Full text of DOI:10.1093/gerona/55.10.M550

Journal of Gerontology: MEDICAL SCIENCES
Copyright 2000 by The Gerontological Society of America
2000, Vol. 55A, No. 10, M550–M553
One-Year Follow-Up of Medication
Management Capacity in Highly
Functioning Older Adults
Helen K. Edelberg,1,2 Elizabeth Shallenberger,2 Jeffrey M. Hausdorff,2 and Jeanne Y. Wei2
1The Henry L. Schwartz Department of Geriatrics and Adult Development, Mount Sinai School of Medicine, New York.
2Division on Aging, Harvard Medical School and Gerontology Division, Beth Israel Deaconess Medical Center,
Boston, Massachusetts.
Background. We tested the hypothesis that impairment in the ability to take medication independently predicts early
functional decline.
Methods. A 12-month, prospective cohort study was performed at two continuing-care retirement facilities using the
Drug Regimen Unassisted Grading Scale (DRUGS). This geriatric screening tool utilizes a stepwise progression of four
tasks: (i) identification, (ii) access, (iii) dosage, and (iv) timing.
Results. Forty-seven (86%) of the eligible participants completed the 12-month follow-up assessment; three were
transferred to skilled nursing facilities. The mean age at study entry was 84.2 Ϯ 5.1 years; 72% of the participants were
women, and 68% were college educated. At 12 months there was a decline in the Mini-Mental State Examination
(MMSE) score (p ϭ .029), an increase in the timed “Up and Go” test (p ϭ .023), and a decline in the DRUGS score (p ϭ
.029). Nine (18%) of the participants resided in assisted- versus independent-living situations compared with three par-
ticipants (5%) at study entry (p ϭ .031). Both 12-month DRUGS score and 12-month self-reported medication manage-
ment capacity were associated with 12-month MMSE (p ϭ .0001 and p ϭ .019, respectively). Baseline DRUGS score
was associated with 12-month MMSE and Geriatric Depression Scale scores (p ϭ .0002 and p ϭ .002, respectively).
Both baseline DRUGS score and self-reported medication management capacity were also associated with residence in
assisted-living communities at 6 months (p ϭ .029 and p ϭ .040, respectively). MMSE was not associated with any of
the clinical outcomes.
Conclusions. The DRUGS tool may predict functional decline in highly functioning older adults.
LTHOUGH population-based screening data suggest
that needing or receiving assistance with medication
METHODS
A
management may predict increased risk of frailty, previous
studies of medication management have not extensively ex-
plored the relationship between medication management,
functional status, and the risk of functional decline (1–5).
Cognitive impairment has been associated with the inability
to take medications independently, as well as with gait dis-
orders, falls, and functional impairment. (2–4,6–10). The on-
set of functional impairment and disability may be heralded
by subtle changes in either cognitive or physical status.
We recently developed and tested a geriatric screening
tool, the Drug Regimen Unassisted Grading Scale (DRUGS)
to assess medication management and various aspects of
function in the ambulatory setting (11). An extension of one
item of the Instrumental Activities of Daily Living (IADL)
scale and the “brown-bag” test, the DRUGS tool was devel-
oped as a performance-based, individualized measure of the
patient’s capacity to manage his or her own medication reg-
imen (11,12). The purpose of this study is to determine if an
impairment in the ability to take medications independently,
measured by the DRUGS tool, correlates with overall func-
tional status and predicts future functional disability.
Geriatric Screening Tools
Self-report.—Modified Katz Activities of Daily Living
(maximum score ϭ 6) and Lawton IADL (maximum score ϭ
8) scales were used to assess basic and instrumental activi-
ties of daily living, respectively (12,13). In addition, each
participant was asked a series of medication management
questions (11): (i) Do you take your own medication with-
out help?; (ii) Does someone remind you to take your medi-
cation on a regular basis?; and (iii) Does someone set up
your medication in advance?
