How Disease Burden
Influences Medication Patterns
for Medicare Beneficiaries:
Implications for Policy
BRUCE C. STUART
PETER LAMY CENTER ON DRUG THERAPY AND AGING
UNIVERSITY OF MARYLAND BALTIMORE
ABSTRACT: This issue brief provides benchmarks, based on empirical analyses
of 2002 data, for evaluating the quality of pharmaceutical care under the Medicare
Part D prescription drug benefit. The analyses cover all major classes of pharma-
ceuticals used by beneficiaries with one of eight chronic conditions. Common
patterns observed include: 1) a mounting disease burden is associated with
increasingly complex medication regimens in every group, and 2) the intensity
and persistence of drug use tend to rise with disease burden up to a point, before
declining for beneficiaries with the greatest morbidity. The study concludes that
neither traditional drug quality indicators nor new quality assurance mechanisms
mandated by law are well aligned to capture suboptimal medication use at either
end of the spectrum of disease burden in the Medicare population. A holistic
approach to medication management is needed to ensure that Part D plans meet
beneficiaries and policymakers’ expectations for high-quality care.
For more information about this
study, please contact:
Bruce C. Stuart, Ph.D.
Professor and Executive Director
Peter Lamy Center on
Drug Therapy and Aging
University of Maryland Baltimore
? ? ? ? ?
It is well known that Medicare beneficiaries are heavy users of prescription medi-
cations. This fact dominated the policy debate over the voluntary prescription drug
This and other Commonwealth
Fund publications are online at
learn more about new publications
when they become available, visit
the Fund’s Web site and register to
receive e-mail alerts.
Commonwealth Fund pub. 1106
This issue brief is based on Medication Use by Aged and Disabled Medicare Beneficiaries
Across the Spectrum of Morbidity: A Chartbook, published by the Peter Lamy Center on
Drug Therapy and Aging at the University of Maryland Baltimore in 2007. To obtain
a copy, please visit the center’s Web site at www.pharmacy.umaryland.edu/lamy.
2 THE COMMONWEALTH FUND
benefit that became available in January 2006. Much
less is known about the quality and effectiveness of
medication use by beneficiaries. This is partly due to the
federal government’s delay in releasing prescription
drug data from Part D plans, but the problem will not
be solved simply by making Part D prescription claims
available to researchers. A more fundamental problem
is that there are no systematic, evidence-based guide-
lines available to evaluate medication regimens of older
patients, particularly those with complex morbidity.
Building the evidence base for appropriate
drug therapy for Medicare beneficiaries must begin
with an empirical, population-level assessment of how
medication regimens vary across the continuum of
disease burden. To accomplish this objective, the
researchers used data from the period prior to the
implementation of the new drug benefit to describe
the breadth, intensity, and persistence of medication
regimens for the Medicare population as a whole and
for beneficiaries with eight common chronic condi-
tions: diabetes, depression, dementia, chronic obstruc-
tive pulmonary disease (COPD), arthritis, hypertension,
ischemic heart disease, and heart failure.
Two features of this study distinguish it from
previous research on medication use by Medicare
beneficiaries. First, it examines drug utilization pat-
terns for beneficiaries with each of these chronic con-
ditions in the larger context of their overall disease
burden. Over their lifetimes, few Medicare beneficiar-
ies suffer from single diseases with well-accepted,
evidence-based treatment recommendations. Rather,
most ultimately develop multiple conditions, for which
treatment guidelines are either lacking or ambiguous in
the presence of significant comorbidity. Geriatricians
have developed various clinical recommendations and
best practice statements for dealing with complex
morbidity in older patients. The development of evi-
dence-based guidelines for treating complex morbidity
has proven elusive, however, in part due to scant
epidemiologic data describing how treatment patterns
for particular chronic conditions change with rising
morbidity. This study was designed to help build this
A second unique feature of the study is that it
takes into account the prescription drugs used to treat
all of the diseases that beneficiaries suffer—not just
those specific to the eight chronic conditions. This
analysis examines how beneficiaries’ medication regi-
mens evolve with accumulating morbidity, highlighting
areas of potential concern regarding underuse, overuse,
and inappropriate use of medication therapy for par-
ticular segments of the chronically ill population.
The sample for the study (N=8,455) was selected from
the 2002 Medicare Current Beneficiary Survey (MCBS).
