Antihypertensive Drug Persistence and Compliance Among Newly Treated Elderly Hypertensives in Ontario
Article Outline
- Abstract
- Materials and Methods
- Results
- Discussion
- Study Limitations
- Conclusions
- Acknowledgment
- Supplementary data
- References
- Copyright
Abstract
Background
Poor medication-taking behaviors are important considerations in the management of hypertension.
Methods
We conducted a retrospective cohort study addressing antihypertensive drug persistence and compliance by linking 4 administrative databases and a province-wide clinical database in Ontario, Canada, to derive a cohort of elderly hypertensive patients, aged 66 years or more, who had received a new prescription for an antihypertensive agent between 1997 and 2005 to determine trends across years and associations with drug class and sociodemographic and other factors.
Results
Our cohort consisted of 207,473 patients (58.4% were women, mean age 74.2 years, 73.1% were comorbid-free), 41,236 of whom had diabetes. Persistence and compliance increased between 1997 and 2005 (all P
<.02) and were greater in those of higher socioeconomic status but lesser in urban residents (all P<.0001). Persistence was lower in comorbid-free patients and greater in older patients, whereas compliance was lower in older patients and greater in women and comorbid-free patients (all P<.0001). Significant differences between the drug classes emerged with initial prescriptions for all drug classes showing greater therapy and class persistence compared with diuretics (all P<.0001). Angiotensin-converting enzyme inhibitors showed the best therapy persistence and compliance, and beta-blockers showed the worst compliance (all P<.0001).
Conclusion
Our data provide evidence of an overall improvement in antihypertensive drug compliance and persistence across years, as well as significant differences across drug classes and other patient-level factors. Awareness of such factors could translate into concerted efforts at optimizing medication-taking behaviors among newly diagnosed elderly hypertensive patients.
Keywords: Antihypertensive drug, Compliance, Hypertension, Persistence
Poor medication-taking behaviors are particularly relevant for the management of chronic asymptomatic diseases, such as hypertension, in which no immediate physical symptoms result from such behaviors.1 In the long-term, however, inadequate blood pressure control significantly increases the risks of target organ damage and death. Not surprisingly, economic analyses reveal considerable excess costs (up to one third of the total costs of the treatment of incident hypertensives) resulting from switching and discontinuation of initial therapy.2 The basis for poor medication-taking behavior is likely multifactorial, stemming from factors at several levels.1, 3
Identification of factors that are associated with poor compliance/persistence could lead to concerted and focused efforts at optimizing medication-taking behaviors. In addition, differences between drug classes in compliance/persistence could result in more substantial differences in real-world effectiveness of medications than would be expected to arise simply on the basis of the modest differences between drug classes in specific end points evident in randomized clinical trials. Indeed, it is well recognized that randomized trials underestimate side effects and problems with poor compliance/persistence, particularly those trials that use run-in phases to optimize likelihood of compliance and at least short-term tolerability.4
Because a recently published systematic review of studies evaluating compliance/persistence revealed significant heterogeneity and methodological flaws among these studies, we designed this study to address these limitations.5 We conducted a retrospective cohort study using administrative databases capturing all prescriptions and health service use by the entire elderly population of a Canadian province with a universal single-payer health care system. As a result, our study is not subject to the selection or measurement biases noted in the systematic review of prior studies in this field.5 We used these databases to identify an inception cohort of newly treated elderly Ontarians to determine whether there has been a change in compliance/persistence over the last decade and whether these parameters are associated with drug class, age, gender, comorbidity burden, socioeconomic status, and urban versus rural residence.
Materials and Methods
Study Population and Data Sources
We used the Ontario Drug Benefit prescription drugs database to identify all Ontario residents aged more than 66 years who had received a new prescription for an antihypertensive agent in the 5 main classes of antihypertensives (Appendix-Supplementary Table 1) between January 1, 1997, and December 31, 2005. To avoid the pitfalls associated with defining patients as medication users on the basis of only 1 prescription, we followed recently published recommendations to define our inception cohort as users of a particular medication on the basis of at least 2 prescriptions within 100 days of completion of the first prescription so as to not erroneously include patients who did not take their initial antihypertensive medication.3 The Ontario Drug Benefit database records all drugs prescribed from a minimally restricted formulary for patients aged more than 65 years in Ontario.
