VA Drug Formulary: Better Oversight Is Required, but Veterans Are Getting
Needed Drugs (Letter Report, 01/29/2001, GAO/GAO-01-183).

Over the past three years, the Department of Veterans Affairs (VA) has
made significant progress in establishing its national drug formulary,
which has generally met with prescriber acceptance. GAO reported that
veterans are generally receiving the drugs they need and that veterans
rarely register complaints concerning prescription drugs. VA has not
provided sufficient oversight, however, to ensure that the Veterans
Integrated Service Networks (VISN) and medical centers comply with
formulary policies and that the flexibility given to them does not
unduly compromise VA's goal of formulary standardization. Contrary to VA
formulary policy, some facilities omitted national formulary drugs or
modified the closed drug classes. While adding a limited number of drugs
to supplement the national formulary is permitted, as more drugs are
added by VISNs, formulary differences among facilities are likely to
become more pronounced, decreasing formulary standardization. While VA
recognizes the trade-off between local flexibility and standardization,
it lacks criteria for determining the appropriateness of adding drugs to
supplement the national formulary and therefore may not be able to
determine whether the decrease in standardization is acceptable.

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

 REPORTNUM:  GAO-01-183
     TITLE:  VA Drug Formulary: Better Oversight Is Required, but
	     Veterans Are Getting Needed Drugs
      DATE:  01/29/2001
   SUBJECT:  Veterans benefits
	     Veterans hospitals
	     Standards and standardization
	     Drugs
	     Health services administration
	     Health care services
	     Internal controls
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-01-183
A

Report to the Ranking Member, Committee on Veterans' Affairs, U. S. Senate

January 2001 VA DRUG FORMULARY

Better Oversight Is Required, but Veterans Are Getting Needed Drugs

GAO- 01- 183

Letter 3 Appendixes Appendix I: Scope and Methodology 24

Appendix II: Comments From the Department of Veterans Affairs 29 Appendix
III: GAO Contact and Staff Acknowledgments 33

Related GAO Products 34 Tabl es Table 1: Percentage of Total Outpatient
Prescriptions Filled for National Formulary Drugs by VISN and Medical
Centers

Within Each VISN, October 1999 Through March 2000 13 Table 2: Ease of
Obtaining Nonformulary Drug Approvals Reported

by Prescribers 17 Table 3: Likelihood of Prescribers' Converting Patients
From

Nonformulary Drug Prescriptions to Formulary Drug Prescriptions 17 Table 4:
Prescribers Reported Current Formulary Helps More Than

Hinders Their Ability to Prescribe Drugs 19 Table 5: Questionnaire Sample
and Response Size 26

Figures Figure 1: Variation in Number of Unique Drugs VISNs Added to
Supplement National Formulary, June 1997- March 2000 11

Figure 2: Medical Center Modifications to Formulary Closed Classes 12 Figure
3: Steps Generally Followed to Obtain Nonformulary and Formulary Drugs 15

Abbreviations

FDA Food and Drug Administration IOM Institute of Medicine PBM Pharmacy
Benefits Management Strategic Healthcare Group P& T pharmacy and
therapeutics VA Department of Veterans Affairs VHA Veterans Health
Administration VISN Veterans Integrated Service Network

Lett er

January 29, 2001 The Honorable John D. Rockefeller IV Ranking Member
Committee on Veterans' Affairs United States Senate

Dear Senator Rockefeller: The large number of drugs manufactured to treat
various medical conditions has led health care organizations to develop
lists of medications that they encourage their health care providers to use
when they write

prescriptions for patients. These lists of medications, grouped by
therapeutic class, are known as drug formularies. Organizations prefer, and
sometimes require, that these drugs be used as the provider's first choice
because organizations believe these drugs represent the best value among
therapeutic choices and because by concentrating their purchases
organizations can secure better prices.

As early as 1955, the Department of Veterans Affairs' (VA) medical centers
began using formularies to manage their pharmacy inventories, and until
recently each of VA's 173 medical centers maintained its own formulary. In
June 1997, VA established its national formulary by consolidating locally

developed and managed formularies into a core list of drugs. The national
formulary was intended to give health care providers who have outpatient
prescription privileges the same list of formulary drugs no matter where in
the VA system they practice medicine, and to assure veterans that they will
have access to the same group of formulary drugs for their treatment at all
VA medical centers no matter where they seek care. VA medical centers were
directed to make all national formulary drugs available to prescribers. To
provide flexibility in meeting local patient needs, VA allows its 22
regional Veterans Integrated Service Networks (VISN) to add drugs locally

to supplement the national formulary and requires them to establish approval
processes for prescribers to obtain drugs not listed on their formularies.
Over the past few years, you and others in the Congress have raised
questions about VA's national formulary. In December 1999, we issued a

report addressing your interest in how VA manages its national formulary. 1
In June 2000, the Institute of Medicine (IOM), as mandated by the Congress
in House Report 105- 610, issued its report on the effect VA's national
formulary has had on the cost and quality of VA health care, the
restrictiveness of VA's national formulary, and how the national formulary
compares with private and other government formularies. 2 To supplement the
findings of our December 1999 report, you asked us to review (1) the impact
of national formulary implementation by VISNs and

medical centers on VA's goal of standardizing its formulary and (2) VISN and
medical center processes for obtaining nonformulary outpatient
prescriptions. You also asked that we examine whether VA's national

formulary allows health care providers to prescribe the medications they
believe their patients need. To address these VA formulary issues, we
reviewed the formulary policies and activities of VA's central office and
its VISNs and compiled and analyzed nationwide VA prescription data. To
assess implementation at the local level, we also discussed formulary issues
with VA officials at three medical centers we visited in three different
VISNs. We selected these medical centers on the basis of their percentage of
prescriptions filled for formulary drugs. At your request, we also conducted
a survey of 2, 000 health care providers who have VA prescription writing
privileges (prescribers) to determine whether they believe implementation of
the

national formulary has adversely affected their ability to prescribe needed
medications for the veterans they treat. In addition, we reviewed the
congressionally mandated IOM assessment of VA's national formulary issued in
June 2000, and we incorporated its pertinent results in our report.

