Prescription Drugs and Medicaid: Automated Review Systems Can Help
Promote Safety, Save Money (Letter Report, 06/11/96, GAO/AIMD-96-72).

Inappropriate use of prescription drugs can lead to drug-induced
illness, hospitalization, and even death. Inappropriate drug use can
also prove expensive for the Medicaid program. As a result, Congress
mandated that states establish utilization review programs--called
prospective reviews--to review Medicaid prescriptions before drugs are
dispensed. Automated prospective drug utilization review systems are
proving a low-cost way for states to help both doctors and pharmacies
safeguard Medicaid recipients from potentially harmful medical
reactions. Although the main emphasis of these
systems--appropriately--has been safety, both safety benefits and dollar
savings accrue from their use. Because results vary on the basis of how
such systems are administered, it is important that states share their
experiences. Absent any analysis of data from the Iowa demonstration
project or any concerted effort by the Health Care Financing
Administration to collect and share other states' experiences, states
have had only limited access to both safety and cost data--information
that is critical to informed decisionmaking and to maximizing the
effectiveness and efficiency of automated prospective drug utilization
review systems.

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

 REPORTNUM:  AIMD-96-72
     TITLE:  Prescription Drugs and Medicaid: Automated Review Systems 
             Can Help Promote Safety, Save Money
      DATE:  06/11/96
   SUBJECT:  Drugs
             Health care cost control
             Health care programs
             Medical information systems
             Patient care services
             Therapy
IDENTIFIER:  Medicaid Program
             Medicaid Management Information System
             Maryland
             Missouri
             New Mexico
             Oregon
             Pennsylvania
             
******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO report.  Delineations within the text indicating chapter **
** titles, headings, and bullets are preserved.  Major          **
** divisions and subdivisions of the text, such as Chapters,    **
** Sections, and Appendixes, are identified by double and       **
** single lines.  The numbers on the right end of these lines   **
** indicate the position of each of the subsections in the      **
** document outline.  These numbers do NOT correspond with the  **
** page numbers of the printed product.                         **
**                                                              **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced.  Tables are included, but    **
** may not resemble those in the printed version.               **
**                                                              **
** Please see the PDF (Portable Document Format) file, when     **
** available, for a complete electronic file of the printed     **
** document's contents.                                         **
**                                                              **
** A printed copy of this report may be obtained from the GAO   **
** Document Distribution Center.  For further details, please   **
** send an e-mail message to:                                   **
**                                                              **
**                                            **
**                                                              **
** with the message 'info' in the body.                         **
******************************************************************


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


Report to Congressional Requesters

June 1996

PRESCRIPTION DRUGS AND MEDICAID -
AUTOMATED REVIEW SYSTEMS CAN HELP
PROMOTE SAFETY, SAVE MONEY

GAO/AIMD-96-72

Automated Prescription Drug Review Systems

(511188)


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

  CFR - code of federal regulations
  DUR - drug utilization review
  FDA - Food and Drug Administration
  HCFA - Health Care Financing Administration
  HHS - Department of Health and Human Services
  MMIS - Medicaid Management Information System
  PRODUR - prospective drug utilization review
  Rx - prescription

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


B-261530

June 11, 1996

The Honorable William F.  Clinger, Jr.
Chairman
Committee on Government Reform and Oversight
House of Representatives

The Honorable Christopher Shays
Chairman
Subcommittee on Human Resources
 and Intergovernmental Relations
Committee on Government Reform and Oversight
House of Representatives

The Honorable Edolphus Towns
Ranking Democratic Member
Subcommittee on Human Resources
 and Intergovernmental Relations
Committee on Government Reform and Oversight
House of Representatives

It is widely accepted in the health care community that inappropriate
use of prescription drugs can cause adverse reactions that can lead
to drug-induced illness, hospitalization, even death.  Such
inappropriate use can also be expensive for the Medicaid program. 
Concerned about this issue, the Congress mandated that states
establish utilization review programs to review Medicaid
prescriptions before drugs are dispensed (called prospective reviews)
in order to prevent potential adverse medical reactions.  The
legislation did not require that prospective screening be automated. 
However, 43 states plus the District of Columbia have implemented or
plan to implement automated prospective drug utilization review
(PRODUR) systems.  In most instances, PRODUR systems are implemented
concurrently with an automated screening capability for Medicaid
eligibility since both depend on automated systems that offer
real-time responses to inquiries.  The five states in our review have
this feature. 

In August 1994 we reported on the potential benefits to the Medicaid
program of automated PRODUR systems, after examining how information
technology could be used in implementing a drug utilization review
(DUR) program.\1 In response to your request, this report examines
states' actual experiences in using automated PRODUR systems in their
Medicaid programs, recognizing that such experience has been brief. 
Specifically, you asked that we focus on how such systems can (1)
improve patient safety by identifying and preventing inappropriate
drug therapy, (2) reduce program costs, and (3) reduce the incidence
of fraud, waste, and abuse.  You also asked that we determine any
other concerns that would hinder the effective implementation of
these systems.  Finally, we discuss the varying ways in which these
systems can be established, and the role that the Department of
Health and Human Services' (HHS) Health Care Financing Administration
(HCFA) and others could play in enhancing states' effective use of
PRODUR systems. 


--------------------
\1 Prescription Drugs:  Automated Prospective Review Systems Offer
Potential Benefits for Medicaid (GAO/AIMD-94-130, Aug.  5, 1994). 


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

Automated prospective drug utilization review (PRODUR) systems
increased patient safety and reduced Medicaid program costs in the
five states whose systems we examined; all five states found the
systems beneficial and worthwhile.  During a 12-month period ending
June 30, 1995, these five states' systems alerted pharmacists to over
6.3 million prescriptions that had the potential to cause adverse
medical reactions from drug therapy problems including drug-drug
interaction, overutilization, and pregnancy conflict.\2 Over 650,000
(10 percent) of these prescriptions were canceled because of the
potential serious risk to patients.  According to state officials,
pharmacists reviewed and eventually filled the other prescriptions on
the basis of the pharmacists' professional judgment and/or
consultation with the recipients or their physicians. 