We prespecified a 1-year washout period before entry into the inception cohort to ensure that our cohort comprised newly treated patients and thus excluded patients with a first claim date within 1 year of the beginning of the study period or within 1 year of their entry into the Ontario Drug Benefit. We linked this cohort to the Ontario Health Insurance Plan physician claims database, the Canadian Institute for Health Information hospitalization database, and the Registered Persons Database using a unique encrypted Ontario health card number that preserved the exact identification of individuals but allowed for the linkage of individuals across the 3 administrative databases. The Ontario Health Insurance Plan database records all fee-for-service billings for physician services rendered in Ontario, including the most responsible diagnosis at each visit. The Canadian Institute for Health Information database records the primary responsible and up to 15 secondary diagnoses for all discharges from acute care hospitals. Studies on the validity of these administrative databases have confirmed their high degree of accuracy and comprehensiveness.6 Our study received ethics approval from the institutional review board at Sunnybrook Health Sciences Centre, Toronto, Ontario.
We created a cohort of incident hypertensives by excluding any patients who had another condition for which an antihypertensive might be prescribed. International Classification of Diseases 9th Revision codes were used for Ontario Health Insurance Plan and Canadian Institute for Health Information before 2002, and International Classification of Diseases 10th Revision codes were used for Canadian Institute for Health Information after 2002 (Appendix-Supplementary Table 2 for International Classification of Diseases codes). Patients were excluded with Ontario Health Insurance Plan or Canadian Institute for Health Information claims within 4 years (or “marker” medications prescribed in the Ontario Drug Benefit within 1 year before the initial antihypertensive prescription date) for these other conditions (Appendix-Supplementary Table 2) before their initiation of antihypertensive treatment. We also stratified our analyses to examine patients with and without diabetes mellitus using presence or absence in the Ontario Diabetes Database7 in the 1 year before the initial antihypertensive prescription.
Data Synthesis
We followed prescriptions filled by our cohort of newly treated hypertensives for 2 years after their initial antihypertensive prescription and examined persistence, referring to continuation of treatment with that drug class (“class persistence”) or with any antihypertensive agent (“therapy persistence”) within 60 days of the end date of the last prescription and compliance, referring to not missing medication doses in the context of ongoing use.3 As such, a switch to a different class of antihypertensive drug would be defined as loss of “class persistence” but continuation of “therapy persistence.” Drug classes were classified as follows: angiotensin-converting enzyme inhibitor, angiotensin II receptor blocker, beta-blocker, calcium channel blocker, thiazide and thiazide-like diuretics, and combination agents. Agents that consisted of a potassium-sparing diuretic plus a thiazide diuretic were counted as 1 prescription in the diuretic class (22.4% of diuretic cohort). Patients with a prescription for 2 concurrent separate agents accounted for 4.9% of the entire cohort and thus were excluded; similarly, patients with a prescription for a combination agent accounted for 2.7% of the entire cohort and also were excluded. Sensitivity analyses were conducted using either 30- or 100-day time windows for refill prescriptions. Compliance was determined using the medication possession ratio among patients persistent with antihypertensive drug therapy. Medication possession ratio similar to the proportion of days covered was calculated using the “days supplied” variable in the Ontario Drug Benefit database and the ratio of total days in which antihypertensive medication was supplied to the total days in the follow-up period, capped at 1.0 and defined as high if more than 0.8.3
Statistical Analysis
Multivariable logistic regression models were used to test specific hypotheses regarding each of the following outcome variables over the first 2 years after initial antihypertensive prescription: therapy persistence, class persistence, and high medication possession ratio. Predictor variables in each of the analyses included age, gender, cohort year, drug class, comorbidity burden (as reflected by the Charlson comorbidity index score [0, 1 or 2+]8), socioeconomic status (using neighborhood income quintiles derived from census data by postal code9, 10), and residence location (urban vs rural using Statistics Canada definitions11). A 2-tailed P value of less than .05 was considered significant. All analyses were conducted using SAS Version 9.1.3 (SAS Institute, Cary, NC).