Appendix I contains further details of our scope and methodology. We
performed our work between January 2000 and January 2001 in accordance with
generally accepted government auditing standards. 1 VA Health Care: VA's
Management of Drugs on Its National Formulary (GAO/ HEHS- 00- 34, Dec. 14,
1999).

2 IOM, Description and Analysis of the VA National Formulary (Washington, D.
C.: IOM, June 2000).

Results in Brief While VA has made significant progress in establishing a
national formulary, it has not provided sufficient oversight to ensure that
it is fully achieving its national formulary goal of standardization. For
example, not all VISNs' medical centers are in compliance with the national
formulary. Specifically, two of the three facilities we visited, each in a
separate VISN, omitted more than 140 required national formulary drugs, and
all three facilities inappropriately modified the national formulary list of
required drugs for certain drug classes by adding or omitting some drugs. As
VA policy allows, VISNs have added drugs to supplement the national
formulary, ranging from 5 drugs at one VISN to 63 drugs at another. However,
the wide variation in the number of drugs added by the VISNs

raises concern that this practice, if not appropriately monitored, may
result in unacceptable decreases in formulary standardization. To address
this issue, VA plans to begin reviewing new drugs approved by the Food and
Drug Administration (FDA) to determine if they will be added to the national
formulary or if VISNs may continue to add them to their formularies to
supplement the national formulary. The approval process for the use of
nonformulary drugs across VA's health

care system does not ensure that all facilities have an efficient and timely
process. We found wide variability in how requests are made, who approves
such requests, and how much time it takes. Similarly, IOM determined that
many variants of the nonformulary approval process were being used. In
addition, VA does not have systematic data to determine the extent to which
nonformulary drugs are being requested, approved, or

denied. Although VISNs are required to establish systems to analyze such
requests, 12 of the 22 VISNs have not done so. As a result, VA does not know
whether approved requests met established criteria or whether denied
requests were appropriate. VA health care providers who prescribe drugs
generally reported that veterans get the drugs they need. These prescribers
reported that the national formulary generally contains the drugs their
patients need or, when necessary, prescribers can usually get nonformulary
drugs. Despite their overall satisfaction, we believe the weaknesses we
identified merit correction. Consequently, we are recommending that VA (1)
ensure compliance with national formulary policy, (2) ensure that drugs
added to supplement the national formulary do not inappropriately decrease
formulary standardization, and (3) improve the nonformulary process. In
responding to a draft of this report, VA agreed with our findings and
concurred with our recommendations.

Background In fiscal year 2000, VA's Veterans Health Administration (VHA)
provided primary and specialty medical care to approximately 3. 2 million
veterans at a cost of about $18 billion. VA's pharmacy benefit cost
approximately $2 billion- about 12 percent of the total VHA budget- and
provided approximately 86 million prescriptions. In contrast, 10 years ago
VA's pharmacy benefit represented about 6 percent of VA's total health care
budget. Health care organizations' efforts to control pharmacy costs and
improve quality of care include (1) implementing formularies that limit the
number of drug choices available; (2) establishing financial incentives,
such as variable copayments, to encourage the use of formulary drugs; (3)
using compliance programs, such as prior authorization, that encourage or
require physicians to prescribe formulary drugs; and (4) developing clinical
guidelines for prescribing drugs. VA does not have authority to use
financial incentives to encourage compliance with its formulary. VA provides
outpatient pharmacy services free to veterans receiving medications for
treatment of service- connected

conditions and to low- income veterans whose incomes do not exceed a
threshold amount. Other veterans who have prescriptions filled by VA may be
charged $2 for each 30- day supply of medication. 3 In 1995, VA began
transforming its delivery and management of health care

to expand access to care and increase efficiency. As part of this
transformation, VA decentralized decision- making and budgeting authority to
22 VISNs, which became responsible for managing all VA health care. VISNs
were given substantial operational autonomy. Although VISN and medical
center directors are held accountable in annual performance agreements for
meeting certain national and local goals, attaining

formulary goals has not been part of their performance standards. VA medical
centers began using formularies as early as 1955 to manage their pharmacy
inventories. Because of the geographic mobility of VA patients, VA officials
believed that a national formulary would improve

3 The Veterans Millennium Health Care and Benefits Act (P. L. 106- 117, Nov.
30, 1999) authorized the Secretary of the Department of Veterans Affairs to
increase the current $2 copayment for each 30- day supply of medication and
to establish maximum monthly and

maximum annual pharmaceutical copayments for veterans who have multiple
outpatient prescriptions. To date, VA has not implemented a change in the
copayment amount.

veterans' continuity of care. In September 1995, VA established a
centralized group to manage its pharmacy benefit on a nationwide basis. In
November 1995, VISNs were established, and the Under Secretary for Health
directed each VISN to develop and implement a VISN- wide formulary. To
develop their formularies, the VISNs generally combined existing medical
center formularies and eliminated rarely prescribed drugs. VISN formularies
became effective on April 30, 1996. Also in 1996, the Congress required VA
to improve veterans' access to care regardless of the

region of the United States in which they live. As part of its response, VA
implemented a national drug formulary on June 1, 1997, by combining the core
set of drugs common to the newly developed VISN formularies. In addition to
the national and VISN formularies, a few medical centers retained their own
formularies.

VA's Pharmacy Benefits Management Strategic Healthcare Group (PBM) is
responsible for managing the national formulary list, maintaining databases
that reflect drug use, and monitoring the use of certain drugs. VISN

directors are responsible for implementing and monitoring compliance with
the national formulary, ensuring that VISN restrictions placed on national
formulary products are appropriate, and ensuring that a nonformulary drug
approval process is functioning in all of their medical centers.