Along with increasing patient safety, these systems also reduced
program costs by over $30 million, according to state and contractor
reports.  Over $5 million of this total was attributable to rejecting
efforts to refill prescriptions before a large portion of the earlier
prescription would have been consumed, potentially causing harm to
the patient; the remaining $25 million resulted from prescriptions
that were denied due to patient ineligibility.  While these direct
benefits are significant, the major dollar savings, in all
likelihood--though more difficult to document--are achieved through
avoided hospitalization due to inappropriate drug therapy.  On the
basis of its review of studies related to drug-induced illnesses, the
Food and Drug Administration (FDA) indicated that 6.4 percent of
hospital admissions nationwide can be traced to inappropriate drug
therapy; some studies cite even higher rates for the elderly. 
Accordingly, avoided hospitalization could potentially save hundreds
of millions of dollars annually.  Savings could also accrue because
PRODUR systems can help identify potential fraud, waste, and abuse. 

The five states in our sample screened for different conditions and
handled prescription cancellations differently.  Consequently,
reported numbers and types of patient safety alerts, prescription
cancellations, and rates of savings varied.  One state, for example,
did not screen for pregnancy conflict.  Three states automatically
deny prescriptions with overutilization alerts, while the other two
states place the responsibility with pharmacists to either deny or
fill the prescriptions following such alerts.  At present, states
have no systematic way to share experiences and best practices.  One
approach toward accomplishing this would entail establishing a
central clearinghouse at the state or federal level to collect and
disseminate information.  This would allow all states to make more
informed decisions, offering citizens the best protection and states
the most savings. 


--------------------
\2 New Mexico's alert data cover only the 6 months between January
and June 1995. 


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

Authorized in 1965 under Title XIX of the Social Security Act,
Medicaid is a federally aided, state-run medical assistance program. 
Over 35 million people received an estimated $152 billion in Medicaid
services during fiscal year 1995.  At the federal level, HCFA is
responsible for administering the Medicaid program, establishing
policies, developing operating guidelines, and ensuring states'
compliance with Medicaid regulations. 

Federal regulations require states to provide certain basic medical
services, such as inpatient hospital and physician care, under their
Medicaid programs.  Federal regulations also authorize states to
provide optional services, such as optometrist services, eyeglasses,
and prescription drugs.  During fiscal year 1995, all states provided
prescription drugs as part of their Medicaid programs; the reported
cost of these drugs during that year was about $8.3 billion. 

Because of growing concern over the increased use and cost of
prescription drugs, in 1990 the Congress amended the Social Security
Act to require states to implement DUR programs by January 1, 1993.\3
The legislation mandated that these reviews include prospective
screening for potential drug problems due to therapeutic duplication,
drug-disease contraindication, drug-drug interaction, incorrect drug
dosage, incorrect duration of treatment, drug-allergy interactions,
and clinical abuse or misuse.  (A glossary of drug utilization review
terms appears at the end of this report.) HCFA was instructed to
issue guidelines to the states on prospective DUR cost and benefit
reporting.  While not requiring that states use automated systems for
prospective screening, the law encouraged their use, and required
HCFA to initiate at least 10 statewide automated PRODUR demonstration
projects.  HCFA was required to report to the Congress by January 1,
1994. 

In addition to promoting patient safety, DUR programs must also be
designed to educate physicians and pharmacists to better identify
patterns of fraud, abuse (including overuse), or other inappropriate
or medically unnecessary care.  Physicians and pharmacists can also
use these systems to increase their knowledge about patterns of use
associated with specific drugs or groups of drugs. 


--------------------
\3 Public Law 101-508, November 5, 1990 (Omnibus Budget
Reconciliation Act of 1990). 


      HOW THE AUTOMATED SYSTEM
      WORKS
---------------------------------------------------------- Letter :2.1

When a Medicaid patient submits a prescription to be filled, the
pharmacist transmits recipient identification and prescription
information to a statewide database via the PRODUR system.\4 In an
on-line, real-time environment, after verifying the recipient's
eligibility, the system screens the prescription against the
recipient's known Medicaid medical and prescription history.  The
system then sends the pharmacy a message indicating whether the claim
is "payable" (valid), and whether any potential drug therapy problem,
such as a drug-drug interaction, exists. 

If a potential drug therapy problem does exist, the pharmacist
consults with the recipient and/or the recipient's physician,
depending upon the seriousness of the problem.  After such
consultation and according to the pharmacist's judgment, the
pharmacist may fill the prescription, resubmit the claim for a
different drug prescribed by the physician, or submit a reversal to
cancel the claim. 

Pharmacies in states that do not use automated, statewide PRODUR
systems are generally limited to comparing the prescription presented
with the patient's medical history and prescription data maintained
at that specific pharmacy or chain of pharmacies.  Such a local
system would not have the benefit of the patient's complete Medicaid
history; moreover, a local system would lack on-line eligibility
verification. 


--------------------
\4 Each state's PRODUR system is integrated with its Medicaid
Management Information System (MMIS), used by the state to process
Medicaid claims; each MMIS includes a statewide database of patients'
drug and medical histories. 


   SCOPE AND METHODOLOGY
------------------------------------------------------------ Letter :3

To determine if states' PRODUR systems have improved patient safety,
we obtained contractor-prepared reports and other data for five of
the eight states that had been operating automated PRODUR systems for
12 months or longer at the time we began our review in May 1995: 
Maryland, Missouri, New Mexico, Oregon, and Pennsylvania.\5 These
reports showed the number and types of drug therapy "alerts"
transmitted by each state's PRODUR system, along with the number of
claims pharmacies reversed as a result of such alerts.  We selected
these five states because (1) a single contractor serviced three of
the states (Maryland, New Mexico, and Oregon) and agreed to perform
the needed analyses, and (2) two states (Missouri and Pennsylvania)
agreed to provide us with 12 months' worth of their Medicaid
prescription drug data for our own analysis. 

To determine the extent to which states' PRODUR systems have provided
measurable savings, we and contractors analyzed the five states'
PRODUR data for all conditions screened for.  To determine actual
savings realized from the denial of overutilization claims (often
called "early refills"\6 ), we analyzed all such transactions denied
between July 1, 1994, and March 31, 1995, identifying whether they
were subsequently refilled within 90 days--including the period April
1995 through June 1995.\7 State PRODUR and HCFA officials agreed with
our assumption that if the prescriptions were not refilled within 90
days after being initially presented, they probably would not be
refilled, and that the value of these prescriptions could reasonably
be viewed as measurable savings. 

