher rate than the Baltimore RO (69 percent)?

    Response: ROs that have difficulty in meeting performance targets 
are predominantly in high-cost metropolitan areas with high employee 
turnover. The Baltimore RO is located in a densely populated area and 
faces significant competition in recruiting and retaining a highly 
skilled workforce.

    Question 3. Please detail the plans that VA has in place to take 
lessons learned from what is being done right in the Des Moines RO and 
apply them to lower performing ROs like Baltimore.

    Response: VA aggressively monitors regional office performance and 
develops specific action plans to improve problem areas. Performance is 
evaluated against both national and individual targets that are 
established at the beginning of each fiscal year. Oversight is provided 
through site visits conducted by both the Compensation and Pension 
Service and the Area Directors. Lessons learned and specific examples 
of ``best practices'' from these visits are provided to assist ROs 
increase their performance.

    Question 4: Please provide the Subcommittee with data about the 
breakdown of claims processing errors reported by STAR and the VA OIG 
report. What percentages of these errors were Veterans being under 
compensated, and what percentage were Veterans being overcompensated?

    Response: STAR performed 21,747 national quality reviews in FY 2009 
and cited 2,676 benefit entitlement errors. Of the benefit entitlement 
errors, 18 percent were potential Veteran/Beneficiary overpayments and 
24 percent were potential Veteran/Beneficiary underpayments.

    Question 5. Please provide the Subcommittee with data for outgoing 
and new RO directors for the past three years, along with RO error data 
for the same time. Does the VA find any correlation between new 
leadership and a change in the accuracy of an RO?

    Response: The RO director is a critical component in the 
performance of a regional office. However, VBA does not believe that 
the STAR accuracy rate, be it positive or negative, can be attributed 
to just one factor or one individual. Data requested is attached 
(Attachment 1).

    Question 6. Please provide the list of all the names of the VA 
Regional Office Directors, the date on which they were hired, their 
current salaries, last bonus, and the 2009 accuracy rate per STAR 
review.

    Response: See spreadsheet. [Attachment 2, which lists the current 
salaries and bonuses of the VA Regional Directors is being retained in 
the Committee files. ]

    Question 7: Please provide the name and location of the STAR team 
director.

    Response: The STAR team director is Edna MacDonald, Assistant 
Director for Quality Assurance. She is located in VA Central Office, 
Washington DC.

    Question 8: Please provide the number of STAR reviewers.

    Response: There are 38 STAR team reviewers, which is an increase of 
66 percent from FY 2009.