As all formularies do, VA's national formulary limits the number of drug
choices available to health care providers. VA's formulary lists more than
1,100 unique drugs that are assigned to 1 of 254 drug classes- groups of
drugs similar in chemistry, method of action, or purpose of use. After
performing reviews of drug classes representing the highest costs and

volume of prescriptions, VA decided that some drugs in 4 of its 254 drug
classes were therapeutically interchangeable- that is, essentially
equivalent in terms of efficacy, safety, and outcomes- and therefore had the
same therapeutic effect. This determination allowed VA to select one or more
of these drugs for its formulary to seek better prices through competitively
bid committed- use contracts. 4 Other therapeutically equivalent drugs in
these classes were then excluded from the formulary. These four classes are
known as “closed” classes. VA has not made clinical decisions
regarding therapeutic interchange in the remaining 250 drug

4 Under committed- use contracts, VA commits to use primarily the contract
drug, instead of other therapeutically interchangeable drugs, to guarantee
drug companies a high volume of use in exchange for lower prices.

classes, and it does not limit the number of drugs that can be added to
these classes. These are known as “open” classes. In some cases,
drugs listed on the national formulary may be restricted. Restrictions are
generally placed on the use of drugs if they have the potential to be used
inappropriately. For example, restrictions are placed on drugs with
potentially serious side effects, such as interferon, which is used to treat
such conditions as hepatitis C. VA has also adopted guidelines to assist
practitioners in making decisions about the diagnosis, treatment,

and management of specific clinical conditions, such as congestive heart
failure. In addition, VA has adopted criteria to help standardize treatment,
improve the quality of patient care, and promote cost- effective drug

prescriptions. Finally, VA limits prescribing privileges for some drugs to
specially trained physicians and requires consultation with a specialist
before certain drugs can be prescribed. National Formulary

VA has made significant progress in establishing a national formulary, with
Standardization Not most drugs being prescribed from the formulary list.
Nevertheless, VA's oversight has not been sufficient to ensure that it is
fully achieving its Yet Achieved

national formulary goal of standardizing its drug benefit nationwide. We
found that some facilities have omitted required national formulary drugs.
In addition, the extent to which VISNs add drugs to supplement the national
formulary has the potential for conflicting with VA's ability to achieve
standardization if not closely managed. Also, we found that some

facilities, contrary to policy, have modified the list of drugs available in
closed classes. National Formulary Drugs

Almost 3 years after VA facilities were directed to make available locally
all Were Omitted From national formulary drugs, two of the three medical
centers we visited did

Formularies in Medical not list all national formulary drugs in the
formularies used by their Centers We Visited

prescribers. VHA's national formulary policy directive states that items
listed on the national formulary shall be made available throughout the VA
health care system and must be available in all VA facilities. 5 While a
physical supply of all national formulary drugs is not required to be
maintained at all facilities, if a clinical need for a particular formulary
drug

5 Items on the national formulary include both drugs and medical and
surgical supply items, such as diabetic supplies and bandages. Our study
focused exclusively on drugs.

arises in the course of treating a patient, it must be made available to the
patient.

Many drugs listed on the national formulary were not available as formulary
choices in two of the three medical centers we visited. In the

first, about 25 percent (286 drugs) of the national formulary drugs were not
available as formulary choices. These included drugs used to treat high
blood pressure and mental disorders, as well as drugs used to treat the
unique medical needs of women. At the second medical center, about 13
percent (147 drugs) of the national formulary drugs were omitted,

including drugs used to treat certain types of cancer and others used to
treat stomach conditions.

Health care providers at these two medical centers were required to seek
nonformulary drug approvals for over 22, 000 prescriptions of national
formulary drugs from October 1999 through March 2000. If the national
formulary had been properly implemented at these medical centers,
prescribers would not have had to use extra time to request and obtain
nonformulary drug approvals for these drugs, and patients could have started
treatment earlier. Our analysis showed that over 14,000 prescriptions were
filled as nonformulary drugs for 91 of the 286 drugs at the first center. 6
No prescriptions were filled for the remaining 195 drugs.

At the other medical center, over 8,000 prescriptions for 23 of the 147
drugs were filled as nonformulary drugs. No prescriptions were filled for
the remaining 124 drugs.

Drugs Added to Supplement VA's policy allowing VISNs to supplement the
national formulary locally has the National Formulary the potential for
conflicting with VA's ability to achieve standardization if Decrease
Standardization

not closely managed. From June 1997 through March 2000, VISNs added 244
unique drugs to supplement the list of drugs on the national formulary. The
number of drugs added by each VISN varies widely, ranging from as many as 63
by VISN 20 (Portland) to as few as 5 by VISN 8 (Bay Pines).

(Fig. 1 shows the number of drugs added by each VISN.) Adding drugs to
supplement the national formulary is intended to allow VISNs to be
responsive to the unique needs of their patients and to allow quicker

6 After our visit, we were informed by a pharmacy official that the medical
center adopted the national formulary as its own on June 30, 2000.

formulary designation of new FDA- approved drugs. 7 However, the wide
variation in the number of drugs added by the VISNs to supplement the
national formulary raises concern that this practice, if not appropriately
monitored, could result in unacceptable decreases in formulary
standardization. VA officials have acknowledged that this variation affects
standardization and told us they plan to address it. For example, the PBM

plans to review new drugs when approved by the FDA to determine if they will
be added to the national formulary or if VISNs may continue to add them to
their formularies to supplement the national formulary. 7 VA formulary
policy provides that a new drug must be on the market for a minimum of 1
year before it can be added to the national formulary.

Figure 1: Variation in Number of Unique Drugs VISNs Added to Supplement
National Formulary, June 1997- March 2000 Number of Drugs Added

0 to 10 11 to 25

Numbered Areas Specify VISN Designations

26 to 50

Number of Drugs Added

More Than 50 70 60 50 40 30 20 10

0 1

2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 VISN Source: GAO
analysis of data from VA's PBM.

Medical Centers The medical centers we visited also inappropriately modified
the national

Inappropriately Modified formulary list of drugs in the closed classes.
Contrary to VA formulary the List of Drugs in Closed

policy, two of three medical centers added two different drugs to two of the
Classes

four closed classes, and one facility did not make a drug available (see
fig. 2).

Figure 2: Medical Center Modifications to Formulary Closed Classes

a Drug purchased under national committed- use contract. Source: VA medical
centers.

While our analysis was performed at the medical center level, the IOM found
similar nonconformity at the VISN level. Specifically, IOM reported

that 16 of the 22 VISNs modified the list of national formulary drugs for
the closed classes. 8 Most Prescribed Drugs Are

From October 1999 through March 2000, 90 percent of VA outpatient on the
National Formulary

prescriptions were written for national formulary drugs. The percentage of
national formulary drug prescriptions filled by individual VISNs varied
slightly, from 89 percent to 92 percent. We found wider variation among
medical centers within VISNs- 84 percent to 96 percent (see table 1).

