Association Between Depressive and Anxiety Disorders and
Adherence to Antihypertensive Medication in Community-
Living Elderly Adults
Lia Gentil, MD,* Helen Maria Vasiliadis, PhD,*†Michel Pre ´ville, PhD,*†Cindy Bosse ´, BSc,* and
Djamal Berbiche, PhD†
OBJECTIVES: To identify the determinants of antihyper-
tensive medication adherence in community-living elderly
DESIGN: Longitudinal observational study.
SETTING: Population-based health survey in the province
of Quebec, Canada.
PARTICIPANTS: Data from
(N = 2,811) of community-dwelling adults in Quebec aged
65 and older participating in the E´tude sur la Sante ´ des
Aı ˆne ´s study. The final study sample analyzed consisted of
926 participants taking antihypertensive drugs during the
2 years of the study.
MEASUREMENTS: Adherence to antihypertensive medi-
cation was measured using days of supply obtained during
a specified time period. Depression and anxiety disorders
were assessed using Diagnostic and Statistical Manual of
Mental Disorders, Fourth Edition, criteria, and physical
health status was measured using the Charlson Comorbidi-
ty Index. Other factors considered were age, education,
marital status, annual family income, and number of anti-
hypertensive drugs that participants used.
RESULTS: Mean antihypertensive proportion (percentage)
of days supplied in was 92.5% in Year 1 and 59.4% in
Year 2. The presence of depression and anxiety disorders
and the number of antihypertensive medications signifi-
cantly predicted medication adherence. The sex by depres-
sion and anxiety disorders interaction term was significant.
CONCLUSION: Adherence to antihypertensive medica-
tion was significantly associated with depression and anxi-
ety disorders in men but not women. The treatment of
depression and anxiety disorders in individuals with hyper-
tension may be helpful in improving medication adherence
rates and healthcare outcomes. J Am Geriatr Soc 60:2297–
depressive and anxiety disorders; older adults
words: antihypertensivemedication adherence;
sion is one of the most important modifiable risk factors
for cardiovascular disease, end-stage renal disease, stroke,
and death.2Pharmacotherapy is an important aspect of the
management of hypertension. Poor adherence to antihyper-
tensive drugs has been acknowledged as a major contribu-
tor to the lack of blood pressure control,3which in turn
significantly affects clinical outcomes and puts a heavy
economic burden on individual patients and society in
Barriers to antihypertensive medication adherence are
multifactorial. Previous studies have shown age, social
support, income, health status, and number of antihyper-
tensive medications to be associated with antihypertensive
drug adherence.5–7Others have also shown that mental
disorders such as depression are associated with poor
medication adherence, especially for those with chronic
disease.8Understanding the determinants of adherence to
antihypertensive medication may help identify subpopula-
tions and interventions to improve adherence and outcomes
in elderly adults.
Using a large representative community sample of
elderly adults and an adaptation of Andersen’s behavioral
model of health care that suggests that predisposing,
enabling, and need factors determine an individual’s access
to and use of health services,9the purpose of this study
was to identify the association between predisposing
factors such as age, sex, and level of education; enabling
factors such as marital status, level of income, and mono-
he prevalence rate of hypertension in Canadian
seniors was estimated to be 65% in 2007.1Hyperten-
From the *Sante ´ Communautaire, University of Sherbrook, and†Research
Center, Charles LeMoyne Hospital, Longueuil, Quebec, Canada.
Address correspondence to Lia Gentil, 150 Place Charles-Le Moyne
bureau 200, C.P. 11, Longueuil, QC, Canada J4K 0A8.
© 2012, Copyright the Authors
Journal compilation © 2012, The American Geriatrics Society
therapy, and needs factors such physical and mental health
and adherence to antihypertensive medication using survey
and administrative data, which has the advantage of
increasing the validity of the results.
The data used in this study were obtained from the
Enque ˆte sur la Sante ´ des Aı ˆne ´s (ESA) survey conducted in
2005 to 2008 using a probabilistic sample (N = 2,811) of
French-speaking community-dwelling adults aged 65 and
older (94% of the Quebec population speaks French). Par-
ticipants living in the northern regions were excluded on
feasibility grounds; in 2005, 10% of the elderly population
resided in the northern regions. A random-digit dialling
method was used to develop the sampling frame of the
study, which included stratification according to three geo-
graphical areas—metropolitan area with a total population
of at least 100,000 inhabitants, urban area with a popula-
tion between 1,000 and 99,999 inhabitants, and rural area
with fewer than 1,000 inhabitants—according to the defi-
nitions of the Quebec Institute of Statistics.10In each geo-
graphical area, a proportional sample of households was
constituted according to the 16 administrative regions of
Quebec. The response rate for this study was 76.5%.