Results
Cohort Characteristics
We identified 207,473 patients (58.4% were women, mean age 74.2 years, 73.1% were comorbid-free) for our inception cohort, 41,236 of whom had diabetes. The characteristics of this cohort, both overall and stratified by diabetic status or initial antihypertensive drug class, are summarized in Table 1, Table 2. The majority of patients were urban residents initially prescribed either an angiotensin-converting enzyme inhibitor or a diuretic. Approximately two thirds of patients were therapy persistent over the first 2 years of their treatment, and more than 90% demonstrated high compliance over the first 2 years after their initial antihypertensive prescription. However, whereas approximately two thirds of patients were class persistent over the first year, this decreased to just more than 50% by the second year (Table 1). Figure 1, Figure 2 graphically display the proportions of patients who were class persistent and compliant according to diabetic status and initial antihypertensive drug class over the first 2 years of their treatment. Hypertensive individuals with diabetes were more likely to be prescribed an angiotensin-converting enzyme inhibitor than nondiabetic hypertensive individuals and less likely to be prescribed a beta-blocker, calcium channel blocker, or diuretic (all P<.0001); moreover, patients with diabetes were more likely to persist with therapy or initial drug class but less likely to be compliant (all P<.05).
Table 1. Cohort Characteristics
| Factor | No. (%) | ||
|---|---|---|---|
| Overall Cohort | Diabetic Cohort | Nondiabetic Cohort | |
| N=207,473 | n=41,236 | n=166,237 | |
| Drug class | |||
| ACEI | 95,773(46.2) | 31,414(76.2) | 64,359(38.7) |
| ARB | 9452(4.6) | 2041(4.9) | 7411(4.4) |
| BB | 21,973(10.6) | 1935(4.7) | 20,038(12.0) |
| CCB | 23,603(11.4) | 2351(5.7) | 21,252(12.8) |
| Diuretic | 56,672(27.3) | 3495(8.5) | 53,177(32.0) |
| Gender | |||
| Female | 121,165(58.4) | 19,948(48.4) | 101,217(60.9) |
| Residence locationa | |||
| Urban | 178,048(85.8) | 35,385(85.8) | 142,663(85.8) |
| Income quintilea | |||
| 1(lowest) | 40,778(19.6) | 9149(22.2) | 31,629(19.0) |
| 2 | 44,208(21.3) | 9172(22.2) | 35,036(21.1) |
| 3 | 41,975(20.2) | 8361(20.3) | 33,614(20.2) |
| 4 | 39,384(19.0) | 7524(18.2) | 31,860(19.2) |
| 5(highest) | 40,258(19.4) | 6846(16.6) | 33,412(20.1) |
| Age, y | |||
| 66-70 | 83,886(40.4) | 18,113(43.9) | 65,773(39.6) |
| 71-75 | 58,957(28.4) | 12,139(29.4) | 46,818(28.2) |
| 76-80 | 37,974(18.3) | 6877(16.7) | 31,097(18.7) |
| 81-85 | 18,893(9.1) | 2945(7.1) | 15,948(9.6) |
| 85+ | 9763(4.7) | 1162(2.8) | 8601(5.2) |
| Cohort year | |||
| 1997 | 18,491(8.9) | 2936(7.1) | 15,555(9.4) |
| 1998 | 18,749(9.0) | 3155(7.6) | 15,594(9.4) |
| 1999 | 21,356(10.3) | 4070(9.9 | 17,286(10.4) |
| 2000 | 25,317(12.2) | 5383(13.0) | 19,934(12.0) |
| 2001 | 24,188(11.7) | 5168(12.5) | 19,020(11.4) |
| 2002 | 25,452(12.3) | 5458(13.2) | 19,994(12.0) |
| 2003 | 25,226(12.2) | 5005(12.1) | 20,221(12.2) |
| 2004 | 25,600(12.3) | 5324(12.9) | 20,276(12.2) |
