Abstract
Depression and anxiety are factors associated with poor adherence to medications that lead to increased healthcare costs. The authors hypothesize that these conditions will moderate the association between adherence and healthcare costs. The aim was to examine the healthcare costs associated with adherence to antihypertensive agents in the elderly with and without depression and anxiety. The sample included participants with hypertension and used hypertensive agents (N=926). Medication possession ratio was used to calculate medication adherence. Mean total healthcare costs included costs for inpatient stays, emergency department visits, outpatient visits, physician fees, and outpatient medications. Mental disorders were assessed using a questionnaire based on Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition criteria. The total healthcare costs were significantly greater for nonadherent participants with depression/anxiety than for adherent participants without depression/anxiety (Δ$1841, P<.0001). This study suggests that treating mental disorders in elderly patients with hypertension will decrease total healthcare costs.
Hypertension is the most common cardiovascular disease (CVD) in the world and affects 65% of Canadian seniors.1 Hypertension in older adults increases the risk for CVD such as coronary events, stroke, heart failure, and peripheral arterial disease caused by vascular damage.2 Studies have shown that CVD is the second leading cause of death in Canada.3
Nonadherence to pharmacotherapy is a major public health concern that has been called the “invisible epidemic,” especially in individuals with chronic conditions.4 As Esposti and colleagues5 reported, poor adherence to antihypertensive agents jeopardizes blood pressure control, which can lead to cardiovascular morbidity, hospitalization, and mortality. Factors associated with nonadherence to antihypertensive agents in the literature have included age, social support, income, and clinical factors such as depression.6, 7
A recent meta‐analysis reported that the prevalence of depression is close to 27% in individuals with hypertension.8 Furthermore, the prevalence of hypertension in individuals with anxiety is higher than in the general population.9 Previous studies have reported a significant association between depression, anxiety, and poor adherence to antihypertensive medications,10 leading to more complications, longer inpatient stays, increased morbidity, and higher mortality.11 A recent study by Pesa and colleagues12 reported that an increase in the proportion of days covered by antihypertensive medication was associated with a decrease in both total healthcare costs and hypertension‐related inpatient stays, outpatient and emergency department visits, and pharmacy costs.
To improve the health and quality of life of people with hypertension and reduce healthcare system costs, it is crucial to further explore the relationship between adherence to antihypertensive medications and common mental disorders such as depression and anxiety. Presently, there is a lack of literature concerning the impact on healthcare costs of the interaction between adherence to antihypertensive agents and the presence of depression and anxiety. We hypothesize that the presence of depression and anxiety will moderate the association between healthcare costs and adherence to antihypertensive agents.
In this population‐based study, we used linked health survey and administrative data on health services use, which increases the validity of the results by allowing for the control of confounders and limiting recall bias regarding healthcare service and drug use. From a healthcare system perspective, we examined healthcare costs associated with adherence to antihypertensive agents in community‐living older adults with and without depression and anxiety in a publicly managed healthcare system.
Methods
Data for this study were obtained from the longitudinal Quebec Survey on Seniors’ Health (Enquête sur la Santé des Ainés [ESA]) conducted in 2005–2008 using a probabilistic sample (n=2811) of French‐speaking community‐dwelling older adults aged 65 years and older. In Quebec, 94% of the population speaks French. Potential participants living in northern regions were not recruited because of the high cost of field work. In 2005, this population compromised 10% of the elderly population in Quebec. The probabilistic sampling method was stratified by three geographical areas based on population density criteria: (1) metropolitan (population ≥100,000), (2) urban (1000–99,999), and (3) rural (<1000), according to the definitions of the Quebec Institute of Statistics. In each geographic 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%. A preliminary study showed that there was no significant difference between participants and nonparticipants for age, sex, and region.
