CLINICAL THERAPEUTICSWOL.
20, NO. 4, 1998
Predictors of Medication Adherence in the Elderly
Rajesh Balkrishnan, MS Pharm
Divisions of
Evaluative
North Carolina
Policy and
University of
and Pharmaceutical
Carolina at
Hill, Chapel
ABSTRACT
As average life expectancy increases, so
do the incidence of chronic diseases and
the number of persons receiving longterm drug therapy. Thus elderly patients’
noncompliance
with medication
regimens has the potential for sweeping medical and economic consequences
and is
likely to become increasingly
important
in the design of disease-management
programs for this population.
The author
conducted a MEDLINE@ search of the
English-language
literature for the years
1962 to 1997 to identify articles conceming predictors of medication compliance
in the elderly. A descriptive analysis of
this literature indicated that there remains
some uncertainty
about the reasons for
noncompliant medication-taking
in the elderly. Clear associations have been established between elderly patients’ medication adherence and race, drug and dosage
form, number of medications, cost of medications, insurance coverage, and physi-
764
and
cian-patient
communication.
However,
the findings are inconsistent with regard
to the effects of patients’ age, sex, socioeconomic
status, living arrangement,
comorbidities, number of physician visits,
and knowledge,
attitudes,
and beliefs
about health. Until the results of further
comprehensive
studies are available, the
current knowledge should be considered
when designing and implementing
disease-management
programs for the elderly. Key words: medication adherence,
elderly, predictors, patient compliance.
INTRODUCTION
Medication adherence can be defined as
the extent to which a person’s medicationtaking behavior coincides with medical
advice. 1 Complete medication adherence
occurs when the patient follows the physician’s instructions
completely. Possible
signs of nonadherence include failure to
fill prescriptions,
loss to follow-up, and
the absence of serum drug concentrations
0149.2918/98/$19.00
R. BALKRISHNAN
on laboratory testing. Partially compliant
patients take incorrect doses of their drugs
regularly or correct doses more or less often than prescribed.2 Studies have reported
medication adherence rates in the elderly
that range from 26% to 59%.3” Medication adherence is likely to be increasingly
important in the management of disease in
this population. As average life expectancy
increases, so will the incidence of chronic
diseases and the number of patients receiving long-term drug therapy.
The medical
and economic
consequences of elderly patients’ nonadherence
to drug therapy may include lack of drug
efficacy, treatment failure, disease progression, emergence of resistant bacteria,
medication overdose, otherwise avoidable
hospitalizations,
and unnecessary medical
expenses.7 Some of the economic consequences of medication nonadherence
in
the elderly are suggested by the results of
a study by Co1 et aL3 who interviewed patients and their families in 315 consecutive admissions of elderly patients to an
acute care hospital in 1987 and found that
approximately
11% of the admissions
were related to medication
noncompliance. The investigators
estimated that
28% of all admissions in this population
were drug related and 40% were the result of medication
nonadherence.
The
mean cost per admission associated with
medication
nonadherence
in this study
was $2150.
In a meta-analysis
by Sullivan et al8
that included several studies of elderly
patients, 5.3% of hospital admissions in
1986 were related to medication nonadherence, at a cost of $8.5 billion in direct
hospital costs.
The elderly tend to have more chronic
conditions than others and thus receive
prescriptions for a greater number of med-
ications. Many of the changes of aging
are known to adversely affect medication
adherence. Studies have shown that elderly patients may have more difficulty
following medical advice; are less likely
to ask for clarification of the physician’s
instructions; tend to be less active in selfcare; and are more susceptible to the physical, social, and psychological
consequences of nonadherence.3,9,10
Disease, or care, management programs
are comprehensive care packages offered
to managed care organizations by pharmaceutical companies and include a combination of drugs, medical devices, diagnostic aids, patient education programs, and
outcomes research. Patient education programs encourage medication adherence to
improve patient outcomes and reduce unnecessary costs resulting from the adverse
clinical consequences of noncompliance,
particularly noncompliance with drug regimens for chronic conditions.
To be effective, disease-management
interventions in the elderly must take factors affecting medication adherence into
account, including known factors that may
influence and predict medication-taking
behavior. This understanding
could then
become the cornerstone
of an effective
patient education program. The purpose
of this paper is to identify and examine
factors that have the potential to enable or
disable medication adherence in elderly
patients.
