Defense Health Care: Claims Processing Improvements Are Underway but
Further Enhancements Are Needed (Letter Report, 08/23/1999,
GAO/HEHS-99-128).

Pursuant to a congressional request, GAO reviewed the timeliness and
accuracy of the Department of Defense's (DOD) health care claims
processing procedures.

GAO noted that: (1) between July 1, 1997, and June 30, 1998, DOD's
contractors processed 86 percent of claims within 21 days; (2) this met
DOD's timeliness standard of processing 75 percent of claims within 21
days; (3) even so, nearly 3 million claims took more than 21 days to
process, which prompted complaints from some providers and beneficiaries
about what they considered to be payment delays; (4) DOD has several
initiatives under way to improve timeliness, including adopting payment
and penalty standards used by the Medicare program; (5) if these
standards are properly implemented and met by contractors, they should
help reduce providers' complaints; (6) while DOD adequately measures
contractors' performance in claims processing timeliness, it does not
know the extent to which contractors are accurately paying claims; (7)
less than half the claims are subject to its payment accuracy audit, and
the methodology used to calculate the payment error rate is
statistically invalid; (8) all contractors experienced problems with
payment accuracy when they began processing TRICARE claims, often
because they did not have enough time to adequately prepare to
administer the program; (9) although contractors addressed these
problems, they acknowledged that many factors affect the accuracy of
claims processing--primarily the complexity of the program, compounded
by numerous program changes; (10) GAO also found that some claims
processing problems were due to mistakes made by providers and
beneficiaries when filing their claims; (11) furthermore, because they
do not always understand the program, sometimes providers and
beneficiaries complain about adjudication decisions on claims that had
actually been processed correctly; (12) to help ensure payment accuracy,
DOD requires its contractors to use ClaimCheck, a commercial software
program designed to ensure that professional providers are appropriately
paid for services rendered; (13) ClaimCheck's use resulted in changes to
only 3.5 percent of claims in fiscal year 1998 and saved about $53
million; (14) nonetheless, some providers complain about its use because
ClaimCheck's review criteria are not published and available to them;
(15) without this information, they expressed doubt that the criteria
comply with industry claims review standards; (16) GAO found that,
although ClaimCheck's review criteria are based on industry standards,
its use has resulted in some inappropriate denials to TRICARE claims;
and (17) these errors occurred because DOD was slow to direct
contractors to incorporate TRICARE policy changes into their claims
processing systems.

--------------------------- Indexing Terms -----------------------------

 REPORTNUM:  HEHS-99-128
     TITLE:  Defense Health Care: Claims Processing Improvements Are
	     Underway but Further Enhancements Are Needed
      DATE:  08/23/1999
   SUBJECT:  Medical expense claims
	     Health insurance
	     Military benefits claims
	     Claims processing
	     Late payments
	     Department of Defense contractors
	     Contractor payments
	     Performance measures
	     Managed health care
	     Internal controls
IDENTIFIER:  DOD TRICARE Program
	     DOD TRICARE ClaimCheck Software

******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO report.  Delineations within the text indicating chapter **
** titles, headings, and bullets are preserved.  Major          **
** divisions and subdivisions of the text, such as Chapters,    **
** Sections, and Appendixes, are identified by double and       **
** single lines.  The numbers on the right end of these lines   **
** indicate the position of each of the subsections in the      **
** document outline.  These numbers do NOT correspond with the  **
** page numbers of the printed product.                         **
**                                                              **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced.  Tables are included, but    **
** may not resemble those in the printed version.               **
**                                                              **
** Please see the PDF (Portable Document Format) file, when     **
** available, for a complete electronic file of the printed     **
** document's contents.                                         **
**                                                              **
** A printed copy of this report may be obtained from the GAO   **
** Document Distribution Center.  For further details, please   **
** send an e-mail message to:                                   **
**                                                              **
**                                            **
**                                                              **
** with the message 'info' in the body.                         **
******************************************************************

Cover
================================================================ COVER

Report to the Chairman and Ranking Minority Member, Subcommittee on
Military Personnel, Committee on Armed Services, and the Honorable
Charles W.  Stenholm, House of Representatives

August 1999

DEFENSE HEALTH CARE - CLAIMS
PROCESSING IMPROVEMENTS ARE UNDER
WAY BUT FURTHER ENHANCEMENTS ARE
NEEDED

GAO/HEHS-99-128

Claims Processing Improvements Needed

(101627)

Abbreviations
=============================================================== ABBREV

  AMA - American Medical Association
  CCI - Correct Coding Initiative
  CHAMPUS - Civilian Health and Medical Program of the Uniformed
     Services
  CPT - Physicians' Current Procedural Terminology
  DOD - Department of Defense
  E&M - evaluation and management
  HBOC - McKesson/HBO & Company
  HCFA - Health Care Financing Administration
  HCSR - health care service record
  HHS - Department of Health and Human Services
  MCSC - managed care support contractor
  TCC - TRICARE ClaimCheck
  TMA - TRICARE Management Activity

Letter
=============================================================== LETTER

B-282389

August 23, 1999

The Honorable Steve Buyer
Chairman
The Honorable Neil Abercrombie
Ranking Minority Member
Subcommittee on Military Personnel
Committee on Armed Services
House of Representatives

The Honorable Charles W.  Stenholm
House of Representatives

Today, about 8.2 million active duty personnel, their dependents, and
retirees are eligible to receive health care through the $15.6
billion-per-year military health system.  Medical care is provided by
Department of Defense (DOD) personnel in military facilities and
through civilian contractors.  Civilian-provided care requires that
providers or beneficiaries submit claims to DOD contractors who, in
turn, adjudicate the claim and pay according to established rules and
policies. 

Concerns about claims processing timeliness and accuracy have plagued
the military health care system since the advent of TRICARE, DOD's
managed care program.  During the 1-year period ending June 1998, the
contractors we reviewed processed approximately 19 million claims
worth over $1.7 billion.  Health care providers and beneficiaries
have frequently complained that claims were being processed too
slowly and that many errors were occurring.  While DOD contractors
have acknowledged that they experienced problems processing claims in
a timely manner during the start-up phase of health care delivery,
they contend that they are now meeting standards.  In response to
your request, we evaluated the timeliness and accuracy of claims
processing.  We also evaluated the effectiveness of DOD's use of
ClaimCheck\TM , a claim editing software package DOD requires its
contractors to use.  We performed our work between April 1998 and
June 1999 in accordance with generally accepted government auditing
standards.  For a further description of our scope and methodology,
see appendix I. 

   RESULTS IN BRIEF
------------------------------------------------------------ Letter :1

Between July 1, 1997, and June 30, 1998, DOD's contractors processed
86 percent of claims (or 16 million) within 21 days.  This met DOD's
timeliness standard of processing 75 percent of claims within 21
days.  Even so, nearly 3 million claims took more than 21 days to
process, which prompted complaints from some providers and
beneficiaries about what they considered to be payment delays.  DOD
has several initiatives under way to improve timeliness, including
adopting the payment and penalty standards used by the Medicare
program.  If these standards are properly implemented and met by
contractors, they should help reduce providers' complaints. 

While DOD adequately measures contractors' performance in claims
processing timeliness, it does not know the extent to which
contractors are accurately paying claims.  Less than half the claims
are subject to its payment accuracy audit, and the methodology used
to calculate the payment error rate is statistically invalid.  All
contractors experienced problems with payment accuracy when they
began processing TRICARE claims, often because they did not have
enough time to adequately prepare to administer the program. 
Although contractors addressed these problems, they acknowledged that
many factors affect the accuracy of claims processing--primarily the
complexity of the program, compounded by numerous program changes. 
We also found that some claims processing problems were due to
mistakes made by providers and beneficiaries when filing their
claims.  Furthermore, because they do not always understand the
program, providers and beneficiaries sometimes complain about
adjudication decisions on claims that had actually been processed
correctly. 

