Veterans' Health Care: VA Needs Better Data on Extent and Causes of
Waiting Times (Letter Report, 05/31/2000, GAO/HEHS-00-90).

Pursuant to a congressional request, GAO provided information on the
Department of Veterans Affairs'(VA) efforts to reduce the waiting times
for outpatient care, focusing on VA's initiatives to: (1) improve its
waiting time; and (2) address its waiting time problems.

GAO noted that: (1) although VA has begun to collect data systematically
on waiting times for outpatient care, it has yet to develop reliable
national waiting time data; (2) over the past 10 months, VA has
initiated two separate efforts for gathering comprehensive outpatient
waiting time data from its facilities; (3) the first effort, initiated
in June 1999, produced data that were incomplete and inaccurate, in part
because of the differences in facilities' scheduling of appointments;
(4) VA's second effort, initiated in December 1999, was designed to
improve the data's reliability by measuring the average time taken to
schedule an appointment for an entire month; (5) VA officials believe
that the new method is significantly better and should provide VA with
more complete and accurate data; (6) to ensure this, however, VA will
need to overcome some problems identified during implementation, such as
inaccurate appointment codes being entered into the system; (7) in
addition to taking steps to improve its waiting time data, VA has
initiated actions to reduce the time veterans must wait for outpatient
appointments; (8) VA hired a private contractor to develop and implement
techniques to reduce waiting times at selected clinics in VA facilities
nationwide; (9) in addition, VA plans to spend $400 million in fiscal
year 2001 to make improvements in the timeliness of service, patient
access to telephone care, and timely access to clinical information;
(10) however, VA's lack of reliable national waiting time data raises
concerns about whether VA has an adequate basis upon which to design
these initiatives; (11) specifically, without accurate data on the
extent of waiting time problems and analyses of the causes of long
waits, VA cannot assess whether its proposed expenditures would reduce
waiting times or determine how best to allocate funds to reduce waiting
times; and (12) without, reliable baseline waiting time data and a
mechanism to track the funds used to improve timeliness of care, VA will
not be able to objectively measure whether these funds have actually
resulted in reduced waiting times.

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

 REPORTNUM:  HEHS-00-90
     TITLE:  Veterans' Health Care: VA Needs Better Data on Extent and
	     Causes of Waiting Times
      DATE:  05/31/2000
   SUBJECT:  Veterans benefits
	     Patient care services
	     Statistical data
	     Health services administration
	     Veterans
	     Data collection
	     Health resources utilization
	     Data integrity
IDENTIFIER:  VA Veterans Integrated Service Network

******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO Testimony.                                               **
**                                                              **
** 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.                                         **
**                                                              **
******************************************************************

GAO/HEHS-00-90

Appendix I: Scope and Methodology

20

Appendix II: Comments From the Department of Veterans Affairs

21

Appendix III: Staff Acknowledgments

24

Table 1: Comparison of Scheduling Clerk Data and VISTA Data on
Number of Days Patients Wait for an Appointment at
Syracuse VA Medical Center Clinics, October 1999 8

Table 2: Clinics Chosen by VA Teams Participating in the IHI Project 11

Table 3: VA's Proposed Initiatives and Funds Requested to Help
Achieve Its 30-30-20 Timeliness Goals 13

IHI Institute for Healthcare Improvement

VA Department of Veterans Affairs

VISN Veterans Integrated Service Network

VISTA Veterans Health Information Systems and Technology
Architecture

Health, Education, and
Human Services Division

B-284876

May 31, 2000

The Honorable Lane Evans
Ranking Democratic Member
Committee on Veterans' Affairs
House of Representatives

Dear Congressman Evans:

Information on how long veterans must wait for outpatient health care--and
the locations and clinics where veterans must wait the longest--has been
largely anecdotal. While officials from the Department of Veterans Affairs
(VA) and representatives from veteran service organizations believe these
waiting times are often too long, VA lacks national data on the number of
days veterans must wait to get outpatient appointments at VA health care
facilities, as we have previously reported.1 Nonetheless, data gathered by
one facility show that, in January 2000, veterans at the facility had to
wait as many as 118 days to see an ophthalmologist. Similarly, a
representative from one veteran service organization reported that, in May
1999, patients were waiting over 190 days for an appointment in another
facility's cardiology clinic. Concerned about such reports of waiting time
problems, you asked us to describe the initiatives VA has under way to (1)
improve its waiting time data and (2) address its waiting time problems.

