Compliance with Antihypertensive Therapy in the Elderly

Download Compliance with Antihypertensive Therapy in the Elderly

Post on 18-Mar-2017




0 download


  • Am J Cardiovasc Drugs 2008; 8 (1): 45-50ORIGINAL RESEARCH ARTICLE 1175-3277/08/0001-0045/$48.00/0 2008 Adis Data Information BV. All rights reserved.

    Compliance with Antihypertensive Therapy inthe ElderlyA Comparison of Fixed-Dose Combination Amlodipine/Benazepril versusComponent-Based Free-Combination Therapy

    Michael Dickson1 and Craig A. Plauschinat2

    1 College of Pharmacy, University of South Carolina, Columbia, South Carolina, USA2 Novartis Pharmaceuticals, East Hanover, New Jersey, USA

    Background: Treatment regimens that require fewer dosage units and less frequent dosing to decrease theAbstractcomplexity and cost of care are among the strategies recommended to improve compliance with antihypertensivetherapy. Simplifying therapy may be particularly important for elderly patients, who are more likely to have co-morbid conditions and to be taking multiple medications.Objective: To determine rates of compliance with antihypertensive therapy and total costs of care among elderlyMedicaid recipients treated with fixed-dose combination amlodipine besylate/benazepril versus a dihydropyri-dine calcium channel antagonist and ACE inhibitor prescribed as separate agents (free combination).Study design: A longitudinal, retrospective, cohort analysis of South Carolina Medicaid claims for ambulatoryservices, hospital services, Medicare crossover, and prescription drug for the years 19972002. Follow-up was12 months from the index date, defined as the first prescription dispensing date for a study drug.Patients: South Carolina Medicaid beneficiaries aged 65 years.Main outcome measure: Outcomes variables included compliance defined as the medication possession ratio(MPR), which was the total days supply of drug (excluding last prescription fill) divided by the length of follow-up (with number of hospital days subtracted from the numerator and denominator). We hypothesized that elderlyindividuals receiving fixed-dose combination amlodipine besylate/benazepril HCl would be more compliantwith therapy than those receiving a dihydropyridine calcium channel antagonist and ACE inhibitor as freecombination.Results: There were 2336 individuals in the fixed-combination group and 3368 in the free-combination group.The mean age was 76.0 7.2 years, and 82.6% were female. Compliance rates were significantly higher withfixed-dose versus free-combination therapy (63.4% vs 49.0%; p < 0.0001). The average total cost of care forpatients receiving the fixed-dose combination was $US3179 compared with $US5236 (2002 values) for the free-combination regimen. In multivariate regression analyses on the log of total cost of care, average total costsincreased by 0.5% for each 1-unit increase in MPR, and for each additional co-morbidity (measured by thechronic disease score) there was an increase of 10.4%. However, average total costs were reduced by 12.5% forpatients using fixed-dose versus free-combination therapy (p < 0.003).Conclusion: Use of fixed-dose amlodipine besylate/benazepril HCl by elderly Medicaid recipients wasassociated with improved compliance and lower healthcare costs compared with a dihydropyridine calciumchannel antagonist and ACE inhibitor prescribed as separate agents.

    According to the National Health and Nutrition Examination quarters of these individuals with hypertension are aged 50Survey (NHANES) 19992000, 65 million adult Americans or years.[1] The risks of inadequately controlled hypertension, includ-31% of the adult population of the US have hypertension. Three- ing increased rates of myocardial infarction, heart failure, stroke,

  • 46 Dickson & Plauschinat

    and kidney disease, are well established, as are the benefits of quiring fewer dosage units and less frequent dosing to reduce thelowering BP to reduce these risks.[2,3] Antihypertensive therapy in complexity and cost of care are among the strategies recommen-patients aged 60 years has been shown to reduce total mortality, ded to improve compliance with antihypertensive therapy and maycardiovascular morbidity/mortality, stroke, and coronary events. be particularly important for elderly patients, who are more likelyThese benefits are even more pronounced in patients aged 70 to have co-morbid conditions and take multiple medications.[2,13,14]

    years.[3] In addition to improving compliance, uninterrupted antihyperten-sive therapy has been shown to reduce total healthcare costs in aA comparison between NHANES III (198894) and NHANESpopulation of Medicaid recipients.[15] The use of lower-dose com-19992002 showed that the use of antihypertensive medicationbination therapy with two complementary agents has the potentialamong hypertensive adults increased significantly during thatto increase efficacy, decrease adverse effects, reduce medicationtime;[4] nevertheless, more than one-third of hypertensive adultscosts, and increase patient compliance with therapy.[12]remain untreated.[5] Recognizing the need for more than one