To help ensure payment accuracy, DOD requires its contractors to use
ClaimCheck\TM , a commercial software program designed to ensure that
professional providers are appropriately paid for services rendered. 
ClaimCheck\TM 's use resulted in changes to only 3.5 percent of
professional claims in fiscal year 1998 and saved over $53 million. 
Nonetheless, some providers complain about its use because
ClaimCheck\TM 's review criteria are not published and available to
them.  Without this information, they expressed doubt that the
criteria comply with industry claims review standards.  We found
that, although ClaimCheck\TM 's review criteria are based on industry
standards, its use has resulted in some inappropriate denials to
TRICARE claims.  These errors occurred because DOD was slow to direct
contractors to incorporate TRICARE policy changes into their claims
processing systems.  This report makes a number of recommendations to
the Secretary of Defense to improve claims processing timeliness and
accuracy. 

   BACKGROUND
------------------------------------------------------------ Letter :2

DOD's primary medical mission is to maintain the health of 1.6
million active duty service personnel and to provide them with health
care during military operations.  DOD also offers health care to 6.6
million non-active duty beneficiaries, including dependents of active
duty personnel, military retirees, and dependents of retirees.  Under
TRICARE, care is provided in military-operated hospitals and clinics
worldwide and is supplemented by civilian providers.\1

TRICARE is a triple-option benefit program designed to give
beneficiaries a choice among a health maintenance organization, a
preferred provider organization, and a fee-for-service benefit.  The
health maintenance organization option, called TRICARE Prime, is the
only option for which beneficiaries must enroll.  TRICARE Extra is
the preferred provider organization option, and TRICARE Standard is
the fee-for-service option.  Contractors, who are referred to as
managed care support contractors (MCSC), must create networks of
providers for the Prime and Extra options.  During network
development MCSCs recruit providers, negotiate reimbursement rates,
and verify professional credentials. 

TRICARE is organized geographically into 11 health care regions
administered by 5 MCSCs.  The MCSCs' many responsibilities include
processing claims, providing customer service, and developing and
maintaining an adequate network of civilian providers.  While the
MCSCs are ultimately responsible for claims processing, all of the
MCSCs have subcontracted with one of two companies to process claims,
as shown in table 1. 

                          Table 1
          
              TRICARE MCSCs and Subcontractors
             Responsible for Claims Processing

                                        MCSCs'
                    TRICARE MCSC        subcontractors
------------------  ------------------  ------------------
Regions included in our review
----------------------------------------------------------
Northwest           Foundation Health   Wisconsin
                    Federal Services,   Physicians Service
                    Inc.

Southwest           Foundation Health   Wisconsin
                    Federal Services,   Physicians Service
                    Inc.

Southern            Foundation Health   Palmetto
California, Golden  Federal Services,   Government
Gate, and Hawaii-   Inc.                Benefits
Pacific                                 Administrators

Southeast and Gulf  Humana Military     Palmetto
South               Healthcare          Government
                    Services, Inc.      Benefits
                                        Administrators

Central             TriWest Healthcare  Palmetto
                    Alliance, Inc.      Government
                                        Benefits
                                        Administrators

Regions not included in our review\a
----------------------------------------------------------
Northeast           Sierra Military     Palmetto
                    Health Services     Government
                                        Benefits
                                        Administrators

Mid-Atlantic and    Anthem Alliance     Palmetto
Heartland           for Health, Inc.    Government
                                        Benefits
                                        Administrators
----------------------------------------------------------
\a These regions were not included because they did not have at least
1 year of claims processing experience as of July 1998. 

Claims processing involves timely, accurate, and appropriate
adjudication of health care claims based on TRICARE rules and
policies.  Claims processing tasks include receipt of the claim form,
data entry, claims adjudication, and claim payment or denial. 

DOD requires MCSCs to meet specific timeliness and accuracy standards
for claims processing.  MCSCs must process 75 percent of claims
within 21 days.  This standard applies to all claims, even when MCSCs
must obtain additional information to process them.  DOD verifies
whether MCSCs are meeting timeliness standards through its database
of health care service records (HCSR), which are the final records of
the claims.  DOD requires the MCSCs to send an electronic HCSR to DOD
for each claim processed to completion.  DOD also requires MCSCs to
maintain a 98-percent payment accuracy rate and a 97-percent data
input accuracy rate.  DOD conducts quarterly external audits to
monitor whether MCSCs meet these standards. 

DOD requires MCSCs to use ClaimCheck\TM , a commercial claims editing
software package that performs a pre-payment review of professional
claims and helps prevent overpayment by analyzing relationships
between medical procedure codes.  For example, ClaimCheck\TM contains
review criteria, known as edits, to prevent unbundling, a process
whereby providers use two or more procedure codes to describe a
service when a single, more comprehensive code exists.  Generally,
providers receive higher reimbursement for unbundled codes compared
to a single, comprehensive code.  The basic ClaimCheck\TM software
package contains approximately 5 million edits.  However, companies
that purchase ClaimCheck\TM may customize the edits to reflect their
plan's benefit structure.  DOD purchased ClaimCheck\TM software in
March 1994 and had it customized to edit for TRICARE's benefit
structure.  DOD refers to its customized version as TRICARE
ClaimCheck (TCC).  DOD does not require the use of TCC for
anesthesia, pharmacy, physical therapy, or institutional claims
(except ambulatory surgery facility claims), or for adjustments to
claims that were processed prior to the use of TCC.  As a result, TCC
affects only about 60 percent of claims. 

In response to beneficiary and provider concerns, DOD intends to make
changes to future TRICARE contracts that could improve the timeliness
and accuracy of claims processing.  However, because the next round
of contracts is not anticipated to be awarded until 2001, DOD
recently decided to implement selected changes in advance by amending
current contracts.  This effort, called work simplification, involves
adopting timeliness standards similar to Medicare's and changing the
way incomplete claims are handled.  In addition, DOD has contracted
with a consulting firm to evaluate its claims processing procedures
and make recommendations for improvement.  The consultant's report is
due by October 1999. 

--------------------
\1 DOD previously provided health care under the Civilian Health and
Medical Program of the Uniformed Services (CHAMPUS), a
fee-for-service program. 

   MCSCS ARE MEETING DOD'S CLAIMS
   PROCESSING TIMELINESS STANDARD,
   BUT COMPLAINTS ABOUT SLOW
   PAYMENTS CONTINUE
------------------------------------------------------------ Letter :3

Each of the MCSCs experienced problems with claims processing
timeliness during the early months of health care delivery.  This was
partially due to a higher-than-expected claims volume--for example,
two contracts received 40 to 50 percent more claims than anticipated. 
As a result, the claims processing subcontractor had to recruit,
hire, and train additional staff--a process that took approximately 4
months.  During this time, the backlog of incoming claims continued
to grow. 

Claims processing timeliness has improved as MCSCs have gained more
experience with the TRICARE program.  We analyzed over 19 million
claim records and determined that during the period between July 1,
1997, and June 30, 1998, MCSCs met DOD's contractual timeliness
standard by processing 86 percent of claims within 21 days.  Despite
this, nearly 3 million claims took longer than 21 days to process and
therefore some providers and beneficiaries experienced what they
considered to be payment delays. 

      TIMELINESS STANDARDS MET
      OVERALL, BUT DIFFERENCES
      EXIST BY CLAIM
      CHARACTERISTICS
---------------------------------------------------------- Letter :3.1

Processing time was affected by characteristics such as type of claim
(professional, pharmacy, or institutional), submission method
(electronic or paper), and amount allowed for payment.\2 We found
that institutional claims did not meet the standard; however, MCSCs
did meet the standard overall because higher-than-required
percentages of claims in other categories were paid in less than 21
days.  To improve claims processing timeliness in the future, DOD has
proposed several initiatives, including the adoption of some Medicare
standards. 

Tables 2 through 4 display various statistics by claim category.  As
table 2 shows, professional and pharmacy claims met the standard, but
only 66 percent of institutional claims were processed within 21
days.  Pharmacy claims are usually for small dollar amounts, as are
many professional claims.  High-dollar claims, often from hospitals,
are usually the most complicated and often require medical review,
adding to processing time.  For example, as shown in table 3, only 30
percent of claims over $10,000 were paid within 21 days.  Because
institutional claims comprise only 4 percent of all claims, MCSCs
were still able to meet standards overall.  And even though
professional claims met the standard, they comprise 83 percent of the
claims that took more than 21 days to process, which may explain why
some providers complain about delinquent payments. 