To evaluate the extent to which on-line eligibility-verification
screening provided tangible savings, we identified the number of
prescriptions denied (and the dollar value of the drugs) because
recipients were not eligible for Medicaid benefits on the day they
submitted their prescriptions.  Eligibility screening is performed by
a system separate from but often utilized with PRODUR systems.  In
addition, our estimates of cost savings derived from avoided
hospitalization are based on FDA data. 

To discern the degree to which PRODUR systems can assist in
identifying potential fraud, waste, and abuse, we reviewed alerts for
overutilization (early refill) and therapeutic duplication; in such
cases, individuals may be trying to obtain for resale a greater
quantity of medication than that prescribed. 

We performed our review from May 1995 through April 1996, in
accordance with generally accepted government auditing standards
except that, as agreed with the requesters, we did not independently
verify the accuracy of state- and contractor-provided
prescription-transaction data.  This was due to the excessive amount
of time that would have been required to (1) obtain specific
prescription information (such as physician and pharmacy names) and
(2) trace these data back to the original prescription documents. 
Accordingly, we cannot verify the accuracy of the information
provided.  Our conclusions regarding patient safety and program
savings are based on these data.  HHS provided written comments on a
draft copy of this report; they are reprinted as appendix I. 


--------------------
\5 Two of the states included in our earlier review (see
GAO/AIMD-94-130, Aug.  5, 1994)--Tennessee and West Virginia--were
not included in this review because of data-availability problems. 
Therefore, two other states--New Mexico and Oregon--were added to
Maryland, Missouri, and Pennsylvania. 

\6 Early refills include prescriptions submitted for the same drug,
for the same person, and/or by the same or a different pharmacy
before a predetermined amount of the original prescription--such as
75 percent--would theoretically have been consumed by taking the
dosage as prescribed. 

\7 For New Mexico, early-refill transactions between January 1 and
September 30 were analyzed--including whether any prescriptions
initially denied were filled within the following 90 days, through
December 31, 1994. 


   STATE DATA SUGGEST THAT
   AUTOMATED STATEWIDE SYSTEMS
   ENHANCE PATIENT SAFETY
------------------------------------------------------------ Letter :4

Data from PRODUR systems operated by five geographically diverse
states--Maryland, Missouri, New Mexico, Oregon, and
Pennsylvania--show inappropriate drug therapy to be an ongoing and
serious problem in the Medicaid program.  These systems reported
screening over 31.7 million prescription drug claims during 12-month
periods between January 1994 and June 1995, and sent pharmacists
alerts of potentially inappropriate drug therapy for about 6.3
million (20 percent) of these claims.  Over 650,000 of these claims
(2 percent of the total) were reported canceled because of serious
risks posed by such conditions as potential drug-drug interaction,
drug-disease conflict, and pregnancy conflict. 

Table 1 shows the potential for adverse medical reactions, as
evidenced by the fact that alerts for drug-drug interaction,
overutilization, and therapeutic duplication occur in great numbers. 
(Appendix II contains detailed data on PRODUR operations for each of
the five states reviewed.)



                                Table 1
                
                    Types of Warnings and Number of
                Prescriptions Identified in Five States
                 as Potentially Causing Adverse Medical
                               Reactions

                     (Rounded to nearest thousand)

Type of warning                                Number of prescriptions
----------------------------------------  ----------------------------
Overutilization                                              1,619,000
Therapeutic duplication                                      1,491,000
Drug-drug interaction                                          752,000
Underutilization                                             403,000\a
Drug-disease interaction                                     398,000\b
Pregnancy alert                                               15,000\b
Others\c                                                     1,692,000
======================================================================
Total                                                        6,370,000
----------------------------------------------------------------------
\a Total represents activity in four states, since one state does not
screen for this condition. 

\b Total represents activity in three states, since two states do not
screen for this condition. 

\c Includes other warnings such as excessive daily dose--see appendix
II for more detailed information on the types of warnings issued by
individual states' systems. 

Source:  State data, which we did not independently verify. 


   PROGRAM COSTS CAN BE
   SIGNIFICANTLY REDUCED
------------------------------------------------------------ Letter :5

Along with increasing patient safety, the PRODUR systems in the
states we reviewed reduced Medicaid program costs by millions of
dollars annually through the cancellation of potentially wasteful
prescriptions and the denial of prescriptions to ineligible
recipients.  Experiences in the five states reviewed show that
program savings can more than offset the cost of these relatively
inexpensive systems.  Further, however, to the extent that these
systems also help prevent unnecessary hospitalization due to adverse
medical reactions from prescribed drugs, actual annual cost savings
(according to a 1995 FDA review) may be appreciable. 

The cancellation of prescriptions due to the overutilization (early
refill) alert has provided the five states in our review with
substantial recurring savings.  While a legitimate need for an early
refill may exist in some situations, early refills can also indicate
drug overutilization, which can threaten patient safety, or increase
the potential for fraud or abuse.  Our analyses showed that most
canceled early refill prescriptions are subsequently filled, but
savings from those not filled within 3 months, and thus presumed to
represent program savings, have been substantial.  As table 2 shows,
states realized savings of about $5 million from canceled
prescriptions not filled, with individual state savings ranging from
about $153,000 to about $2.3 million. 



                                     Table 2
                     
                           States' Cost Reductions from
                        Prescription Cancellations Due to
                        Overutilization (Early Refill) and
                                  Ineligibility

                                                                        One-time
           Overutilization     Ineligibility        Total cost      installation
State            denials\a           denials        reductions             costs
--------  ----------------  ----------------  ----------------  ----------------
Maryland        $1,168,665        $4,532,395        $5,701,060          $165,000
Missouri           152,577         1,423,229         1,575,806           508,000
New                559,809         1,514,068         2,073,877           174,000
 Mexico
Oregon           2,252,043         1,532,566         3,784,609           360,000
Pennsylv           765,124        16,637,789        17,402,913           675,000
 ania
================================================================================
Total           $4,898,218       $25,640,047       $30,538,265        $1,882,000
--------------------------------------------------------------------------------
\a Transactions for all states except New Mexico were analyzed for
the period July 1994 through June 1995; this included analyzing
whether denied prescriptions were refilled within 90 days, through
June 1995.  In the case of New Mexico, data cover January through
December 1994, likewise including whether denied prescriptions were
filled within 90 days, through December 1994. 