Data Collection
The research procedure was previously reviewed and authorized by the ethics committee of the Sherbrooke Geriatric University Institute. The data were collected for health professional in‐home interviews, which lasted 90 minutes on average. Volunteers were offered $15 compensation for their participation. Written consent to conduct the interview was obtained at the beginning of the interview from all participants. Participants presenting with severe or moderate cognitive problems based on the Mini‐Mental State Examination (score <22) were excluded from answering the questionnaire (n=27).13, 14, 15 Thereafter, individuals presenting no moderate or severe cognitive problems were invited to respond to the ESA‐Qs (n=2784). At the end of the interview, respondents were asked to provide written informed consent for our research team to access their health and pharmaceutical services data from the Régie d'Assurance‐Maladie du Québec (RAMQ) (agency responsible for Quebec's health insurance plan). People who turn 65 are automatically registered for the public plan, administered by the RAMQ, and opt out if they have private drug insurance. Participants with a private drug insurance plan were excluded from this study (n=208). Self‐reported data from the ESA survey were linked to individual level information from the RAMQ's medical and pharmaceutical services databases and from the health ministry's MED‐ECHO database on hospitalizations using the respondents’ health insurance number or, if that number were missing, using the name, sex, address, and month and year of birth of the respondent.
The RAMQ databases contain all claims for medical services rendered and physician fees paid by the Quebec Medicare system. The RAMQ pharmaceutical services database contains the Drug Identification Number of each pharmaceutical, the quantity, dose, date the prescription was filled, length of treatment, and cost including the dispensing fee. The MED‐ECHO database provides information on inpatient stays such as diagnoses, hospital length of stay, and services provided. A success rate of 99.6% (n=2494) was obtained in matching the data, which made up our analytical sample.
The study sample comprised participants from the ESA study with a diagnosis of hypertension if they met the following case criteria as defined by the Canadian Chronic Disease Surveillance System (CCDSS)1: two or more physician claims within 2 years, or one inpatient hospital discharge report listing hypertension as a diagnosis with International Classification of Diseases, 9th Edition (ICD‐9), hypertension codes (ICD‐9 or ICD‐9‐CM: 401‐405) and taking antihypertensive agents registered in RAMQ or MedEcho databases. A total of 926 individuals met the study criteria and were included in the final analysis. The observation period for all participants was 1 year.
Measures
Independent Variables
The medication possession ratio (MPR) has been widely used and should be the first method considered by researchers in the definition of medication adherence.16 The MPR is the day's supply of medication dispensed during a specified follow‐up period divided by the number of days from the first dispensing to the end of the follow‐up period.17 The observation period for the MPR calculation was 1 year. The number of days a participant was hospitalized was subtracted from the denominator because medications dispensed in hospital are not included in the RAMQ database. The MPR (%) was calculated as (total day's supply of medication/number of days in the evaluation period) Χ 100. The MPR cutoff point used was 80%, which is often considered the “standard” for categorizing individuals as adherent (≥80%) or nonadherent (<80%).18 The discontinuation of antihypertensive agents was identified as the presence of a continuous gap of 30 or more days between an expected refill and the actual refill.
The antihypertensive agents identified in this study sample included the following classes: diuretics, β‐adrenergic–blocking agents, angiotensin‐converting enzyme inhibitors, angiotensin II receptor antagonists, α‐adrenergic–blocking agents, calcium channel blockers, nitrates and nitrites, central α agonists, direct vasodilators, and dihydropyridines. More than half of the respondents used more than one antihypertensive medication. Participants using two or more antihypertensive agents or who switched or added an antihypertensive agent were considered as being on one medication in the calculation (no duplication). Further, a preliminary analysis showed that the presence of mental disorders was not associated with the switching of antihypertensive medication (Fisher's test P>.05) or the receipt of more than two antihypertensive agents (chi‐square, P=.21).