Table 1: Percentage of Total Outpatient Prescriptions Filled for National
Formulary Drugs by VISN and Medical Centers Within Each VISN, October 1999
Through March 2000

Range of percentages of Percentages of

prescriptions filled using national VISN

prescriptions filled using formulary drugs, by medical

(office location) national formulary drugs

centers within VISNs

1 (Boston) 91 89- 93 2 (Albany) 89 88- 91 3 (Bronx) 89 87- 94 4 (Pittsburgh)
91 87- 94 5 (Baltimore) 90 87- 94 6 (Durham) 90 88- 92 7 (Atlanta) 90 88- 92
8 (Bay Pines) 90 87- 93 9 (Nashville) 92 89- 95 10 (Cincinnati) 92 90- 94 11
(Ann Arbor) 91 87- 94 12 (Chicago) 92 89- 94 13 (Minneapolis) 91 90- 93 14
(Omaha) 89 84- 92 15 (Kansas City) 92 89- 96 16 (Jackson) 90 87- 94 17
(Dallas) 90 88- 91 18 (Phoenix) 92 85- 95

8 IOM, Description and Analysis of the VA National Formulary, pp. 32- 3.

(Continued From Previous Page)

Range of percentages of Percentages of

prescriptions filled using national VISN

prescriptions filled using formulary drugs, by medical

(office location) national formulary drugs

centers within VISNs

19 (Denver) 92 90- 94 20 (Portland) 91 84- 94 21 (San Francisco) 92 90- 93
22 (Long Beach) 92 90- 94 VA systemwide 90 84- 96 Source: VA's PBM
outpatient prescription database.

The remaining 10 percent of prescriptions filled systemwide were for drugs
VISNs and medical centers added to supplement the national formulary or for
nonformulary drugs. VA's PBM and IOM estimate that drugs added to supplement
the national formulary probably account for about 7 percent of all
prescriptions filled and nonformulary drugs account for approximately 3
percent of all prescriptions filled. However, at the time of our review,
VA's nationwide data could identify a filled prescription only as either a
national formulary drug or not. Without specific information, VA does not
know if the additions are resulting in an appropriate balance between local
needs and national formulary standardization. VA officials told us that they
are

modifying the database to enable it to identify which drugs are added to
supplement the national formulary and which are nonformulary. Approval
Processes for

Medical center approval processes for nonformulary drugs are not always
Nonformulary Drugs timely, and the amount of time needed to obtain such
approvals varied widely across medical centers. In addition, some VISNs have
not Have Weaknesses established processes to collect and analyze data on
nonformulary requests. As a result, VA does not know if approved requests
met its established criteria or if denied requests were appropriate.

Time to Process Although the national formulary directive requires certain
criteria for Nonformulary Drug

approval of nonformulary drugs, it does not dictate a specific nonformulary
Approvals Varies Among approval process. As a result, the processes health
care providers must Facilities follow to obtain nonformulary drugs differ
among VA facilities regarding how requests are made, who receives them, who
approves them, and how

long it takes. In addition, IOM documented wide variations in the
nonformulary drug approval process. Figure 3 shows the steps prescribers
must generally follow to obtain nonformulary and formulary drugs.

Figure 3: Steps Generally Followed to Obtain Nonformulary and Formulary
Drugs

Source: GAO survey and interviews with VA officials.

The person who first receives a nonformulary drug approval request may not
be the person who approves it. For example, 61 percent of prescribers
reported that nonformulary drug requests must first be submitted to a
facility pharmacist, 14 percent said they must first be submitted to
facility pharmacy and therapeutics (P& T) committees, and 8 percent said
they must first be sent to service chiefs. In contrast, 31 percent of
prescribers

reported that it is a facility pharmacist who approves nonformulary drug
requests, 26 percent said that the facility P& T committee approves them,

and 15 percent told us that the facility chief of staff approves them. The
remaining 28 percent reported that various other facility officials or
members of the medical staff approve nonformulary drug requests.

The time required to obtain approval for use of a nonformulary drug varied
greatly depending on the local approval processes. The majority of
prescribers (60 percent) we surveyed reported that it took an average of 9
days to obtain approval for use of nonformulary drugs. 9 But many
prescribers also reported that it took only a few hours (18 percent) or

minutes (22 percent) to obtain such approvals. During our medical center
visits, we observed that some medical center approval processes are less
convenient than others. For example, to obtain approval to use a
nonformulary drug in one facility we visited, prescribers were required to
submit a request in writing to the P& T committee for its review and
approval. Because the P& T committee met only once a month,

the final approval to use the requested drug was sometimes delayed as long
as 30 days. The requesting prescriber, however, could write a prescription
for an immediate 30- day supply if the medication need was urgent.

In contrast, in another medical center we visited, a clinical pharmacist was
assigned to work directly with health care providers to help with drug
selection, establish dose levels, and facilitate the approval of
nonformulary

drugs. In that facility, clinical pharmacists were allowed to approve the
use of nonformulary drugs. If a health care provider believed that a patient
should be prescribed a nonformulary drug, the physician and pharmacist could
consult at the point of care and make a final decision with virtually

no delay. Prescribers in our survey were almost equally divided on the ease
or difficulty of getting nonformulary drug requests approved (see table 2).
9 In emergencies, exceptions are made to allow the patient to obtain the
drug more quickly.

Table 2: Ease of Obtaining Nonformulary Drug Approvals Reported by
Prescribers Response categories Percentage reporting

“Easy” or “very easy” 29

“About as easy as difficult” 40

“Difficult” or “very difficult” 32

Note: Percentages do not total 100 because of rounding. Source: GAO survey.

Regardless of whether the nonformulary drug approval process was perceived
as easy or difficult, the vast majority of prescribers told us such requests
were generally approved. According to our survey results, 65 percent of
prescribers sought approval for a nonformulary drug in 1999. These
prescribers reported that they made, on average, 25 such requests (the
median was 10 requests). We estimated that 84 percent of all prescribers'
nonformulary requests were approved.