                          Table 2
          
            Processing Time by Category of Claim

                          Claims processed
          ------------------------------------------------
             0 to 21 days     More than 21 days
          ------------------  ------------------
Category
of                  Percenta            Percenta       All
claims      Number        ge    Number        ge    claims
--------  --------  --------  --------  --------  --------
Professi  9,480,98        81  2,265,09        19  11,746,0
 onal            3                   3                  76
Pharmacy  6,506,86        97   215,252         3  6,722,11
                 7                                       9
Institut   473,964        66   243,382        34   717,346
 ional
==========================================================
All       16,461,8        86  2,723,72        14  19,185,5
 claims         14                   7                  41
----------------------------------------------------------

                          Table 3
          
              Processing Time by Cost of Claim

                          Claims processed
          ------------------------------------------------
             0 to 21 days     More than 21 days
          ------------------  ------------------
Cost of             Percenta            Percenta       All
claim       Number        ge    Number        ge  claims\a
--------  --------  --------  --------  --------  --------
Less      13,913,0        89  1,750,31        11  15,663,3
 than           61                   1                  72
 $100
$100 to   2,335,39        75   781,886        25  3,117,27
 $999            1                                       7
$1,000     205,395        54   178,397        46   383,792
 to
 $9,999
$10,000      5,149        30    12,120        70    17,269
 or more
==========================================================
All       16,458,9        86  2,722,71        14  19,181,7
 claims         96                   4                  10
----------------------------------------------------------
\a The total number of claims for this table does not match that of
table 2 because it excludes claims with missing cost data. 

The method of submission--paper or electronic--also affected
timeliness.  Forty-three percent of all claims were submitted
electronically, three-fourths of which were pharmacy claims.  As
shown in table 4, we found that 95 percent of electronic claims met
the timeliness standard compared with 79 percent of paper claims. 
Institutional and professional claims can be harder to submit
electronically because they sometimes require additional
documentation that cannot be submitted with the electronic form. 
Furthermore, providers may choose not to invest in the software
needed to submit TRICARE claims electronically if TRICARE is a small
percentage of their business. 

                          Table 4
          
          Electronic and Paper Claims Processed in
                21 Days by Category of Claim

                    Claims processed in 21 days
          ------------------------------------------------
             Paper claims     Electronic claims
          ------------------  ------------------
Category
of                  Percenta            Percenta       All
claims      Number        ge    Number        ge  claims\a
--------  --------  --------  --------  --------  --------
Professi  7,829,36        80  1,651,61        87  9,480,98
 onal            8                   4                   2
Pharmacy   548,386        84  5,958,48        98  6,506,86
                                     1                   7
Institut   332,525        65   141,439        70   473,964
 ional
==========================================================
All       8,710,27        79  7,751,53        95  16,461,8
 claims          9                   4                  13
----------------------------------------------------------
\a The total number of claims for this table does not match that of
table 2 because it excludes claims for which the method of submission
was unknown and all claims that took longer than 21 days to process. 

We also analyzed the effect on timeliness when MCSCs needed to obtain
information from other health insurers or liable third parties before
processing claims to completion.\3 We found that compensation from
other health insurers was obtained for about 10 percent of claims and
that MCSCs met the 75-percent timeliness standard even when they had
to obtain this information from the insurers.  In contrast, the
timeliness standard was not met for claims that involved third-party
liability.  There were fewer than 3,000 of these claims in the 19
million that we evaluated.  Although few claims were actually found
to involve third-party liability, many more were investigated to
determine whether they fell into this category.  These investigations
are one reason claims may be paid after 21 days. 

--------------------
\2 Professional claims represent care rendered by physicians and
other health care providers, such as physical therapists.  Most
institutional claims represent care provided by hospitals.  Pharmacy
claims are claims for prescription drugs. 

\3 When a beneficiary has additional health insurance, TRICARE is
usually the secondary payer.  The only time TRICARE is not the
secondary payer is when Medicaid is involved, or if the beneficiary
has a health insurance policy that is specifically designated as a
TRICARE supplemental policy.  Third-party liability claims involve
treatment for injury or illness resulting from circumstances that
created a legal liability for a third party to pay damages for the
care. 

      EFFORTS UNDER WAY TO IMPROVE
      TIMELINESS
---------------------------------------------------------- Letter :3.2

Although MCSCs have been meeting timeliness standards overall,
beneficiaries and providers have expressed concerns about claims
processing timeliness.  DOD and MCSC officials have identified
several initiatives they believe have the potential to improve claims
processing timeliness.  One of the proposed changes will adopt
revised timeliness standards similar to those used by Medicare.\4
Under these revised standards, MCSCs will be required to pay 95
percent of complete claims within 30 days and 100 percent of them
within 60 days.  MCSCs will be required to pay interest on claims
taking longer than 30 days to process to completion.  As shown in
table 5, MCSCs are already close to meeting this standard because
they processed 92 percent of claims within 30 days.  Although DOD
expects to implement these revised standards in September 1999, they
will require changes to each MCSC contract--a time-consuming process
that could result in delays.  Nonetheless, it is important that DOD
follow through with this initiative, which will help improve
providers' view of TRICARE by mirroring a more familiar program. 

                          Table 5
          
              Number of Days to Process Claims

                                                Cumulative
Number of          Number of     Percentage     percentage
days                  claims      processed      processed
-------------  -------------  -------------  -------------
0 to 13           13,533,876             71             71
14 to 21           2,927,938             15             86
22 to 30           1,146,999              6             92
31 to 60           1,108,031              6             98
61 or more           468,697              2            100
==========================================================
All claims        19,185,541            100
----------------------------------------------------------
Another of the proposed changes, which was implemented in June 1999,
allows MCSCs to return incomplete claims for needed information
without counting them against the timeliness standard.  Previously,
DOD required claims processors to permit claimants 35 days to provide
the information needed to process their claim.  If information was
not received within this time, the claim was denied and would need to
be resubmitted in order to be processed.  This requirement
automatically forced some claims to exceed DOD's 21-day timeliness
standard. 

In addition to DOD's proposed changes, impending changes in industry
standards should also improve timeliness by making it easier for
providers to submit claims electronically.  The Health Insurance
Portability and Accountability Act of 1996 (P.L.  104-191) requires
the industrywide adoption of uniform standards for electronic
transactions, including claims filing.  Uniform standards for
electronic filing will enable providers to submit claims for any
health insurance plan in the same format, eliminating the need for
plan-specific software.  The Department of Health and Human Services
(HHS), the agency responsible for implementing the act, reported that
this effort should be under way in late 1999. 

--------------------
\4 This proposal is contained in draft legislation for DOD's fiscal
year 2000 authorization bill. 

   EXTENT OF CLAIMS PROCESSING
   ACCURACY IS UNKNOWN
------------------------------------------------------------ Letter :4

DOD does not know the extent to which MCSCs are meeting contractual
requirements for claims processing payment accuracy because its
primary assessment tool yields statistically invalid results.  As
with timeliness, all MCSCs experienced problems with claims
processing accuracy during the early months of health care delivery
and subsequently improved.  However, even when problems are
identified and corrected, several factors--such as TRICARE's complex
program structure and frequent program changes--add to the difficulty
of processing claims accurately.  TRICARE's complex rules can also
cause providers and beneficiaries to misunderstand requirements and
submit incorrect information. 

      AUDIT METHODS DO NOT
      ADEQUATELY MEASURE
      PROCESSING ACCURACY
---------------------------------------------------------- Letter :4.1

A DOD contractor conducts quarterly audits of claims processing
accuracy for each TRICARE contract to assess the rate of incorrect
payments and data input errors.  The payment error rate, which is a
combined rate for both denied and paid claims, is computed by adding
the absolute value of underpayments and overpayments and dividing
this amount by the total billed charges for the sampled claims.  The
data input error rate, called the occurrence error rate, is based on
the total number of errors found in the audited claims, divided by
the total number of data fields.  DOD has established standards of 2
percent for payment error rates and 3 percent for occurrence error
rates.  DOD gives financial rewards to MCSCs who achieve a payment
error rate of 1 percent or less, and penalizes them for a rate of 4
percent and above.  Likewise, DOD financially rewards contractors if
their occurrence error rate is 2.4 percent or less and penalizes them
if it is 5 percent or more. 