Source:  State data, which we did not independently verify. 

The five states also realized substantial savings from PRODUR systems
with companion on-line screening capabilities to ensure that
recipients are eligible for Medicaid benefits at the time a
prescription is presented.  Without on-line eligibility verification,
Medicaid recipients can continue to obtain benefits for up to a full
month, even if they lose eligibility during the middle of the month. 
The annual savings from this function, reported at over $25 million
for the five states, more than offset--by a considerable margin--each
state's one-time system installation costs, which ranged from a
reported $165,000 to $675,000.\8 For example, as shown in table 2,
Pennsylvania's PRODUR system denied about $16.6 million in
prescriptions because of ineligibility, from July 1994 through June
1995; in contrast, its one-time installation cost was $675,000. 

On the basis of its review of studies related to drug-induced
illnesses, FDA estimates that 6.4 percent of all hospital admissions
nationally are caused by inappropriate drug therapy--5 percent due to
patient noncompliance with drug regimens, such as overutilization,
and an additional 1.4 percent due to adverse drug reactions. 
Numerous other studies have been conducted worldwide to determine the
extent to which inappropriate drug therapy results in
hospitalization, with estimates ranging from 3 percent for the
general population to as high as 28 percent for the elderly. 
Considering that Medicaid's fiscal year 1995 inpatient
hospitalizations totaled about $42 billion, even limited
implementation of PRODUR systems could have a significant impact on
reducing overall Medicaid program costs. 


--------------------
\8 The states' one-time installation costs included the costs to add
on-line eligibility-verification capabilities to their PRODUR
systems.  We were unable to compare before and after operating costs
because data were not available indicating operating expenses before
the installation of the PRODUR systems.  Individual states' annual
maintenance costs, we were told, run about $5,000 for updating the
systems. 


   POTENTIAL FRAUD AND ABUSE CAN
   BE IDENTIFIED AND PREVENTED
------------------------------------------------------------ Letter :6

As we have pointed out in past reports,\9 potential fraud and abuse
in the Medicaid prescription drug program is a serious problem, with
dollar losses widely estimated to be as high as 10 percent of total
program costs.  In 1992 we reported how 10 states had used
retrospective drug monitoring programs to detect and deter the theft
of controlled substances\10 within the Medicaid program.\11 However,
our analysis of the five states' systems shows that automated PRODUR
systems offer a much more cost-effective alternative to a
"pay-and-chase" approach--in which only retrospective reviews are
undertaken--by detecting potential fraud and abuse at the time a
recipient submits a prescription, and by denying a potentially
invalid claim before the drugs are dispensed. 

The PRODUR systems' alerts for early refills and therapeutic
duplication provide states with the tools needed to detect potential
fraud and abuse and prevent them before they occur.\12 These alerts
can be a clue to detecting potential fraud or abuse because an
individual may be obtaining a greater quantity of medication than
medically necessary, with intentions of selling the drugs--often at
inflated street prices.  For example, our analysis of one of the five
states' data identified over 2,200 recipients who, during a 15-month
period, each obtained a 20-months' supply or greater of controlled
substances, such as Darvon and Valium, in the same therapeutic drug
class.  Further analysis of the 2,200 recipients showed that

  -- 180 of them went to 10 or more different physicians to obtain
     their prescriptions; 1 went to 108;

  -- 252 of the recipients obtained a 3-years' supply or greater of
     controlled substances; 6 of them obtained a 10-years' supply or
     greater; and

  -- 219 of the recipients obtained controlled drugs costing the
     Medicaid program $1,500 or more; 8 of the recipients obtained
     drugs costing $10,000 or more. 


--------------------
\9 Prescription Drug Monitoring:  States Can Readily Identify Illegal
Sales and Use of Controlled Substances (GAO/HRD-92-115, July 21,
1992) and Medicaid Drug Fraud:  Federal Leadership Needed to Reduce
Program Vulnerabilities (GAO/HRD-93-118, Aug.  2, 1993). 

\10 Controlled substances are drugs or other substances that have the
potential for abuse, where such abuse can lead to physical or
psychological dependence.  The Controlled Substance Act of 1970
(Public Law 91-513) places controlled substances into one of five
classes (high to low) on the basis of their potential for abuse or
dependence. 

\11 GAO/HRD-92-115, July 21, 1992. 

\12 Federal regulation (42 CFR 456.705) defines therapeutic
duplication as the prescribing and dispensing of two or more drugs
from the same therapeutic drug class such that the combined daily
dose puts the recipient at risk of adverse medical result or incurs
additional program costs without additional therapeutic benefit. 


   STATES IMPLEMENT SYSTEMS
   DIFFERENTLY, OFTEN NOT SHARING
   EXPERIENCES
------------------------------------------------------------ Letter :7

The states in our review have all implemented their automated PRODUR
systems differently, in large part because no overall source of
information or guidance for implementing PRODUR systems exists.  Some
states issue alerts for a greater number and variety of conditions
relating to patient age, duration of prescription, and other medical
conditions (see appendix III).  Each state has its own DUR board,
which independently sets screening criteria and policies. 

Although all of the systems in the states we reviewed screen for
drug-drug interaction, overutilization, and therapeutic duplication,
we found significant differences in the types of drug therapy
problems the states' PRODUR systems screen for.  One state's system
did not screen for pregnancy conflict, and two of the states' systems
did not screen for underutilization (an indication of noncompliance
with a prescribed drug regimen)--both problems that could have
dramatic effects on patient safety.  Underutilization data we
obtained from three states show that patient noncompliance with
drug-therapy instructions could be a significant problem in the
Medicaid program.  In Maryland alone, for example, between July 1994
and June 1995, over 300,000 underutilization alerts were sent. 