When a nonformulary drug request was disapproved, 60 percent of prescribers
reported that they switched to a formulary drug. However, more than one-
quarter of the prescribers who had nonformulary drug requests disapproved
resubmitted their requests with additional information. The majority of
prescribers we surveyed told us they were more likely to

convert VA patients who were on nonformulary drugs obtained at another VA
facility to formulary drugs than to request a nonformulary drug (see table
3). Table 3: Likelihood of Prescribers' Converting Patients From
Nonformulary Drug

Prescriptions to Formulary Drug Prescriptions Response categories Percentage
reporting

“Likely to convert” or “very likely to convert” 64
“As likely to convert as to seek approval for the nonformulary
drug” 18

“Likely to seek approval for the nonformulary drug” or
“very likely to seek approval of nonformulary drug” 18 Source:
GAO survey.

Consequently, patients who move from one area of the country to another or
temporarily seek care in a different VA facility are likely to be switched
from a nonformulary drug to a formulary drug.

Some VISNs Are Not VA's national formulary policy requires that a request to
use a nonformulary Collecting and Analyzing drug be based on at least one of
six criteria: (1) the formulary agent is Nonformulary Drug Data as
contraindicated, (2) the patient has had an adverse reaction to the

Required formulary agent, (3) all formulary alternatives have failed
therapeutically, (4) no formulary alternative exists, (5) the patient has
previously responded to the nonformulary agent and risk is associated with
changing

to the formulary agent, and (6) other circumstances involving compelling
evidence- based reasons exist. Each VISN is responsible for establishing a
process to collect and analyze data concerning nonformulary drug requests.

Contrary to the national formulary policy, not all VISNs have established a
process to collect and analyze nonformulary request data at the VISN and
local levels. Twelve of VA's 22 VISNs reported that they do not collect
information on approved and denied nonformulary drug requests. Three VISNs
reported that they collect information only on approved

nonformulary drug requests, and seven reported that they collect information
for both approved and denied requests. Consequently, data that could help
VISNs, medical centers, and the PBM offices are not always

collected and analyzed for trends in a systematic manner. Such information
could help VA at all levels to determine the extent to which nonformulary
drugs are being requested and whether medical center processes for approving
these requests meet established criteria. In its report, IOM noted

that inadequate documentation could diminish confidence in the nonformulary
process.

Prescribers Report Seventy percent of VA prescribers in our survey reported
that the formulary

Having Access to the they use contains the drugs their patients need either
to a “great extent” or to a “very great extent.”
Twenty- seven percent reported that the formulary Drugs Veterans Need meets
their patients' needs to a “moderate extent,” with 4 percent
reporting that it meets their patients' needs to “some extent.”
No VA prescribers reported that the formulary meets their patients' needs to
a “very little or no extent.” This is consistent with IOM's
conclusion that the VA formulary

“is not overly restrictive.”

Overall, two and one- half times as many prescribers indicated that the
formulary they currently use “helps” or “greatly
helps” their ability to prescribe drugs as those who said it
“hinders” or “greatly hinders” them

(see table 4).

Table 4: Prescribers Reported Current Formulary Helps More Than Hinders
Their Ability to Prescribe Drugs Response categories Percentage responding

“Helps” or “greatly helps” 48 “Helps as much
as hinders” 34 “Hinders” or “greatly hinders”
18 Source: GAO survey.

Some prescribers reported that the formulary they use helps them keep
current with new drugs and helps remove some of the pressures created by
direct- to- consumer advertising. Other prescribers reported that newly

approved drugs are not made available on the national formulary as soon as
they would like, and some reported their frustration with delays experienced
when certain formulary drugs must be approved by specially trained
physicians before they can be prescribed.

Prescribers we surveyed reported they were generally satisfied with the
national formulary. We asked prescribers who said that they had worked for
VA before the national formulary was established whether the current
formulary does a better job of keeping the list of drugs in the drug classes
from which they most frequently prescribe up to date, as compared with the
formulary they used to use. Three- quarters told us that they had worked

for VA before the national formulary was implemented in June 1997.
Thirtyeight percent of these prescribers reported that the national
formulary was “better” or “considerably better” than
previous formularies. About half (48

percent) indicated that the current formulary was “about the
same” as the one it replaced. Seven percent reported that it was
“worse” or “considerably worse” than previous
formularies.

Few veterans have complained about not being able to obtain the drugs they
believe they need. At the VA medical centers we visited, patient

advocates 10 told us that veterans made very few complaints concerning their
prescriptions. In its analysis of the patient advocates' complaint
databases, IOM found that less than one- half of one percent of veterans'
complaints were related to drug access. 11 IOM further reported that

complaints involving specific identifiable drugs often involved drugs that
are marketed directly to consumers, such as sildenafil (Viagra), which is
used to treat erectile dysfunction. 12 Fifty- one percent of the prescribers
in our survey reported that over the past 3 years, an increasing number of
their patients have requested a drug they have seen or heard advertised in
the media. Our review also indicated that the few prescription complaints
made were often related to veterans' trying to obtain
“lifestyle” drugs or refusals by VA physicians and pharmacists
to fill prescriptions written by

non- VA health care providers. 13 VA officials told us that VA does not fill
prescriptions written by non- VA- authorized prescribers, in part to ensure
that one practitioner manages a patient's care.

Conclusions Over the past 3ï¿½ years, VA has made significant progress in
establishing its national formulary, which has generally met with prescriber
acceptance. Prescribers reported that veterans are generally receiving the
drugs they need and that veterans rarely register complaints concerning
prescription drugs.

VA has not provided sufficient oversight, however, to ensure that VISNs and
medical centers comply with formulary policies and that the flexibility
given to them does not unduly compromise VA's goal of formulary
standardization. Contrary to VA formulary policy, some facilities omitted
national formulary drugs or modified the closed drug classes. While adding

10 Patient advocates are VA employees who are responsible for receiving and
acting on complaints from veterans. 11 IOM obtained formulary- related
complaints from a nationwide database of veteran complaints for over 90
percent of all VA facilities representing all 22 VISNs. IOM determined that
only 2,385 of 570, 937 veteran complaints were attributed to the national
formulary. No VISN had significantly more complaints than any other. (IOM,
Description and Analysis of

the VA National Formulary, p. 145). 12 Sildenafil is available within VA
only through the nonformulary drug approval process.