| 2005 | 23,094(11.1) | 4737(11.5) | 18,357(11.0) |
| Charlson comorbidity index score | |||
| 0 | 15,1759(73.1) | 0(0) | 151,759(91.3) |
| 1 | 43,694(21.1) | 37,078(90.0) | 6616(4.0) |
| 2+ | 12,020(5.8) | 4158(10.1) | 7862(4.9) |
| Therapy persistence | |||
| Over first year | 152,057(73.3) | 31,072(74.2) | 120,985(72.8) |
| Over first 2 y | 137,124(66.1) | 27,770(67.3) | 109,354(65.8) |
| Class persistence | |||
| Over first year | 141,026(68.0) | 28,606(69.4) | 112,420(67.6) |
| Over first 2 y | 113,750(54.8) | 23,529(57.1) | 90,221(54.3) |
| MPR≥0.8) | |||
| Over first year | 141,777(94.4) | 28,469(93.8) | 113,308(94.5) |
| Over first 2 y | 130,409(96.9) | 26,387(96.0) | 104,022(96.8) |
aResidence location and socioeconomic status are missing in 0.2% and 0.4% of overall cohort, respectively. |
Table 2. Cohort Characteristics Stratified by Antihypertensive Drug Class (n=207,473)
| Factor | No. (%) | ||||
|---|---|---|---|---|---|
| ACEI n=95,773 | ARB n=9452 | BB n=21,973 | CCB n=23,603 | Diuretic n=56,672 | |
| Gender | |||||
| Female | 50,970(53.2) | 5390(57.0) | 12,826(58.4) | 13,944(59.1) | 39,786(70.2) |
| Residence locationa | |||||
| Urban | 81,926(85.5) | 8414(89.0) | 18,379(83.6) | 20,576(87.2) | 47,002(82.9) |
| Income quintilea | |||||
| 1(lowest) | 19,205(20.0) | 1677(17.7) | 4186(19.0) | 4837(20.5) | 10,873(19.2) |
| 2 | 20,670(21.6) | 1995(21.1) | 4573(20.8) | 5216(22.1) | 11,754(20.7) |
| 3 | 19,434(20.3) | 1895(20.0) | 4496(20.5) | 4932(20.9) | 11,218(19.8) |
| 4 | 18,015(18.8) | 1942(20.5) | 4213(19.2) | 4452(18.9) | 10,762(19.0) |
| 5(highest) | 18,034(18.8) | 1909(20.2) | 4404(20.0) | 4076(17.3) | 11,835(20.9) |
| Age, y | |||||
| 66-70 | 39,569(41.3) | 4481(47.4) | 9095(41.4) | 8958(37.9) | 19,783(34.9) |
| 71-75 | 27,632(28.8) | 2520(26.7) | 6290(28.6) | 6668(28.2) | 15,847(28.0) |
| 76-80 | 16,877(17.6) | 1514(16.0) | 3922(17.8) | 4435(18.8) | 11,226(19.8) |
| 81-85 | 7868(8.2) | 673(7.1) | 1815(8.3) | 2278(9.6) | 6259(11.0) |
| 85+ | 3827(4.0) | 264(2.8) | 851(3.9) | 1264(5.4) | 3557(6.3) |
| Charlson comorbidity index score | |||||
| 0 | 58,497(61.2) | 6826(72.2) | 18,463(84.0) | 19,571(82.9) | 48,994(86.4) |
| 1 | 30,846(32.2) | 2113(22.4) | 2230(10.1) | 2897(12.3) | 5016(8.8) |
| 2+ | 6430(6.7) | 513(5.4) | 1280(5.8) | 1135(4.8) | 2662(4.7) |
| Therapy persistence | |||||
| Over first year | 72,602(75.8) | 7066(74.8) | 15,568(70.8) | 16,995(72.0) | 39,826(70.3) |
| Over first 2 y | 66,185(69.1) | 6345(67.1) | 13,980(63.6) | 15,225(64.5) | 35,389(62.4) |
| Class persistence | |||||
| Over first year | 65,392(68.3) | 6655(70.4) | 14,919(67.9) | 16,373(69.4) | 37,687(66.5) |
| Over first 2 y | 54,109(56.5) | 5285(55.9) | 12,022(54.7) | 13,242(56.1) | 29,092(51.3) |
| MPR≥0.8 | |||||
| Over first year | 68,108(93.8) | 6614(93.6) | 14,281(91.7) | 15,853(93.3) | 37,086(93.1) |
| Over first 2 y | 63,355(95.7) | 6078(95.8) | 13,125(93.9) | 14,518(95.4) | 33,672(95.1) |
aResidence location and socioeconomic status are missing in 0.2% and 0.4% of overall cohort, respectively. |
Figure 1.
