Adherence Trends for 3 Chronic Disease Medication Classes Among Differently Insured Populations

, , ,
AJPB® Translating Evidence-Based Research Into Value-Based Decisions®, January/February 2014, Volume 6, Issue 1

Adherence trends were examined for commercial, Medicaid, Medicare Part D, and self-insured health plans for oral antidiabetic medications, statins, and direct renin inhibitors.

Cardiovascular disease is the leading cause of death in the United States.1 By 2030, the prevalence of cardiovascular diseases, including hypertension, is projected to reach more than 40% of US adults and cost the United States $1.1 trillion annually (unadjusted 2008 dollars).2 Hypertension, dyslipidemia, and diabetes mellitus are key risk factors for cardiovascular disease and cardiovascular events. Several observational studies have demonstrated that adherence to medications indicated for the treatment of hypertension, diabetes mellitus, or dyslipidemia is associated with better cardiovascular outcomes, decreased cardiovascularrelated and all-cause hospitalizations and emergency department visits, and lower total healthcare costs.3-8 Although the validity of these results has been called into question by demonstration of the “healthy user bias,”9,10 it stands to reason that for a drug to be effective, it needs to be taken regularly. Furthermore, adherence to oral antidiabetic drugs (OADs), antihypertensives, and HMGCoA reductase inhibitors (statins) has been shown to be associated with the cardiovascular intermediate outcomes of lower glycosylated hemoglobin, blood pressure, and low-density lipoprotein cholesterol levels.11,12

Medication adherence is defi ned as “the extent to which a patient acts in accordance with the prescribed interval and dose of a dosing regimen.”13 Payers with large claims databases measure patient adherence with “medication possession” methodologies that use prescription fill dates and days’ supply. These methodologies are based on the number of days a patient has possession of medication rather than if the patient has actually taken the medication. The most commonly used measure in the literature is the Medication Possession Ratio (MPR).14,15 There are many variations of MPR used in the literature; however, MPR is generally calculated as the sum of the days’ supply of medication available in a given period of time divided by the sum of the number of days in the period or the time between the first and last fill of medication.14,15 The latter denominator may infl ate adherence measurements as it does not consider potential discontinuation of chronic therapy.15,16 Also, when shorter observation periods are used, the likelihood of adherence is increased.15

The Pharmacy Quality Alliance (PQA) developed and tested a medication possession adherence measure known as the Proportion of Days Covered (PDC) in several chronic medication classes and concluded that it is to be preferred over MPR methodologies.17 PDC is calculated as the number of days within the calendar year with medication supply divided by the number of days the patient is eligible for benefits. If fills overlap, the start date of the next fill is adjusted to be the day after the previous fill ended. Also, if patients have supply for more than 1 medication within the measured class on the given day, the day is only counted once as being “covered.” Thus, the PDC provides a more conservative estimate of adherence when there may be switching and concomitant therapy.17,18 Unfortunately, there are variations in PDC methodology in the literature and PDC sometimes is used as a misnomer for calculations consistent with MPR methodology.15,19 The definition of 80% PDC or MPR as a threshold for adherence is common in the literature.15 The PQA PDC methodology includes reporting the number of patients reaching 0.80 or greater PDC. While somewhat arbitrary, the 80% threshold for OAD, direct renin inhibitors (DRIs), and statin adherence has been shown to be predictive of clinical outcomes.20,21

Since the PQA endorsed a specific PDC methodology, there has been movement toward using it as the “gold standard.” The National Quality Forum (NQF) endorsed the PDC measure for 5 medication types in 2009.22 The Centers for Medicare & Medicaid Services (CMS) adopted this methodology in 2010 for measurement of adherence to OADs, DRIs, and statins as part of the Five-Star Ratings Quality Rating System.23 The National Committee for Quality Assurance (NCQA) has adopted PDC for adherence measures in the Healthcare Effectiveness Data and Information Set (HEDIS),24 and URAC—a healthcare accreditation agency—has adopted the methodology as part of the mandatory performancemeasures for pharmacy benefit management (PBM) and health plan accreditations.25,26 Also, PDC will be included in the battery of clinical quality measures for Qualified Health Plans (QHPs) accreditation necessary for participation on the Health Insurance Exchanges initiated by the Affordable Care Act.27,28