We identified three problems with DOD's method for determining claim
payment error rates.  First, more than half of the claims are
excluded from the audit process.  DOD does not sample from claims
under $100 for the payment audit because they represent a relatively
small percentage (about 12 percent) of the dollars paid on TRICARE
claims.  However, about 60 percent of all claims fall into this
category and therefore are not subject to this quality assurance
procedure.  Including these claims in the audit would better describe
the quality of MCSCs' claims processing operations because the error
rate would apply to the entire population of claims, regardless of
claim amount. 

Second, the calculation of the payment error rate is not properly
adjusted to account for DOD's stratified sampling and, as a result,
its error rates are statistically invalid.  DOD samples claims from
defined dollar ranges of claim payments.  Each range contains a
different number of claims.  However, DOD does not use statistical
adjustments in its error rate calculation to account for these
differences.\5 As a result, DOD's calculated error rate may be higher
or lower than the actual payment error rate.  Table 6 illustrates the
effect these statistical adjustments would have on the error rates
for the quarterly audits we reviewed for the MCSCs included in our
evaluation.  The third column contains the error rate as computed
with DOD's methodology.  The fourth column shows the error rate
recomputed with statistical adjustments.  A comparison (fifth column)
shows that all but one rate changed. 

                          Table 6
          
            Effect of Statistical Adjustment on
             Error Rates for the Most Recently
          Available Finalized Audits for a 1-Year
                           Period

                                     Corrected
                                    error rate
                                          with  Comparison
                     DOD error     statistical     between
Contra                  rate\b   adjustments\c       rates
ct\a       Quarter   (percent)       (percent)   (percent)
------  ----------  ----------  --------------  ----------
A                1         5.7             5.7         0.0
                 2         5.5             7.8         2.3
                 3        11.3             7.3        -4.0
                 4         4.7             3.5        -1.2
B                1         5.0             4.9        -0.1
                 2         4.0             3.7        -0.3
                 3         6.1             4.1        -2.0
                 4         6.1             3.6        -2.5
C                1         1.1             2.5         1.4
                 2         3.2             3.8         0.6
                 3         1.5             2.3         0.8
                 4         1.4             2.6         1.2
D                1         3.6             3.2        -0.4
                 2         4.6             5.0         0.4
                 3         3.1             3.4         0.3
                 4         3.7             3.5        -0.2
E                1         4.6             4.7         0.1
                 2         3.0             3.9         0.9
                 3         3.2             2.9        -0.3
----------------------------------------------------------
Note:  The earliest audit began in November 1996 and the latest ended
in December 1997.  For one of the TRICARE contracts, only three
finalized audits were available. 

\a The letters in this column represent five contracts for the three
MCSCs we reviewed. 

\b DOD audit reports. 

\c GAO calculations based on the same data used in DOD's audit
reports. 

Third, DOD inappropriately uses billed charges as the denominator to
calculate payment error rates instead of actual payment amounts. 
Because providers' billed charges are typically much higher than the
corresponding payment amounts, DOD's practice of using billed charges
instead of paid amounts for error calculations results in payment
error rates that are artificially low.  For example, suppose a claim
was billed at $500, and the amount paid on the claim was $300.\6
During the audit, a $50 payment error was discovered.  Calculating
the error rate with the billed charges, as DOD does, results in a
10-percent error rate.  Calculating it using the paid amount results
in a 17-percent error rate.  We found that paid charges were also
used in calculating payment error rates for some commercial industry
audits as well as in audits of Medicare claims conducted by HHS'
Inspector General.  A common method used in industry audits for
calculating this type of payment error is to divide the total dollars
in error by the total dollars actually paid.  This calculation is
illustrated in the fourth column of table 7.  These error rates are
3.6 to 12.7 percentage points higher than DOD's calculated rates. 

                          Table 7
          
            Comparison Between Quarterly Payment
           Error Rates Calculated by Contract for
           the Most Recently Available Finalized
                 Audits For a 1-Year Period

                                 Statistically
                                accurate error
                                 rate based on  Comparison
                     DOD error  actual dollars     between
Contra                  rate\b          paid\c       rates
ct\a       Quarter   (percent)       (percent)   (percent)
------  ----------  ----------  --------------  ----------
A                1         5.7            13.5         7.8
                 2         5.5            18.2        12.7
                 3        11.3            17.0         5.7
                 4         4.7             8.9         4.2
B                1         5.0            14.3         9.3
                 2         4.0            10.2         6.2
                 3         6.1            11.5         5.4
                 4         6.1            10.6         4.5
C                1         1.0             5.2         4.1
                 2         3.2             8.5         5.3
                 3         1.5             5.3         3.8
                 4         1.4             6.3         4.9
D                1         3.6             8.6         5.0
                 2         4.6            14.0         9.4
                 3         3.1             9.0         5.9
                 4         3.7             9.4         5.7
E                1         4.6            11.8         7.2
                 2         3.0             9.7         6.7
                 3         3.2             6.8         3.6
----------------------------------------------------------
Note:  The earliest audit began in November 1996 and the latest ended
in December 1997.  For one of the TRICARE contracts, only three
finalized audits were available. 

\a The letters in this column represent five contracts for the three
MCSCs we reviewed. 

\b DOD audit reports. 

\c GAO calculations based on the same data used in DOD's audit
reports. 

Beyond these technical weaknesses, DOD's measures for payment
accuracy and data input, or occurrence, accuracy give only a partial
picture of MCSCs' performance.  These error rates provide some
information on the extent of error but not on the percentage of
claims affected.  Therefore, a useful companion measure, which could
easily be calculated from the same data, is an error rate
representing the percentage of claims processed incorrectly.  For
payment error, this calculation is shown in the fourth column of
table 8.  As illustrated by the first entry for Contract A, when the
error rate is computed correctly using paid amounts, the error rate
is 13.5 percent.  When we calculated the corresponding percentage of
claims affected, the error rate is 16.5 percent.  Together, these two
measures--the statistically accurate error rate based on actual
dollars paid and the corresponding percentage of claims processed
incorrectly--provide a more complete picture of payment errors. 
Although we did not find methodological flaws in the occurrence
audit, a corresponding measure of the percentage of claims affected
could also be calculated for it.  Collectively, these measures, which
are also used in some industry audits, would give a more
comprehensive indication of the quality of MCSCs' claims processing
performance. 

                          Table 8
          
              Error Rates Calculated with GAO-
          Proposed Measures for the Most Recently
          Available Finalized Audits for a 1-Year
                           Period

                           Statistically
                     accurate error rate    Sampled claims
                         based on actual         processed
Contra                    dollars paid\b       incorrectly
ct\a       Quarter             (percent)         (percent)
------  ----------  --------------------  ----------------
A                1                  13.5              16.5
                 2                  18.2              25.3
                 3                  17.0              15.0
                 4                   8.9              14.2
B                1                  14.3              15.6
                 2                  10.2              14.4
                 3                  11.5              17.6
                 4                  10.6              16.1
C                1                   5.2              14.3
                 2                   8.5              14.7
                 3                   5.3              13.4
                 4                   6.3              10.0
D                1                   8.6               8.0
                 2                  14.0              11.8
                 3                   9.0              10.6
                 4                   9.4              10.4
E                1                  11.8              11.7
                 2                   9.7              11.1
                 3                   6.8               8.2
----------------------------------------------------------
Note:  The earliest audit began in November 1996 and the latest ended
in December 1997.  For one of the TRICARE contracts, only three
finalized audits were available. 

\a The letters in this column represent five contracts for the three
MCSCs we reviewed. 

\b GAO calculations based on the same data used in DOD's audit
reports. 

--------------------
\5 These adjustments, called weights, are necessary to correct for
the fact that some ranges may be over-represented in the sample while
others may be under-represented. 

\6 This example is based on TRICARE allowable charges being about 60
percent of billed charges on average. 