Respondent mental health status was measured using a computer‐assisted questionnaire, the ESA Diagnostic Questionnaire (ESA‐Q), which is based on DSM‐IV criteria.19, 20, 21 The ESA‐Q is similar to the Diagnostic Interview Schedule (DIS) and Composite International Diagnostic Interview (CIDI), which demonstrated satisfactory reliability and validity.22 In the ESA, a DSM‐IV diagnosis over a 12‐month period was included for the following disorders: major depression, minor depression, mania, specific phobia, social phobia, agoraphobia, panic disorder, obsessive‐compulsive disorder, and generalized anxiety disorder. The complete definition of the disorders studied in the ESA survey has been previously reported.23 For the analysis, respondents were classified as having (1) at least one probable DSM‐IV depressive or anxiety disorder, or (2) no probable DSM‐IV depressive or anxiety disorder during the observation period.
In this study, the sociodemographic factors considered included age (65–74 years and ≥75 years), education (<10 years and ≥10 years), and marital status (married or living in a couple and single/separated/divorced/widowed).
Respondent physical health condition was measured using the Charlson Comorbidity Index,24 which is a measure predicting mortality by classifying and weighting comorbid conditions (comorbidities). This index was calculated using medical claims with ICD‐9‐CM codes for the 12‐month period prior to the interview date.
Dependent Variable
Total healthcare costs were defined as the sum of costs incurred for inpatient stays, ambulatory visits (outpatient clinic visits and emergency department visits), physician fees, and outpatient medications during the 1‐year observation period, matching the MPR calculation period. Health service use was identified from the RAMQ and MED‐ECHO databases. Physician fees are not included in any of the unit costs and are captured separately through the RAMQ medical services database. The calculation of unit costs was based on summary annual reports for the province of Quebec using a direct allocation method.25 Data were obtained from the Quebec Ministry of Health and Social Services annual budget and activity reports (AS‐471, AS‐478) submitted by each institution in the province of Quebec. Average provincial costs and activity levels (visits, per inpatient day present) for the 2009–2010 fiscal year were used. A hospitalization was valued on the basis of a cost per day and an ED visit as a cost per visit. Outpatient visits in non‐private offices, ie, in public institutions, were also valued at a cost per visit. A detailed and complete description of the calculation of unit and overall healthcare costs (inpatient stays, outpatient visits, ED visits, physician fees, and outpatient medication) was previously reported by Vasiliadis and colleagues.26
Statistical Analysis
Data were weighted to ensure that the true proportions of older adults in each geographical area were reflected in the analysis. Weights were determined based on: (1) the probability of selection of the administrative region in the geographical area (π(a)); (2) the conditional probability of selection of the household in the administrative region (π(b|a)); and (3) the conditional probability of selection of the subject in the household (π(c|ab)). The weight (w) attributed to each subject represented the inverse of its probability of selection (1/(π(abc))). The weighted sample included 2798 older adults living at home. The mean and median sampling design effect were .94 and .95, respectively.27, 28
A generalized linear model with an identify link and gamma function distribution was used to account for the non‐normal distribution of costs and to estimate the association between healthcare costs and the interaction between medication adherence and the presence of mental disorders.29 The analyses controlled for the following potential confounders: age, sex, education, marital status, and physical health condition. Chi‐square statistics were also employed to detect any significant difference in proportions. Four dummy variables were created and included in the model: (1) adherent participants without depression and anxiety, (2) nonadherent participants without depression and anxiety, (3) adherent participants with depression and anxiety, and (4) nonadherent participants with depression and anxiety (reference group). Statistical analyses were performed using SAS version 9.1 (SAS Institute, Inc, Cary, NC).
Results
A total of 926 individuals with hypertension who were taking antihypertensive medications were identified. The prevalence of adherence to antihypertensive agents amounted to 52.8%. Table 1 shows a comparison of participant characteristics by adherence to antihypertensive agents. The prevalence of depression and anxiety was similar in both groups (Table 1).
Table 1.