State data also indicate that how alerts regarding early refill and
therapeutic duplication claims are administered can affect savings,
and the identification and prevention of potential fraud, waste, and
abuse.  As table 3 shows, one state--which does not automatically
deny early refill claims but that requires the pharmacist to initiate
such denials--had a total 9-month claims volume of about 5.8 million
and about 6,100 early refill claims that were not later refilled,
resulting in savings of about $153,000.  In contrast, another
state--which does automatically deny early refill claims--had a
lesser total 9-month claims volume (about 3.2 million) but had about
30,000 early refill claims not later refilled, resulting in
significantly more savings--about $1.2 million.  Thus, it appears
that automatic denial can offer a better chance at detecting
prescriptions that could be dangerous, potentially fraudulent, or
both. 



                                Table 3
                
                 Rates of Prescription Cancellation for
                    Two States, Pharmacist Denial v.
                            Automatic Denial

                                          State 1:  State 2: Automatic
                                 Pharmacist denial              denial
------------------------------  ------------------  ------------------
Total number of claims                   5,836,000           3,162,000
Number initially canceled                   21,000             146,000
Number remaining canceled                    6,100              30,000
 after 90 days
Savings from canceled claims              $153,000          $1,200,000
----------------------------------------------------------------------
Source:  State data, which we did not independently verify. 

Three of the states' systems automatically deny early refills; none
automatically deny claims for therapeutic duplication.  A fourth
state, Maryland, is, however, considering changing its system to
automatically deny claims for therapeutic duplication. 


   HCFA'S INFORMATION SHARING HAS
   BEEN LIMITED
------------------------------------------------------------ Letter :8

As stated earlier, the Social Security Act, as amended by the Omnibus
Budget Reconciliation Act of 1990, encourages states to use automated
systems for prospective DUR, and requires HCFA to conduct
demonstration projects of PRODUR systems and report project results
to the Congress.  HCFA is currently conducting a PRODUR demonstration
project in cooperation with the state of Iowa and has provided
project status reports to the Congress.  No other demonstration
projects are planned at this time. 

Since the inception of the Iowa demonstration project in 1994, HCFA
has provided annual reports for 1994 and 1995 to the Congress
presenting the project goals and objectives, evaluation design, plans
for collecting and analyzing data and reporting project results, and
baseline data on Iowa's past Medicaid program expenditures and
prescription volumes.  The 1995 report,\13 transmitted to the
Congress on March 1, 1996, stated that data are not yet available
with which to evaluate the potential benefits of the Iowa PRODUR
system.  Reports planned for 1996 and 1997 are to begin to provide
analyses of claims and eligibility data from Iowa, and a final
evaluation report is to be prepared in 1998 after the project's
scheduled conclusion in March of that year.  As a result, the Iowa
demonstration project will provide little information to the states
and the District of Columbia as they implement their PRODUR systems. 
Among the 43 states plus the District of Columbia that plan to
implement systems, 38 are scheduled to have theirs in place by the
end of 1996. 

In addition to HCFA's statutory responsibilities, in our 1994
report\14 we urged HCFA to gather information on the costs and
benefits of automated review systems, develop guidance on features
and capabilities, and make such information available to all states. 
HCFA has only partially responded to these recommendations.  In
August 1994 HCFA issued guidelines to assist the states in estimating
and reporting the costs and benefits of both retrospective and
prospective drug utilization review.  However, according to HCFA
officials, while they have gathered information on PRODUR programs
from required annual reports submitted by the states, they have not
developed "best practices" or disseminated what information they do
have, due to a lack of resources.  Further, HCFA officials state that
they have no plans at this time to increase their role, since efforts
to reform health care being discussed in the Congress may affect the
ways in which the Medicaid program operates. 


--------------------
\13 Medicaid Drug Use Review Demonstration Projects - Report to
Congress 1995, Office of Research and Demonstrations, Health Care
Financing Administration, U.S.  Department of Health and Human
Services. 

\14 GAO/AIMD-94-130, Aug.  5, 1994. 


   CONCLUSIONS
------------------------------------------------------------ Letter :9

With automated PRODUR systems, most states are recognizing an
opportunity to use low-cost technology to help both physicians and
pharmacists safeguard Medicaid recipients from inappropriate drug
therapy and its potential adverse medical reactions.  While the
primary emphasis of such systems--appropriately--has been safety,
both safety benefits and dollar savings accrue from their use.  Since
results vary on the basis of how such systems are administered, it is
important that states share their experiences.  Absent any analysis
of data from the Iowa demonstration project or any concerted effort
by HCFA to collect and share other states' experiences, states have
had only limited access to both safety and cost data--information
that is critical to informed decisionmaking and to maximizing PRODUR
system effectiveness and efficiency. 


   RECOMMENDATION
----------------------------------------------------------- Letter :10

Given the substantial safety benefits that can accrue to Medicaid
recipients and the strong potential for immediate savings to the
Medicaid program through the effective use of automated PRODUR
systems, we recommend that the Secretary of Health and Human Services
direct the Administrator, Health Care Financing Administration, to
actively facilitate state sharing of information on the most
efficient use of PRODUR systems.  One way would be to quickly help
establish a working group or other such forum for coordinating the
collecting and sharing of information on best practices for automated
prospective drug utilization review programs, on a nationwide basis. 


   AGENCY COMMENTS AND OUR
   EVALUATION
----------------------------------------------------------- Letter :11

In commenting on a draft of our report, the Department of Health and
Human Services generally agreed with our recommendation and the facts
presented.  It stated that automated prospective drug utilization
review systems are extremely beneficial in ensuring that
prescriptions are medically necessary and unlikely to cause adverse
drug reactions, and concurred with our recommendation that HCFA
facilitate state information sharing on best practices for automated
PRODUR programs.  It noted several prior and ongoing initiatives
related to this objective, including (1) the Office of Inspector
General's May 1995 report\15 focusing on states' progress in
implementing statutory DUR requirements, (2) participation in a
project funded by the American Pharmaceutical Association to analyze
PRODUR best practices,\16 (3) continuing participation in the annual
American Drug Utilization Review Symposium, and (4) the gathering of
information on states' best practices that HCFA plans to publish in
its Drug Utilization Review Newsletter in June 1996.  In connection
with these activities, the Department noted that while resources are
limited, it plans to continue and improve its role as a clearinghouse
dedicated to collecting information and promoting the exchange of
information among states on the operation of effective automated
PRODUR systems. 

