Associating Medication Adherence With Improved Outcomes: A Systematic Literature Review

This literature-based analysis supports association between medication adherence and better clinical, economic, and utilization outcomes in chronic disease populations, with variability across specific diseases.
Published Online: July 31, 2012
Kimberly A. Boswell, MD; Christopher L. Cook, PharmD, PhD; Steven P. Burch, RPh, PhD; Michael T. Eaddy, PharmD, PhD; and C. Ron Cantrell, PhD
Objectives: To evaluate the strength of the evidence addressing the relationship between medication adherence and outcomes across selected common chronic diseases and to identify gaps in the literature.

Study Design: Systematic literature review.

Methods: Original research articles relating medication adherence to clinical, economic, and/or utilization outcomes in North American study populations were examined for 12 diseases: coronary artery disease (CAD)/hyperlipidemia, heart failure (HF), hypertension, post–myocardial infarction (post-MI), bipolar disorder, depression, schizophrenia, diabetes, migraine, seizures, asthma, and chronic obstructive pulmonary disease (COPD). A 10-point rating system was developed to further quantify the level of evidence (range, -5 [lowest] to 5 [highest]).

Results: A total of 105 articles (none for migraines) were eligible. Most clinical outcomes were positively related to adherence, ranging from 64.3% to 100% positive across the diseases (81.0% positive overall); neutral relationships accounted for all remaining clinical outcomes. Although most economic and utilization outcomes were positively related to adherence (56.6% and 73.6%, respectively), results were based on fewer studies and demonstrated greater variability (including some negative relationships) relative to clinical outcomes. The level-of-evidence rating system demonstrated overall positive scores for most clinical, economic, and utilization outcomes, with exceptions being HF- and asthma-related economic outcomes, and COPD-related clinical and economic outcomes.

Conclusions: The highest levels of evidence were achieved for studies pertaining to post-MI, CAD/hyperlipidemia, schizophrenia, and diabetes, all disease states for which >10 studies assessed the adherence-outcomes association. Few studies were identifi ed for seizures, migraines, and COPD, illustrating gaps in the literature where future investigation is needed.

(Am J Pharm Benefits. 2012;4(4):e97-e108)
Results of this literature-based analysis indicate that improved medication adherence is associated with better clinical, economic, and utilization outcomes among patients with common chronic diseases.

  • Overall, the findings from the available studies suggest that efforts to improve medication adherence in clinical practice settings may translate into clinical, economic, and/or utilization benefits.

  •  Some diseases had stronger relationships than others, supporting the need for disease state–specific investigations to characterize the strength of the adherence-outcomes relationship.

  •  As the amount of literature addressing the adherence-outcomes link was limited, further research is needed on this clinically important topic.
Drug spending, estimated to account for 10% of the $2.2 trillion annual healthcare expenditure in the United States, has undergone an 89% increase since 2000,1 attributable to numerous factors that include not only the higher prices for certain types of medications,2 but also greater overall utilization across a wider range of conditions and for longer periods to treat chronic diseases. Increased drug utilization for managing chronic diseases, coupled with the potential for increasingly prolonged treatment, elevates the criticality of maintaining comprehensive medication management, of which adherence is a major part. Within the published literature, adherence to a medication regimen is generally defi ned as the extent to which patients take medications as prescribed by their healthcare providers,3 or the level of patient participation in an agreed-on medication regimen.4 The term “adherence” is generally preferred over “compliance” due to the passivity and potential lack of patient agreement implied by the latter terminology.3 Despite increased focus on developing treatments with the potential for improved adherence (such as simplifi ed regimens for hypertension and better-tolerated medications for mental health disorders),3 poor medication adherence continues to represent a major clinical practice issue, particularly in the context of multiple chronic morbidities.5

It is well appreciated that poor adherence to pharmacologic treatment regimens on the part of patients/caregivers is common, and therefore is a major barrier to improving overall morbidity and mortality,3 with expectedly high variability when considering specific types of diseases and treatments. In a review of 569 studies published over a 50-year period (1948 to 1998) that addressed adherence across 17 chronic and acute disease populations (excluding mental health disorders), the mean rate of nonadherence to therapeutic or preventive health measures (including not only medications but also lifestyle modifi cations and adherence to appointments, recommended screenings, and vaccinations) was 25%—ranging from 12% for human immunodefi ciency virus to 35% for sleep disorders.6 Nonadherence rates exceeded 30% in patients with pulmonary diseases (31%) and diabetes (33%).