Although VA has begun to collect data systematically on waiting times for
outpatient care, it has yet to develop reliable national waiting time data.
Over the past 10 months, VA has initiated two separate efforts for gathering
comprehensive outpatient waiting time data from its facilities. The first
effort, initiated in June 1999, produced data that were incomplete and
inaccurate, in part because of the differences in facilities' scheduling of
appointments. VA's second effort, initiated in December 1999, was designed
to improve the data's reliability by measuring the average time taken to
schedule an appointment for an entire month. VA officials believe and we
agree that the new method is significantly better and should provide VA with
more complete and accurate data. To ensure this, however, VA will need to
overcome some problems identified during implementation, such as inaccurate
appointment codes being entered into the system.

In addition to taking steps to improve its waiting time data, VA has
initiated actions to reduce the time veterans must wait for outpatient
appointments. First, VA hired a private contractor to develop and implement
techniques to reduce waiting times at selected clinics in VA facilities
nationwide. In addition, VA plans to spend $400 million in fiscal year 2001
to make improvements in the timeliness of service, patient access to
telephone care, and timely access to clinical information. However, VA's
lack of reliable national waiting time data raises concerns about whether VA
has an adequate basis upon which to design these initiatives. Specifically,
without accurate data on the extent of waiting time problems and analyses of
the causes of long waits, VA cannot assess whether its proposed expenditures
would reduce waiting times or determine how best to allocate funds to reduce
waiting times. Further, without reliable baseline waiting time data and a
mechanism to track the funds used to improve timeliness of care, VA will not
be able to objectively measure whether these funds have actually resulted in
reduced waiting times. We recommended that VA take actions to identify the
extent and causes of waiting time problems and to monitor and track
expenditures for addressing these problems. VA agreed with our
recommendations.

VA operates one of the nation's largest health care systems, providing
services in over 600 patient care facilities in all 50 states, Guam, and
Puerto Rico. These facilities include 172 medical centers as well as
numerous community-based outpatient clinics and domiciliaries.2 Twenty-two
regional Veterans Integrated Service Networks (VISN) manage the facilities
and serve as the basic budgetary and decisionmaking units for determining
how best to provide services to veterans within their geographic boundaries.
Over the past several years, VA has begun to adopt managed care practices,
treating more of its patients in outpatient settings and reducing its heavy
reliance on inpatient care. VA reports that, during fiscal year 1999, it
provided care to about 3.5 million veterans, resulting in almost 38 million
outpatient visits.

To serve veterans appropriately, the Veterans' Health Care Eligibility
Reform Act of 1996 required VA to ensure that veterans enrolled in VA's
health care system receive timely care. For outpatient care, VA's goals are
that patients (1) receive an initial, nonurgent3 appointments with their
primary care or other appropriate provider within 30 days of requesting one;
(2) receive specialty appointments within 30 days when referred by a primary
care provider; and (3) be seen within 20 minutes of their scheduled
appointments.4 VA refers to these goals as the "30-30-20" goals.

To schedule outpatient appointments, VA relies on the Veterans Health
Information Systems and Technology Architecture (VISTA). VISTA automates
major clinical, management, and administrative functions and includes a
scheduling component. The scheduling component has been upgraded several
times since VA developed it about 2 decades.