                                                                                          Attachment 1
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
                                                                                                                   Time in                               STAR Rating   STAR Rating   STAR Rating
                Station                            Area                    Director          Effective Date of    Position        Previous Director       Accuracy      Accuracy      Accuracy
                                                                                                Appointment         (Yrs)                                Rate (FY09)   Rate (FY08)   Rate (FY07)
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Philadelphia                            Eastern                    Lastowka, Thomas                   7/1/1990         19.8   Cary, Robert                    79.9%         87.4%         91.9%
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Columbia                                Southern                   Hawkins, Carl                     10/1/1998         11.6   Sloan, Stedman                  88.1%         87.6%         88.4%
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Boise                                   Western                    Vance, James                      1/30/2000         10.2   Barker, Barry                   82.4%         88.2%         88.9%
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Waco                                    Central                    Lowe, Carl                        3/12/2000         10.1   McRae, Jerry                    84.5%         81.7%         90.2%
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Baltimore                               Eastern                    Wolohojian, George               12/15/2002          7.4   Quinton, Newell                 67.9%         82.7%         88.3%
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Phoenix                                 Western                    Flint, Sandy                       6/1/2003          6.9   Nappi, Patrick                  86.1%         86.8%         91.2%
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
St. Louis                               Central                    Unterwagner, Dave                 9/21/2003          6.6   Graves, Kim                     84.0%         84.0%         93.2%
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
New Orleans                             Central                    Christian, Rowland                 2/8/2004          6.2   Jackson, Barry                  81.9%         88.4%         93.9%
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
San Diego                               Western                    Fetzer, Lily                       8/8/2004          5.7   Dusenbery, Michael              83.5%         82.2%         86.4%
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Buffalo                                 Eastern                    Terrell, Donna                   10/31/2004          5.5   McCoy, Jack                     82.9%         87.5%         90.9%
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Montgomery                              Southern                   Randle, Ric                       6/26/2005          4.8   Watson, Montgomery              81.6%         84.4%         84.9%
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Muskogee                                Central                    Jarvis, Sam                       10/2/2005          4.6   Burks, Larry                    89.2%         86.9%         94.2%
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Manila                                  Western                    Skelly, Jon                        1/8/2006          4.3   Barker, Barry                   89.7%         89.7%         95.0%
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Louisville                              Southern                   Thompson, Keith                   4/16/2006          4.0   Wardle, Jimmy                   87.6%         86.5%         89.4%
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Manchester                              Eastern                    Woolford, Charles                 8/20/2006          3.7   Cully, Maribeth                 76.0%         81.0%         85.8%
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
White River Junction                    Eastern                    Woolford, Charles                 8/20/2006          3.7   Cully, Maribeth                 79.1%         83.3%         81.6%
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Los Angeles                             Western                    Witty, Kerrie                      9/3/2006          3.6   Liff, Stewart                   75.5%         81.5%         81.0%
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Denver                                  Western                    Jacobs, Janice                   11/26/2006          3.4   Alger, Jeffrey                  86.0%         89.7%         90.4%
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Detroit                                 Eastern                    Walker, Linda                    12/24/2006          3.3   Thompson, Keith                 78.1%         76.1%         90.2%
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Roanoke                                 Southern                   Nicholas, William                  2/4/2007          3.2   Smith, John                     85.5%         89.5%         90.8%
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Cleveland                               Eastern                    Cange, Joyce                       7/8/2007          2.8   Mayes, Brad                     88.6%         84.9%         95.0%
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Oakland                                 Western                    Flint, Lynn                        7/8/2007          2.8   Smith, Catherine                82.3%         89.6%         91.4%
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Hartford                                Eastern                    Leonard, Dave                     7/22/2007          2.8   Walker, Linda                   83.1%         87.4%         86.1%
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Little Rock                             Central                    Rawls, Cheryl                      8/5/2007          2.7   Nicholas, William               87.9%         86.9%         86.2%
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Providence                              Eastern                    Navaratnasingam, Pritz             9/2/2007          2.6   Witty, Kerrie                   86.0%         90.3%         91.6%
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Huntington                              Southern                   Commens, Zita                     3/30/2008          2.1   Rawls, Cheryl                   82.9%         86.0%         94.1%
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Atlanta                                 Southern                   Bocchicchio, Al                   4/27/2008          2.0   Burks, Larry                    79.0%         77.5%         84.3%
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Winston-Salem                           Southern                   Umlauf, Dan                       6/22/2008          1.8   Montgomery, John                86.0%         90.6%         89.6%
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Salt Lake City                          Western                    Bilosz, Mark                      7/20/2008          1.8   Wadsworth, Douglas              80.4%         88.2%         88.4%
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Lincoln                                 Central                    Miller, Loren                      8/3/2008          1.7   Smith, Peter                    92.9%         94.9%         93.3%
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Wichita                                 Central                    Waller, Antione                  10/12/2008          1.5   Umlauf, Dan                     79.4%         84.7%         86.7%