13 We asked prescribers in our survey how often in 1999 their patients asked
them to rewrite a prescription from a non- VA prescriber so that it could be
filled by VA. Thirty- one percent said “often” or “very
often,” 34 percent reported that it occurred
“occasionally,” and 21 percent said “seldom.”
Fourteen percent said that they never received such requests.

a limited number of drugs to supplement the national formulary is permitted,
as more drugs are added by VISNs, formulary differences among facilities are
likely to become more pronounced, decreasing formulary standardization.
While VA recognizes the trade- off between local flexibility and
standardization, it lacks criteria for determining the appropriateness of
adding drugs to supplement the national formulary. Consequently, VA cannot
determine whether the resulting decrease in standardization is acceptable.
Not all VISN directors have met their responsibilities for implementing
national formulary policy.

Inefficiencies that exist in the nonformulary drug approval processes across
the system can cause delays in making final treatment decisions. In
addition, the processes require health care provider time and energy that
might be better used for direct patient care. We believe a more efficient
nonformulary drug approval process could enable facilities to benefit from
lessons learned in other locations. Finally, VISNs lack the data needed to

analyze nonformulary drug requests to determine whether all approved
requests met approval criteria and all denied requests were appropriate.

Recommendations for In order to ensure more effective management of the
national formulary, Executive Action

we recommend that the Secretary of Veterans Affairs direct the Under
Secretary for Health to take the following actions:

? Establish a mechanism to ensure that VISN directors comply with national
formulary policy. ? Establish criteria that VISNs should use to determine
the

appropriateness of adding drugs to supplement the national formulary and
monitor the VISNs' application of these criteria. ? Establish a nonformulary
drug approval process for medical centers

that ensures appropriate and timely decisions and provides that veterans for
whom a nonformulary drug has been approved will have continued access to
that drug, when appropriate, across VA's health care

system. ? Enforce existing requirements that VISNs collect and analyze the
data needed to determine that nonformulary drug approval processes are
implemented appropriately and effectively in their medical centers,
including tracking both approved and denied requests.

Agency Comments and In commenting on a draft of this report, VA agreed with
our findings and

Our Evaluation concurred with our recommendations. VA highlighted key
improvements planned or already in progress that should further enhance the
process.

VA's actions to address our recommendations are summarized below. ? VA plans
to improve oversight at all organizational levels to help facilitate
consistent compliance with national formulary policy. In its

comments, VA discussed important components of improving compliance with the
national formulary, including examining data to identify outliers. However,
VA did not articulate a mechanism for ensuring that its oversight results in
consistent compliance, which may reduce the effectiveness of its planned
actions.

? VA plans to establish criteria for VISNs to use to determine the
appropriateness of adding drugs to supplement the national formulary. ? VA
plans to establish steps for its nonformulary drug approval process that all
medical centers and VISNs must follow. However, in its comments, VA did not
specifically address how veterans would have

continued access to previously approved nonformulary drugs across VA's
health care system. We believe such access is important. ? VA plans to
establish steps for reporting its nonformulary approval activities. In its
comments, VA did not explicitly include tracking of denied requests as part
of the nonformulary approval activities. We expect that its nonformulary
approval activities will include tracking

denied requests, as well as approved nonformulary drug requests, to
determine the appropriateness of all medical center prescribing decisions.
VA plans to implement these corrective actions by June 2001. Its comments
are included in appendix II.

We are sending copies of this report to the Honorable Anthony J. Principi,
Secretary of Veterans Affairs; appropriate congressional committees; and
other interested parties. We will also make copies available to others upon
request.

Please call me at (202) 512- 7101 if you or your staff have questions about
this report or need additional assistance. Another contact and staff
acknowledgments are listed in appendix III.

Sincerely yours, Cynthia A. Bascetta, Director Health Care- Veterans' Health
and Benefits Issues

Appendi Appendi xes x I

Scope and Methodology To obtain policies and procedures from the 22 Veterans
Integrated Service Networks (VISN), we mailed a questionnaire to each of the
22 VISN formulary leaders- pharmacists or physicians who serve on the
Department of Veterans Affairs' (VA) Pharmacy Benefits Management advisory
board. To determine the extent to which VA health care providers

write prescriptions for national formulary drugs, we analyzed data from VA's
national outpatient prescription database. To assess the implementation of
the national formulary and obtain firsthand opinions about it, we
interviewed medical and administrative staff at three VA medical centers
located in three different VISNs. To obtain VA health care

providers' views on VA's formulary, including whether or not it is
restrictive, we mailed a questionnaire to a nationally representative sample
of 2,000 VA health care prescribers. We also used information contained in
the Institute of Medicine's Description and Analysis of the VA National
Formulary, issued in June 2000. VISN Survey To obtain policies and
procedures from the 22 VISNs, we mailed a questionnaire to VISN formulary
leaders. 1 We asked if there were VISNwide policies for several areas,
including adding drugs to the VISN

formulary, requesting nonformulary drugs, converting patients from one drug
to another, and tracking requests for nonformulary drugs. In addition, we
sought information on the number of drugs added to and dropped from the VISN
formulary, the number of requests for nonformulary drugs, and the number of
requests that were approved and denied. All 22 VISN formulary leaders
completed and returned questionnaires.

1 Each VISN has a formulary leader- typically a senior pharmacy manager or
physician. Collectively, the 22 leaders compose a national advisory board
that assists VA's Pharmacy Benefits Management Strategic Healthcare Group
(PBM) and its medical advisory panel. Formulary leaders are knowledgeable in
pharmacological matters, including drugs and treatment practices, and in
national and local formulary policies and procedures. The role of a
formulary leader includes helping PBM decide what drugs will be added to or
deleted from the national formulary, developing and adopting national
policies affecting the use of drugs on the national formulary, and assisting
with other matters the leaders or PBM officials raise during their periodic
meetings.

Outpatient VA's national database on outpatient prescriptions contains
information for

Prescription Database each outpatient prescription filled at each VA medical
center, including the

drug prescribed, date of the prescription, patient and prescriber
identifiers, Analysis medical center responsible for filling the
prescription, and whether the prescribed drug is a national formulary drug.
We used this database to

? develop a sample of VA health care providers who wrote prescriptions, ?
determine the total number of outpatient prescriptions filled at VISNs

and VA medical centers, ? determine the number of filled outpatient
prescriptions written for

national formulary drugs within a certain time frame, 2 and ? determine how
many VISN formulary drug prescriptions were filled in the three VISNs where
we performed site visits.