LITERATURE
REVIEW
To identify studies that examined predictors of medication adherence in the elderly, the author conducted a MEDLINE@
search of the English-language
literature
from 1962 to 1997. Both “and” and “or”
searches were performed using the fol-
765
CLINICAL THERAPEUTICS’
lowing key words: patient compliance,
aged, cooperative behavior, patient acceptance of health care, drug prescription,
and drug utilization. Searching by alternative terms such as “medication adherence” led back to “patient compliance.”
The reference lists of the articles identified were used to find additional articles,
and the key words of all articles were used
to perform further MEDLlNE searches.
All studies that described predictors of
medication adherence and nonadherence
in the elderly were reviewed for three
types of information:
(1) the factors affecting medication adherence; (2) clinical
features associated with variations in medication adherence; and (3) methods used
to quantify the effect of each predictor of
adherence. Meta-analytic
methods were
not used to aggregate study findings because of the diversity of the study designs,
measures of clinical end points, and medical conditions studied. Fourteen studies
were identified and are included in the
present review.3”*g-18
FACTORS PREDICTING
MEDICATION
ADHERENCE
The published literature investigates the
relationship between elderly patients’ nonadherence with drug therapy and various
demographic,
medical, behavioral, economic, social, and medication- or medical
practice-related
variables. Some of the
demographic
variables studied include
age, race, sex, income level, occupation,
education level, social class, and marital
status. Medical variables that have been
studied include severity and duration of
illness, number of comorbid conditions,
frequency of use of medical services, and
patient satisfaction with health-care provider and quality of care. Medication-
related variables examined include type
of medication, drug-delivery system, therapeutic regimen, and adverse effects. Economic variables include type of insurance
coverage and costs of drugs and medical
care. Behavioral variables include physician-patient interaction, patients’ knowledge about their medical condition, selfreported compliance,
and attitudes and
beliefs about health.
The present review examines selected
variables separately, although in most
cases it is the interaction of several factors that allows prediction of elderly patients’ medication-adherence
behavior.
Demographic
Variables
Age was not an important predictor of
medication adherence in most of the published studies. However, in a prospective
study of 220 randomly selected hospitalized, chronically
ill elderly patients,
Schwarz et al6 reported that 68% of 53
patients aged 275 years made medication
errors, compared with 57% of 123 patients aged ~75 years. More recently,
Coons et al” found no relationship
between age and medication adherence in
1028 older adults who were interviewed
about their medication-adherence
behavior. Sharkness and Snow’* reported similar results in a study of 125 hypertensive
veterans. However, a recent study by Monane et all3 employing data from 4068
New Jersey Medicaid enrollees demonstrated that advanced age (285 years) was
associated with good compliance (280%)
(odds ratio [OR] = 2.1; 95% confidence
interval [CI], 1.72-2.60).
Race has been shown to be significantly
associated with medication adherence in
the elderly. In the retrospective cohort study
by Monane et al, I3 white race was signifi-
R. BAJ_MUSHNAN
cantly associated with better medication
adherence (OR = 0.55 for blacks; 95% CI,
0.44-0.68).
In their study of veterans’
views on hypertension
and compliance,
Sharkness and Snow’* also found white
race to be significantly associated with better medication adherence (P = 0.10).
Schwarz et a1,6 studying a randomly selected population of chronically ill elderly
patients, found that virtually
identical
numbers of males and females made medication-taking
errors. Similarly, no relationship between medication
adherence
and sex was found by Coons et al” in a
population of elderly adults taking various prescription medications or by Monane et alI3 in a cohort of elderly outpatients taking antihypertensive
medication.
Only Co1 et al3 in their study of 315 elderly inpatients, reported a significant association between sex and medication adherence, finding that elderly females were
3.3 times more likely than elderly males
to be hospitalized for complications
resulting from medication
nonadherence
(P = 0.04).