Characteristics such as age, sex, and region were similar in
nonrespondents and respondents.
A health professional first contacted potential respondents
on the telephone to describe the objectives and length of
the study, answer questions, and ask them to participate in
an in-home interview. Next, a letter describing the study
was sent to reassure potential participants about the credi-
bility of the investigation and of the interviewer. Appoint-
ments were then made with those who volunteered. Data
were collected through semistructured in-home interviews
conducted by trained research nurses (n = 20), all staff
members of a national polling firm, to whom the principal
investigator (MP) gave 2 days of training on administration
of the computer-assisted ESA Diagnostic Questionnaire
(ESA-Q). Respondents were offered $15 compensation for
The in-home interviews, which lasted 90 minutes on
average, took place within 2 weeks of initial contact. Writ-
ten consent to conduct the interview was obtained at the
beginning of the interview from all volunteers. Because
memory problems affect the accuracy of the information
given andperformance on
naires,11,12people with severe or moderate cognitive prob-
(score < 22)13were excluded at the beginning of the inter-
view (n = 27). Thereafter, individuals with no moderate or
severe cognitive problems were invited to respond to the
ESA-Q (n = 2,784). At the end of the interview, respon-
dents were asked to provide written informed consent for
the research team to access their health and pharmaceuti-
cal services data from the Re ´gie d’Assurance-Maladie du
Que ´bec (RAMQ) (Quebec health insurance plan). Self-
reported data from the ESA survey were linked to individ-
ual-level information from the RAMQ medical and phar-
maceutical services databases and from the health ministry
respondent’s health insurance number or, if that number
was missing, using the respondent’s name, sex, address,
and month and year of birth. A success rate of 99.6%
(n = 2,494) was obtained in the matching of the data,
which made up the sample for analysis. Participants with
private drug insurance were excluded (n = 208). The ethics
committee of the Sherbrook Geriatric University Institute
reviewed and authorized the research procedure.
The sample included all participants with a diagnosis of
hypertension (codes 401.x–405.x) according to the Inter-
national Classification of Diseases, Ninth Revision, Clini-
cal Modification (ICD-9-CM), in RAMQ or MED-ECHO
databases. Individuals who did not have prescriptions
for antihypertensive drugs (?30 days) were excluded
(n = 40). The final study sample analyzed consisted of 926
participants diagnosed with
antihypertensive drugs during the 2 years of the study.
Various methods have been used to estimate antihy-
pertensive medication adherence using retrospective data-
bases.14In this study, medication adherence was measured
by estimating the proportion (percentage) of days supply
of medication dispensed during a specified follow-up per-
iod divided by the number of days from the first dispens-
ing to the end of the follow-up period.15The index date of
the study corresponds to the date when the first antihyper-
tensive medication was filled during the study period. The
period of exposure studied was 2 years divided into two
periods: 12 months before and 12 months after the ESA
interview. The number of days a patient was hospitalized
was subtracted from the denominator because the medica-
tions dispensed in hospital are not included in the RAMQ
database. A preliminary analysis showed that depression
and anxiety were not associated with switching of antihy-
pertensive drugs (Fisher test P = .06).
The predisposing and enabling factors included in this
study were age (65–74 vs
?10 years), marital status (married or living as a couple
vs single, separated, divorced, widowed), and annual
family income (<CDN $15,000 vs ?CDN $15,000). The
number of antihypertensive drugs that participants used
was classified in two categories (monotherapy: being trea-
ted with a single antihypertensive drug, and polytherapy:
being treated with two or more antihypertensive drugs).
A previously proposed method of imputation for missing
categorical data was used to estimate income for participants
for whom that information was missing (n = 90).16
Need factors were respondent mental and physical
health status. Respondent mental health status was mea-
sured using a computer-assisted questionnaire, the ESA-Q,
which the research team developed and was based on
Diagnostic and Statistical Manual of Mental Disorders,
Fourth Edition (DSM-IV) criteria.17The ESA-Q is an
adapted version of the mood and anxiety disorders sec-
tions of the Diagnostic Interview Schedule and the Com-
posite International Diagnostic Interview, which have
demonstrated satisfactory reliability and validity to estab-
lish psychiatric diagnoses.18–20The complete definition of
?75), education (<10 vs
GENTIL ET AL.