Despite the established role of adherence-related education and monitoring in clinical research and practice settings, the extent to which different levels of adherence ultimately impact patient outcomes is a poorly understood phenomenon, due in part to the complexities inherent in designing studies with this research question in mind.7 In a published meta-analysis of the relationship between adherence and outcomes in patients receiving medical treatment (including medications and nonpharmacologic management such as dietary restrictions), data were analyzed for 19,000 patients with acute or chronic diseases from 63 eligible studies published over 30 years (1968 to 1998).7 The risk difference for outcomes (based on various types of disease-specific and nonspecific measures) when comparing high versus low adherence was 0.26 (95% confidence interval [CI], 0.20-0.32; P <.001), meaning that the proportion of patients achieving a good outcome significantly increased by 26% in adherent versus nonadherent populations. There was a stronger relationship between adherence and outcomes in chronic diseases compared with acute diseases (risk difference of 0.31 [95% CI, 0.22- 0.39] for 33 chronic diseases studies versus 0.20 [95% CI, 0.12-0.28] for 30 acute disease studies),7 an important finding considering the higher likelihood for low adherence among patients with chronic conditions.3

While it may be expected that poor patient adherence would translate into worse clinical outcomes, greater utilization of healthcare services, and higher economic costs,3,4 there have been few attempts to systematically assess the level of evidence supporting the relationship between adherence and outcomes across a variety of common chronic diseases. To address this issue, we conducted a systematic literature review to evaluate the strength of evidence within the literature for selected types of cardiovascular disease (CVD), and mental health, metabolic, neurologic, and pulmonary disorders associated with substantial disease burden in the United States,8 with consideration of clinical, utilization, and economic outcomes. This analysis also allowed the opportunity to identify gaps in the literature where there is a need for additional research.


Literature Search

An electronic search was conducted using resources from the US National Library of Medicine and National Institutes of Health PubMed ( within the following 12 disease states: coronary artery disease (CAD), heart failure (HF), hypertension, post–myocardial infarction (post-MI), bipolar disorder, depression, schizophrenia, diabetes, migraine, seizures, asthma, and chronic obstructive pulmonary disease (COPD). In addition, supplementary searches were conducted through Ovid MEDLINE, Web of Science, and Google Scholar. The search terms that were used, either alone or in combination, were “adherence,” “compliance,” “costs,” “noncompliance,” “outcomes,” “hospitalizations,” “utilization,” “economics,” “Medicaid,” “Medicare,” “persistence,” “prescription drugs,” and various terms for the selected disease states.

The search was limited to English-language articles that evaluated one of the disease states of interest and the association between adherence and outcomes. Articles published between 1974 and May 2010 were included. Exclusion criteria included articles that (1) did not address >1 of the aforementioned disease states; (2) did not describe original empirical research, such as review articles; (3) did not relate medication adherence to outcomes; and/or (4) described results for study populations

outside of the United States or Canada. Two independent reviewers were utilized to evaluate the articles for study selection. Any scoring discrepancies were resolved between the reviewers.

Data Extraction and Analysis

The 12 identified disease states were divided into 5 broad disease categories: CVD (including CAD, HF, hypertension, and post-MI), mental health (bipolar disorder, depression, and schizophrenia), metabolic (diabetes), neurological (migraine and seizures), and pulmonary (asthma and COPD). Within each disease category, the eligible studies identified were further classified determine whether clinical, economic, and/or utilization outcomes were assessed. The overall relationship between medication adherence and clinical, economic, and utilization outcomes was evaluated, as well as across all individual disease states. Since many studies evaluated and reported multiple outcomes as well as multiple disease states, the results were based on the total number of outcomes, rather than the total number of studies.

Determining and Rating Levels of Evidence

The eligible articles were assigned a level of evidence, as described by the Agency for Healthcare Research and Quality of the US Department of Health and Human Services: level I, evidence from >1 properly designed, randomized, controlled trial; level II-1, evidence from well-designed, controlled trials without randomization; level II-2, evidence from well-designed cohort or case-control analytic studies, preferably from >1 center or research group; level II-3, evidence from multiple time series with or without the intervention (dramatic results in uncontrolled trials could also be regarded as being of this level); and level III, opinions of respected authorities based on clinical experience, descriptive studies, or reports of expert committees.

PDF is available on the last page.
Diabetes Compendium
Oncology Compendium