    The objectives of this study, therefore, were to determine theantihypertensive agent to achieve BP goals in the majority ofrates of compliance and total costs of care among elderly Medicaidpatients, the use of ACE inhibitor-based combination therapy alsorecipients treated with a fixed-dose combination of amlodipinesignificantly increased over the past decade.[4] In both NHANESbesylate and benazepril HCl versus a dihydropyridine calciumsurveys, antihypertensive medication use was greater among olderchannel antagonist (DHP-CCA) and ACE inhibitors prescribed asthan younger adults and increased significantly in the groups agedseparate agents.5059 years and 70 years. Hypertension treatment and rates of

    control both increased significantly for men (p < 0.001 for bothdimensions), but did not change significantly for women. In spite Methodsof the evidence for increased use of antihypertensive therapy in theelderly population, individuals aged >60 years still showed the This was a longitudinal, retrospective, cohort study using thelowest rate of BP control, including those receiving antihyperten- South Carolina Medicaid database for the period 19972002. Thesive treatment.[2,6] number of Medicaid beneficiaries varied by year and month but, in

    Both physician-related and patient-related factors contribute to general, there was an average of about 700 000 Medicaid benefi-poor BP control in the elderly. Physicians may be reluctant to treat ciaries throughout the study period. Claims for the study includedelderly patients because of concern about adverse events and drug ambulatory care, hospital care, prescription drug claims, andinteractions, and consequently may not prescribe the most effec- Medicare crossover claims for patients selected during the studytive antihypertensive agents at the most effective doses.[7] Accord- period. Although patients may have had claims in more than 1ing to a study of medication use in a representative US sample, year, each patient was included only once in the study and fol-23% of women and 19% of men aged 65 years reported taking lowed for 12 months (the follow-up period) from the date of thefive or more prescription drugs.[8] The presence of common co- first prescription of study drug (the index date). Prescriptions maymorbid conditions, including asthma/chronic obstructive pulmon- be for any quantity of medication up to a 3-month supply. Allary disease, depression, gastrointestinal disorders, and osteoarthri- personal identifiers were removed before data analyses to protecttis, has been associated with reduced use of antihypertensive the confidentiality of study participants in conformance with themedications in the elderly, even in the presence of compelling Health Insurance Portability and Accountability Act (HIPAA).indications for therapy (e.g. coronary artery disease or stroke).[9] In Medicaid is a state-federal jointly funded program to provideaddition, physicians may accept a higher SBP level in their pa- healthcare services to those who qualify based on a means test. Atients than is currently recommended by JNC 7 (Seventh Report of full range of healthcare services is included largely through privatethe Joint National Committee on Prevention, Detection, Evalua- sector providers who submit claims for payment. Medicaid pa-tion, and Treatment of High Blood Pressure) guidelines because tients may be asked for a modest copayment (but they are notthey consider DBP to be more clinically relevant.[10] However, required to pay), but they do not pay for services and then seekSBP continues to increase with increasing age, whereas DBP tends reimbursement. For example, pharmacies submit claims for pre-to decrease or normalize; as a result, more than 90% of individuals scriptions dispensed for Medicaid beneficiaries. While there wereaged 50 years may be staged on the basis of SBP alone.[11] some limitations on the drug benefit during the time of this study,

    these did not influence the availability of antihypertensive medica-Patient compliance with treatment regimens decreases as thetions.number of drugs taken and the cost of medication increase.[12] The

    inconvenience and confusion associated with taking multiple Patients were included in the study if they were aged at least 65drugs also reduces compliance. Simpler treatment regimens re- years but

  • Compliance with Antihypertensive Therapy 47

    was used as a measure of compliance, defined as the percentage oftime in days that a patient had the study drugs available (excludingthe last prescription fill) during the 1-year follow-up period. Thenumber of hospital days were subtracted from the numerator anddenominator. MPR is one of several commonly used metrics formeasuring medication adherence. It is used in this study because itis the most generic of adherence measures.[16,17] The number ofdays of hospitalization, if any, was subtracted from the number ofdays of drug supply (numerator) and the number of days of follow-up (denominator) to account for non-use of drug while hospital-ized and possible prescription changes upon discharge. For pa-