We interviewed PBM headquarters officials who had either oversight or
maintenance responsibility for the database to help assess the validity and
reliability of the outpatient prescription data. We also performed our own
analytic checks of the data. We found that data critical to our analysis-
the data field indicating whether a prescription had been written for a
national formulary drug- contained errors. We worked with PBM officials to
correct the data, and they implemented a monthly routine to detect and

correct these errors in the future. We reran our data checks, verified that
the database had been corrected, and concluded that the data were acceptable
for the purposes of our work.

Site Visits To assess formulary implementation at the local level, we
interviewed medical and administrative staff at three different VA medical
centers- one

located in Biloxi, Mississippi (VISN 16); one in Gainesville, Florida (VISN
8); and one in Omaha, Nebraska (VISN 14). We selected these VISNs and
medical centers on the basis of formulary drug use from October through
December 1999, the period for which the most recent and complete data were
available at the time we did our work. For example, VISN 8 had the

highest percentage of prescriptions for national formulary drugs (93
percent), VISN 16's percentage of national formulary drug prescriptions was
at the national average (90 percent), and VISN 14 had the lowest 2 October
1, 1999, to March 31, 2000, was the time frame for which the most recent and

complete data were available at the time of our work.

percentage of prescriptions filled using national formulary drugs (88
percent).

Questionnaire to We mailed questionnaires to a representative sample of
2,000 VA health

Prescribers care prescribers whose prescriptions had been dispensed from
October 1

through December 31, 1999, to obtain their opinions and experiential data on
various aspects of VA's national formulary. We drew this random sample from
VA's most recent national outpatient prescription database- a data file that
contains information, including a prescriber identifier, on all

outpatient prescriptions filled in the VA health care system. We mailed
questionnaires to the entire sample of prescribers on April 17, 2000, with
follow- up mailings on May 17 and June 21 to those who had not responded by
those dates. We accepted returned questionnaires through

September 1, 2000. Some prescribers' responses indicated that they did not
write prescriptions for drugs; their prescription privileges were limited to
medical and surgical supplies, such as diabetic strips and food

supplements. Other returned questionnaires indicated that the addressee had
either left or retired from VA. These providers were thus considered
ineligible for our purposes and were removed from the sample. Approximately
11 percent of the questionnaires were returned as undeliverable, and we
received no response from approximately 16 percent

of those to whom we mailed questionnaires. After adjusting the sample
accordingly, we determined the number of useable returned questionnaires to
be 1, 217- a response rate of about 69 percent. (See table 5.)

Table 5: Questionnaire Sample and Response Size Number Total questionnaires
mailed 2, 000

Returned by nonprescribers a 225

Adjusted sample size 1, 775

Not deliverable by Postal Service 226 Delivered but not returned 328
Received after closing date 4

Useable questionnaires returned 1, 217 Final response rate 69%

a These are uncompleted, but returned, questionnaires. They represent
individuals who indicated that they either did not have drug prescribing
privileges (50 people, or 2.6 percent of the sample) or had left or had
retired from VA (175 people, or 8.7 percent of the sample.) These
individuals would have been excluded from the original sample if that
information had been known in advance. Because this was a simple random
sample, we believe that our results are projectable to all of VA's health
care providers who have outpatient drug prescribing privileges.

Sampling Errors Surveys based on a sample are subject to sampling errors.
Sampling error represents the extent to which a survey's results differ from
what would have been obtained had everyone in the universe of interest
received and returned the same questionnaire- in this case, all VA health
care providers who have outpatient drug prescribing privileges. Sampling
errors have two elements: the width of the confidence interval around the
estimate (sometimes referred to as the precision of the estimate) and the
confidence level at which the confidence interval is computed.

The confidence interval reflects the fact that estimates actually encompass
a range of possible values, not just a single value, or a “point
estimate.” The interval is expressed as a point, plus or minus some
value. For example, in our questionnaire, we asked prescribers, “To
what extent does your VA formulary contain the drugs you believe your
patients need?” The percentage of respondents who reported a
“great extent” or “very great

extent” was 69.1. This particular question had a confidence interval
of plus or minus 2.6 percentage points. Thus, the “true” answer
for this question may or may not be 69.1 percent, but it has a high
probability of falling between 66. 5 and 71.7 percent (69.1- percent point
estimate, plus or minus 2.6 percentage points). Confidence intervals vary
for individual questions (depending upon how many of the individuals who
could have answered a question actually did so), but, unless otherwise
noted, all percentages presented in this report are within a range of plus
or minus 3. 5 percentage points.

The confidence level is a measure of how certain we are that the
“true” answer lies within a confidence interval. We used a 95-
percent confidence level, which means that if we repeatedly took new samples
of prescribers

from the October through December prescription database and performed the
same analysis of their responses each time, 95 percent of these samples
would yield estimates that would fall within the confidence interval stated.

In the previous example, this means that we are 95- percent certain that
between 66. 5 and 71.7 percent of prescribers believe that the VA formulary

contains to a “great extent” or “very great extent”
the drugs they believe their patients need.

Nonsampling Errors Surveys can also be subject to other types of systematic
error or bias that can affect results, known as nonsampling errors. One
potential source of nonsampling error can be the questionnaire itself. To
ensure that questions were clear and unbiased, we consulted with subject
matter and questionnaire experts within GAO and obtained comments from
individuals representing VA's PBM and medical advisory panel, a working
group of 11 practicing VA physicians and 1 practicing Department of Defense
physician who help manage VA's national formulary, as well as

individuals representing the Institute of Medicine. Finally, the
questionnaire was tested with 14 VA prescribers in VA medical centers in
four locations: Phoenix, Arizona; Washington, D. C.; Hampton, Virginia; and
Cincinnati, Ohio.