Class persistence (%) by antihypertensive drug class. ACEI
=angiotensin-converting inhibitor; ARB=angiotensin II receptor blocker; BB=beta-blocker; CCB=calcium channel blocker; DM=diabetic; Non-DM=nondiabetic.
Figure 2.
Medication possession ratio greater than 0.8 (%) by antihypertensive drug class. ACEI
=angiotensin-converting inhibitor; ARB=angiotensin II receptor blocker; BB=beta-blocker; CCB=calcium channel blocker; DM=diabetic; Non-DM = nondiabetic.
Multivariable Analyses
Urban residents and those with no or minimal comorbidity were less likely to persist with therapy or with the initial drug class; on the other hand, those of higher socioeconomic status, older patients (even after adjusting for gender), and those first treated in the later years of our study (compared with the earlier years) were more likely to remain therapy persistent and class persistent over the first 2 years (Table 3). Women were more likely to be therapy persistent but less likely to be class persistent within the first 2 years of their initial prescription, suggesting that women demonstrated more frequent switching between drug classes. Patients initiated on all antihypertensive drug classes showed better therapy persistence and class persistence than those initiated on a diuretic, even after adjusting for baseline covariates. In particular, patients initially prescribed angiotensin-converting enzyme inhibitors (P<.0001) were the most likely to remain therapy persistent compared with patients initiated on all other antihypertensive drug classes, with angiotensin II receptor blocker following second (P<.01). Patients initially prescribed diuretics were least likely to be class persistent (P<.0001), with patients initiating on angiotensin-converting enzyme inhibitors, angiotensin II receptor blockers, or calcium channel blockers all showing similar class persistence.
Table 3. Determinants of Therapy Persistence (n=206,603), Class Persistence (n=206,603), and Medication Possession Ratio≥0.8(n=136,673) after Initial Antihypertensive Drug Prescription
| Variable | Therapy Persistence | Class Persistence | MPR≥0.8 | |||
|---|---|---|---|---|---|---|
| OR (95% CI) | P Value | OR (95% CI) | P Value | OR (95% CI) | P Value | |
| Drug class | ||||||
| ACEI | 1.37(1.34-1.41) | <.0001 | 1.23(1.20-1.26) | <.0001 | 1.17(1.10-1.25) | <.0001 |
| ARB | 1.22(1.16-1.28) | <.0001 | 1.22(1.16-1.27) | <.0001 | 1.07(0.93-1.22) | .35 |
| BB | 1.08(1.04-1.11) | <.0001 | 1.15(1.12-1.19) | <.0001 | 0.79(0.73-0.86) | <.0001 |
| CCB | 1.13(1.09-1.17) | <.0001 | 1.23(1.19-1.27) | <.0001 | 1.08(0.99-1.19) | .078 |
| Diuretic | 1.0(ref) | N/A | 1.0(ref) | N/A | 1.0(ref) | N/A |
| Gender | ||||||
| Female | 1.09(1.07-1.11) | <.0001 | 0.95(0.93-0.97) | <.0001 | 1.12(1.06-1.18) | <.0001 |
| Male | 1.0(ref) | N/A | 1.0(ref) | N/A | 1.0(ref) | N/A |
| Residence location | ||||||
| Urban | 0.78(0.76-0.80) | <.0001 | 0.79(0.77-0.81) | <.0001 | 0.82(0.76-0.88) | <.0001 |
| Rural | 1.0(ref) | N/A | 1.0(ref) | N/A | 1.0(ref) | N/A |
| Income (per quintile) | 1.04(1.03-1.04) | <.0001 | 1.03(1.02-1.03) | <.0001 | 1.10(1.08-1.12) | <.0001 |
| Age (per 5-y increment) | 1.02(1.01-1.03) | <.0001 | 1.05(1.05-1.06) | <.0001 | 0.88(0.87-0.90) | <.0001 |
| Cohort year (per 2-y increment) | 1.04(1.03-1.05) | <.0001 | 1.01(1.00-1.02) | .013 | 1.07(1.05-1.10) | <.0001 |
| Charlson comorbidity index score | ||||||
| 0 | 0.76(0.73-0.79) | <.0001 | 0.72(0.69-0.74) | <.0001 | 1.29(1.17-1.41) | <.0001 |
| 1 | 0.88(0.83-0.93) | <.0001 | 0.90(0.85-0.95) | <.0001 | 0.99(0.87-1.13) | .89 |
| 2+ | 1.0(ref) | N/A | 1.0(ref) | N/A | 1.0(ref) | N/A |
Urban residents and older patients were less likely to be compliant (ie, medication possession ratio>0.8), whereas women, those of higher socioeconomic status, those free of comorbidity, and those first treated in the later years of our study (compared with the earlier years) were more likely to comply. Compliance was significantly better with angiotensin-converting enzyme inhibitors and significantly worse with beta-blockers compared with diuretics.