The inconsistency and variations in the use of MPR and PDC in the literature encumbers the comparison of adherence measurements in different medication classes and in different patient populations when not conducted in the same study. Several studies have shown differences in the adherence to various chronic therapies within the same population.5,7,11 Insurance type has been shown to be a significant predictor of adherence,29 likely because of the differences in demographic and clinical characteristics of the patients eligible for each type of benefit. It is of interest to compare adherence rates in these disparate populations. To the authors’ knowledge, such a comparative analysis does not exist. The objective of this analysis is to compare adherence rates for 3 key chronic medication categories in populations enrolled in different insurance types using the PQA-endorsed PDC methodology and to understand whether adherence may be trending over time.


A retrospective, cross-sectional analysis was performed using de-identified data from a sample of participating clients of MedImpact Healthcare Systems, Inc, a national pharmacy benefits manager, during 2012 and 2013. Four distinct cohorts were defined by type of insurance: commercial, Medicaid (a mix of Medicaid benefi ciaries and enrollees in programs for the indigent and other statefunded public programs), Medicare Part D beneficiaries (a mix of Medicare Advantage Prescription Drug [MA-PD] plans and Prescription Drug Plans [PDPs]), and selfinsured commercial health plans.

Within each insurance type, adherence was calculated as the Proportion of Days Covered (PDC) for each target medication class using specifications from the CMS Medicare 2014 Part C & D Star Rating Technical Notes (released September 27, 2013) and Pharmacy Quality Alliance Technical Specifications.30,31 Specifically, for patients 18 years and older with at least 2 claims for the target medication class, PDC was measured from first claim in the measurement period (index date) to end of the measurement period or member disenrollment. Days of medication coverage was calculated using fill dates and days of supply elements of prescription claims within each patient’s measurement period.32 Adherence was calculated for each cohort by using patients’ adherence rates while adjusting for length of member enrollment, or member years. Member years were calculated as number of months enrolled divided by months eligible in each measurement period. Target medication classes included OADs, DRIs, and statins—the 3 classes used in the CMS Part D Patient Safety measures. Patients utilizing insulin during the measurement period were excluded from the OAD measurement.

PDC and percent of patients meeting the definition of adherence of PDC greater than or equal to 0.80 were measured over the 2012 and 2013 calendar years as specified by PQA and CMS. We calculated these metrics for rolling 1-year periods at the conclusion of each quarter to observe possible temporal trends or seasonality. Analyses were performed using SAS (Cary, North Carolina) version 9.3.


Patients enrolled in Medicare Part D demonstrated greater adherence compared with the other insurance types in all 3 measured medication classes (

Figure 1

). The insurance types listed in order of decreasing observed adherence were Medicare Part D, commercial, self-insured health plans, and lastly, Medicaid. This was consistent across all medication classes. The percentage of Medicare Part D patients meeting the threshold of PDC greater than or equal to 0.80 ranged from 81.2% to 82.3% for OADs; 80.9% to 82.4% for DRIs; and 77.1% to 79.1% for statins. Of patients enrolled in commercial plans, 77.5% to 78.7% of patients were adherent to DRIs; 74.2% to 75.1% of patients were adherent to OADs; and 72.1% to 73.8% of patients were adherent to statins. The proportions of patients in self-insured plans who were adherent to therapy were 70.5% to 72.5% for DRIs; 67.0% to 68.2% for OADs; and 64.4% to 67.1% for statins. The Medicaid cohorts consistently had the smallest proportion of adherent patients compared with other insurance types. The proportion of adherent Medicaid patients ranged from 66.3% to 66.7% for OADs; 60.6% to 62.1% for DRIs; and 57.2% to 58.6% for statins. The proportion of adherent patients in the commercial and Medicare Part D populations increased from calendar year 2012 to calendar year 2013 for all 3 categories (