In June 1999, VA began for the first time--in response to a congressional
request for information on outpatient waiting times--to systematically
collect data on the length of time veterans must wait for outpatient health
care appointments.5 According to VA officials, collecting these data
required software modifications to VISTA's appointment-scheduling component
because the system was not designed to collect information on waiting times.
However, because VA clinics used this scheduling component in varying and
often inconsistent ways, the data generated from the first data collection
effort were not comparable from facility to facility. In an effort to
address this problem, VA implemented in December 1999 a new data collection
method that it plans to use to evaluate VISN directors' performance in
improving waiting times, beginning in September 2000. Although VA officials
told us they believe the new method is a better one, implementation
problems--such as inaccurate appointment types being entered into the
system--could reduce data reliability.

In its first attempt to collect outpatient waiting time information, VA
designed a software program to extract data from VISTA's
appointment-scheduling component once a month at every VA outpatient clinic
nationwide. The program, run on the last day of each month, determined for
each clinic the number of days veterans had to wait for a scheduled primary
care appointment from that 1 day. For primary care appointments, the third
available appointment was used because, according to VA, the first and
second appointments are often held open for urgent care and would not be
given to veterans calling for routine care. Waiting times for all clinics of
the same type--such as orthopedics or primary care--were then averaged
within a facility. For example, if a facility had 10 primary care clinics,
the number of days to the third available appointment for each clinic was
gathered, summed, and the total divided by 10 to establish the facility
average waiting time for primary care appointments.

However, clinics varied in how they used the appointment-scheduling
component, and data generated for the third available appointment could not
be aggregated to obtain an overall picture of waiting times. Several other
problems with this first data collection effort resulted in VA not being
able to determine whether waiting times in a given clinic were understated
or overstated.

ï¿½ Some clinics showed a higher number of available appointment slots than
could actually be staffed and relied on the scheduling clerks to know which
slots were available and which slots should not be filled--a distinction the
software program could not make. Therefore, by averaging "available" slots,
the software program probably understated the time a veteran would wait for
an appointment.

ï¿½ Conversely, some clinics scheduled more patients than the number of
appointment slots shown in the scheduling component--perhaps to account for
patients who fail to appear for their appointments or because the number of
appointment slots shown were fewer than the number of patients that the
clinic could actually accommodate during that time. In such cases, the
software program probably overstated waiting times.

ï¿½ Finally, the software program only extracted information on clinic
availability on the last day of the month for which it was run. Because of
day-to-day variations in scheduling, the data collected for that 1 day might
not have been representative of general clinic availability on the other
days of the month.

We attempted to validate VISTA's automated data by comparing it with data
collected manually at one location in VISN 2--Syracuse, New York. To measure
adherence to VA's 30-day waiting time standard, scheduling clerks at
facilities in VISN 2 check appointment availability for 1 day during the
second week of every month, counting the number of days to the third
available appointment using VISTA scheduling package information. Although
this technique parallels that used by VA's software program, the manual
collection of waiting time data allows clerks in VISN 2 facilities to
identify nuances associated with a particular location or clinic. For
example, one doctor out of several in a clinic may not be accepting new
patients, so when a new patient asks for an appointment, the clerk would not
check that doctor's availability for appointments but instead would check
the availability of others. In this situation, VISTA's count of available
appointments would be overstated because it would count all doctors.

Comparing data from both approaches for 12 clinics located at the Syracuse
VA medical center, we noted large differences between waiting time data
reported by clerks and data extracted from VISTA. For five clinics, clerks
reported waiting times ranging from 17 to 56 days shorter than the
computer-extracted data; for seven clinics, waiting times reported by clerks
ranged from 2 to 47 days longer. Waiting times were closest for neurology
and farthest apart for orthopedics and cardiology (see table 1).

                     Days as          Days as
 Clinic              reported by      extracted by        Difference in
                     clerks           software program    days reported
 Cardiology          10               66                  -56
 Ear, nose, and
 throat              47               69                  -22
 Gastroenterology    3                30                  -27
 General surgery     56               22                  +34
 Gynecology          38               10                  +28
 Neurology           15               13                  +2
 Ophthalmology       70               41                  +29
 Optometry           99               61                  +38
 Orthopedics         81               34                  +47
 Podiatry            37               54                  -17
 Primary care        5                36                  -31
 Urology             44               29                  +15

Some variations may be attributable to the data's being collected on
different days in October. However, even in the unlikely event that all
variations could be attributed to different collection days, the variations
still serve to illustrate the problems inherent in collecting data for only
1 day of the month.