Sensitivity Analyses
Using 30- or 100-day time windows to define persistence did not appreciably change our main findings reported above (Appendix-Supplementary Tables 3 and 4). Analyses stratified by diabetic status yielded similar relationships among the diabetic and nondiabetic cohorts compared with the overall cohort, except that some of the associations did not achieve statistical significance in the diabetic cohort, although they remained consistent in direction (Appendix-Supplementary Tables 5 and 6).
Discussion
We found that therapy persistence and compliance were highest in those elderly patients who were initiated on an angiotensin-converting enzyme inhibitor compared with other antihypertensive drug classes. We also confirmed our earlier observations that persistence with antihypertensive therapy is improving over time12 and extend our earlier work by also demonstrating that compliance is improving over time. We found that persistence was higher in older individuals, rural residents, people with higher income, and individuals with higher burdens of comorbidity. However, in women, therapy persistence was better whereas class persistence was worse, implying greater switching between antihypertensive drug classes in women. Moreover, compliance improved over time and was greater in rural residents, women, and those with higher income; however, contrary to our findings on persistence, compliance was higher in younger individuals and those individuals with lower burdens of comorbidity. Our study differs from prior work on this topic by reporting medication use by an inception cohort of incident hypertensives who had filled at least 2 prescriptions and were followed for a prolonged period, and by using current consensus-guided definitions for persistence and compliance.3 Notably, the odds ratio as provided in Table 3 tends to yield an estimate of risk further away from 1.0 than the relative risk, particularly given that the prevalence of therapy persistence, class persistence, and high medication possession ratio were not low.
A population-based cohort study of incident hypertensives found that therapy persistence at 1 year and median time of persistence were 71.5% and 3.07 years, respectively. Unlike our data, class-specific persistence was highest for angiotensin II receptor blockers at 70.6% (compared with 62.2% for angiotensin-converting enzyme inhibitors) at 1 year and median time of persistence was longest for angiotensin II receptor blockers at 2.90 years (compared with 2.24 years for angiotensin-converting enzyme inhibitors).13 Several retrospective cohort studies of incident hypertensives in the Saskatchewan Health databases over varying time periods demonstrate comparable findings to ours. In an earlier analysis, female gender and older age were associated with persistence.14 Such findings were reproduced in a later study of similar design.15 However, in that same study, therapy persistence was found to be associated with less severe health status (as reflected by the chronic disease score), in contrast with our findings.15 In addition, therapy persistence has been found to differ according to the initial drug class (predating the introduction of angiotensin II receptor blockers) in the following decreasing order: angiotensin-converting enzyme inhibitors, calcium channel blockers, beta-blockers, and diuretics, even after adjustment for age, gender, and surrogate measures of health status.16 Likewise, another analysis revealed that sociodemographic factors and drug class were determinants of persistence; specifically, women, elderly patients, and, distinct from our findings, those prescribed angiotensin II receptor blockers initially were more likely to persist with therapy.17 The apparent superiority of angiotensin II receptor blockers in persistence and compliance also has been confirmed in more recent examinations.15, 18, 19, 20, 21, 22 Contrary to the recently published Ongoing Telmisartan Alone and in Combination With Ramipril Global Endpoint Trial, we found no substantial difference in class persistence between angiotensin-converting enzyme inhibitors and angiotensin II receptor blockers.23 Finally, our study confirms one other study that also reported higher socioeconomic status was associated with greater therapy persistence.24
Unlike our data, compliance was associated with older age and male gender in a retrospective analysis of incident and prevalent hypertensives;25 further, among incident elderly hypertensives, a proportion of days covered greater than 0.8 was associated with comorbid cardiac disease.26 Similarly, older age27 and greater comorbidity28 also were associated with compliance in 2 other studies using the Department of Veterans Affairs' pharmacy database. On the other hand and in line with our findings, women were more likely to be compliant in a telephone questionnaire administered to prevalent hypertensives.29 Initial drug choice also was found to be an important factor such that angiotensin-converting enzyme inhibitors26, 30 and angiotensin II receptor blockers20, 22 demonstrated the greatest likelihood of compliance.