We had reviewed the documents the Department referred to and agree
that these efforts have provided some information to the states on
the operation of effective automated PRODUR systems.  The
Department's future such efforts will need to provide an ongoing and
continuous exchange of information as most states move rapidly toward
operational PRODUR systems by the end of 1996.  Information in the
Inspector General's report and in the report resulting from the
project funded by the American Pharmaceutical Association cited by
the Department in its response is based on the states' early
experiences with automated PRODUR systems--primarily 1993 data.  And
HCFA's stated plans to issue a DUR newsletter in June 1996 will be
only the second newsletter since the states were required to
implement DUR programs in January 1993. 

While agreeing that PRODUR systems are often capable of deterring
potentially fraudulent activity, the Department said the primary
emphasis of these systems is on patient safety and quality of care. 
It stated that PRODUR systems are not specifically designed to detect
Medicaid fraud and abuse and that retrospective DUR systems may be
more capable of detecting potentially fraudulent activity and
referring this activity to states' Medicaid Surveillance Utilization
Review units for follow-up.  While we acknowledge the safety and
quality-of-care orientation, it is being clearly established in
states with automated PRODUR systems that tremendous benefits accrue
for program integrity--detecting and preventing potentially
fraudulent or abusive activity before drugs are dispensed.  Relying
on retrospective systems alone is needlessly restrictive and
expensive, resulting in a "pay-and-chase" approach in which
recipients have already received the drugs and the states' Medicaid
Surveillance Utilization Review units must use their resources to
conduct investigations and seek whatever recovery they can
attain--usually minimal. 

Finally, the Department recommended caution in attributing the $30
million in cost savings cited in our report to denials due to early
refills and ineligibility, and noted that additional methodological
issues should be considered.  We discussed these issues with HCFA
program managers responsible for DUR program implementation.  They
said that our methodology could overstate savings because our early
refill analysis searched for 90 days after an initial denial to see
whether a prescription was subsequently filled, while states can
allow recipients to get up to a 100-days' supply of drugs.  HCFA
stated that with 100-day-supply prescriptions, situations could arise
in which more than 90 days would elapse between an initial denial and
a subsequent refill, thus eroding program savings identified under
our methodology. 

While such situations could occur, they appear unlikely because of
the series of events that would be involved, that is, a recipient's
seeking to refill a 100-day-supply prescription several days after
initially filling it, being denied, then waiting over 90 days before
attempting a refill.  Moreover, further analysis of the data obtained
for two of the states in our review showed that only about 1 percent
of their total prescription drug claims were for prescriptions with
over a 90-days' supply.  In reference to cost savings attributed to
ineligibility, HCFA described a specific example in which recipients
under states' medically needy programs may not become eligible for
Medicaid benefits until certain portions of their incomes are spent
on medical expenses.  Thus, some ineligible recipients denied
prescriptions may subsequently obtain these prescriptions after they
become eligible.  Our methodology did not consider ineligible
recipients' later becoming eligible and obtaining previously denied
prescriptions.  However, this possibility assumes that ineligible
recipients would postpone filling denied prescriptions until after
they become eligible for Medicaid benefits.  Further, HCFA's example
involves those receiving benefits under states' "medically needy"\17

programs--recipients that accounted for only about 11 percent of the
total Medicaid population during fiscal year 1994--the latest data
available. 


--------------------
\15 Medicaid Drug Use Review Programs:  Lessons Learned by States,
Office of Inspector General, Department of Health and Human Services,
OEI-01-92-00800, May 1995. 

\16 Medicaid DUR Programs:  1993 Final Report to American
Pharmaceutical Association Foundation, Earlene E.  Lipowski, R.Ph.,
Ph.D., College of Pharmacy, University of Florida; and Ted Collins,
R.Ph., Center for Health Systems Research and Analysis, University of
Wisconsin-Madison, August 1995. 

\17 The term "medically needy" is defined differently by each state,
but in general applies to those whose incomes are too high to qualify
for Medicaid but are nonetheless in need of assistance.  In one
state, for example, those whose incomes are too high for unqualified
Medicaid assistance but fall within 133 percent of the poverty level
can receive Medicaid benefits after spending a specified amount for
medical services during the fiscal year. 


--------------------------------------------------------- Letter :11.1

We are sending copies of this report to the Secretary of Health and
Human Services; the Administrator, Health Care Financing
Administration; the Director, Office of Management and Budget; and
other interested parties.  Copies will also be made available to
others upon request. 

We plan no further distribution of this report until 15 days after
the date of this letter, unless you publicly announce its contents
earlier.  At that time, we will release copies to interested parties. 

Please contact me at (202) 512-5539 if you have any questions
concerning this report.  Major contributors to this report are listed
in appendix V. 

Patricia T.  Taylor
Director, Health, Education, and
 Human Services Information Systems




(See figure in printed edition.)Appendix I
COMMENTS FROM THE DEPARTMENT OF
HEALTH AND HUMAN SERVICES
============================================================== Letter 



(See figure in printed edition.)


ALERTS AND CLAIM CANCELLATIONS
RESULTING FROM STATES' PRODUR
SYSTEMS
========================================================== Appendix II

The following tables provide detailed contractor- and state-provided
data on the results of automated PRODUR systems during 1994 and 1995. 
These data, which we did not independently verify, include (1) the
number and cost of claims processed, (2) the types and numbers of
drug therapy alert messages sent via the states' PRODUR systems, (3)
the number of overutilization/early refill prescriptions denied and
not later filled (highlighted in bold in each table), and (4) the
number of claims canceled. 

Tables for Missouri and Pennsylvania also include data on numbers of
claims canceled for each drug therapy alert condition; similar data
were not available for the other three states. 