Recognizing the shortcomings of its initial data collection effort, VA
introduced in December 1999 a new method of collecting data on outpatient
appointment waiting times. While this new method also uses information from
the VISTA scheduling package, it measures the average number of days between
the date that an appointment was requested and the scheduled date of the
appointment. These data are collected on a monthly basis for every patient,
not just for 1 day. VA most recently extracted data on the average time from
request to appointment for the entire month of April 2000. VA now has 4
months of data to monitor waiting times for the next available
appointment--January, February, March, and April--since the new software
program was initiated in December 1999.

The amount of time between the day an appointment was requested and the date
for which it was scheduled is then summed for all patients and a per-patient
average is calculated. This calculation is performed once a month for all
clinics within each VA facility, then summed for all clinics of the same
type in each facility. By examining only actual scheduled appointments, the
new method is designed to overcome the problems posed by the variations in
clinic scheduling practices. Similarly, obtaining data on all appointments
scheduled in a month provides VA with complete information.

This second approach also takes into account the possibility that patients
may not want or need the next available appointment. For example, some
patients may schedule a follow-up visit in 6 months at their physicians'
direction or may choose an appointment further in the future rather than the
next available appointment because of their work or personal demands. VA has
chosen to omit scheduled appointments such as these from its calculations
because it is interested only in determining the time that patients have no
choice but to wait for the next available appointment.

To exclude these types of appointments from its calculations, VA introduced
with this second effort a procedure under which scheduling clerks must enter
in a particular field on the computer whether an appointment is (1) the next
available (indicated by an "N"), (2) other than the next available at the
request of the clinician (indicated by a "C"), or (3) other than the next
available at the request of the patient (indicated by a "P"). In calculating
the average time to appointment, only those appointments marked with an "N"
are considered.

VA field officials, who have oversight responsibilities for appointment
scheduling, told us that the new data collection method is conceptually
superior to VA's initial approach because the data collected is more
accurate. However, these officials raised concerns that early problems in
implementing the method could reduce the data's reliability. For example, if
a patient initially requests the next available appointment but then decides
to select an appointment further in the future, the clerk must exit and then
reenter the system and rekey all of the appointment information--simply to
change the appointment code from "N" to "P." According to these field
officials, requiring scheduling clerks to exit and then reenter the
scheduling system in such events is cumbersome and may cause busy clerks to
leave the "N" they initially entered rather than changing it to a "P." As a
result, the appointment would be inappropriately included in the clinic's
waiting time data and would overstate the actual waiting time at the clinic.

In addition, VA's ability to collect comparable waiting time data for all
facilities may be limited because some clinics schedule appointments only 2
months into the future. In such cases, patients would be asked to call back
if no appointments were available during that 2-month time period. As a
result, data collected from these clinics would always show average waiting
times of 2 months or less. To the extent patients in the clinics must wait
longer than 2 months to obtain an appointment, data collected from these
clinics would understate waiting times.

Believing its new approach will provide adequate and accurate data on
waiting times, VA intends to use this information to evaluate VISN
directors' performance in their efforts to reach waiting time goals,
beginning in September 2000. According to VA's "Fiscal Year 2000 Network
Performance Plan," VISN directors' performance will be measured on their
ability to provide patient appointments within 30 and 45 days for six
clinics: audiology, cardiology, ophthalmology, orthopedics, primary care,
and urology. According to an official in VA's Chief Network Office, VA
identified these six clinics by asking clinical managers from each of the 22
VISNs to survey their facilities on what clinics had the most problems with
waiting times. VA plans to collect waiting time data for all clinics and
might replace one or more of the six clinics with different clinics if the
data show that their waiting times are worse.