Associations between antihypertensive medication persistence or compliance and residence location have not been previously presented. The reasons why patients do not comply or persist with their antihypertensive therapy and why such behaviors differ across drug classes are likely multifaceted. Differences in drug tolerability, dosing frequency, and patient perceptions of side effects are potential contributors. Medication costs would not likely be significantly accountable given that the Ontario Drug Benefit program provides drug benefits for all Ontarians aged 65 years and older from a minimally restricted formulary; parenthetically, the Ontario Drug Benefit pays for the lowest priced interchangeable generic drug listed in the formulary when both generic and brand name drugs are available. Variable methods for monitoring, definitions, study populations, and durations of follow-up may explain, at least in part, discrepant findings across studies. In fact, a recent cross-national study revealed that although the percentages of therapy persistence among incident elderly hypertensives were similar across 3 countries (United States, Canada, The Netherlands), predictor variables demonstrated associations with therapy persistence that were either consistent or inconsistent in direction across the populations.24
Study Limitations
There are several limitations inherent in our study. First, medication possession ratio fails to capture poor compliance beyond unfilled prescriptions as in the cases of filled prescriptions not taken or inappropriately used; further, we did not account for medication hoarded over from previous prescriptions in determining persistence. Second, given the dynamic nature of both compliance and persistence,1 our results apply only to the first 2 years of treatment; as well, we did not address other potential predictors (eg, ethnicity27) of these parameters. Third, we were unable to relate compliance or persistence to blood pressures, because administrative databases do not contain information on actual blood pressures. Fourth, apparently nonpersistent patients might still return to treatment at the end of observation; despite extending the time window for defining persistence to 100 days in one of our sensitivity analyses, between 54% and 75% of patients return to treatment in the year after discontinuation.21 Fifth, although previous studies suggest that patients who exhibit better compliance/persistence have fewer deaths and cardiovascular events,31, 32 we did not specifically analyze this relationship.
Conclusions
Overall, our data provide evidence of a general improvement in antihypertensive drug persistence and compliance across years, as well as significant differences across drug classes and other patient-level factors. Targeted efforts directed toward men, urban residents, and those of low socioeconomic status to increase antihypertensive drug persistence and compliance should be considered. Because all antihypertensive drug classes seem to exert similar benefits in cardiovascular risk reduction in randomized clinical trials,33 any differences between classes in persistence/compliance when used in clinical practice might be expected to translate into differences in the real-world effectiveness of each drug class in reducing cardiovascular event rates.34 Thus, close attention to persistence and compliance with antihypertensive drug therapy is important to reduce hypertension-related cardiovascular disease.
Acknowledgment
The authors thank Nadia Khan for helpful comments.