Prescribers' Demographics Prescribers were asked to provide demographic and
VA employment information as well as opinions about the relevance and
usefulness of VA's formulary. On average, VA prescribers in our sample have
worked for VA for 11 years, with most of those years at their current
medical facility. Physicians and nurses constitute the largest groups of
prescribers (65 and 15 percent, respectively), followed by physician
assistants (7 percent) and other allied health professionals, such as
dentists (14 percent). Most of the prescribers' time working in VA is spent
treating patients- on average, 26 hours each week. According to the national
prescription file from which we drew our sample, VA prescribers who
completed our questionnaire averaged 849 prescription fills from October
through December 1999, the 3month period we chose as the basis of our
survey. The median number of filled prescriptions was relatively low- 252-
because a few prescribers had a large number of prescriptions filled during
the period, while many

prescribers had only a few prescriptions filled.

Comments From the Department of Veterans

Appendi x II Affairs

Appendi x I II

GAO Contact and Staff Acknowledgments GAO Contact Walter Gembacz, (202) 512-
6982 Staff

George Poindexter, Stuart Fleishman, Mike O'Dell, and Kathie Kendrick
Acknowledgments

made key contributions to this report.

Related GAO Products

State Pharmacy Programs: Assistance Designed to Target Coverage and Stretch
Budgets (GAO/ HEHS- 00- 162, Sept. 6, 2000).

Prescription Drug Benefits: Applying Private Sector Management Methods to
Medicare (GAO/ T- HEHS- 00- 84, Mar. 22, 2000).

VA Health Care: VA's Management of Drugs on Its National Formulary

(GAO/ HEHS- 00- 34, Dec. 14, 1999).

Prescription Drug Benefits: Implications for Beneficiaries of Medicare HMO
Use of Formularies (GAO/ HEHS- 99- 166, July 20, 1999).

Defense Health Care: Fully Integrated Pharmacy System Would Improve Service
and Cost- Effectiveness (GAO/ HEHS- 98- 176, June 12, 1998).

(406187) Lett er

GAO United States General Accounting Office

Page 1 GAO- 01- 183 VA Drug Formulary Oversight

Page 2 GAO- 01- 183 VA Drug Formulary Oversight

Page 3 GAO- 01- 183 VA Drug Formulary Oversight United States General
Accounting Office

Washington, D. C. 20548 Page 3 GAO- 01- 183 VA Drug Formulary Oversight

Page 4 GAO- 01- 183 VA Drug Formulary Oversight

Page 5 GAO- 01- 183 VA Drug Formulary Oversight

Page 6 GAO- 01- 183 VA Drug Formulary Oversight

Page 7 GAO- 01- 183 VA Drug Formulary Oversight

Page 8 GAO- 01- 183 VA Drug Formulary Oversight

Page 9 GAO- 01- 183 VA Drug Formulary Oversight

Page 10 GAO- 01- 183 VA Drug Formulary Oversight

Page 11 GAO- 01- 183 VA Drug Formulary Oversight

Page 12 GAO- 01- 183 VA Drug Formulary Oversight

Page 13 GAO- 01- 183 VA Drug Formulary Oversight

Page 14 GAO- 01- 183 VA Drug Formulary Oversight

Page 15 GAO- 01- 183 VA Drug Formulary Oversight

Page 16 GAO- 01- 183 VA Drug Formulary Oversight

Page 17 GAO- 01- 183 VA Drug Formulary Oversight

Page 18 GAO- 01- 183 VA Drug Formulary Oversight

Page 19 GAO- 01- 183 VA Drug Formulary Oversight

Page 20 GAO- 01- 183 VA Drug Formulary Oversight

Page 21 GAO- 01- 183 VA Drug Formulary Oversight

Page 22 GAO- 01- 183 VA Drug Formulary Oversight

Page 23 GAO- 01- 183 VA Drug Formulary Oversight

Page 24 GAO- 01- 183 VA Drug Formulary Oversight

Appendix I

Appendix I Scope and Methodology

Page 25 GAO- 01- 183 VA Drug Formulary Oversight

Appendix I Scope and Methodology

Page 26 GAO- 01- 183 VA Drug Formulary Oversight

Appendix I Scope and Methodology

Page 27 GAO- 01- 183 VA Drug Formulary Oversight

Appendix I Scope and Methodology

Page 28 GAO- 01- 183 VA Drug Formulary Oversight

Page 29 GAO- 01- 183 VA Drug Formulary Oversight

Appendix II

Appendix II Comments From the Department of Veterans Affairs

Page 30 GAO- 01- 183 VA Drug Formulary Oversight

Appendix II Comments From the Department of Veterans Affairs

Page 31 GAO- 01- 183 VA Drug Formulary Oversight

Appendix II Comments From the Department of Veterans Affairs

Page 32 GAO- 01- 183 VA Drug Formulary Oversight

Page 33 GAO- 01- 183 VA Drug Formulary Oversight

Appendix III

Page 34 GAO- 01- 183 VA Drug Formulary Oversight

Ordering Information The first copy of each GAO report is free. Additional
copies of reports are $2 each. A check or money order should be made out to

the Superintendent of Documents. VISA and MasterCard credit cards are
accepted, also. Orders for 100 or more copies to be mailed to a single
address are discounted 25 percent.

Orders by mail: U. S. General Accounting Office P. O. Box 37050 Washington,
DC 20013

Orders by visiting: Room 1100 700 4th St. NW (corner of 4th and G Sts. NW)
U. S. General Accounting Office Washington, DC

Orders by phone: (202) 512- 6000 fax: (202) 512- 6061 TDD (202) 512- 2537

Each day, GAO issues a list of newly available reports and testimony. To
receive facsimile copies of the daily list or any list from the past 30
days, please call (202) 512- 6000 using a touchtone phone. A recorded menu
will provide information on how to obtain these lists.

Orders by Internet: For information on how to access GAO reports on the
Internet, send an e- mail message with “info” in the body to:

info@ www. gao. gov or visit GAO's World Wide Web home page at: http:// www.
gao. gov

To Report Fraud, Waste, or Abuse in Federal Programs

Contact one: ? Web site: http:// www. gao. gov/ fraudnet/ fraudnet. htm ? e-
mail: fraudnet@ gao. gov ? 1- 800- 424- 5454 (automated answering system)

United States General Accounting Office Washington, D. C. 20548- 0001

Official Business Penalty for Private Use $300

Address Correction Requested Presorted Standard

Postage & Fees Paid GAO Permit No. GI00
*** End of document. ***