VA has begun several initiatives to reduce its waiting times. For example,
in June 1999, VA retained a private contractor to help it develop and
implement techniques to reduce waiting times in selected VA clinics
nationwide. Recognizing that this project will not completely solve its
waiting time problems, VA plans to spend $400 million in fiscal year 2001 to
help it achieve its timeliness goals. However, VA currently has little
detailed knowledge of the extent or causes of outpatient waiting times
systemwide. Further, without baseline waiting time data and a method for
tracking and monitoring funds, VA will not be able to objectively measure
whether any funds spent to reduce waiting times have produced the desired
improvements. Given VA's lack of reliable national waiting time data and
lack of detailed plans specifying which facilities or clinics would
implement the improvements, we cannot assess whether VA's proposed
expenditures would reduce waiting times.

In July 1999, VA contracted with the Boston-based Institute for Healthcare
Improvement (IHI) to develop ways to reduce waiting times in specific
clinics selected by VA facilities nationwide. As part of this project, 134
teams from VA facilities across the nation developed strategies to reduce
waiting times for appointments in 160 primary or specialty care clinics.
While some teams worked to reduce waiting times in only one clinic, several
chose two or more clinics. Table 2 summarizes the clinics that the VA teams
chose for their IHI projects.

 Clinic                                          Number of clinics
 Primary care                                    88
 Orthopedics                                     11
 Eye (including ophthalmology and optometry)     9
 Urology                                         9
 Gastrointestinal                                8
 General medicine/surgery                        6
 Neurology                                       5
 Cardiology                                      3
 Dermatology                                     3
 Mental health/psychiatry                        3
 Rheumatology                                    3
 Endocrinology                                   2
 Physical therapy                                1
 Women's health                                  1
 Other specialty care (specialty not identified) 8
 Total                                           160

Source: Institute for Healthcare Improvement.

Since the project began in July 1999, some VA clinics have reported notable
successes in reducing waiting times, often by increasing provider
availability and office efficiency. One VA facility, for example, reports
that it has reduced the number of days veterans must wait for a primary care
appointment from 240 days to about 14 days--that is, veterans who once had
to wait almost 8 months for primary care are now able to obtain appointments
within 2 weeks. VA staff attribute this success, in part, to scheduling
changes, such as reducing the number of unnecessary return visits scheduled
for veterans and no longer automatically rescheduling new patients who do
not show up for their appointment.

Other improvements reported to us include an orthopedic clinic's having
reduced its waiting time for consultations from 130 days to 7 days, and an
eye clinic's having reduced its waiting time from 120 days to 50 days. Some
teams, however, have struggled with reducing waiting times, citing barriers
such as physician and team member resistance to the changes and difficulties
in sustaining improvements made to reduce waiting times.

The IHI project ended in March 2000, and the final report summarizing the
results of this project is due in July 2000. According to VA officials,
reduced waiting times in other clinics could be expected to follow from the
success of these initial projects as VA facilities implement the techniques
they learned by participating in IHI.

In its fiscal year 2001 budget submission to the Congress, VA is proposing
to spend $400 million toward meeting its 30-30-20 timeliness goal (that is,
appointments within 30 days for initial primary care and specialty care
referrals and patients seen by providers within 20 minutes of scheduled
appointments). VA plans to spend these funds to improve the timeliness of
service, patient access to telephone care, and timely access to clinical
information.6 VA expects that these improvements will help it achieve its
30-30-20 goal by fiscal year 2003. Table 3 shows the fiscal year 2001 funds
VA plans to spend on each of these improvements.7

                                       Projected expenditure
 Proposed initiative
                                       (in millions)
 Timeliness of service                 $223
 Patient access to telephone care      21
 Timely access to clinical information 156
 Total                                 $400

Source: VA Acting Director for Resource Formulation.

To meet its timeliness of service initiative, VA plans to spend a total of
$223 million to implement the following three efforts:

ï¿½ Scheduling package: VA plans to spend $34.7 million to redesign or replace
its outpatient appointment-scheduling package. VA believes that difficulties
with the current system reduce both access to and timeliness of service
because it does not allow flexibility in appointment length or scheduling
across different facilities. VA officials told us that they recently formed
a team to review whether it should redesign or replace the scheduling
package. After its review, the team is expected to provide a detailed
proposal of project costs, time frames, and implementation strategies. VA
anticipates this project to be a multiyear effort and expects the $34.7
million to help fund the project for about 2 years.