Supplementary data
Appendix-Supplementary Tables
References
- . Medication adherence and persistence as the cornerstone of effective antihypertensive therapy. Am J Hypertens. 2006;19:1190–1196
- . The direct costs to the NHS of discontinuing and switching prescriptions for hypertension. J Hum Hypertens. 1998;12:533–537
- Recommendations for evaluating compliance and persistence with hypertension therapy using retrospective data. Hypertension. 2006;47:1039–1048
- . Completeness of safety reporting in randomized trials: an evaluation of 7 medical areas. JAMA. 2001;285:437–443
- . A review of studies of adherence with antihypertensive drugs using prescription databases. Ther Clin Risk Manag. 2005;1:93–106
- . A summary of studies on the quality of health care administrative databases in Canada. In: Goel V, William J, Anderson G, et al. editor. Patterns of Health Care in Ontario: The ICES Practice Atlas. 2nd ed.. Ottawa: Canadian Medical Association; 1996;p. 339–345
- . Diabetes in Ontario: determination of prevalence and incidence using a validated administrative data algorithm. Diabetes Care. 2002;25:512–516
- . Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol. 1992;45:613–619
- . Assessing ecologic proxies for household income: a comparison of household and neighbourhood level income measures in the study of population health status. Health Place. 1999;5:157–171
- . PCCF+Version 4J User's Guide: Automated Geographic Coding Based on the Statistics Canada Postal Code Conversion Files, Including Postal Codes through September 2006. Ottawa: Health Analysis and Measurement Group, Statistics Canada; 2007;Report No.: 82F0086-XDB
- Definitions of “Rural”. Ottawa: Statistics Canada; 2002;Agriculture and Rural Working Paper Series, No. 61
- . Hypertension management in the elderly has improved: Ontario prescribing trends, 1994 to 2002. Hypertension. 2005;45:1113–1118
- Discontinuation of antihypertensive drugs among newly diagnosed hypertensive patients in UK general practice. J Hypertens. 2006;24:1193–1200
- Persistence with treatment for hypertension in actual practice. CMAJ. 1999;160:31–37
- Patterns of persistence with antihypertensive medications in newly diagnosed hypertensive patients in Italy: a retrospective cohort study in primary care. J Hypertens. 2005;23:2093–2100
- Effect of initial drug choice on persistence with antihypertensive therapy: the importance of actual practice data. CMAJ. 1999;160:41–46
- . Antihypertensive persistence and drug class. Can J Cardiol. 2002;18:649–656
- . Continuation of initial antihypertensive medication after 1 year of therapy. Clin Ther. 1998;20:671–681
- A retrospective, population-based analysis of persistence with antihypertensive drug therapy in primary care practice in Italy. Clin Ther. 2002;24:1347–1357discussion 1346
- Patient adherence with amlodipine, lisinopril, or valsartan therapy in a usual-care setting. J Manag Care Pharm. 2003;9:424–429
- Persistence and discontinuation patterns of antihypertensive therapy among newly treated patients: a population-based study. J Hum Hypertens. 2005;19:607–613
- . Persistence, adherence, and risk of discontinuation associated with commonly prescribed antihypertensive drug monotherapies. J Am Board Fam Med. 2007;20:72–80
- Telmisartan, ramipril, or both in patients at high risk for vascular events. N Engl J Med. 2008;358:1547–1559
- A cross-national study of the persistence of antihypertensive medication use in the elderly. J Hypertens. 2008;26:145–153
- Adherence to antihypertensive drug treatment and blood pressure control: a real practice analysis in Italy. J Hum Hypertens. 2008;22:51–53
- The effects of initial drug choice and comorbidity on antihypertensive therapy compliance: results from a population-based study in the elderly. Am J Hypertens. 1997;10(7 Pt 1):697–704
- . Antihypertensive medication adherence in the Department of Veterans Affairs. Am J Med. 2007;120:26–32
- . Medication acquisition and self-reported adherence in veterans with hypertension. Med Care. 2009;47:474–481
- Prevalence and predictors of poor antihypertensive medication adherence in an urban health clinic setting. J Clin Hypertens (Greenwich). 2007;9:179–186
- . Variations in compliance among hypertensive patients by drug class: implications for health care costs. Clin Ther. 1997;19:1446–1457discussion 1424-1425
- Effect of medication nonadherence on hospitalization and mortality among patients with diabetes mellitus. Arch Intern Med. 2006;166:1836–1841
- . Relationship between adherence to evidence-based pharmacotherapy and long-term mortality after acute myocardial infarction. JAMA. 2007;297:177–186
- Effects of different regimens to lower blood pressure on major cardiovascular events in older and younger adults: meta-analysis of randomised trials. BMJ. 2008;336:1121–1123
- . Patient adherence and medical treatment outcomes: a meta-analysis. Med Care. 2002;40:794–811
Funding: Alberta Heritage Foundation for Medical Research.
Conflict of Interest: None of the authors have any conflicts of interest associated with the work presented in this manuscript.
Authorship: All authors had access to the data and played a role in writing this manuscript.
PII: S0002-9343(09)00797-9
doi:10.1016/j.amjmed.2009.08.008
© 2010 Elsevier Inc. All rights reserved.