                               Table II.1
                
                  Maryland's PRODUR System Alerts and
                Cancellations, July 1, 1994 Through June
                                30, 1995

----------------------------------------  ----------------------------
======================================================================
Total number of claims processed                             4,270,169
======================================================================
Total cost of claims processed                            $136,708,110
Drug therapy alert condition                          Number of alerts
Drug-age conflict                                               10,193
Drug-disease interaction                                        28,782
Drug-drug interaction                                          182,164
Excessive daily dose                                            66,762
Excessive daily dose/children                                      696
Excessive daily dose/over age 65                                22,373
Excessive quantity dispensed                                    15,289
Insufficient daily dose for age                                 12,134
Overutilization (early refill)                                 187,505
Pregnancy conflict                                               3,657
Therapeutic duplication                                        143,062
Underutilization                                               339,937
======================================================================
Total number of alerts                                       1,012,554
======================================================================
Total number of claims canceled due to                         182,681
 alerts
----------------------------------------------------------------------
Source:  State data, which we did not independently verify. 



                               Table II.2
                
                  Missouri's PRODUR System Alerts and
                Cancellations, July 1, 1994 Through June
                                30, 1995

----------------------------------------  -------------  -------------
======================================================================
Total number of claims processed                             6,780,901
======================================================================
Total cost of claims processed                            $173,370,503
Drug therapy alert condition                  Number of      Number of
                                                 alerts         claims
                                                              canceled
Above maximum daily dose                        250,415         12,187
Above maximum dose range                         17,934            690
Additive side effect                              2,566             66
Below minimum daily dose                        240,336          9,827
Below minimum dose range                         36,515          1,328
Current Rx applies to 3 90-day therapy          9,868 1           ,177
Current Rx exceeds 7 90-day therapy             8,775 1           ,573
Current Rx initiates 2 90-day therapy           2,401 6             88
Drug-disease interaction                         14,825            738
Drug-drug interaction                           108,940          3,441
Drug-indicated disease conflict                  59,052          2,411
Indicated for prior drug's side effect              831             13
Maintenance dose                                 17,993            209
Overutilization (early refill)                  394,795         25,409
Side effect medical condition                       823             10
Significant side effect                          14,008            247
Therapeutic duplication                         157,615          6,970
======================================================================
Total number of alerts                                       1,457,692
======================================================================
Total number of claims canceled due to                          66,984
 alerts
----------------------------------------------------------------------
Source:  State data, which we did not independently verify. 



                               Table II.3
                
                 New Mexico's PRODUR System Alerts and
                 Cancellations, January 1, 1995 Through
                             June 30, 1995

----------------------------------------  ----------------------------
======================================================================
Total number of claims processed                             1,084,912
======================================================================
Total cost of claims processed                             $26,286,725
Drug therapy alert condition                          Number of alerts
Drug-age conflict                                               16,609
Drug-disease interaction                                        23,008
Drug-drug interaction                                           38,373
Excessive daily dose                                             8,508
Excessive daily dose/children                                    1,974
Excessive daily dose/over age 65                                 7,148
Excessive quantity dispensed                                     1,273
Insufficient daily dose for age                                 56,349
Overutilization (early refill)                                  45,526
Pregnancy conflict                                               3,641
Therapeutic duplication                                         11,905
Underutilization                                                18,364
======================================================================
Total number of alerts                                         232,678
======================================================================
Total number of claims canceled due to                           5,397
 alerts
----------------------------------------------------------------------
Note:  New Mexico changed contractors in January 1995; prior alert
data are not available. 

Source:  State data, which we did not independently verify. 



                               Table II.4
                
                   Oregon's PRODUR System Alerts and
                Cancellations, July 1, 1994 Through June
                                30, 1995

----------------------------------------  ----------------------------
======================================================================
Total number of claims processed                             2,584,672
======================================================================
Total cost of claims processed                             $74,767,962
Drug therapy alert condition                          Number of alerts
Drug-age conflict                                                8,668
Drug-disease interaction                                        40,451
Drug-drug interaction                                          223,665
Excessive daily dose                                            64,238
Excessive daily dose/children                                      694
Excessive daily dose/over age 65                                19,744
Excessive quantity dispensed                                     5,439
Insufficient daily dose for age                                 11,765
Overutilization (early refill)                                 309,081
Pregnancy conflict                                               6,887
Therapeutic duplication                                        196,014
Underutilization                                                25,367
======================================================================
Total number of alerts                                         912,013
======================================================================
Total number of claims canceled due to                          23,810
 alerts
----------------------------------------------------------------------
Source:  State data, which we did not independently verify. 



                               Table II.5
                
                Pennsylvania's PRODUR System Alerts and
                             Cancellations,
                   July 1, 1994 Through June 30, 1995

----------------------------------------  -------------  -------------
======================================================================
Total number of claims processed                            16,590,943
======================================================================
Total cost of claims processed                            $503,963,506
Drug therapy alert condition                  Number of      Number of
                                            alerts sent         claims
                                                              canceled
Drug-age conflict                                 2,004            172
Drug-drug interaction                           198,905         13,799
High dose alert                                 348,991         36,134
Low dose alert                                  539,836         36,121
Overutilization (early refill)                  682,276        205,256
Pregnancy conflict                                  451             41
Therapeutic duplication                         982,852         84,078
======================================================================
Total number of alerts                                       2,755,315
======================================================================
Total number of claims canceled due to                         375,601
 alerts
----------------------------------------------------------------------
Source:  State data, which we did not independently verify. 


DRUG THERAPY PROBLEMS SCREENED FOR
BY STATES' PRODUR SYSTEMS
========================================================= Appendix III

Drug therapy alert          Maryla  Missou     New          Pennsylvan
condition                       nd      ri  Mexico  Oregon          ia
----------------------  --  ------  ------  ------  ------  ----------
Above maximum dose                       X
 range
Additive Side Effect                     X
Below minimum dose                       X                           X
 range
Current Rx applies to                    X
 90-day therapy
Current Rx exceeds 90                    X
 day-therapy
Current Rx Initiates                     X
 90-day therapy
Drug-age conflict                X               X       X           X
Drug-disease                     X       X       X       X
 interaction
Drug-drug interaction            X       X       X       X           X
Drug-indicated disease                   X
 conflict
Excessive daily dose             X       X       X       X           X
Excessive daily dose/            X               X       X
 children
Excessive daily dose/            X               X       X
 over age 65
Excessive quantity               X               X       X
 dispensed
Indicated for prior                      X
 drug's side effect
Insufficient daily               X               X       X
 dose for age
Maintenance dose                         X
Overutilization/                 X       X       X       X           X
 early refill
Pregnancy conflict               X               X       X           X
Side effect medical                      X
 condition
Significant side                         X
 effect
Therapeutic                      X       X       X       X           X
 duplication
Underutilization                 X               X       X
----------------------------------------------------------------------
Source:  State data, which we did not independently verify. 