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Chicago                                 Central                    Honeycutt, Duane                  11/9/2008          1.4   Olson, Michael                  81.6%         86.5%         88.7%
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Indianapolis                            Eastern                    Stephens, Michael                  1/4/2009          1.3   Kuewa, Dennis                   84.6%         86.0%         88.1%
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
St. Paul                                Central                    Fillman, Carol                    1/18/2009          1.3   Alger, Jeffrey                  86.1%         91.7%         90.9%
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Togus                                   Eastern                    Karczewski, Scott                 2/25/2009          1.2   Bilosz, Mark                    90.4%         92.9%         90.3%
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Des Moines                              Central                    Davis, Dave                       3/15/2009          1.1   Braley, Richard                 92.7%         91.5%         93.0%
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Honolulu                                Western                    Betts, Tracy                      3/29/2009          1.1   Reed, Greg                      75.9%         83.7%         78.0%
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Jackson                                 Southern                   Moore, Craig                      7/19/2009          0.8   Adair, Joseph                   81.8%         82.3%         87.1%
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
San Juan                                Southern                   Murphy, Thomas                     8/2/2009          0.7   Moreno, Sonia                   72.1%         75.7%         82.2%
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Pittsburgh                              Eastern                    McCoy, Beth                      10/11/2009          0.5   Cully, Maribeth                 80.6%         88.7%         92.4%
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Albuquerque                             Western                    Singleton, Grant                 10/25/2009          0.5   Moore, Craig                    79.8%         90.9%         93.3%
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
New York                                Eastern                    Malley, Sue                      10/25/2009          0.5   Amberg-Blyskal, Patricia        75.9%         76.0%         83.0%
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Seattle                                 Western                    Prieb, Patrick                   10/25/2009          0.5   Fillman, Carol                  81.6%         83.5%         88.2%
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Reno                                    Western                    Russell, Ed                       11/8/2009          0.5   Iddings, Donald                 90.7%         89.2%         89.8%
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Portland                                Western                    Marshall, Chris                  11/22/2009          0.4   McClellan,                      86.6%         85.8%         87.0%
                                                                                                                              Christopher
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Sioux Falls                             Central                    Brubaker, James                    1/4/2010          0.3   Smith, John                     86.9%         94.5%         90.8%
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Washington                              Southern                   Reyes-Maggio, Judy               10/31/2004          5.5   Wilson, Keith                     N/A         59.7%         75.8%
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Appeals Mgmt Cntr                       OFO                        Burke, Ronald                      2/1/2009          1.2   Russo, Arnold                     N/A           N/A           N/A
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
St. Louis RMC                           Central                    Hawthorne, Corey                  3/28/2010          0.1   Prieb, Patrick                    N/A           N/A           N/A
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Anchorage                               Western                    Remotely managed by Salt                                                                   86.5%         84.9%         86.3%
                                                                    Lake City Director
                                                                    (Bilosz, Mark)
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Boston                                  Eastern                    VACANT                                                     Braley, Richard                 77.5%         81.3%         81.1%
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Cheyenne                                Western                    Remotely managed by the                                                                      N/A           N/A           N/A
                                                                    Denver
                                                                   Director (Jacobs,
                                                                    Janice)
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Fargo                                   Central                    Remotely managed by                                                                        91.6%         91.3%         86.7%
                                                                    Sioux Falls
                                                                   Director (Brubaker,
                                                                    James)
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Fort Harrison                           Western                    Remotely managed by Salt                                                                   91.4%         91.0%         92.6%
                                                                    Lake City Director
                                                                    (Bilosz, Mark)
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Houston                                 Central                    VACANT                                                     Henderson, Ursula               82.4%         82.5%         84.3%
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Milwaukee                               Central                    VACANT                                                     Olson, Michael                  92.7%         93.0%         92.6%
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Nashville                               Southern                   VACANT                                                     Corley, Brian                   88.2%         92.6%         93.2%
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Newark                                  Eastern                    VACANT                                                     McCourt, John                   81.0%         83.7%         82.8%
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
St. Petersburg                          Southern                   VACANT                                                     Barker, Barry                   84.7%         87.0%         91.7%
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Wilmington                              Eastern                    Remotely managed by the                                                                    82.4%         83.6%         83.6%
                                                                    Philadelphia Director
                                                                    (Lastowka, Thomas)
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------