      INADEQUATE CONTRACT
      TRANSITION TIME CONTRIBUTED
      TO EARLY CLAIM DIFFICULTIES
---------------------------------------------------------- Letter :4.2

A major factor contributing to early claims processing inaccuracies
was the short transition period allowed for MCSCs to prepare for
delivering health care.  For its initial TRICARE contracts, DOD tried
to recover time lost in procurement delays by reducing the scheduled
8- to 9-month transition period to 6 months.  Previously, we reported
that DOD had experienced serious problems with contractors' inability
to process claims by the start-work date of the contract because the
6-month transition period was too short.\7 In August 1995, we
recommended that DOD adhere to the 8- to 9-month scheduled transition
period and discontinue reducing such periods.\8 However, DOD did not
extend the transition period to 9 months, and MCSCs continued to
experience problems completing the preparatory tasks needed to
deliver health care and accurately process claims by the health care
delivery start date.  DOD officials have recently stated that,
because MCSCs have been struggling to fully prepare for health care
delivery, they now believe a longer transition period--9 to 12
months--is needed. 

During the transition period, MCSCs are required to build complete
networks of physicians and others for providing medical care. 
Typically, these networks consist of thousands of providers and
hundreds of hospitals and pharmacies, and the network has to be in
place 30 to 60 days prior to the start of health care delivery. 
MCSCs generally did not assemble a complete network in the allotted
time.  In addition to recruiting providers, DOD required MCSCs to
conduct an extensive verification of providers' credentials, a
process that sometimes took months to complete.  Because health care
delivery began before providers' professional credentials could be
verified and entered into the claims processing system, some claims
were erroneously paid as non-network.  These errors sometimes took
months to rectify.  Not only did this irritate providers, but it also
created additional, unnecessary work for the claims
processors--especially since the vast majority of providers were
eventually certified to provide care. 

--------------------
\7 CHAMPUS Has Improved Its Methods for Procuring and Monitoring
Fiscal Intermediary Services to Process Medical Claims
(GAO/HRD-85-56, Aug.  23, 1985); Implementation of the CHAMPUS Reform
Initiative (GAO/T-HRD-89-25, June 5, 1989). 

\8 Despite TRICARE Procurement Improvements, Problems Remain
(GAO/HEHS-95-142, Aug.  3, 1995). 

      TRICARE'S COMPLEXITY AND
      FREQUENT PROGRAM CHANGES
      AFFECT ACCURACY
---------------------------------------------------------- Letter :4.3

Many claims processing errors are caused by program complexities and
frequent changes.  MCSCs told us that, of the many programs they
administer--including Medicare and private plans--TRICARE is unique
and the most complicated, contributing to claims processing
difficulties.  The following features contribute to TRICARE's
complexity: 

  -- Each of TRICARE's three options has a different array of
     benefits, copayments, deductibles, and adjudication procedures. 
     For example, each option has different cost shares, provider
     payments, and authorization requirements, creating added
     difficulty in processing claims and increasing the potential for
     processing errors.  Sometimes, even within an option, different
     claims processing rules apply.  For example, a Prime beneficiary
     could elect to use a provider without authorization and pay a
     higher cost share for the care. 

  -- For the Prime and Extra options, it is difficult to maintain
     accurate provider reimbursement information because payment
     agreements are complicated and individual providers may belong
     to multiple practices with different agreements. 

  -- Claims submitted under the Standard option can be complex to
     process because providers can either accept TRICARE's allowable
     amount as payment in full or charge up to an additional 15
     percent on a claim-by-claim basis. 

  -- For each claim, MCSCs' subcontractors must connect with and rely
     on selected DOD databases to verify eligibility, deductibles,
     and enrollment.  MCSCs stated that this requirement complicates
     claims processing and increases the likelihood of errors.  In
     contrast, most private insurers maintain their own files for
     these purposes. 

  -- TRICARE is almost never the primary payer when other health
     insurance is involved.  Thus, MCSCs' subcontractors must
     understand the requirements of many other programs' benefit
     structures and obtain reimbursement information before a claim
     can be processed to completion. 

  -- TRICARE is subject to many special demonstration programs, such
     as TRICARE Prime Remote and TRICARE Senior Prime, which have
     different claims processing requirements.\9

TRICARE's frequent program changes further complicate claims
processing.  Program changes, which include changes to health care
benefits as well as administrative changes, are generally
communicated throughout the year in the form of contract
modifications.  As of October 1998, DOD had instructed the MCSCs we
reviewed to implement about 650 contract modificationsan average of
about 130 per contract since 1995.  DOD and subcontractor officials
stated that most contract modifications have an impact on claims
processing.  MCSCs stated that their ability to process claims
accurately is impeded because most changes affect claims processing
and require system reprogramming and testing as well as staff
retraining within a relatively short timegenerally a month or less. 

DOD's recently established work simplification initiative calls for
program benefit changes to be implemented on an annual basis, with 8
to 9 months of lead time provided prior to implementation.  In
addition, DOD plans to implement administrative changes on a
quarterly basis with the same amount of lead time as benefit
changes.\10 This should reduce claims processing errors resulting
from frequent program changes. 

--------------------
\9 The TRICARE Senior Prime program is a 3-year demonstration project
under which Medicare will reimburse DOD for care provided to
Medicare-eligible beneficiaries under the TRICARE Prime option.  The
TRICARE Prime Remote program provides medical care comparable to
coverage under the TRICARE Prime program to active duty members
assigned to remote locations. 

\10 A similar proposal is contained in draft legislation for DOD's
fiscal year 2000 authorization bill. 

      MCSCS ARE NOT RESPONSIBLE
      FOR ALL CLAIM ERRORS
---------------------------------------------------------- Letter :4.4

Although DOD and its MCSCs are responsible for the majority of claims
processing errors, about 16 percent of adjustments to claims were due
to filing errors.  If providers and their office staff do not
understand the TRICARE program, their claims may be submitted with
inaccurate or incomplete information.  After these claims are
processed to completion, the providers may disagree with the outcomes
and submit additional information.  Once this information is
provided, the claims must be reprocessed. 

MCSCs are required to conduct educational seminars and to publish
provider handbooks and newsletters communicating TRICARE issues,
including claims filing.  We found that MCSCs were providing training
seminars semiannually for their network providers and annually for
their non-network providers.  However, they told us that because
TRICARE is usually a small percentage of providers' businesses,
providers have little incentive to participate in educational
seminars or to read the many bulletins and updates to stay current on
the frequent program changes.  For example, in some urban areas
providers may accept patients from 20 different health insurers--and
need to understand all their requirements--with TRICARE often being a
small portion of their practices.  MCSCs stated that TRICARE is the
most complicated plan in which providers participate.  Consequently
some providers do not express an interest in learning about the
program until they have questions about their claims. 

Because beneficiaries and providers do not always understand the
TRICARE program, they may file their claims incorrectly or complain
about adjudication decisions on claims that have been processed
correctly.  For example, misunderstandings can arise when a covered
service is processed but no check or a smaller-than-expected check is
issued.  This could happen when annual deductibles have not been met,
and beneficiaries do not understand that they are responsible for
paying for the covered services.  This could also occur when other
health insurance has paid as much as TRICARE allows, but the provider
expects additional payment from TRICARE as the secondary carrier.  In
addition, because of negotiated discounts, providers are sometimes
paid less under TRICARE than under DOD's previous civilian health
program, CHAMPUS.  While these differences are the result of policy
changes and not processing errors, some providers may not recognize
this. 

   DOD'S SLOWNESS IN IMPLEMENTING
   POLICY CHANGES HAS LED TO
   COMPLAINTS ABOUT TCC
------------------------------------------------------------ Letter :5

TCC software, which is used to prevent overpayments on professional
claims, saved DOD over $53 million during fiscal year 1998.  While
providers have frequently complained about TCC determinations, TCC
determinations changed only a small percentage (3.5 percent) of
professional claims during this time.  Providers have also expressed
concern that they have no assurance that the software's edits comply
with industry standards.  We found that the basic product was
developed based on industry standards and that TCC--DOD's modified
version--essentially mirrors the standard commercial product. 
Nonetheless, in spite of its effectiveness, TCC inappropriately
denied procedures on some claims because DOD has been slow to direct
MCSCs to reflect policy changes affecting TCC outcomes in their
claims processing systems.  MCSCs also occasionally provided
incomplete and inaccurate information, which led providers to believe
they had no recourse over TCC outcomes. 