DRUGS tool.—The DRUGS tool was employed as pre-
viously described (11). Briefly, this tool involves a stepwise
progression of four medication management tasks: (i) iden-
tification, or showing the appropriate medications; (ii) ac-
cess, or opening the appropriate containers; (iii) dosage, or
taking out the correct number per dose; and (iv) timing, or
demonstrating the timing of doses. The subject is asked to
perform these tasks for each of the prescription and nonpre-
scription medications that he or she plans to take on the day
M550

MEDICATION MANAGEMENT IN OLDER ADULTS
M551
of the evaluation, including PRN or “as needed” medica-
tions, using the DRUGS tool. A visual aid, in the form of a
sheet of paper marked with a grid labeled “time” (7 am–11
pm), “meal,” and “medications,” is employed to standardize
the process. The DRUGS tool is an easy-to-administer indi-
vidualized measure that examines medication self-adminis-
tration. It takes approximately 4 to 5 minutes, and interrater
and test-retest reliability of the DRUGS tool are Ͼ.90 (11).
The Spearman correlation, Fisher’s exact, chi-square, and
Wilcoxon rank sum tests were employed to examine the re-
lationships between predictors.
New variables were derived to control for baseline values
by calculating the mathematical difference between values
at either 6 or 12 months and baseline values. Multivariate
models were derived using stepwise linear regression in
which the dependent variable was either baseline DRUGS
summary score or 6- or 12-month change in DRUGS sum-
mary score. Age, sex, and covariates with p Ͻ .2 on univari-
ate analysis were entered into the models, and variables
with p Ͻ .05 were retained in the models. Additional statis-
tical analyses were performed to explore whether changes
in self-reported medication management capacity or MMSE
were also associated with the standard measures or clinical
outcomes. Stepwise logistic regression models were derived
using baseline self-reported medication management capac-
ity as the dependent variable. Collinearity diagnostics were
performed on the final models. Nonsignificant predictors
were entered into the final models to check for uncontrolled
confounding (change in ␤ coefficient of Ͼ10%). Residual
and regression diagnostics were performed.
Standard Measures of Functional Status
The results of these geriatric screening tools were com-
pared with a complement of previously validated standard
measures of functional status, which were categorized for
clinical relevance. There were tests of cognitive status
(Mini-Mental State Examination [MMSE]), affective status
(15-item Geriatric Depression Scale), physical function
(timed “Up and Go” test), medical conditions (Jaeger card to
test near vision, Charlson Comorbidity Index), medication-
specific factors (number of medications, number of doses),
and social factors (living alone or with a partner) (14–18).
Covariates included age, gender, and level of education.
Patients and Setting
RESULTS
This prospective cohort study was conducted at two con-
tinuing-care retirement facilities in the greater Boston area.
Patients were recruited from the Beth Israel Deaconess Ger-
ontology Group practice between October 10, 1996, and Jan-
uary 27, 1997. All ambulatory community-dwelling patients
aged 70 years and older who presented to one of the study
sites were eligible for the study. Patients were excluded if
they were currently not taking any prescription or nonpre-
scription medications or if they refused to participate.
Study Population
A total of 67 patients were approached. Four were ineligi-
ble, two did not take any medications, one lived in a differ-
ent facility at the time of the study, and five declined to par-
ticipate (these individuals were more cognitively and
functionally impaired than those who agreed to participate).
A total of 58 subjects were enrolled in the study and com-
pleted the baseline assessment. At 6 months, three patients
had moved from independent apartments to an on-site
skilled nursing facility (study endpoint). Fifty-three of the
55 community-dwelling patients (96%) completed the
6-month follow-up assessment; one refused to participate
and another did not complete the DRUGS test. Forty-seven
(86%) of the 55 eligible participants completed the 12-
month follow-up assessment; five refused to complete the
study, and three were lost to follow-up.
The mean age at study entry was 84.2 Ϯ 5.1 years; 72%
of the participants were women, 68% were college edu-
cated. There was no significant difference between the
patients who participated in the baseline, 6-month, and 12-
month assessments in terms of age, gender, living arrange-
ment, or level of education (Table 1).
Data Collection
Data for all participants were collected at the time of the
initial outpatient office visit (HE, ES) and at a 6-month fol-
low-up visit. The initial contact was used to obtain written
informed consent and collect clinical data. The chart review
data included the list of medications and the 19 variables
that constitute the Charlson Comorbidity Index (18). Test
status was ascertained by an investigator (HE, ES) who was
blinded to health status and sociodemographic factors. An
additional office visit was scheduled at 6 months. Data were
obtained regarding predefined clinical outcomes: interim
clinic visits (0, Ն2), emergency room visits (0, Ն1), acute
care hospitalizations (0, Ն1), institutionalization, and death.