Participant Characteristics by Adherence Level to Antihypertensive Medication
| All Patients (N=926) | OR (95% CI) | ||||
|---|---|---|---|---|---|
| Nonadherent Participants, No. (%) | Adherent Participants, No. (%) | Χ 2 Test | P Value | ||
| No. (%) | 437 (47.2) | 489 (52.8) | |||
| Age | |||||
| 65–74 | 227 (51.9) | 252 (51.5) | 0.02 | 1 | 1 (0.79–1.32) |
| 75+ | 210 (48.1) | 237 (48.5) | |||
| Sex | |||||
| Male | 109 (24.9) | 122 (24.9) | 0 | .9 | 1 (0.74–1.35) |
| Female | 328 (75.1) | 367 (75.1) | |||
| Marital status | |||||
| Married | 183 (41.9) | 218 (44.6) | 0.69 | .4 | 1 (0.78–1.34) |
| Single/divorced/separated/widowed | 254 (58.1) | 271 (55.4) | |||
| Education | |||||
| <10 y | 122 (27. 9) | 115 (23. 5) | 2.23 | .13 | 1.07 (0.73–1.56) |
| ≥10 y | 315 (72.1) | 374 (76.5) | |||
| Probable depressive or anxiety disorder | |||||
| Yes | 58 (13.3) | 69 (14.1) | 0.14 | .7 | 1.25 (0.93–1.68) |
| No | 379 (86.7) | 420 (85.9) | |||
| Charlson Comorbidity Index | |||||
| 0 | 283 (64.8) | 284 (58.1) | |||
| 1–2 | 128 (29.3) | 157 (32.1) | 7.5 | .06 | 0.77 (0.33–1.79) |
| 3–5 | 16 (3.7) | 35 (7.2) | 0.94 (0.40–2.22) | ||
| ≥6 | 10 (2.2) | 13 (2.7) | 1.68 (0.61–4.64) | ||
Abbreviations: CI, confidence interval; OR, odds ratio.
The adjusted mean healthcare costs are presented in Table 2 and the absolute difference in healthcare costs is presented in Table 3. The results showed that the average total healthcare costs incurred were significantly higher for nonadherent participants with depression/anxiety than for adherent participants without depression/anxiety (Δ$1841, P<.0001). The findings also showed that among participants with depression and anxiety, nonadherence was associated with increased healthcare costs reaching on average of $1658 (P<.0001). Further, among nonadherent participants, the presence of depression and anxiety was associated with increased costs reaching $2868 (P<.001) on average. Among participants without depression and anxiety, adherence was associated with higher healthcare costs reaching on average $1028 (P<.001). Furthermore, the presence of depression/anxiety in adherent participants was associated with higher healthcare costs than in nonadherent patients without depression/anxiety (Δ$1210, P<.0001). Finally, the interaction between the presence of depression and anxiety and adherence on healthcare costs was significant (P=.04).
Table 2.
Adjusteda Mean Healthcare Costs By Adherence to Antihypertensive Agent and the Presence of Mental Disorders
| Nonadherent Patients With Depression/Anxiety (n=58) | Adherent Patients Without Depression/Anxiety (n=420) | Adherent Patients With Depression/Anxiety (n=69) | Nonadherent Patients Without Depression/Anxiety (n=379) | |||||
|---|---|---|---|---|---|---|---|---|
| Mean | 95% CI | Mean | 95% CI | Mean | 95% CI | Mean | 95% CI | |
| Ambulatory costs | $2209 | $2090–$2328 | $1220 | $1176–$1243 | $1225 | $1117–$1332 | $979 | $924–$1024 |
| Inpatient | $1544 | $1440–$1648 | $723 | $685–$761 | $639 | $545–$732 | $693 | $653–$773 |
| Medications | $1799 | $1608–$1990 | $1987 | $1917–$2057 | $2221 | $2050–$2394 | $1388 | $1314–$1462 |
| Physician fees | $702 | $666–$737 | $482 | $469–$495 | $510 | $478–$542 | $325 | $311–$339 |
| Total costs | $6253 | $5820–$6686 | $4412 | $4254–$4571 | $4595 | $4205–$4985 | $3385 | $3218–$3551 |
Abbreviation: CI, confidence interval.