    Table I. Patient characteristics

    Characteristic Fixed-dose Free Totalcombination combination (n = 5704)(n = 2336) (n = 3368)

    Age (y) [mean SD] 75.6 7.1 76.2 7.3 76.0 7.2

    Sex (% female) 82.5 82.7 82.6

    Race (% African 72.0 70.1 70.9American)

    Residency (% urban) 89.3 90.8 90.2

    Chronic disease score 4.8 2.3 5.0 2.3 4.9 2.3(mean SD)

    tients in the free-combination group, days of drug possession wereprescriptions for study drugs in one of the selection years counted only if both drugs were available on the same day. Total(19972001), and were continuously eligible for Medicaid for 12 cost of care was defined as the sum of payments for Medicaidmonths following the index date. Patients were excluded if they claims for ambulatory care (Healthcare Financing Administrationhad >180 days of hospitalization,

  • 48 Dickson & Plauschinat


    Patient characteristics of the study population as of the indexdate are shown in table I. A total of 5704 Medicaid recipients aged65 years were included in the study, including 2336 in the fixed-dose combination group and 3368 in the free-combination group.The mean age of patients was 76.0 7.2 years, 82.6% werefemale, and 70.9% were African American. The demographics ofthe two groups were similar.


    As shown in table II and graphically presented in figure 1,compliance with antihypertensive therapy measured by MPR wassignificantly higher for patients in the fixed-dose combinationgroup compared with those in the free-combination group (63.4%vs 49.0%; p < 0.05). These compliance rates are lower than aredesired, but the fixed-dose combination group did demonstrate a







    Fixed-dose combination therapyFree-combination therapy


    R (


    Fig. 1. Compliance rates, calculated as medication possession ratio(MPR), in the fixed-dose combination and free-combination therapygroups. MPR was significantly higher for patients receiving fixed-dosecombination therapy compared with free-combination therapy (63.4% vs49%; p < 0.05). MPR was calculated as total days drug supply (excludinglast prescription fill) divided by 365 days (with number of hospital dayssubtracted from the numerator and denominator). significantly higher rate of compliance than the free-combination

    group.pre-index period, each of these variables was standardized by the

    Total Healthcare Costsnumber of months of pre-period eligibility.

    The CDS has been implemented by researchers in several ways. The average total cost of care for the fixed-dose combinationWe implemented the chronic disease concept using the method group was $US3179 (2002 value) compared with $US5236 (2002described by Clark et al.[19] Clarks method is a modification of value) for the free-combination group (p < 0.0001) [figure 2]. Forwork by VonKorf and colleagues,[20] who created a CDS that was each type of claim, including ambulatory, drug, and hospitalbased on pharmacy claims and weighted by expert panel judg- claims, costs were lower for the fixed-dose combination than forment. Clarks modification of VonKorfs measure was to derive the free-combination group. Ambulatory, drug, and other costsweights using a regression model rather than expert opinion.Because nonstandardized regression coefficients cannot be gener-alized to other situations, it was not possible to use those weights,and the work of Schneeweiss and others suggests that simplecounts of different prescription drugs received during the past year appear to be well performing comorbidity measures in epide-miological studies.[21] Therefore, we used an unweighted CDSderived using the therapeutic categories described in the 1995Clark et al.[19] paper. Patients may have multiple prescriptions forthe same chronic condition, but each condition is counted onlyonce in the CDS for this study.

    Patient-level data were used to calculate the MPR; a two-sample t-test was used to compare compliance rates between thetwo treatment groups. Multivariate analyses were performed toassess the relationship between compliance and total healthcarecost in the 12-month follow-up period (log of total healthcarecost), controlling for study year of enrollment, MPR, CDS, race,sex, and study group (fixed combination vs free combination). Alldata management and statistical analyses were performed usingSAS, version 8.2 (SAS Institute, Inc., Cary, NC, USA).

    7 500

    5 000

    2 500



    ts o

    f car

    e ($








    Fig. 2. Average total cost of care for patients receiving fixed-dose combina-tion therapy vs free-combination therapy. The average total cost of care forthe fixed-dose combination therapy group was $US3179 compared with$US5236 (2002 values) for the free-combination therapy group. Compo-nent costs for fixed-dose vs free-combination therapy groups were$US1120 vs $US1646 for ambulatory claims, $US1322 vs $US1952 fordrug costs, $US334 vs $US410 for hospital costs and $US402 vs $US1229for other costs (Medicare cross-over costs).