As to other demographic
variables,
Coons et al” reported that nonadherence
in older adults was significantly
associated with higher socioeconomic
status
(P < 0.01). Co1 et al3 found that compared
with persons aged >65 years who were eligible for Medicaid, the risk of hospitalization as a result of nonadherence
was
lower (OR = 0.002) among those having a
high income level ($15,000-$20,000)
and
was greatly increased (OR = 8.3) among
those
at a middle
income
level
($lO,OOO-$15,000) (P = 0.02). These investigators also reported that nonadherence rates were significantly higher in elderly persons living alone, although Coons
et al” did not find living alone to be a predictor of medication nonadherence.
Medical Variables
Coons et al” found no association between elderly patients’ self-reported physical health and medication
adherence.
However, they did find a highly significant inverse relationship between psychological stress and medication adherence
(P < 0.01).
Sharkness and Snow’* reported that elderly patients having more than one
chronic illness requiring the use of multiple drugs were more likely to believe
themselves in need of treatment and therefore were more likely to be adherent to
their medication regimens than those requiring therapy with only one drug. Similar findings were reported by Monane et
all4 in their study of 7247 outpatients aged
65 to 99 years who were receiving Medicaid. However, in another study,15 the
same group found that patients taking
multiple agents for the same condition
were as likely to be compliant as those
taking single agents but that patients who
were prescribed multiple medications for
separate conditions showed significantly
poorer compliance (OR = 0.8 for >8 prescriptions; 95% CI, 0.7-0.9). These findings were not supported by the results of
German et al5 and Coons et al,” who
found no relationship between number of
comorbidities and medication adherence.
Medication-Related Variables
Monane et all5 taking thiazide users
as their reference group in a cohort of
elderly hypertensive
patients, found a
significantly
greater likelihood of medication adherence
among users of angiotensin-converting
enzyme inhibitors
(OR = 1.9; 95% CI, 1.6-2.1)
calcium
channel blockers (OR = 1.7; 95% CI,
767
CLINICAL THERAPEUTICS”
1S-2. l), and beta-adrenergic
blockers
(OR = 1.4; 95% CI, 1.2-1.7). Sclar et all6
have shown that using a sustained-release
formulation of antihypertensive
therapy
in Medicaid recipients
would improve
adherence to therapy and result in significant cost savings. Similarly, other studies17-19 have shown that use of sustainedrelease formulations results in significant
decreases in health-care service utilization and costs (P < 0.05).
Co1 et al3 found that the greater the
number of medications prescribed for elderly patients, the greater the medication
nonadherence (P = 0.02). Similar results
were obtained by Coons et al,” who
showed that nonadherence among the elderly was significantly
associated with
greater numbers of prescribed medications (P = 0.008). However, Sharkness
and Snow12 found that patients who took
more than one hypertensive
medication
were less likely to depart from the prescribed regimen than were those taking
only one drug (P = 0.02). They also reported that patients who were taking other
medications for comorbid conditions were
more compliant with their antihypertensive therapy.
Sclar et all9 reported that patients who
are initially prescribed an antihypertensive medication requiring once-daily or
weekly dosing rather than multiple daily
doses have infrequent changes in their
therapeutic regimens and far lower use of
concomitant
therapy for blood pressure
control (6% of patients taking once-daily
therapy, compared with 11% to 16% of
patients taking multiple daily doses; P <
0.05).
Skaer et a120~21
have shown that the
use of pharmacy-based value-added utilities such as prescription-refill
reminders
and unit-of-use
packaging significantly
improves adherence.
768
Economic Variables
Soumerai et a122demonstrated that limiting reimbursement
for effective drugs
puts low-income
elderly patients at increased risk of institutionalization
in nursing homes. Co1 et al3 found that higher
monthly medication costs were generally
associated with higher rates of nonadherence in the elderly. These investigators
found that elderly patients who thought
that their medications were expensive had
a much higher rate of nonadherence
(37.9%) than did those who did not think
their medications were expensive or did
not express an opinion (19.6%) (P = 0.03).
Co1 et al3 also reported that 52% of elderly patients who did not have prescription drug coverage were likely to depart
from a prescribed regimen, compared with
3 1% of those who had such coverage (P =
0.04). Soumerai et a12*reported that limiting reimbursement
to three drugs in elderly Medicaid recipients was associated
with an increased risk of admission to
nursing homes (relative risk [RR] = 2.2;
95% CI, 1.2-4.1) and hospitalization (RR
= 1.2; 95% CI, 0.8-1.6). This effect was
absent when the cap was discontinued.