STATE IMPLEMENTATION OF PRODUR
SYSTEMS:  STATUS AS OF APRIL 1,
1996
========================================================== Appendix IV

   Figure IV.1:  Nationwide View

   (See figure in printed
   edition.)



                               Table IV.1
                
                  Twenty-nine States Operating PRODUR
                 Systems as of April 1, 1996, and Dates
                                 Begun

State                                             Date operation began
----------------------------------------  ----------------------------
Alaska                                                       June 1995
California                                                 August 1995
Delaware                                                   August 1994
Florida                                                      July 1995
Illinois                                                  January 1993
Indiana                                                     March 1996
Kentucky                                                September 1994
Louisiana                                                   April 1996
Maine                                                      August 1995
Maryland                                                  January 1993
Massachusetts                                             October 1995
Minnesota                                                February 1996
Mississippi                                               October 1995
Missouri                                                 February 1993
Montana                                                 September 1994
Nebraska                                                     June 1995
New Hampshire                                              August 1995
New Mexico                                                October 1993
New York                                                    March 1995
Oregon                                                      March 1994
Pennsylvania                                                 June 1993
Rhode Island                                             December 1994
Texas                                                       March 1995
Utah                                                        April 1995
Vermont                                                  November 1993
Virginia                                                     July 1994
Washington                                                  March 1996
West Virginia                                                July 1992
Wyoming                                                   October 1995
----------------------------------------------------------------------


                               Table IV.2
                
                  Sixteen States Planning to Implement
                PRODUR Systems, as of April 1, 1996, and
                          Planned Start Dates

State                                                    Planned start
----------------------------------------  ----------------------------
Alabama                                                           1996
Arkansas                                                          1997
Colorado                                                          1996
Connecticut                                                       1996
District of Columbia                                              1996
Georgia                                                           1996
Idaho                                                             1997
Iowa                                                              1996
Kansas                                                            1996
New Jersey                                                date not set
North Carolina                                                    1996
North Dakota                                                      1996
Ohio                                                              1996
Oklahoma                                                          1996
South Dakota                                              date not set
Wisconsin                                                         1997
----------------------------------------------------------------------


                               Table IV.3
                
                  Six States Not Planning to Implement
                             PRODUR Systems

----------------------------------------------------------------------
State

Arizona\a

Hawaii

Michigan

Nevada

South Carolina

Tennessee\a
----------------------------------------------------------------------
\a These states are exempt from the 1990 Social Security Act mandate
to implement DUR programs because they operate managed care programs. 


MAJOR CONTRIBUTORS TO THIS REPORT
=========================================================== Appendix V

ACCOUNTING AND INFORMATION
MANAGEMENT DIVISION, WASHINGTON,
D.C. 

Frank W.  Reilly, Director, Information Systems Methods and Support
William B.  Ritt, Assistant Director
Michael P.  Fruitman, Communications Analyst

KANSAS CITY REGIONAL OFFICE

John B.  Mollet, Evaluator-in-Charge
Donald L.  Ficklin, Computer Specialist
David R.  Solenberger, Senior Evaluator
John G.  Snavely, Staff Evaluator
Karen S.  Sifford, Staff Evaluator


GLOSSARY
============================================================ Chapter 0

Definitions of the following drug utilization review terms are from
42 CFR 456.702/705 and state PRODUR manuals. 


      ADVERSE DRUG-DRUG
      INTERACTION
-------------------------------------------------------- Chapter 0:0.1

The potential for, or the occurrence of, an adverse medical effect as
a result of the recipient's using two or more drugs together. 


      ADVERSE MEDICAL RESULT
-------------------------------------------------------- Chapter 0:0.2

A significant undesirable effect experienced by a patient due to the
prescribed course of drug therapy. 


      DRUG-AGE CONTRAINDICATION
-------------------------------------------------------- Chapter 0:0.3

Use of a drug that is not recommended for use in the age group of the
patient.  This can occur when the patient is too old or too young for
the given medication. 


      DRUG-ALLERGY INTERACTION
-------------------------------------------------------- Chapter 0:0.4

The significant potential for, or the occurrence of, an allergic
reaction as a result of drug therapy. 


      DRUG-DISEASE
      CONTRAINDICATION
-------------------------------------------------------- Chapter 0:0.5

The potential for, or occurrence of, an undesirable alteration of the
therapeutic effect of a given prescription because of the presence of
an existing disease condition--such as an ulcer drug exacerbating a
patient's high blood pressure. 


      GROSS OVERUSE
-------------------------------------------------------- Chapter 0:0.6

Repetitive overutilization without therapeutic benefit. 


      INCORRECT DRUG DOSAGE
-------------------------------------------------------- Chapter 0:0.7

A dosage lying outside the standard daily dosage range necessary to
achieve therapeutic benefit. 


      INCORRECT DURATION OF DRUG
      TREATMENT
-------------------------------------------------------- Chapter 0:0.8

The number of days of prescribed therapy exceeding or falling short
of the standard recommendation for the condition for which it was
prescribed. 


      OVERUTILIZATION
-------------------------------------------------------- Chapter 0:0.9

Use of a drug in quantities or for durations that put the recipient
at risk of an adverse medical result. 


      PREGNANCY CONFLICT
------------------------------------------------------- Chapter 0:0.10

Use of the prescribed drug is not recommended during pregnancy. 


      THERAPEUTIC DUPLICATION
------------------------------------------------------- Chapter 0:0.11

The prescribing and dispensing of two or more drugs from the same
therapeutic class, such as analgesics (pain relievers), resulting in
a combined daily dose that puts the recipient at risk of an adverse
medical result, or that incurs additional program cost without
additional therapeutic benefit. 


      UNDERUTILIZATION
------------------------------------------------------- Chapter 0:0.12

Use of a drug in insufficient quantity to achieve a desired
therapeutic goal. 

*** End of document. ***