DECEMBER 2012–VOL. 60, NO. 12 JAGS
the disorders studied in the ESA survey has been reported
previously.21For the analysis, the respondents were classi-
fied as having at least one or no probable DSM-IV disor-
der (depressive or anxiety disorder) during the observation
Respondent physical health was measured using the
Charlson Comorbidity Index.22Each medical condition
has a weight assigned from 1 to 6 derived from relative
risk estimates of proportional hazard regression models
using clinical data. This index has been shown to have a
strong monotonic association of approximately a doubling
in mortality per increment in index level.22This index was
calculated using medical claims (ICD-9-CM) codes for the
12-month period before the interview.
Frequencies were calculated for categorical variables and
means for continuous variables. Pearson correlations were
determined between all variables and medication adher-
ence to identify possible linear and nonlinear correlates of
A repeated-measures regression model was used to
identify change in adherence between the two 12-month
periods of the study as a function of the predisposing,
enabling, and needs factors. The association between the
interaction term sex by anxiety and depression and adher-
ence was also tested. Statistical analyses were performed
using SAS version 9.1 (SAS Institute, Inc., Cary, NC) and
a 95% confidence interval.
Nine hundred twenty-six respondents diagnosed with
hypertension who were taking antihypertensive medica-
tions were studied (Table 1). The majority of the respon-
dents were female (75.1%) and not married (56.7%),
reported an annual income of greater than $15,000
(81.2%), and had at least 10 years of education (74.2%).
The reported prevalence of a depressive or anxiety disorder
was 13.7% in this sample. More than half of the respon-
dents used more than one antihypertensive medication
(diuretics, beta-blockers, angiotensin-converting enzyme
inhibitors, calcium channel blockers). Diuretics (13.4%)
and angiotensin-I converting enzyme inhibitors (17%)
were the most commonly prescribed initial agents. The
mean antihypertensive proportion (percentage) of days
supply obtained was 92.5% in Year 1 and 59.4% in Year 2
(P < .001).
Table 2 shows the correlation between medication
adherence and study variables. Medication adherence in
Year 1 was significantly correlated with sex, a depressive
or anxiety disorder, and income. In Year 2, medication
adherence was correlated with monotherapy.
Table 1. Respondents’ Sociodemographic and Health
Characteristics (N = 926)
CharacteristicN % (95% Confidence Interval)
Single, divorced, widowed
Depressive or anxiety disorder
Charlson Comorbidity Index
and Study Variables
Correlations Between Medication Adherence
Year 1 Year 2
Depressive or anxiety disorder
Charlson Comorbidity Index
P < .05.
Table 3. Comparison of Medication Adherence Between
Year 1 and Year 2 as a Function of the Predisposing,
Enabling, and Needs Factors and Interaction Term
Beta (95% Confidence
Single, separated, divorced,
Education < 10 years
Absence of anxiety or
Charlson Comorbidity Index 0
Income < $15,000
Diagnostic and Statistical
Manual of Mental Disorders
by sex interaction term
0.006 (?0.023 to ?0.037)
?0.007 (?0.040 to ?0.024)
?0.008 (?0.118 to ?0.05)
JAGSDECEMBER 2012–VOL. 60, NO. 12
MEDICATION ADHERENCE IN OLDER ADULTS
The repeated-measures regression analysis (Table 3)
showed that, of the variables studied in the model, a
depressive or anxiety disorder and number of antihyper-
adherence. Greater adherence was observed in participants
without a depressive or anxiety disorder and those taking
only one antihypertensive medication.
With regard to the interaction between sex and
depressive or anxiety disorder, the results showed a signifi-
cant association. Women reporting a depressive or anxiety
disorder were more adherent than men reporting the pres-
ence of a mental disorder (B = 0.087, P = .046), although
there was no significant difference in medication adherence
between women and men without probable depressive or
anxiety disorder (B = 0.024, P = .20).
To the knowledge of the authors, this is the first study of
the association between depression and anxiety disorder
and adherence to antihypertensive medication in a large,
representative, community-living, older-adult population,
adjusting for a number of predisposing, enabling, and
needs factors such as depression and anxiety and physical
status (Charlson Comorbidity Index).