      CLAIMCHECK\TM IS USED BY
      MANY COMMERCIAL PLANS AND IS
      BASED ON INDUSTRY STANDARDS
---------------------------------------------------------- Letter :5.1

ClaimCheck\TM is a leader in the claim editing software industry and
has more than 200 customers nationwide, including the Department of
Veterans Affairs and over 60 percent of Blue Cross Blue Shield
carriers.  In October 1998, HHS' Health Care Financing Administration
(HCFA) started supplementing its Correct Coding Initiative (CCI)
edits with selected ClaimCheck\TM edits to prevent overpayments in
the Medicare program.\11

Despite ClaimCheck\TM 's general acceptance in the insurance
industry, the providers we spoke with expressed an overall concern
about commercial code-editing software.  They stated that because the
edits are not published and available to them, they have no way of
ensuring that the edits comply with the American Medical
Association's (AMA) Physicians' Current Procedural Terminology (CPT)
coding guidelines, which are the industry standard.  Officials of
McKesson/HBO & Company (HBOC), who market the software, stated that
its edits are based upon CPT guidelines published by the AMA as well
as guidelines published by HCFA and medical specialty societies.  In
addition, physicians retained by the HBOC Clinical Consulting Network
were involved in the development of ClaimCheck\TM and are also
involved in the yearly software updates.\12

ClaimCheck\TM can be modified to reflect any health care plan's
benefit structure and reimbursement policies.  However, because
purchasers of such software can customize the edits, some providers
argue that they have no assurances that such modifications comply
with industry standards.  We found that TCC essentially mirrors the
commercial product because DOD has made only 12 customizations to the
software to reflect its benefit structure and reimbursement policies. 
DOD's customizations are described in appendix III.  Furthermore,
according to DOD officials, DOD centrally directs all TCC
modifications, and MCSCs cannot independently customize it. 

--------------------
\11 The CCI was developed by Administar specifically for Medicare to
help reduce provider overpayments. 

\12 HBOC's Clinical Consulting Network, which currently consists of
more than 180 members, represents a cross-section of physicians with
extensive clinical practice, academic, and medical management
experience. 

      DOD HAS BEEN SLOW TO MAKE
      POLICY CHANGES AFFECTING TCC
      DETERMINATIONS
---------------------------------------------------------- Letter :5.2

MCSCs were unanimous that the biggest problem with TCC was the length
of time it took for DOD to direct implementation of changes to
reimbursement policies.  Most program changes, including those
affecting TCC, must be communicated and implemented through contract
modifications.  Policy changes can take a long time to issue because
they must be drafted and priced, sent to MCSCs for comment, and then
finalized and issued.  Additional time is also needed for
implementation. 

DOD's decision to reimburse dermatologists for surgical pathology
provides an example of this problem.\13 In April 1996--early into the
implementation of TCC--DOD realized that the software's edits
resulted in denials to dermatologists for surgical pathology
procedures.  Initially, DOD's policy supported this determination,
but DOD subsequently decided that, unlike other providers,
dermatologists were qualified to perform surgical pathology and
should be reimbursed accordingly.  Because ClaimCheck\TM 's auditing
logic does not accommodate physician specialties, this change had to
be accommodated within the MCSCs' claims processing systems in order
to prevent inappropriate TCC denials.  However, it took DOD almost 2
years to finalize the modification and provide it to MCSCs.  One MCSC
stated that dermatologists left its network solely because of DOD's
inability to react quickly to this needed change. 

--------------------
\13 Surgical pathology is the gross and microscopic examination of
sampled tissue. 

      CONFUSION ABOUT THE ABILITY
      TO CHALLENGE TCC
      DETERMINATIONS ADDS TO
      PROVIDERS' FRUSTRATION
---------------------------------------------------------- Letter :5.3

Misleading communication regarding the proprietary nature of TCC
edits has fueled providers' frustration because they have sometimes
been unable to obtain explanations from MCSCs concerning the edits
that affected their claims.  However, HBOC officials stated that
ClaimCheck\TM is not a "black box" because purchasers receive
narrative descriptions on how every edit works.  DOD officials added
that providers can request and receive information on specific edits
from MCSCs.  MCSCs have on-line access to explanations about the
edits that result in the most frequent adjustments and denials.  HBOC
also provides a toll-free telephone number MCSCs can call to obtain
explanations for all other types of edits.  However, DOD officials
acknowledged that MCSCs have incorrectly told providers that the
edits cannot be explained to them.  To ensure that MCSCs share
appropriate information with health care providers, DOD stated that
it recently reminded them of the availability of the on-line
rationale and the toll-free hotline.  The extent to which DOD's
reminder addresses this problem remains to be seen. 

Providers' frustration was further compounded by DOD's and MCSCs'
poor communication regarding the available recourse over TCC
determinations.  As part of its allowable charge review process, DOD
has established a process for reconsidering claims denied by software
edits; however, this process has not been well communicated to
providers and beneficiaries.  As a result, many providers and
beneficiaries who questioned TCC determinations were incorrectly
informed that these determinations accurately reflected TRICARE
policy and that no recourse for review was available to them.  DOD's
Medical Director for the Southwest Region said that the failure to
inform providers of the TCC determination review process created
significant problems for the network, including some providers'
decisions to leave it. 

Beneficiaries' and providers' complaints that DOD and its MCSCs did
not make a review process available to them prompted the Congress to
mandate, in the Strom Thurmond National Defense Authorization Act for
Fiscal Year 1999 (P.L.  105-261), that DOD establish an appeals
process for TCC denials.  In response, DOD has proposed a two-level
appeals process for TCC determinations.  DOD has informed MCSCs that
they are to advise beneficiaries and providers that they can request
a TCC appeal if they are dissatisfied with a TCC determination.  If
beneficiaries or providers are dissatisfied with the results of the
initial review, DOD has proposed a second level of TCC appeals. 

      CLAIM-EDITING SOFTWARE MAY
      NOT BE REQUIRED IN FUTURE
      CONTRACTS
---------------------------------------------------------- Letter :5.4

In order to be less prescriptive and to allow MCSCs to use best
industry practices, DOD is considering eliminating the requirement
that MCSCs use TCC or any other claim-editing software from the next
round of TRICARE contracts.  DOD officials stated that, in the
future, interested companies would probably offer to use code-editing
software whether or not they are required to do so.  They would most
likely choose ClaimCheck\TM because it is the industry leader, and it
is already being used by current MCSCs for TRICARE as well as by many
other potential MCSCs for their commercial health care plans. 

DOD officials added that, even though MCSCs would be permitted to use
different code-editing software, the claim outcomes would be required
to accurately reflect the TRICARE benefit.  Because differences in
the types of software used and individual MCSC customization could
result in inconsistently processed claims, DOD will need to closely
monitor claim outcomes to ensure that MCSCs adhere to the TRICARE
benefit. 

   CONCLUSIONS
------------------------------------------------------------ Letter :6

MCSCs are meeting DOD's timeliness standard for processing claims. 
However, the overall timeliness measure masks weaker performance in
processing certain types of claims, such as those submitted by
hospitals and other institutions.  Furthermore, many providers and
beneficiaries continue to complain about slow claims payment, perhaps
because some 3 million claims took more than 21 days to process.  DOD
has proposed initiatives to improve claims processing timeliness. 
These initiatives include adopting timeliness standards similar to
Medicare's, paying interest on claims unresolved after 30 days, and
not including incomplete claims in measuring performance against the
timeliness standard.  These initiatives appear to be steps in the
right direction as they mirror standards in both Medicare and the
health insurance industry.  If these initiatives improve payment
timeliness, DOD will enhance TRICARE's image to providers and
encourage more confidence in the program. 

Although DOD attempts to assess claims processing accuracy, we found
limitations in its methodology, which currently yields statistically
invalid results.  It is imperative that DOD accurately measure
payment error rates to better identify and correct problems as well
as assess MCSCs' performance.  However, the TRICARE program
structure, with its many complexities, means that claims processing
difficulties are not always easily resolved.  Inappropriate claim
denials have sometimes been made because of DOD's slowness to direct
MCSCs to make policy changes.  In addition, impediments such as
inadequate startup time and frequent program changes can cause claims
processing errors.  Expediting the policy change process, providing
additional startup time, and consolidating program changes could help
improve claims processing accuracy. 