Adjusted for sociodemographic and clinical factors.
Table 3.
Absolute Difference in Adjusted Healthcare Costs By Study Group (Adherence Mental Health Status)
| Ambulatory Costs (Δ, P, 95% CI) | Inpatient (Δ, P, 95% CI) | Physician Fees (Δ, P, 95% CI) | Medications (Δ, P, 95% CI) | Total Costs (Δ, P, 95% CI) | |
|---|---|---|---|---|---|
| |Δ| $ Nonadherent with depression/anxiety (n=58) vs adherent without depression/anxiety (n=420) | $989a ($862–$1116) | $820a ($710–$931) | $219a ($182–$257) | −$188 (−$391 to $15) | $1841a ($1380–$2302) |
| |Δ| $ Nonadherent with depression/anxiety (n=58) vs adherent with depression/anxiety (n=69) | $984a ($824–$1144) | $905a ($765–$1044) | $192a ($144–$240) | −$428a (−$680 to $‐166) | $1658a ($1075–$2241) |
| Nonadherent with depression/anxiety (n=58) vs Δ|$ nonadherent without depression/anxiety (n=379) | $1230a ($1103–$1358) | $850a (739–962) | $377a (339–415) | $410a ($206 to $615) | $2868a ($2404–$3332) |
| Adherent with depression/anxiety (n=69) vs |Δ| $ nonadherent without depression/anxiety (n=379) | $246a ($129–$363) | −$54 (−$47 to $156) | $184a ($ 150–$219) | $834a ($647–$1021) | $1210a ($786–$1634) |
| |Δ| $ Adherent without depression/anxiety (n=420) vs nonadherent without depression/anxiety (n=379) | $241a ($178 to 304) | $30 ($25 to $85) | $157a ($138 to $176) | $599a ($ 497 to $700) | $1028a ($797 to $1258) |
| Δ| $ Adherent without depression/anxiety (n=420) vs adherent with depression/anxiety (n=69) | −$5 (−$120 to $111) | $84 (−$17 to $185) | −$27 (−$62 to $7) | −$234b (−$420 to −$49) | −$182 (−$603 to $238) |
Abbreviation: CI, confidence interval.
P<.0001.
P<.01.
Discussion
To our knowledge, this is the first study to explore the impact of the association between mental disorders and poor adherence to antihypertensive medications on overall healthcare costs from the healthcare system perspective in Canada, using linked health survey and administrative databases in an older adult population. The prevalence of adherence was 52.8%, while a previous study showed adherence rates for antihypertensive agents varying from 29% to 91% using different methods to evaluate adherence to antihypertensive agents.10 In contrast with other studies, the prevalence of depression and anxiety was similar in adherent and nonadherent participants. Our hypothesis is that 1 year is not long enough to observe the effect of depression and anxiety on medication adherence. A previous study based on a longer follow‐up of this sample showed that the presence of depression/anxiety was associated with a decrease in medication adherence over a 2‐year follow‐up.10 Furthermore, although not reaching statistical significance, adherence individuals were more likely to have a higher level of comorbidities. Our results concord with those presented by Cumming and colleagues,30 which show that individuals perceiving their health status to be poorer are more likely to be adherent to hypertensive agents.