The MCBS collects extensive information on prescrip-
tion drug utilization and spending from self-reports and
reviews of medication containers. The MCBS also
provides claims data for Part A (inpatient and outpa-
tient hospital and post-acute services) and Part B (phy-
sician and other provider services) for each respondent.
Diagnostic indicators from the claims data, together
with self-reports of chronic conditions, were used to
define the eight disease cohorts. In order to capture the
full spectrum of rising disease burden, each cohort was
stratified into 10 equal-size deciles, based on cumula-
tive health care spending during the year from all payer
sources, including beneficiaries’ out-of-pocket ex-
penses.1 Medication statistics were computed at the in-
dividual beneficiary level, aggregated by disease co-
hort and decile of disease burden, and then weighted to
be nationally representative of the community-dwelling
Medicare population. Beneficiaries enrolled through
the Medicare+Choice program (privately administered
plans, now called Medicare Advantage) and those in
long-term care facilities were excluded due to lack of
critical data elements.
The study focused on four sets of indicators
designed to benchmark the quality and effectiveness
of beneficiaries’ medication use before the advent of
Part D drug coverage. The first indicator is the fraction
of total health care spending devoted to prescription
drugs. This captures changes in the allocation of
health resources used to treat beneficiaries with rising
HOW DISEASE BURDEN INFLUENCES MEDICATION PATTERNS FOR MEDICARE BENEFICIARIES: IMPLICATIONS FOR POLICY 11
1 The dementia and depression cohorts were divided into
spending quintiles rather than deciles because of small
2 USP, Medicare Prescription Drug Benefit Model Guide-
lines, 2004. http://www.usp.org/healthcareInfo/mmg/
initialGuidelines.html, accessed Dec. 5, 2006.
3 Medication-intensive conditions are defined as disease
clusters that significantly predict spending on prescription
drugs in the RxHCC model. CMS, Part D Payment and
Risk Adjustment. http://www.cms.hhs.gov/DrugCoverage
accessed Dec. 4, 2006.
4 B. Stuart, L. Simoni-Wastila, I. Zuckerman et al., Medica-
tion Use by Aged and Disabled Medicare Beneficiaries
Across the Spectrum of Morbidity: A Chartbook (Balti-
more: Peter Lamy Center on Drug Therapy and Aging,
University of Maryland Baltimore, 2007). The Chartbook
is available at: www.pharmacy.umaryland.edu/lamy.
5 Combination drug use was common within the Medicare
population in 2002. The number of unadjusted prescrip-
tion fills for the community-dwelling population was 28.9
compared with 36.5 adjusted fills.
6 The MCBS does not provide data on the number of days’
supply for drug fills, so it is possible that some of the
differences in fill counts may mask differences in pre-
scription size. A sensitivity analysis using pill counts
found no evidence of systematic bias in prescription size
by spending decile. See Appendix A in the Chartbook.
7 S. T. Fleming, H. G. Pursley, B. Newman et al., “Comor-
bidity as a Predictor of Stage of Illness for Patients with
Breast Cancer,” Medical Care, Feb. 2005 43(2):132–40.
8 See C. R. Jaén, K. C. Strange, and P. A. Nutting, “Com-
peting Demands of Primary Care: A Model for the Deliv-
ery of Clinical Preventive Services,” Journal of Family
Practice, Feb. 1994 38(2):166–71; D. T. Ko, M. Mandani,
and D. A. Alter, “Lipid-Lowering Therapy with Statins in
High-Risk Elderly Patients,” Journal of the American
Medical Association, Apr. 21, 2004 291(15):1864–70;
and D. A. Redelmeier, S. H. Tan, and G. L. Booth, “The
Treatment of Unrelated Disorders in Patients with
Chronic Medical Diseases,” New England Journal of
Medicine, May 21, 1998 338(21):1516–20.
9 S. C. Durso, “Using Clinical Guidelines Designed for
Older Adults with Diabetes Mellitus and Complex Health
Status,” Journal of the American Medical Association,
Apr. 26, 2006 295(16):1935–40.
10 D. M. Kirking, J. A. Lee, J. J. Ellis et al., “Patient-Reported
Underuse of Prescription Medications: A Comparison of
Nine Surveys,” Medical Care Research and Review,
Aug. 2006 63(4):427–46; and Boston Consulting Group,
The Hidden Epidemic: Finding a Cure for Unfilled
Prescriptions and Missed Doses, BCG: Boston, 2004.