This information was confirmed by chart review. The insti-
tutional review board of the Beth Israel Deaconess Medical
Center approved the protocol.
Table 1. Demographic Data at Baseline, 6, and 12 Months
Data Analysis
Baseline
6 Mo
12 Mo
Univariate and multivariate analyses were performed us-
ing the SAS statistical package for Windows, version 6.12
(SAS Institute Inc., Carey, NC). Univariate associations be-
tween the continuous outcomes (DRUGS summary score,
change in DRUGS summary score) and continuous vari-
ables were examined by Spearman correlation. Student’s t
test was used to compare the DRUGS score and change in
DRUGS score between two groups for binary predictors.
Bonferroni correction was applied to multiple comparisons.
Number of participants (%)
Mean age, y (SD)
Gender, number women (%)
Living arrangement, number
alone (%)
Level of education, number
completed college (%)
Assisted living, number (%)
Nursing home, number (%)
59
84.2 (5.1)
43 (72)
53 (90)
84.1 (4.9)
37 (70)
47 (81)
85.2 (5.3)
31 (70)
44 (73)
37 (70)
31 (70)
41 (69)
3 (5)
0 (0)
36 (68)
5 (8)
3 (5)
30 (4)
9 (18)
3 (6)

M552
EDELBERG ET AL.
Change in Functional Status
Change in DRUGS score between baseline and 6 months
was associated with an increased frequency of emergency
department visits ( p ϭ .024), whereas change in self-
reported medication management capacity was associated
with an increased frequency of ambulatory clinic visits
( p ϭ .024). Change in DRUGS summary score and in self-
reported medication management capacity between baseline
and 12 months was associated with residence in assisted-
versus independent-living situations at 12 months ( p ϭ
.049, p ϭ .030).
The baseline, 6-, and 12-month mean values for standard
measures of cognitive, affective, physical, and functional
status, medication-specific factors, and DRUGS summary
scores are presented in Table 2. There was a decline in
MMSE (27.2 to 26.4, p ϭ .029), an increase in the timed
“Up and Go” test (13.1 to 18.6 s, p ϭ .023), and a decline in
the DRUGS score (93.2 to 81.1, p ϭ .029). The increase in
the timed “Up and Go” was also evident at 6 months (13.1
to 17.5 s, p ϭ .0008). There was no change in the Geriatric
Depression Scale, Jaeger score, or Charlson Comorbidity
Index. Self-reported IADL and activities of daily living
(ADL) scores did not change over the 12-month follow-up
period. At 12 months, nine (18%) of the participants resided
in assisted-living rather than independent-living housing
units, compared with three participants (5%) at study entry
( p ϭ .031).
Predictive Value
Both baseline DRUGS score and self-reported medica-
tion management capacity were also associated with resi-
dence in assisted-living situations at 6 months (summary
score, 80.2 vs 94.4 for those in assisted- vs independent-liv-
ing situations at 6 months; p ϭ .029, p ϭ .040, respec-
tively). There was no association between either the per-
formance-based or self-reported measures of medication
management capacity and the 12-month clinical outcomes.
MMSE was not associated with any of the 6- or 12-month
clinical outcomes.
Multivariate analyses, adjusted for age, gender, living ar-
rangement, and medication-specific factors, demonstrated a
statistically significant association between baseline DRUGS
summary score (dependent variable), change in MMSE (␤ ϭ
1.69, p ϭ .0002), and change in Geriatric Depression Scale
(␤ ϭ 1.93, p ϭ .002) over time. The logistic regression
model with baseline self-reported medication management
capacity as the dependent variable, adjusted for age, gender,
living arrangement, and medication-specific factors, in-
cluded change in Geriatric Depression Scale (odds ratio ϭ
0.99, 95% confidence interval ϭ 0.62, 1.57; p ϭ .965).
There was no evidence of collinearity between the vari-
ables in the model by Belsey-Kuh-Welch criteria. Outliers
(data points with jackknife residuals Ͼ3 or ϽϪ3) were re-
moved, and the model was rerun without significant change
in the ␤ coefficients. There were no influential points
(Cook’s distance Ͼ1).