Overall, our findings revealed that the presence of depression and anxiety influenced the association observed between healthcare costs and adherence. Previous studies have shown that depression and anxiety in the elderly are associated with healthcare costs26 and adherence throughout time.10
The results also showed that among individuals with depression and anxiety, nonadherence was associated with higher healthcare costs. The lower medication costs observed were offset by increased inpatient and outpatient costs and physician fees paid out. These results are consistent with others reporting on increased healthcare costs associated with nonadherence to oral antihyperglycemics in diabetic older adults with common mental disorders.31 Previous studies have also shown that low adherence is associated with higher risk of vascular events and hospitalization and consequently increase healthcare costs in individuals with hypertension.11
Among individuals without depression/anxiety, however, adherence was significantly associated with higher healthcare costs. When looking more closely at the data, the main cost drivers in these adherent individuals were costs associated with medication use (58.3% [$599/$1028]) followed by outpatient visits (23.4% [$241/$1028]). There were no increased costs associated with inpatient stays, suggesting that costs in adherent individuals without depression and anxiety are mainly associated with physician follow‐up and prescriptions and not complications leading to hospitalizations.
Among adherent individuals, the presence of depression and anxiety did not lead to increased healthcare costs. Additional analyses showed that 18% of those without depression/anxiety and 19% of those with depression/anxiety were also using antidepressants, which may, in part, explain the results. This suggests that this subgroup of the sample may be more alike than not. Further, antidepressant use, may, to some extent, have improved the symptoms of depression and anxiety and consequently the health outcomes.
These results showed that depression and anxiety are significant independent predictors of healthcare costs in older adults with hypertension after controlling for age, sex, education level, marital status, and physical health condition. Previous studies found that comorbid depression and anxiety symptoms were predictors of cardiovascular events,32 and cardiovascular events increase healthcare costs for the health system. In the management of hypertension it was also been shown that psychotherapy and the use of antidepressants combined with antihypertensive drugs helped prevent severe attacks of hypertension in elderly patients.33
Study Limitations
This study has several limitations. First, medication adherence was measured through pharmacy claims data on medications dispensed. This method does not measure the actual consumption of medications. This could lead to an overestimation of adherence; however, the MPR has been recommended for evaluating adherence with hypertension therapy.11 Second, the RAMQ databases cannot be used to identify patients’ severity level since clinical information is not provided. Third, it is possible that some people are being dispensed an antihypertensive agent for an illness other than hypertension with different associated prescribing practices. Fourth, the presence of self‐reported depression and anxiety can be subject to a social desirability bias. We assume, however, this not to be differential among subgroups. Fifth, the severity, duration of the symptoms, diagnoses from the administrative data, and treatment for depression and anxiety was not considered in the analyses, which may have had a greater impact on adherence and health system costs. Sixth, in this study we do not report healthcare costs associated specifically with CVDs. Seventh, the majority of respondents used more than one antihypertensive medication. Although this was not considered in the analyses, we did control for the presence of other comorbidities and severity. Eighth, given the sample size, it was not possible to report healthcare costs for mental health vs other reasons such as cardiovascular complications. Administrative data are not reliable for coding mental health–related visits; the majority are for chronic physical conditions. Finally, the healthcare cost data may not be generalizable to systems that differ significantly from the Canadian healthcare system.
Conclusions
Inadequate medication adherence and depression/anxiety significantly increases healthcare costs. Comorbid hypertension and depression/anxiety are prevalent, and it has been demonstrated that depressive symptoms can interfere with ideal control of blood pressure. It is important to manage depression and anxiety during the process of treating hypertension. Depression and anxiety treatment and follow‐up services for individuals with hypertension would likely improve blood pressure control, which may, in turn, reduce the impact on the use of healthcare system resources.
Author Contributions
G.L. contributed to the design, data analyses, data interpretation, and manuscript preparation. V.H.M. contributed to the concept and design and manuscript preparation. P.M. contributed to manuscript preparation and acquisition of participants and data. B.D. contributed to data analyses.
Conflict of Interest
None.
Sponsor's Role
None.
Acknowledgments
This study was supported by a Canadian Institutes of Health Research (CIHR) operating grant (200683MOP). The ESA study was supported by a CIHR operating grant (200403MOP) and the Quebec Health Research Fund (Fonds de Recherche en santé du Québec [FRSQ]) (ref. 9854). G. L. read and approved the final manuscript.