Feb. 23, 2007.
11 This may be changing. For example, the American Diabe-
tes Association standards of diabetes care in specific
populations recommend that aggressive glycemic control
may not be appropriate for frail older persons. See ADA
guidelines in Diabetes Care, Jan. 2006 29(Suppl. 1):S26–
S29; and A. F. Brown, C. M. Mangione, D. Saliba et al.,
“Guidelines for Improving the Care of the Older Person
with Diabetes Mellitus,” Journal of the American Geriat-
rics Society, May 2003 51(5 Suppl.):S265–S280.
12 ACOVE (Assessing Care of Vulnerable Elders) is a col-
laborative project between RAND and Pfizer, designed to
identify evidence-based indicators for quality of care for
elderly individuals that encompass various domains.
including medication use. Together with the American
College of Physicians Task Force on Aging, the ACOVE
group developed 43 medication quality measures covering
prescribing indications, avoidance of inappropriate medi-
cations, patient education, and medication monitoring.
None of the 43 indicators considers patient disease
burden. See T. Higashi, P. G. Shekelle, D. H. Solomon
et al., “The Quality of Pharmacologic Care for Vulnerable
Older Patients,” Annals of Internal Medicine, May 4, 2004
140(9):714–22. SCRIPT (Study of Clinically Relevant
Indicators of Pharmacologic Therapy) is a collaboration
of various health care trade organizations and governmen-
tal agencies representing the Coalition for Quality in
Medication Use. The purpose of SCRIPT is to develop
operational quality measures for medication use in outpa-
tient settings focusing on six disease states: coronary ar-
tery disease/post-MI, atrial fibrillation, heart failure,
dyslipidemia, hypertension, and diabetes. A unique aspect
of SCRIPT is that the medication measures are evaluated
longitudinally to capture outcome indicators. However,
like ACOVE, there is no explicit consideration for patient
disease burden. See www.ahqa.org/pub/uploads/Kogut.ppt,
accessed Feb. 16, 2007. PQA (Pharmacy Quality Alli-
ance) is a membership organization representing over 60
public and private organizations with a stake in measuring
performance of pharmacy services. As of this writing
PQA has approved performance measures that include
medication quality indicators for selected cardiovascular
drugs, diabetes, and respiratory disorders, plus indicators
for exposure to inappropriate medications and drug–drug
interactions. The PQA medication quality indicators are
all well-established standards and break no new ground.
None of the PQA measures to date takes patients’ disease
burden into consideration.
accessed Feb. 23, 2007.
14 http://www.cms.hhs.gov/pqri/, accessed Jan. 25, 2008.
ABOUT THE AUTHOR
Bruce C. Stuart, Ph.D., is professor and executive director of the Peter Lamy Center on Drug Therapy
and Aging at the University of Maryland Baltimore. In 1997 he joined the faculty of the University of
Maryland School of Pharmacy as the Parke-Davis endowed chair in geriatric pharmacotherapy and
was selected as a Maryland Eminent Scholar for his work in geriatric drug use. In 2007, Dr Stuart
was appointed a commissioner to the Medicare Payment Advisory Commission (MedPAC). He
received his economics training at Whitman College and Washington State University.
The author would like to thank his fellow coauthors for the report, Medication Use by Aged and
Disabled Medicare Beneficiaries Across the Spectrum of Morbidity: A Chartbook. These include
Linda Simoni-Wastila, Ilene Zuckerman, Jalpa Doshi, Dennis Shea, Thomas Shaffer, and Lirong
Zhao. Thanks are also due to members of the benchmark study Advisory Board who made critically
helpful comments on earlier drafts of the work: Stuart Guterman, Scott Smith, Julianne Howell,
Penny Mohr, Steve Blackwell, John Poisal, and Joan Sokolovsky. Amy Davidoff, Ruth Lopert, and
Richard Stefanacci also made important contributions to the benchmark project. Crystal Weaver and
Reba Cornman were responsible for getting the report ready for production and distribution.
The mission of The Commonwealth Fund is to promote a high performance health care system.
The Fund carries out this mandate by supporting independent research on health care issues and
making grants to improve health care practice and policy. The views presented here are those of the
author and not necessarily those of The Commonwealth Fund or its directors, officers, or staff.