Physician-Patient Interaction
Co1 et al3 found that the greater the
number of physicians patients saw regularly, the greater the number of hospital
admissions
associated with medication
nonadherence (P = 0.07). Monane et ali5
reported that medication adherence was
significantly
better in patients who had
more physician visits (OR = 2.2 for eight
or more recent visits; 95% CI, 1.8-2.5).
In a study involving
46 practicing
physicians and 357 patients with diabetes
mellitus or congestive heart failure, Hulka
R. BALKRISHNAN
et a123 found that neither patient characteristics nor severity of disease had an influence on the occurrence of medication
errors. In the patients with congestive
heart failure, good communication
of instructions and information by the physician was associated with low levels of all
types of medication error (P < 0.05).
Patients’ Health-Related
and Beliefs
Knowledge
Patients’ knowledge and beliefs about
health have been shown to affect their
medication-taking
behavior. Sharkness
and Snow12 reported that male veterans
who knew they would require lifelong
treatment
for hypertension
were 1.3
times less likely to depart from the prescribed regimen than were those who did
not know this (P = 0.04). Since publication of the study by Becker and Maiman
in 1975,24 the Health Belief Model has
frequently been used to explain medication adherence behavior. According
to
this model, there are certain sociobiological determinants
of patient compliance,
including
health beliefs, health-related
motivations,
perceptions of the psychological and other costs of the recommended action, and aspects of the physician-patient
relationship.
The model is
used to account for compliance in general, although it is not specific for medication adherence. According to Cramer,2
“no convincing
model for predicting
medication
adherence has been developed. The poor and inconsistent correlation between compliance and health beliefs suggests that the research focus
should be shifted from the prediction of
compliance problems to their identification and the development of ways to improve compliance.”
DISCUSSION
Although many studies have been published since, one still has to agree with
Becker and Maiman that “patient compliance has become the best documented,
but least understood, health behavior.“24
The present review of the published literature confirms that there is an association
between medication adherence and race,
drug and dosage form selected, number of
medications, cost of medications, insurance coverage,
and physician-patient
communication.
However, the findings on
the effects of age, sex, socioeconomic status, living arrangement, comorbidity, contact with physician, and health knowledge
and beliefs are inconsistent.
Although it has not been proved empirically, medication adherence seems essential to the success of any disease-management program. Thus the design and
development of any disease-management
program for the elderly must take into account the factors that clearly affect medication adherence. Some of these factors, such
as the choice of drug or dosage form and
physician-patient communication, are more
amenable to control through patient education than are ones such as insurance coverage and cost of medication. In particular,
the elderly nonwhite population should be
targeted by patient-education
programs to
improve medication adherence.
CONCLUSIONS
Our understanding
of medication adherence in the elderly is still incomplete. As
this review demonstrates, no single factor
predicts medication adherence in the elderly, and multiple predictors and their
interactions
need to be examined more
carefully. The role of several potential
769
CLINICAL THERAPEUTICS’
predictors of adherence behavior remains
unexplained or unclear. There is a need
for more research in this area to help design disease-management
programs that
specifically aim educational interventions
at the elderly patients who are most likely
to be noncompliant.
Improved adherence
should, in turn, lead to improved health
outcomes.
Address
correspondence
to: Rajesh
Balkrishnan,
MS Pharm, Division
of
Pharmaceutical
Policy and Evaluative
Sciences,
of
School
North
CB#7360,
of Pharmacy,
Carolina
Beard
Hall,
at
University
Chapel
Chapel
Hill,
Hill,
NC
27599-7360.
REFERENCES
6. Schwarz D, Wang M, Z&z L, et al. Medication errors made by elderly, chronically
ill patients. Am J Public Health. 1962;
52:2018-2029.
7. Greenberg RN. Overview of patient compliance with medication dosing: A literature review. Clin Thel: 1984;6:592-599.
8. Sullivan SD, Kreling DH, Hazlet TK.
Noncompliance with medication regimens
and subsequent hospitalizations: A literature analysis and cost of hospitalization
estimate. J Res Pharm Econ. 1990;2:
19-33.