    2008 Adis Data Information BV. All rights reserved. Am J Cardiovasc Drugs 2008; 8 (1)

  • Compliance with Antihypertensive Therapy 49

    were significantly different between the two treatment groups patients aged 18 years.[22] In an earlier retrospective study of a(p < 0.0001), but there was no significant difference in hospital managed-care patient population that was younger on averagecost (p = 0.5022). Results of the multivariate analyses on the log of (mean age 53 years, 50% female), patients prescribed fixed-dosetotal cost of care for the 1-year follow-up period are presented in combination amlodipine besylate/benazepril also demonstratedtable III. Controlling for year of study enrollment, MPR, CDS, significantly higher compliance rates and lower per-patient aver-race, and sex, patients using the fixed-dose combination had an age costs of cardiovascular-related care than those prescribedaverage total cost of care increase of 0.5% (p < 0.0001) for each concomitant DHP-CCA and ACE inhibitors as separate agents.[23]

    one unit increase in MPR, and a 10.4% (p < 0.0001) increase for This study is subject to several limitations. The potential foreach additional co-morbidity (as measured by the CDS). We selection bias exists because of the retrospective, nonrandomizedwould expect increased medication compliance to increase the study design; however, propensity scores and covariates theoreti-average cost of the drug used, and patients with more co-morbidi- cally associated with the outcomes of interest were used to controlties to have, in general, higher healthcare costs. However, the for potential confounding factors. Using a Medicaid population aseffect of being in the fixed-combination group mitigated these a source of data offers the advantages of completeness of informa-increases. There was a 12.5% lower total cost of care for the fixed- tion and stable enrollment, but may limit the ability to generalizecombination group compared with the free-combination group (p results to the general population; in this analysis, the population< 0.003). There was no statistically significant relationship be- studied was predominantly African American and female. Thetween race and sex and the cost of care (p = 0.6343 and p = 0.7614, study also excluded patients with extended hospital or nursingrespectively). home stays and thus may not be representative of all elderly

    people. Finally, assessment of compliance was based on prescrip-tion records and there was no confirmation that the study drugsDiscussionwere taken, nor was there documentation of clinical outcomes.

    In this study of elderly Medicaid recipients, patients prescribeda fixed-dose combination of amlodipine besylate/benazepril had Conclusionhigher rates of compliance and lower total healthcare costs thanthose receiving a free-combination of the same component drug This analysis shows that elderly Medicaid patients receiving aclasses. These results are consistent with those seen in a larger fixed-dose combination of amlodipine besylate/benazepril werestudy of the South Carolina Medicaid population that included all significantly more compliant with antihypertensive therapy and

    Table III. Regression on the log of total cost of carea

    Dependent variable: log(total cost of care), mean = 7.8837, n = 5704

    independent variable coefficientsb SE t p > t

    Intercept 7.2018 0.0607 118.69

  • 50 Dickson & Plauschinat

    9. Wang PS, Avorn J, Brookhart MA, et al. Effects of noncardiovascular comorbidi-had lower total healthcare costs over a period of 1 year thanties on antihypertensive use in elderly hypertensives. Hypertension 2005; 46

    individuals receiving a free combination of a DHP-CCA and an (2): 273-9ACE inhibitor prescribed concomitantly as separate agents. Ef- 10. Oliveria SA, Lapuerta P, McCarthy BD, et al. Physician-related barriers to theforts to enhance compliance that are tailored to elderly patient effective management of uncontrolled hypertension. Arch Intern Med 2002;

    162 (4): 413-20populations, including simplification of treatment regimens, may11. Franklin SS, Jacobs MJ, Wong ND, et al. Predominance of isolated systolicultimately improve patient outcomes and reduce healthcare costs.

    hypertension among middle-aged and elderly US hypertensives: analysis basedon National Health and Nutrition Examination Survey (NHANES) III. Hyper-tension 2001; 37 (3): 869-74