This study found a relationship between depressive or
anxiety disorder and hypertension in this older adult popu-
lation, which is consistent with previous studies in the gen-
eral population.23Adherence to medication decreased over
time, particularly within the first 12 months. Early discon-
tinuation of treatment is a major problem with long-term
antihypertensive treatment. According to a previous study,
half of individuals stop treatment within 1 year despite
having been prescribed longer-term treatment.24In this
study, the adherence rates observed with antihypertensive
medications were high in Year 1 (92.5%) but decreased
significantly in Year 2 (59.2%).
Other studies show antihypertensive medication adher-
ence rates ranging from 35% to as high as 97%.25,26One
study found that 40% of the sample reported some form
of medication nonadherence during the past year.27
The results also show that the use of numerous antihy-
pertensive medications is associated with poor adherence.
The number of antihypertensive drug classes is a predictor
of better adherence.28The results of the current study also
showed that the association between depressive or anxiety
disorder and antihypertensive medication adherence was
stronger in men than women.29
Finally, the findings of the current study should be
interpreted in light of some limitations. First, the adminis-
trative databases do not contain information on blood
pressure, so only changes in prescribing frequency and
patterns of antihypertensive use can be used as proxy
measures for changes in detection and control of hypertension.
Second, it is possible that some of the participants charac-
terized as totally adherent did not take their medications
despite filling their prescriptions, leading to overestimation
of adherence in this study. Third, the RAMQ has some
important advantages, including the ability to document
all healthcare services use without recall bias or incomplete
history information, but the limitations of this claims-
based information must be considered.30It is possible that
participants discontinued use of their drugs on the advice
of their doctors because of side effects, although it
is expected that these physicians would start therapy
with another antihypertensive medication. Fourth, an
MMSE score greater than 22 does not exclude cognitive
impairment significant enough to compromise the validity
of the interview, but only 30 individuals (N = 926) with
an MMSE score less than 24 were included. In addition,
self-reported data were combined with administrative data
from RAMQ, which contains information about medica-
These findings highlight the need to consider individual
risk factors when prescribing antihypertensive medication.
Addressing problems such as depression or anxiety with
hypertension may be helpful in improving medication
adherence rates and health are outcomes.
Conflict of Interest: This study was supported by Canadian
Institutes of Health Research (CIHR) Operating Grant
200683MOP. The ESA study was supported by CIHR
Operating Grant 200403MOP and a Quebec Health
Research fund—Fonds de recherche en Sante ´ du Que ´bec
Author Contributions: Gentil and Vasiliadis: Design of
analyses, interpretation of data, preparation of manuscript.
Pre ´ville: Concept and design, acquisition of participants
and data, preparation of manuscript. Berbiche: Data analy-
sis. Bosse ´: Preparation of manuscript.
Sponsor’s Role: None.
1. Public Health Agency of Canada. Report from the Canadian Chronic
Disease Surveillance System: Hypertension in Canada, 2010. Ottawa,
ON, Canada: Public Health Agency of Canada [on-line]. Available at
CCDSS_HTN_Report_FINAL_EN_20100513.pdf Accessed July 8, 2010.
2. MacMahon S, Peto R, Cutler J, et al. Blood pressure, stroke, and coronary
heart disease. Part 1, Prolonged differences in blood pressure: Prospective
observational studies corrected for the regression dilution bias. Lancet
3. Krousel-Wood M, Thomas S, Muntner P, et al. Medication adherence: A
key factor in achieving blood pressure control and good clinical outcomes
in hypertensive patients. Curr Opin Cardiol 2004;19:357–362.
4. Hsu YH, Mao CL, Wey M. Antihypertensive medication adherence among
elderly Chinese Americans. J Transcult Nurs 2010;21:297–305.
5. Monane M, Bohn RL, Gurwitz JH, et al. Compliance with antihypertensive
therapy among elderly Medicaid enrollees: The roles of age, gender, and
race. Am J Public Health 1996;86:1805–1808.
6. Morris AB, Li J, Kroenke K, et al. Factors associated with drug adherence
and blood pressure control in patients with hypertension. Pharmacotherapy
7. Vawter L, Tong X, Gemilyan M, et al. Barriers to antihypertensive medica-
tion adherence among adults—United States, 2005. J Clin Hypertens
8. Wang PS, Bohn RL, Knight E, et al. Noncompliance with antihypertensive
medications: The impact of depressive symptoms and psychosocial factors.