ï¿½ Contract physicians: VA plans to use additional contract specialist
physicians to reduce the backlog of patients waiting for specialty care and,
as a result, reduce the amount of time patients must wait to receive such
care. According to VA budget officials, VA has not yet determined how much
of the fiscal year 2001 funds it will use for this purpose, nor which
facilities will receive additional specialists. VA hopes to contract for
specialists based on patient needs and facility capabilities. Because VA
does not know at this time which facilities or clinics will contract with
physicians, what specialties are experiencing the longest waiting times, and
how much it intends to spend on this initiative, it is unclear to what
extent using contract physicians will reduce patient waiting times.

ï¿½ Outpatient clinics: VA plans to improve timeliness by establishing 63
additional community-based outpatient clinics. VA believes that by offering
patients more convenient access to care--and possibly alleviating workload
at other clinics--these patients might receive more timely initial and
subsequent appointments for their primary care.8 However, VA has not yet
determined the cost or locations of these new clinics or to what extent they
would reduce waiting times. Further, it is unclear whether this initiative
would reduce waiting times at all. Specifically, in 1999, 12 of the 22 VISN
directors we surveyed told us that by increasing veterans' access to VA
health care, community-based outpatient clinics could stimulate demand. If
this occurs, waiting times could actually lengthen. VA is currently studying
whether existing community-based outpatient clinics have resulted in shorter
waiting times.

VA also believes that telephone care will help it achieve its timeliness
goals. Specifically, VA officials cited studies that found that a
significant percentage of visits to emergency departments, urgent care, and
doctors' offices bring no medical benefits, and a percentage of unnecessary
visits could be avoided by providing patients the option to speak to a nurse
over the telephone. To improve patient access to telephone care, VA plans to
spend $21 million to begin providing nurse advice lines nationwide, 24 hours
a day, 7 days a week. According to an official in VA's Office of Primary and
Ambulatory Care, VA plans to identify basic systemwide standards and
measures as part of this nationwide initiative. To ensure that the advice
provided over the phone is standardized at all locations, for example, this
official explained that VA will need to purchase a license for automated
medical and legal protocols. Along with this telephone care program, VA is
also considering using a portion of the $21 million to purchase self-care
manuals to provide to patients to improve their medical knowledge and
possibly reduce unnecessary visits. VA expects the program to be fully
implemented at the end of fiscal year 2002.

To improve timely access to clinical information, VA plans to spend $156
million to primarily provide the information technology support it believes
is necessary to ensure timely and quality care. VA officials believe that
improvements such as access to clinical information across sites of care
will help ensure prompt service and improve continuity of quality care. For
example, to improve access to clinical information across sites of care, VA
plans to improve software capabilities so that facilities can access and
exchange patient health summary information across sites of care. In
addition, VA also plans to expand several national telemedicine pilot
programs and study the results of these programs to establish best practices
for new telemedicine programs. Further, VA plans to purchase software and
hardware--such as personal computers, network servers, and X-ray
equipment--and hire additional technical staff to support these functions.
VA has not yet determined how much of the $156 million it will spend on each
of these initiatives or which locations will receive these improvements.
According to VA's Acting Director for Resource Formulation, these
determinations will be based on individual needs and the capabilities of the
facilities within the 22 VISNs.

Given VA's lack of reliable national waiting time data, VA cannot reasonably
allocate its proposed $400 million to facilities that have the worst
problems. VA also cannot measure program progress, as it currently does not
have a baseline of reliable national data from which to measure such
progress. Moreover, without a thorough, data-driven analysis of the cause of
the long waits and a method for tracking and monitoring funds, VA cannot
objectively measure whether funds appropriated to reduce waiting times would
produce these desired reductions.