9. DiMatteo MR, Hays RD. Adherence to
cancer regimens: Implications for treating
the older patient. Oncology. 1992;6(Suppl):
50-57.
1. Haynes RB. Introduction. In: Haynes RB,
Taylor DW, Sackett DL, eds. Compliance
in Health Care. Baltimore: The Johns
Hopkins University Press; 1979:1-7.
10. Breemhaar B, Visser AP, Kleijnen JGVM.
Perceptions and behavior among elderly
hospital patients: Descriptions and explanation of age differences in satisfaction,
knowledge, emotions and behavior. Sot
Sci Med. 1990;31:1377-1385.
2. Cramer JA. Relationship between medication compliance and medical outcomes.
Am J Health-Syst Pharm. 1995;52(Suppl
3):S27-S29.
11. Coons SJ, Shehan SL, Martin SS, et al.
Predictors of medication noncompliance
in a sample of older adults. Clin Ther.
1994;16:110-117.
3. Co1 N, Fanale JE, Komhom P. The role of
medication noncompliance
and adverse
drug reactions in hospitalizations
in the
elderly. Arch Intern Med. 1990;150:841845.
12. Sharkness CM, Snow DA. The patient’s
view of hypertension and compliance. Am
J Prev Med. 1992;8:141-146.
4. Cooper JK, Love DW, Raffoul PR. Intentional prescription
nonadherence
(noncompliance) in the elderly. J Am Geriutr
Sot. 1982;30:329-333.
13. Monane M, Bohn RL, Gurwitz JH. Compliance with antihypertensive
therapy
among elderly Medicaid enrollees: The
roles of age, gender and race. Am J Public Health. 1996;86:1805-1808.
5. German PS, Klein LE, McPhee SJ, et al.
Knowledge and compliance with drug regimens in the elderly. J Am Geriatr Sec.
1982;30:568-571.
14. Monane M, Bohn RL, Gurwitz JH. Noncompliance with congestive heart failure
therapy in the elderly. Arch Intern Med.
1994;154:433-437.
770
R. BALKRISHNAN
15. Monane M, Bohn RL, Gurwitz JH. The
effect of initial drug choice and comorbidity on antihypertensive
therapy compliance. Results from a population-based
study in the elderly. Am J Hypertens. 1997;
10:697-704.
20. Skaer TL, Sclar DA, Markowski DJ, et al.
Effect of value-added
utilities on prescription refill compliance and health care
expenditures for hypertension. J Hum Hypertens. 1993;7:515-518.
21
16. Sclar DA, Tessier GC, Skaer TL, et al. Effect of pharmaceutical formulation of diltiazem on the utilization of Medicaid and
health maintenance organization services.
Curr Ther Rex 1994;55:1136-1149.
17. Skaer TL, Sclar DA, Markowski DJ, et al.
Utility of a sustained release formulation
for antihypertensive
therapy. J Hum Hypertens. 1993;7:519-522.
Skaer TL, Sclar DA, Markowski DJ, et al.
Effect of value-added
utilities on prescription refill compliance and Medicaid
health expenditures-a
study of patients
with non-insulin-dependent
diabetes mellitus. J Clin Pharm Thel: 1993; 18:295-299.
22. Soumerai SB, Ross-Degnan D, Avron .I, et
al. Effects of Medicaid drug-payment limits on admission to hospitals and nursing
homes. NEJM. 1991;325:1072-1077.
18. Sclar DA, Skaer TL, Chin A, et al. Utility
of a transdermal delivery system for antihypertensive therapy. Part I. Am J Med.
1991;91(Suppl lA):50S-56s.
23. Hulka BS, Cassel JC, Kupper LL, et al.
Communication, compliance, and concordance between physicians and patients
with prescribed medications. Am J Public
Health. 1976;66:847-853.
19. Sclar DA, Skaer TL, Robison LM, et al.
Effect of antihypertensive formulation on
health services expenditures. Clin Auton
Res. 1993;3:363-368.
24. Becker MH, Maiman LA. Sociobehavioral
determinants
of compliance with health
and medical care recommendations.
Med
Care. 1975;13:10-24.
771