J Clin Hypertens (Greenwich). 2017;19:75–81. DOI: 10.1111/jch.12869. © 2016 Wiley Periodicals, Inc.
References
- 1. Public Health Agency of Canada . Report from the Canadian Chronic Disease Surveillance System: Hypertension in Canada, 2010. Ottawa, ON: Public Health Agency of Canada; [online]; 2010. http://www.phac-aspc.gc.ca/cd-mc/cvd-mcv/ccdss-snsmc-2010/pdf/. Accessed July 8, 2010. [Google Scholar]
- 2. Aronow WS, Fleg JL, Pepine CJ, et al. ACCF/AHA 2011 expert consensus document on hypertension in the elderly: a report of the American College of Cardiology Foundation Task Force on Clinical Expert Consensus Documents, developed in collaboration with the American Academy of Neurology, American Geriatrics Society, American Society for Preventive Cardiology, American Society of Hypertension, American Society of Nephrology, Association of Black Cardiologists, and European Society of Hypertension. J Am Coll Cardiol. 2011;57:2037–2114. [DOI] [PubMed] [Google Scholar]
- 3. Statistics Canada . The 10 leading causes of death, 2011. Health Fact Sheets. http://www.statcan.gc.ca/pub/82-625-x/2014001/article/11896-eng.htm. Accessed April 10, 2015. Updated 2014.
- 4. World Health Organization . Adherence to Long‐Term Therapies: Evidence for Action. Geneva, Switzerland: World Health Organization; 2003. www.who.int/chronic-conditions/en/adherence-report.pdf. Accessed December 14, 2014. [Google Scholar]
- 5. Esposti LD, Saragoni S, Batacchi P, et al. Antihypertensive therapy among newly treated patients: an analysis of adherence and cost of treatment over years. Clinicoecon Outcomes Res. 2010;2:113–120. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. 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. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Morris AB, Li J, Kroenke K, et al. Factors associated with drug adherence and blood pressure control in patients with hypertension. Pharmacotherapy. 2006;26:483–492. [DOI] [PubMed] [Google Scholar]
- 8. Li Z, Li Y, Chen L, et al. Prevalence of depression in patients with hypertension: a systematic review and meta‐analysis. Medicine (Baltimore). 2015;94:1317. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Player MS, Peterson LE. Anxiety disorders, hypertension, and cardiovascular risk: a review. Int J Psychiatry Med. 2011;41:365–377. [DOI] [PubMed] [Google Scholar]
- 10. Gentil L, Vasiliadis HM, Préville M, et al. Association between depressive and anxiety disorders and adherence to antihypertensive medication in community‐living elderly adults. J Am Geriatr Soc. 2012;60:2297–2301. [DOI] [PubMed] [Google Scholar]
- 11. Dragomir A, Côté R, Roy L, et al. Impact of adherence to antihypertensive agents on clinical outcomes and hospitalization costs. Med Care. 2010;48:418–425. [DOI] [PubMed] [Google Scholar]
- 12. Pesa JA, Van Den Bos J, Gray T, et al. An evaluation of the impact of patient cost sharing for antihypertensive medications on adherence, medication and health care utilization, and expenditures. Patient Prefer Adherence. 2012;6:63–72. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. 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. 1989;37:29–34. [DOI] [PubMed] [Google Scholar]
- 14. Crum RM, Anthony JC, Bassett SS, Folstein MF. Population‐based norms for the Mini‐Mental State Examination by age and educational level. JAMA. 1993;269:2386–2391. [PubMed] [Google Scholar]
- 15. 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. [DOI] [PubMed] [Google Scholar]