Baseline, 6-, and 12-Month Correlates
At baseline, 6, and 12 months, there was a positive associa-
tion between the DRUGS tool and MMSE scores (r ϭ .42, p ϭ
.0008; r ϭ .34, p ϭ .003; and r ϭ .35, p ϭ .018, respectively).
Controlling for medication-specific factors in a multivariate
model, both 12-month DRUGS score and 12-month self-
reported medication management capacity were associated
with 12-month MMSE (p ϭ .0001, p ϭ .019, respectively).
There was a significant association between the 6-month
DRUGS tool and 6-month self-reported ADL capacity ( p ϭ
.025). Both 12-month DRUGS scores and 12-month self-
reported medication management capacity were associated
with 12-month IADL capacity ( p ϭ .026, p ϭ .007, respec-
tively). Neither DRUGS performance nor self-reported
medication management capacity was associated with the
total number or total dose of medication at baseline, 6, or 12
months.
Table 2. Standard Measures, Self-Report and DRUGS Tool at
Baseline, 6, and 12 Months
Baseline
DRUGS
Mean (SD)
6-Mo
DRUGS
Mean (SD)
12-Mo
DRUGS
Mean (SD)
DISCUSSION
This study had three major findings. First, the DRUGS
tool has face validity as a measure of medication manage-
ment capacity over time. Second, the DRUGS tool provides
information about functional correlates, particularly cogni-
tive status. Finally, the DRUGS score is apparently associ-
ated with subsequent transfer from an independent-living to
an assisted-living situation.
Standard
Charlson Index
1.3 (1.2)
5.4 (3.7)
27.2 (2.4)
3.6 (3.1)
13.1 (5.9)
1.4 (1.4)
4.6 (3.8)
26.8 (3.5)
3.7 (2.8)
1.2 (1.2)
5.0 (3.9)
26.4 (4.1)*
3.9 (2.9)
Jaeger score
MMSE
Geriatric Depression Scale
Timed “Up and Go,” s
17.5 (10.8)*
18.6 (16.3)*
Self-report
The DRUGS tool has implied functional significance. It
has intrinsic value in assessing medication management ca-
pacity. Short-term follow-up data confirm the relationship
between self-reported ADL and IADL scores and self-
reported and performance-based medication management
capacity. Irrespective of the actual summary score, the
DRUGS tool is designed to provide valuable information
regarding the patient’s ability to manage his or her own
medication regimen without assistance.
ADLs capacity
IADLs capacity
5.8 (0.4)
6.4 (1.9)
48 (81.4)
5.8 (0.5)
6.3 (2.3)
41 (75.9)
5.7 (0.7)
6.6 (2.3)
38 (80.9)
MM capacity, number (%)†
DRUGS tool
Total medications
Total doses
Summary score
6.6 (3.0)
9.3 (5.3)
93.2 (11.2)
4.2 (2.9)
6.4 (3.7)
9.1 (6.0)
87.9 (20.3)
3.3 (2.3)
7.6 (3.6)
10.3 (5.4)
81.1 (29.3)*
4.4 (4.4)
Time to completion, min
Note: MMSE ϭ Mini-Mental State Examination; ADLs ϭ activities of daily
living; IADLs ϭ instrumental activities of daily living; MM ϭ medication
management.
*Denotes statistically significant change compared with baseline, p Յ .05.
†Able to take medication independently by self-report.
There is a robust association between the DRUGS tool
and cognitive status as measured by the MMSE. Unlike the
MMSE, however, change in DRUGS score is associated

MEDICATION MANAGEMENT IN OLDER ADULTS
M553
3. Fitten LJ, Coleman L, Siembieda DW, Yu M, Ganzell S. Assessment
of capacity to comply with medication regimens in older patients. J
Am Geriatr Soc. 1995;43:361–367.
4. Meyer ME, Schuna AA. Assessment of geriatric patients’ functional
ability to take medication. DICP. 1989;23:171–174.
5. Smith NL, Psaty BM, Heckbert SR, Tracy RP, Cornell ES. The reli-
ability of medication inventory methods compared to serum levels of
cardiovascular drugs in the elderly. J Clin Epidemiol. 1999;52:143–
146.