Overall, claims processing problems have caused some providers to
become disillusioned with the TRICARE program.  DOD and MCSCs are
taking steps to address these problems.  If these steps are not
successful, DOD could face increasing problems attracting the number
of civilian providers necessary to ensure that beneficiaries have
adequate access to health care. 

   RECOMMENDATIONS
------------------------------------------------------------ Letter :7

In order to better measure and improve claims processing accuracy,
the Secretary of Defense should direct the Assistant Secretary of
Defense for Health Affairs to do the following: 

  -- Restructure the methodology used for claims processing accuracy
     audits so that performance measures more accurately and
     completely reflect MCSCs' performance and are more comparable to
     those generally used in the industry.  This restructuring should
     include (1) ensuring that claims of all dollar amounts are
     subject to the payment accuracy audit, (2) ensuring that error
     rate computations are statistically accurate and meaningful, and
     (3) adding additional measures of program performance, such as
     the percentage of claims processed with errors. 

  -- Grant new MCSCs a longer transition period--9 to 12
     months--between contract award and the start of health care
     delivery. 

To ensure that needed program changes are made in a timely manner, we
recommend that the Secretary of Defense direct the Assistant
Secretary of Defense for Health Affairs to expedite the process used
to direct MCSCs to implement program changes.  To help eliminate
confusion resulting from frequent program changes, we also recommend
that the Secretary consolidate contract modifications and direct
MCSCs to implement them on a quarterly basis. 

   AGENCY COMMENTS AND OUR
   EVALUATION
------------------------------------------------------------ Letter :8

In commenting on a draft of this report, the Assistant Secretary of
Defense for Health Affairs stated that DOD concurs with the report's
findings regarding past problems associated with processing TRICARE
claims.  DOD also stated that the report is supportive of its efforts
to improve the accuracy and timeliness of claims payment and the
implementation of program changes.  In response to our
recommendations, DOD agreed to provide new MCSCs a longer transition
period between contract award and the start of health care delivery,
to expedite the process used to direct MCSCs to implement program
changes, and to consolidate contract modifications and direct MCSCs
to implement them on a quarterly basis.  However, DOD only partially
concurred with our recommendation that it restructure the methodology
used for claims processing accuracy audits. 

We recommended that DOD ensure that claims of all dollar amounts,
including those under $100, be subject to the payment accuracy audit. 
In response, DOD stated that because of the significant amount of
expense involved with auditing these small claims, the return on
investment would be very low and would not affect the overall impact
of errors.  In our opinion, the expense involved in sampling these
claims should not be prohibitive because the low variance in this
category (the size of errors can range only from 1 cent to $99.99)
means that it could be sampled at a much lower rate compared with the
higher-dollar claim categories.  In fact, when DOD recalculates the
required sample size, it may find the existing sample could be
redistributed to include low-dollar claims so that the number of
claims sampled overall remains the same.  While we agree that
including these claims may not result in a large financial effect on
the government, it is an important quality assurance procedure
because these low-dollar claims comprise 60 percent of the claims
paid and consequently affect a large number of beneficiaries and
providers.  Sampling claims under $100 is also important in
describing the quality of operations because the resulting error rate
would include the entire population of claims.  Surprisingly, despite
its concerns about the value of auditing low-dollar claims, DOD said
it would review its current quarterly sampling methodology to
determine the costs and benefits of reviewing claims of all dollar
amounts. 

DOD stated that there are other mechanisms in place to ensure payment
accuracy, such as internal quality assurance audits conducted by each
MCSC and on-site surveillance by TRICARE Management Activity
representatives.  However, while these mechanisms provide some useful
information, DOD does not use them to measure MCSC's performance
against contract standards. 

DOD also disagreed with our recommendation that it use paid amounts
rather than billed amounts to calculate payment error rates, stating
that while it might result in higher error rates, no additional
information would be gained.  Our point is not that the use of paid
charges results in a higher payment error rate, but that paid amounts
are a more logical and meaningful measure that will provide better
information on MCSCs' performance.  Payments under TRICARE are
usually based on a fee schedule or negotiated amounts, not billed
amounts.  Therefore, when computing payment error rates, using actual
amounts paid seems more appropriate and useful. 

---------------------------------------------------------- Letter :8.1

As agreed with your offices, we are sending copies of this report to
the Honorable William S.  Cohen, Secretary of Defense, and will make
copies available to others upon request.  Please contact me on (202)
512-7101 if you or your staff have any questions concerning this
report.  Staff contact and other contributors are listed inappendix
IV. 

Sincerely,

Stephen P.  Backhus
Director, Veterans' Affairs and
 Military Health Care Issues

SCOPE AND METHODOLOGY
=========================================================== Appendix I

To assess claims processing timeliness, we obtained a health care
service record (HCSR) file from the Department of Defense (DOD)
containing 19,185,541 records of completed claims that were processed
between July 1, 1997, and June 30, 1998, for the managed care support
contractors (MCSC) that had at least 1 year's experience in
processing claims as of July 1998.  Thus, we included claims from 8
of the 11 regions but did not include claims processed in the 3
regions that began health care delivery in 1998.  (See table 1 (page
4) for a list of the regions that were and were not included in our
analyses.  See appendix II for timeliness statistics on the 3 regions
that did not have at least 1 year's experience processing claims as
of July 1998.) The information for each claim represented the status
of the claim at the time we received it and did not contain all data
that may have been used to process the claim.  For example, if the
claim was adjusted multiple times, only the most recent adjustment
information was on the database.  In addition, while we did not
independently verify the accuracy of the data, we conducted
reliability tests to ensure the consistency of the information with
DOD's internal reports.  We also reviewed the computer programs used
to prepare their timeliness reports. 

To identify the time taken to process a claim, we used DOD's formula
for calculating the number of days between the date the claim was
filed and the date it was processed to completion.  We performed this
calculation for all claims and summarized the calculations for
several groups of claims.  These groups were claim category
(professional, pharmacy, and institutional), method of submission
(electronic or paper), amount allowed for payment, and whether other
health insurers or third parties were liable for health care costs. 
To identify DOD's activities to improve timeliness, we also met with
TRICARE Management Activity (TMA) officials to discuss the work
simplification initiatives relating to claims processing. 

We assessed DOD's process for determining claims processing accuracy
by analyzing the four most recently completed audit reports for each
of the TRICARE contracts we reviewed.\14 We gathered information from
officials at DOD and from its external auditor, Meridian Resource
Corporation, about the audit process, including methods used to draw
the samples and calculate the error rates.  We also acquired from DOD
both the audit reports and the corresponding sample data.  To
calculate sampling weights, we obtained the files containing
necessary data on the populations from which the samples were drawn. 
To ensure that the correct files were received, we replicated
findings on the audit reports from the data we received; however, we
did not verify the accuracy of the audit process itself. 

To assess the effect of contract modifications on claims processing,
we met with the TMA officials responsible for developing,
implementing, and monitoring them.  We also met with representatives
from MCSCs and their claims processing subcontractors to learn how
they were affected by contract modifications.  We obtained and
analyzed schedules of these modifications to TRICARE contracts to
determine their volume.  We obtained information from DOD on MCSCs'
responsibilities for provider education to assess their efforts to
teach correct claims filing.  We interviewed and obtained information
from each of the MCSCs to determine what efforts were under way to
educate providers and to identify the effect of provider education on
claims processing accuracy.  We also interviewed the claims
processing subcontractors, who sometimes assist the MCSCs with
education efforts. 

To assess the magnitude of filing errors, we obtained computerized
files from Wisconsin Physicians Service and Palmetto Government
Benefits Administrators, the two claims processing subcontractors. 
These files contained records of all adjustments to claims submitted
between July 1, 1997, and June 30, 1998, in the eight regions with at
least 1 year's experience in processing claims as of July 1998.  The
records identified whether an error(s) was made by the contractor or
by the person filing the claim. 