Figure 2

). The proportion of adherent patients decreased between the 2012 and 2013 calendar years for Medicaid benefi ciaries. The self-insured commercial population exhibited an increase in the proportion of patients who were adherent to OADs, a decrease in the proportion of patients who were adherent to DRIs, and no change in the proportion of patients who were adherent to statins between the 2012 and 2013 calendar years.


This analysis observed the greatest adherence in a Medicare Part D population, followed by commercial, self-insured health plans, and then Medicaid populations. Although copays are typically lowest for Medicaid patients, the lower prevalence of adherent patients is not unexpected, given the characteristics of the eligible beneficiaries. The Medicaid population has a greater prevalence of chronic diseases, mental disorders, and higher utilization of acute care services compared with commercial populations.33 In addition to greater disease burden, Medicaid patients are recipients of lower quality of care, similar to the uninsured.34,35

Higher rates observed among Part D beneficiaries are likely impacted by the Five-Star Quality Rating System that rewards health plans for increases in quality performance. The observed high adherence in Medicare Part D is also consistent with our understanding that adherenceimproves with age. Increasing age is a signifi cant predictor of the increased likelihood of being adherent to statins, antihypertensives, and antidiabetics.36-41 However, adherence to these medication classes has been shown to increase with age only up to a certain point, with the “younger old” being more adherent.42-46 Older age is associated with more comprehension errors, inconsistent preferences, and decreases in working memory that are associated with decreased adherence.47,48 Therefore, a younger Medicare Part D population would be more likely to have a higher proportion of adherent patients compared with an older one. Our observations may not be representative of the nation’s Medicare Part D population, as MedImpact’s Medicare population contains 21% low-income subsidized eligible patients compared with 34% of Part D benefi ciaries nationwide.49

We observed the least adherence and the smallest proportion of adherent patients with the statin medication class (


, Figure 1). This is consistent with other observational studies examining diabetes, hypertension, nd dyslipidemia medication adherence.29,50,51 A limitation of the PDC methodology is the possibility of inflation when combination therapy is used. For example, a patient prescribed dual OAD therapy who is adherent to 1 OAD (eg, glipizide) and not adherent to the other (eg, metformin) will be classifi ed as adherent. Since PDC methodology requires supply of only 1 medication on each given day, multi-drug regimens, as is common with OAD, provide more opportunities to be adherent compared with monotherapy, as is the case with statins. The decreased prevalence of adherence to statins compared with DRIs may refl ect a greater prevalence of intolerance or negative patient beliefs, such as concerns regarding possible side effects. Patient beliefs of lower perceived risk of myocardial infarction, expected short treatment duration, and concern about potential harm from statins have been shown to be negative predictors of statin adherence.38

The generalizability of these observations to other populations may be limited due to the complex nature of adherence. Adherence is impacted by many factors not previously mentioned or described here, such as gender,36,42 ethnicity,36-38,41,52 comorbities,37,39,41 pharmacy type,37 and level of cost sharing.42,53-56 However, this study selected data from a large geographically diverse population.


Using the CMS-adopted PDC adherence methodology, adherence rates differ by insurance type. Medicare Part D and commercial plans have a greater proportion of patients who are adherent to the major chronic therapies for diabetes, hypertension, and dyslipidemia compared with Medicaid and self-insured health plans. OAD and DRI adherence measurements were greater than adherence to statins for all insurance types. Measurements were stable over time, with commercial and Medicare Part D populations showing improvements from 2012 to 2013. The measurements indicate that many patients are nonadherent to these medication categories. Health plan sponsors should start or continue to increase promotion of intervention efforts aimed at improving medication adherence and health outcomes.