J Gen Intern Med 2002;17:504–511.
9. Andersen R, Newman JF. Societal and individual determinants of medical
care utilization in the United States. Milbank Mem Fund Q Health Soc
GENTIL ET AL.
DECEMBER 2012–VOL. 60, NO. 12 JAGS
10. Institut de la statistique du Que ´bec. Lexique du syste `me du code ge ´ographi- Download full-text
que du Que ´bec. 2007. Available at http://www.stat.gouv.qc.ca/publica-
tions/referenc/pdf2008/lexique_sept2007.pdf Accessed December 13, 2010.
11. Crum RM, Anthony JC, Bassett SS, et al. Population-based norms for the
Mini-Mental State Examination by age and educational level. JAMA
12. Kafonek S, Ettinger WH, Roca R, et al. Instruments for screening for
depression and dementia in a long-term care facility. J Am Geriatr Soc
13. 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.
14. Andrade SE, Kahler KH, Frech F, et al. Methods for evaluation of medica-
tion adherence and persistence using automated databases. Pharmacoepi-
demiol Drug Saf 2006;15:565–574; discussion 575–577.
15. Steiner JF, Prochazka AV. The assessment of refill compliance using pharmacy
records: Methods, validity, and applications. J Clin Epidemiol 1997;50:105–116.
16. Van Buuren S, van Rijckevorsel JLA. Imputation of missing categorical
data by maximizing internal consistency. Psychometrika 1992;57:567–580.
17. Diagnostic and Statistical Manual of Mental Disorders, 4th Ed. Washing-
ton, DC: American Psychiatric Association, 1994.
18. Bucholz KK, Marion SL, Shayka JJ, et al. A short computer interview for
obtaining psychiatric diagnoses. Psychiatr Serv 1996;47:293–297.
19. Murphy JM, Monson RR, Laird NM, et al. A comparison of diagnostic
interviews for depression in the Stirling County Study: Challenges for
psychiatric epidemiology. Arch Gen Psychiatry 2000;57:230–236.
20. Robins LN, Helzer JE, Croughan J, et al. National Institute of Mental
Health Diagnostic Interview Schedule. Its history, characteristics, and valid-
ity. Arch Gen Psychiatry 1981;38:381–389.
21. Preville M, Boyer R, Grenier S, et al. The epidemiology of psychiatric disor-
ders in Quebec’s older adult population. Can J Psychiatry 2008;53:822–832.
22. Charlson ME, Pompei P, Ales KL, et al. A new method of classifying
prognostic comorbidity in longitudinal studies: Development and validation.
J Chronic Dis 1987;40:373–383.
23. Krousel-Wood M, Islam T, Muntner P, et al. Association of depression
with antihypertensive medication adherence in older adults: Cross-sectional
and longitudinal findings from CoSMO. Ann Behav Med 2010;40:
24. Vrijens B, Vincze G, Kristanto P, et al. Adherence to prescribed antihyper-
tensive drug treatments: Longitudinal study of electronically compiled
dosing histories. BMJ 2008;336:1114–1117.
25. Hyre AD, Krousel-Wood MA, Muntner P, et al. Prevalence and predictors
of poor antihypertensive medication adherence in an urban health clinic
setting. J Clin Hypertens 2007;9:179–186.
26. Schroeder K, Fahey T, Ebrahim S. How can we improve adherence to
blood pressure—lowering medication in ambulatory care? Arch Intern Med
27. Safran DG, Neuman P, Schoen C et al. Prescription drug coverage and
seniors: Findings from a 2003 national survey. Health Aff (Millwood)
2005; Suppl Web Exclusives:W5-152–W5-166.
28. Siegel D, Lopez J, Meier J. Antihypertensive medication adherence in the
Department of Veterans Affairs. Am J Med 2007;120:26–32.
29. Nau DP, Aikens JE, Pacholski AM. Effects of gender and depression on
oral medication adherence in persons with type 2 diabetes mellitus. Gend
30. Tamblyn R, Lavoie G, Petrella L, et al. The use of prescription claims data-
bases in pharmacoepidemiological research: The accuracy and comprehen-
siveness of the prescription claims database in Quebec. J Clin Epidemiol
JAGS DECEMBER 2012–VOL. 60, NO. 12
MEDICATION ADHERENCE IN OLDER ADULTS