- 16. Karve S, Cleves MA, Helm M, et al. An empirical basis for standardizing adherence measures derived from administrative claims data among diabetic patients. Med Care. 2008;46:1125–1133. [DOI] [PubMed] [Google Scholar]
- 17. Steiner JF, Prochazka AV. The assessment of refill compliance using pharmacy records: methods, validity, and applications. J Clin Epidemiol. 1997;50:105–116. [DOI] [PubMed] [Google Scholar]
- 18. Karve S, Cleves MA, Helm M, et al. Good and poor adherence: optimal cut‐point for adherence measures using administrative claims data. Curr Med Res Opin. 2009;25:2303–2310. [DOI] [PubMed] [Google Scholar]
- 19. Wittchen HU, Robins LN, Cottler LB, et al. Cross‐cultural feasibility, reliability and sources of variance of the Composite International Diagnostic Interview (CIDI). The Multicentre WHO/ADAMHA Field Trials. Br J Psychiatry. 1991;159:658. [DOI] [PubMed] [Google Scholar]
- 20. Erdman HP, Klein MH, Greist JH, et al. A comparison of two computer‐administered versions of the NIMH diagnostic interview schedule. J Psychiatr Res. 1992;26:85–95. [DOI] [PubMed] [Google Scholar]
- 21. American Psychiatric Association . Diagnostic and Statistical Manual of Mental Disorders, 4th ed. Washington, DC: American Psychiatric Association; 1994. [Google Scholar]
- 22. Bucholz KK, Marion SL, Shayka JJ, et al. A short computer interview for obtaining psychiatric diagnoses. Psychiatr Serv. 1996;47:293–297. [DOI] [PubMed] [Google Scholar]
- 23. Préville M, Boyer R, Grenier S, et al. The epidemiology of psychiatric disorders in Quebec's older adult population. Can J Psychiatry. 2008;53:822–832. [DOI] [PubMed] [Google Scholar]
- 24. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40:373–383. [DOI] [PubMed] [Google Scholar]
- 25. Drummond MF. Methods for the Economic Evaluation of Health Care Programmes, 3rd edn. Toronto, ON: Oxford University Press; 2005. [Google Scholar]
- 26. Vasiliadis HM, Dionne PA, Preville M, et al. The excess healthcare costs associated with depression and anxiety in elderly living in the community. Am J Geriatr Psychiatry. 2013;21:536–548. [DOI] [PubMed] [Google Scholar]
- 27. Cox BG, Cohen SB. Methodological Issues for Health Care Surveys. Statistics, Textbooks and Monographs. New York, NY: M. Dekker; 1985. [Google Scholar]
- 28. Trudel G, Courtemanche R, Tarte F. Validation, Coding, Entry and Weighting of Data for the Québec Health Survey, Technical Manual 87‐07, Québec, Canada: Ministère de la Santé et des Services sociaux. Gouvernement du Québec [Quebec Ministry of Health and Social Services]; 1992. [Google Scholar]
- 29. McCullagh P, Nelder J. Generalized Linear Models. New York, NY: Chapman & Hall; 1983. [Google Scholar]
- 30. Cummings KM, Kirscht JP, Binder LR, Godley AJ. Determinants of drug treatment maintenance among hypertensive persons in inner city Detroit. Public Health Rep. 1982;97:99–106. [PMC free article] [PubMed] [Google Scholar]
- 31. Gentil L, Vasiliadis HM, Préville M, Berbiche D. Adherence to oral antihyperglycemic agents among older adults with mental disorders and its effect on health care costs, Quebec, Canada, 2005‐2008. Prev Chronic Dis. 2015;12:E231. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Strik JJ, Denollet J, Lousberg R, Honig A. Comparing symptoms of depression and anxiety as predictors of cardiac events and increased health care consumption after myocardial infarction. J Am Coll Cardiol. 2003;42:1801–1807. [DOI] [PubMed] [Google Scholar]
- 33. Simon JD, Peter RJ, John P, David JN. Treatment of anxiety and depressive disorders in patients with cardiovascular disease. BMJ. 2004;328:939–943. [DOI] [PMC free article] [PubMed] [Google Scholar]