6. Fulmer T, Gurland B. Evaluating the caregiver’s intervention in the el-
der’s task performance: capacity versus actual behavior. Int J Geriatr
Psychiatry. 1997;12:920–925.
7. Ruscin JM, Semla TP. Assessment of medication management skills
in older outpatients. Ann Pharmacother. 1996;30:1083–1088.
8. Gambassi G, Landi F, Lapane KL, Sgdari A, Mor V, Bernabei R. Is
drug use by elderly with cognitive impairment influenced by type of
dementia? Pharmacotherapy. 1999;19:430–436.
9. Nakamura T, Meguro K, Sasaki H. Relationship between falls and
stride length variability in senile dementia of the Alzheimer type. The
Gerontologist. 1996;36:108–113.
10. Tinetti ME, Inouye SK, Gill TM, Doucette JT. Shared risk factors for
falls, incontinence, and functional dependence. Unifying the approach
to geriatric syndromes. JAMA. 1995;273:1348–1353.
11. Edelberg HK, Shallenberger E, Wei JY. Medication management ca-
pacity in highly functioning community-living older adults: detection
of early deficits. J Am Geriatr Soc. 1999;47:592–596.
with the need for increased home services (i.e., assisted liv-
ing) and multiple emergency room visits. This study did not
explore the specific reasons for the transfer. The DRUGS
score may reflect deterioration in health or functional status.
Alternatively, mismanagement of medications, with or
without subsequent adverse drug events, may have precipi-
tated the move. Medication management capacity as mea-
sured by the DRUGS tool may provide a key to understand-
ing the patient’s ability to live independently.
There are several limitations to this study. This was a
small cohort of highly functioning, well-educated individu-
als. The power of the study was insufficient to detect differ-
ences in clinical outcomes, such as institutionalization or
death. In addition, the results are not readily generalizable
to the population at large, and the DRUGS tool needs to be
validated in a larger, ethnically and socioeconomically di-
verse population of older adults. Nevertheless, the DRUGS
tool appeared to identify highly functioning individuals who
are at risk of functional decline in this sample population.
These findings may actually strengthen the validity and
comparative value of the DRUGS tool as a measure of med-
ication management capacity and functional status.
12. Lawton MP, Brody EM. Assessment of older people: self-maintaining
and instrumental activities of daily living. The Gerontologist. 1969;9:
179–186.
13. Katz S, Ford AB, Moskowitz RW, Jackson BA, Jaffe MW. Studies of
illness in the aged. The index of ADL: a standardized measure of bio-
logical and psychosocial function. JAMA. 1963;185:914–919.
14. Folstein MF, Folstein SE, McHugh PR. Mini-Mental State: a practical
method for grading the cognitive state of patients for the clinician. J
Psychiatr Res. 1975;12:189–198.
In conclusion, the DRUGS tool, a performance-based
screening tool that relies on direct observation, may com-
plement self-report measures of medication management
capacity and may be useful in predicting functional decline.
Acknowledgments
This research was supported, in part, by the Department of Health and
Human Services (Grants AG00294, AG608812, and AG00251).
15. Yesavage JA. Geriatric Depression Scale. Psychopharmacol Bull.
1988;24:709–711.
Address correspondence to Dr. H. Edelberg, The Henry L. Schwartz
Department of Geriatrics and Adult Development, Box 1070, Mount Sinai
School of Medicine, 1 Gustave L. Levy Place, New York, NY 10029-6574.
E-mail: helen.edelberg@mssm.edu
16. Podsiadlo D, Richardson S. The timed “Up and Go”: a basic test of
functional mobility. J Am Geriatr Soc. 1991;39:142.
17. Leitman MW. Manual for Eye Examination and Diagnosis. Oradell,
NJ: Medical Economics Company; 1946.
18. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of
classifying prognostic comorbidity longitudinal studies: development
and validation. J Chron Dis. 1987;40:373–383.
References
1. Brody KK, Johnson RE, Douglas Ried L. Evaluation of a self-report
screening instrument to predict frailty outcomes in aging populations.
The Gerontologist. 1997;37:182–191.
2. Burns E, Austin CA, Bax ND. Elderly patients’ understanding of their
drug therapy: the effect of cognitive function. Age Ageing.1990;19:
236–240.
Received March 7, 2000
Accepted March 13, 2000
Decision Editor: William B. Ershler, MD