    Acknowledgments12. Neutel JM, Smith DH. Improving patient compliance: a major goal in the manage-

    ment of hypertension. J Clin Hypertens 2003; 5 (2): 127-32

    Dr Dickson received an unrestricted research grant from Novartis Pharma- 13. Schroeder K, Fahey T, Ebrahim S. How can we improve adherence to bloodpressure-lowering medication in ambulatory care? Systematic review of ran-ceuticals to conduct this study. Dr Plauschinat is a full-time employee ofdomized controlled trials. Arch Intern Med 2004; 164 (7): 722-32Novartis Pharmaceuticals Corp. in the Evidence Based Medicine Department,

    a division of US Medical and Regulatory Affairs. 14. Elliott WJ. Compliance -and improving it -in hypertension. Manag Care 2003; 12(8 Suppl. Hypertension): 56-61Dr Dickson was responsible for the conceptualization, design, and statisti-

    cal analysis of this study as well as drafting the manuscript. Dr Plauschinats 15. McCombs JS, Nichol MB, Newman CM, et al. The costs of interrupting antihyper-tensive drug therapy in a Medicaid population. Med Care 1994; 32 (3): 214-26role includes assistance with the design of the study, and assistance with

    preparation of the manuscript (including assistance with revisions). 16. Sclar DA, Chin A, Skaer TL, et al. Effect of healthcare education in promotingprescription refill compliance among patients with hypertension. Clin TherThe results of this study were previously presented as an abstract at 58th1991; 13: 489-95Scientific Meeting of The Gerontological Society of America, 1822 Novem-

    ber 2005, Orlando, FL, USA. 17. Steiner JF, Prochazka AV. The assessment of refill compliance using pharmacyrecords: methods, validity, and application. J Clin Epidemiol 1997; 50 (1):105-16

    18. DAgostino RB. Propensity score methods for bias reduction in the comparison ofReferences a treatment to non-randomized control group. Stat Med 1998; 17: 2265-81

    1. Fields LE, Burt VL, Cutler JA, et al. The burden of adult hypertension in the United19. Clark DO, Von Korff M, Saunders K, et al. A chronic disease score withStates 1999 to 2000: a rising tide. Hypertension 2004; 44 (4): 398-404

    empirically derived weights. Med Care 1995; 33: 783-952. Chobanian AV, Bakris GL, Black HR, et al. The seventh report of the Joint

    20. VonKorf M, Wagner EH, Saunders K. A chronic disease score from automatedNational Committee on Prevention, Detection, Evaluation, and Treatment ofpharmacy data. J Clin Epidemiol 1992; 45: 197-203High Blood Pressure: the JNC 7 report. JAMA 2003; 289 (19): 2560-72

    3. Staessen JA, Gasowski J, Wang JG, et al. Risks of untreated and treated isolated 21. Schneeweiss S, Wang PS, Avorn J, et al. Consistency of performance ranking ofsystolic hypertension in the elderly: meta-analysis of outcome trials. Lancet comorbidity adjustment scores in Canadian and US utilization data. J Gen2000; 355 (9207): 865-72 Intern Med 2004; 190: 444-50

    4. Gu Q, Paulose-Ram R, Dillon C, et al. Antihypertensive medication use among US22. Dickson M, Plauschinat CA. Antihypertensive therapy compliance and total cost of

    adults with hypertension. Circulation 2006; 113 (2): 213-21care in a Medicaid population: fixed-dose combination versus free combination

    5. Cheung BM, Ong KL, Man YB, et al. Prevalence, awareness, treatment, and treatment [abstract]. American Society of Hospital Pharmacists Summer Meet-control of hypertension: United States National Health and Nutrition Examina- ing; 2005 Jun 11-15; Boston (MA). Am Soc Hosp Pharm 2005 Jun; 62: P71Etion Survey 2001-2002. J Clin Hypertens (Greenwich) 2006; 8 (2): 93-8

    23. Taylor AA, Shoheiber O. Adherence to antihypertensive therapy with fixed-dose6. Hajjar I, Kotchen TA. Trends in prevalence, awareness, treatment, and control ofamlodipine besylate/benazepril HCl versus comparable component-based ther-hypertension in the United States, 1988-2000. JAMA 2003; 290 (2): 199-206apy. Congest Heart Fail 2003; 9 (6): 324-32

    7. Garg JP, Elliott WJ, Folker A, et al., for the RUSH University HypertensionService. Resistant hypertension revisited: a comparison of two university-basedcohorts. Am J Hypertens 2005; 18 (5): 619-26

    Correspondence: Dr Michael Dickson, University of South Carolina, College8. Kaufman DW, Kelly JP, Rosenberg L, et al. Recent patterns of medication use inof Pharmacy, Columbia, SC 29208, USA.the ambulatory adult population of the United States: the Slone survey. JAMA

    2002; 287 (3): 337-44 E-mail:

    2008 Adis Data Information BV. All rights reserved. Am J Cardiovasc Drugs 2008; 8 (1)


View more >