We met with officials of McKesson/HBO & Company (HBOC), the
distributors of ClaimCheck\TM , to discuss the development and
features of their claims editing software and to obtain statistics on
its market penetration.  To identify specific physician complaints
about TRICARE ClaimCheck (TCC), we reviewed extensive documentation
of physicians' complaints provided by various medical societies,
individual physician practices, and TMA.  We also interviewed
officials from the American Medical Association, the Texas Medical
Association, and the American Academy of Dermatology, who were
identified as having specific concerns about the software.  In
addition, we contacted individual physician practices, which were
referred to us by the various advocacy groups, to discuss their
concerns and to obtain supporting claim documentation.  To assess
whether physicians' complaints were valid, we met with DOD's TCC
policy officials to discuss the implementation and customization of
ClaimCheck\TM software for the TRICARE program.  We obtained
documentation on DOD's policy for using the software, including
instances in which a specific edit could be overridden by a
contractor to allow payment in certain circumstances. 

To determine how the TCC software is actually working, we met with
MCSC officials as well as their claims processing subcontractors.  We
discussed the yearly updates as well as notifications of interim
changes to TCC decisions, such as policy changes, that DOD would like
for contractors to make within their own claims processing systems. 
We obtained information on how the contractors communicate with
providers about TCC.  We also discussed the process through which a
provider can question TCC decisions on specific claims as well as how
MCSCs' customer service representatives are trained to respond to
these inquiries. 

We performed our work between April 1998 and June 1999 in accordance
with generally accepted government auditing standards. 

--------------------
\14 At the time we initiated our review, the earliest audit began in
November 1996 and the latest ended in December 1997.  For one of the
TRICARE contracts, only three finalized audits were available. 

CLAIMS PROCESSING TIMELINESS FOR
THE NORTHEAST, MID-ATLANTIC, AND
HEARTLAND REGIONS
========================================================== Appendix II

This appendix provides information on claims processing timeliness
for the three regions that did not have at least 1 year of processing
experience.  We obtained data from DOD's HCSR database to determine
the timeliness of claims processing in the Northeast region, managed
by Sierra Military Health Services, and the Mid-Atlantic and
Heartland regions, managed by Anthem Alliance for Health, Inc. 
However, we could not use this file to independently verify
timeliness because approximately 20 percent of the records were
missing.  Therefore, to assess timeliness for these regions, we used
DOD's monthly analyses of MCSCs' claims records, which are based on a
more complete version of this same file. 

As shown in table II.1, the MCSC for the Northeast region met the
timeliness standard of processing 75 percent of claims within 21 days
in 5 of their first 9 months.  However, during this time, nearly half
a million claims took longer than 21 days to process. 

                         Table II.1
          
          Claims Processing Time in the Northeast
                           Region

                                    Percentage paid within
Month and year    Claims processed                 21 days
----------------  ----------------  ----------------------
July 1998                   87,692                   79.57
August 1998                100,823                   81.33
September 1998             178,700                   73.61
October 1998               211,376                   78.36
November 1998              120,661                   70.85
December 1998              364,582                   76.95
January 1999               294,538                   70.35
February 1999              375,865                   84.45
March 1999                 219,082                   71.48
==========================================================
Total for 9              1,953,319                   76.60
 months
----------------------------------------------------------

As table II.2 shows, the MCSC for the Mid-Atlantic and Heartland
regions met the timeliness standard for 4 of the first 10 months of
processing claims.  About 1 million of the over 4 million claims
processed during this time took longer than 21 days to process.

                         Table II.2
          
             Claims Processing Time in the Mid-
               Atlantic and Heartland Regions

                                    Percentage paid within
Month and year    Claims processed                 21 days
----------------  ----------------  ----------------------
June 1998                  153,888                   89.40
July 1998                  356,405                   76.83
August 1998                359,420                   74.47
September 1998             514,561                   72.73
October 1998               420,357                   70.77
November 1998              245,086                   72.08
December 1998              667,272                   70.50
January 1999               504,915                   72.77
February 1999              550,796                   77.89
March 1999                 547,471                   83.55
==========================================================
Total for 10             4,320,171                   75.27
 months
----------------------------------------------------------

DOD'S CUSTOMIZATION OF TRICARE
CLAIMCHECK
========================================================= Appendix III

To ensure that ClaimCheck\TM 's edits reflected TRICARE policy, DOD
officials compared the auditing logic in the ClaimCheck\TM manual to
TRICARE policy.\15 When conflicts were identified, DOD officials
either adopted the ClaimCheck\TM determination as policy or
customized the ClaimCheck\TM determination to conform to TRICARE
policy.  For example, the generic version of ClaimCheck\TM always
denies reimbursement for procedures billed with modifiers -24, -25,
and -79, which are used in conjunction with procedure codes to better
describe the circumstances under which medical services were
performed.\16 During its review of the auditing logic, DOD decided to
always allow payment for procedures correctly billed with these
modifiers.  DOD calls the customized product TCC.  Contractors
receive annual TCC updates, which are customized centrally by HBOC
based on DOD direction.  To ensure uniformity, MCSCs are not
permitted to individually customize TCC except by direction from DOD. 
DOD's customizations to date are listed in this section. 

--------------------
\15 DOD's TCC policy officials stated that, because ClaimCheck\TM 's
software logic was well documented and supported, they did not
perform an edit-by-edit review for each of the 5 million edits. 

\16 Modifier 24 is used to describe an unrelated evaluation and
management service by the same physician during a postoperative
period.  Modifier 25 is used to describe a significant, separately
identifiable evaluation and management service performed by the same
physician on the same day of a procedure or other service.  Modifier
79 describes an unrelated procedure or service by the same physician
during a postoperative period. 

      DOD-DIRECTED CUSTOMIZATIONS
----------------------------------------------------- Appendix III:0.1

1.  Deleted the incidental edit for Physicians' Current Procedural
Technology (CPT) 76818 (fetal biophysical profile) with CPT 76805
(complete fetal and maternal evaluation) so that both procedures will
be paid when billed together. 

2.  Added the following CPT codes for payment of the following
cosmetic/experimental procedures:  15775 (skin graft), 15776 (skin
grafts), 89329 (sperm evaluation), 65771 (radial keratotomy), 95961
(functional cortical mapping), and 52510 (dilation of prostatic
urethra). 

3.  Customized the mutually exclusive edit to allow reimbursement for
the most clinically intensive procedure as opposed to the procedure
with the highest charges or the procedure with the lowest charges. 

4.  Added TRICARE-specific procedure codes for payment. 

5.  Customized to always allow reimbursement for modifiers -24, -25,
and -79. 

6.  Added all Health Care Financing Administration Common Procedural
Coding System modifiers for system recognition.\17

7.  Customized CPT 94150 (vital capacity) to be found incidental to
all evaluation and management (E&M) procedure codes since payment for
this code is included in the allowable amount of the E&M codes. 

8.  Deleted the edit that found CPT 90887 (interpretation of
psychiatric exam) incidental to CPT 90845 (psychoanalysis) so that
they will both be paid when billed together. 

9.  Customized system to recognize modifiers -26, -27, -59, and -90. 

10.  Deleted incidental edit associated with CPT 62278 and 62279
(epidural codes) when billed with maternity codes so that they will
be paid. 

11.  Effective January 1, 1998, deleted the incidental edit
associated with CPT 54150 (newborn circumcision) and E&M codes to
allow payment for the circumcision when billed with an E&M code.\18

12.  Effective December 1, 1998, added TRICARE-specific codes
W0002-W0019 for automated multi-channel laboratory tests so that they
will be paid. 

(See figure in printed edition.)Appendix IV

--------------------
\17 System recognition does not mean that these procedure codes will
be paid.  It means that the claims will be able to pass through the
system without having to stop for manual review. 

\18 In January 1999, DOD directed MCSCs to make this change by
February 1999.  However, some MCSCs did not make the change until
March 1999.  With a retroactive effective date, MCSCs may adjust
claims, when brought to their attention, back to January 1, 1998. 

COMMENTS FROM THE DEPARTMENT OF
DEFENSE
========================================================= Appendix III

(See figure in printed edition.)

(See figure in printed edition.)

(See figure in printed edition.)

(See figure in printed edition.)

GAO CONTACT AND STAFF
ACKNOWLEDGMENTS
=========================================================== Appendix V

GAO CONTACT

Michael T.  Blair, Jr., (404) 679-1944

ACKNOWLEDGMENTS

In addition to the contact named above, Bonnie Anderson, Deborah
Edwards, Art Kendall, Robert DeRoy, Dayna K.  Shah, Cynthia Forbes,
and Lois Shoemaker made key contributions to this report. 

*** End of document. ***