VA's lack of reliable national waiting time data not only limits its ability
to identify which facilities and clinics have the longest waiting times but
also limits VA's ability to know whether waiting time problems are
systemwide or isolated to particular facilities. VA officials explained that
they hope to collect more reliable appointment waiting time data in the near
future but, in the interim, are encouraging VISN directors to monitor their
facilities' waiting times by reviewing the data extracted from the VISTA
software program for the facilities within their VISN. According to a VA
official, VISN directors are also encouraged to take appropriate actions to
reduce waiting times, including implementing techniques learned through the
IHI project, such as targeting resources to clinics where they are most
needed and improving office efficiencies.

Although VA has budgeted funds expressly to help it achieve its timeliness
goals, it does not plan to track these funds by project and thus will be
unable to monitor whether the funds have contributed to reduced waiting
times. Specifically, VISN directors have the authority to allocate funds to
the facilities within their VISNs. VA officials told us that, for this
reason, VA does not have detailed spending plans for much of the $400
million of planned obligations. For example, while VA has proposed spending
$223 million on three initiatives to improve the timeliness of service, it
could only provide funding estimates for one: the $34.7 million to redesign
or replace its scheduling package.

Because VA will not be designating the exact uses for most of the $400
million, each VISN director has discretion over how the funds will be used.
Consequently, VA will not be able to track whether these funds were actually
spent on improving timeliness or whether their use had the intended effect
on reducing waiting times.

Anecdotal information has led to the perception that VA is not meeting its
30-30-20 timeliness goals and that veterans in different locations do not
have equal access to timely care. In order to correct these perceived
problems, VA managers must make decisions about where and how to apply the
agency's finite resources in order to have the greatest effect. To make such
decisions, we believe that VA managers must have reliable data on the extent
and causes of waiting times in different locations and in different clinics.

Although VA has made significant strides in collecting waiting time data
systemwide, some problems must be overcome before the data can be reliably
used. Once improvements are made, we believe the data can be used to give VA
managers an estimate of national waiting times and those for every
individual VA clinic. However, VA has not analyzed the causes of waiting
times and is still waiting for final data from the IHI project.
Consequently, it does not have an adequate basis for determining which
initiatives could best improve the timeliness of care. Further, VA does not
plan to ensure that its initiatives for improving timeliness will result in
decreased waiting times or even if these funds will be spent on timeliness
initiatives.

We recommend that the Secretary of Veterans Affairs direct the Under
Secretary for Health to first determine the extent of waiting times and
their causes and then develop a spending plan that will result in solving
the identified waiting time problems. In addition, VA should develop a
mechanism for monitoring and tracking expenditures for improving timeliness
to evaluate how well targeted funds have reduced waiting times.

We provided VA a draft copy of our report for its review and comment.
Overall, VA agreed with the report's conclusions and recommendations and
indicated that it will use it as a guide for developing a national plan for
addressing waiting time issues. As a result of our report, VA is also
planning to require VISNs to develop systematic spending plans for waiting
time expenditures and is planning to track projected goals for each waiting
time initiative, including a description of anticipated and measurable
benefits in meeting its goals.

VA said that it believes its February 2000 clinic appointment waiting time
data are adequate for use as a baseline to measure progress. While the
February 2000 data are better than that previously available, VA has not yet
tested the validity of each clinic's data and, thus, cannot be assured that
these data can be used as an adequate baseline. As noted in the report, we
found errors at some locations--specifically, inaccurate appointment types
being entered into the system--that indicate the data may not be reliable.
Moreover, VA points out that it is continuing to collect and improve upon
its monthly data and that its understanding will increase as the data are
refined.

VA also noted that it is developing a systematic spending plan for
addressing identified waiting time problems. It also intends to integrate
existing and proposed initiatives--which have been fragmented--into a
cohesive national plan. VA appears to be committed to carrying out these and
other initiatives aimed at achieving its waiting time goals. However, we
cannot assess the effectiveness of these actions until they are completed.
The full text of VA's comments is presented in appendix II.

As arranged with your office, unless you announce its contents earlier, we
plan no further distribution of this report until 30 days after the date of
this letter. At that time, we will send copies to the Honorable Togo D.
West, Jr., Secretary of Veterans Affairs; appropriate congressional
committees; and other interested parties. We will also make copies available
to others upon request.

Please contact me at (202) 512-7101 or Ronald J. Guthrie at (303) 572-7332
if you or your staff have any questions. Other key contributors to this
report are listed in appendix III.

Sincerely,

Cynthia Bascetta, Associate Director
Veterans' Affairs and Military
Health Care Issues

Scope and Methodology

To determine what initiatives VA has under way to improve its waiting time
data, we reviewed and analyzed VA data extracted under both of its recent
approaches and VA supporting documentation for these approaches. We
interviewed VA headquarters and field officials to gain their perspectives
on these data extracts and whether they could be used as a reliable measure
of waiting times systemwide.

To determine what initiatives VA has under way to address its waiting time
problems, we reviewed VA's fiscal year 2001 budget submission to the
Congress, VA's Fiscal Year 2000 Network Performance Plan, and other
documents detailing VA's goals to reduce its waiting times for appointments.
In addition, we interviewed VA headquarters and field officials and IHI
officials and faculty to gain their perspectives on the project to reduce
waiting times. To learn more about the VA/IHI project, we also attended two
of the three learning sessions and reviewed and analyzed relevant documents
obtained from these sessions. Finally, we interviewed veteran service
organization representatives and obtained data related to their reports of
waiting times in several clinics.

We conducted our work between January 2000 and May 2000 in accordance with
generally accepted government auditing standards.

Comments From the Department of Veterans Affairs

Staff Acknowledgments

James Espinoza, Lisa Gardner, Steve Gaty, George Lorenzen, Karen Sloan, and
Alan Wernz also made key contributions to this report.

(406185)

Table 1: Comparison of Scheduling Clerk Data and VISTA Data on
Number of Days Patients Wait for an Appointment at
Syracuse VA Medical Center Clinics, October 1999 8

Table 2: Clinics Chosen by VA Teams Participating in the IHI Project 11

Table 3: VA's Proposed Initiatives and Funds Requested to Help
Achieve Its 30-30-20 Timeliness Goals 13
  

1. See VA Health Care: Progress and Challenges in Providing Care to Veterans
(GAO/T-HEHS-99-158 , July 15, 1999).

2. Domiciliaries are residential rehabilitation and health maintenance
centers for veterans who do not require hospital or nursing home care but
are unable to live independently because of medical or psychiatric
disabilities.

3. VA's timeliness standard for urgent care requires that veterans have
access to such care 24 hours a day.

4. For purposes of this review, we focused on how long it takes veterans to
obtain a scheduled appointment for outpatient health care, not the length of
time spent in the waiting room to see a provider.

5. According to an official from VA's Office of Primary and Ambulatory Care,
from 1993 to 1998, VA required facilities to self-report waiting time data
for several specialty clinics. However, VA recognized that self-reported
data were not reliable and that such data did not capture waiting time
information on a national basis.

6. According to VA's Acting Director for Resource Formulation, VA identified
these initiatives by summarizing and reviewing proposals from each of VA's
22 VISNs and from headquarters program officials. Although the estimated
funding needed to accomplish these goals by fiscal year 2003 exceeds the
$400 million, VA determined that it was a reasonable investment in fiscal
year 2001 to help reach this future goal.

7. Of this $400 million, VA plans to obtain $77 million through new
appropriated dollars from the Congress. The remaining $323 million is
expected to come from fiscal year 2000 activities. Specifically, $200
million is expected to come from management savings and $123 million is
expected to be available as a result of a reduction in staff through
retirement incentives.

8. According to VA, it determines the need for community-based outpatient
clinics based on a demographic analysis that includes an evaluation of the
proposed area, its proximity to the parent facility, the existing and
projected veteran population, and the characteristics of the population to
be served.
*** End of document. ***