Meta-Analysis: Chronic Disease Self-Management Programs for Older Adults

Context Do self-management programs improve outcomes of adults with chronic conditions? Contribution This meta-analysis summarizes data from 53 randomized, controlled trials of self-management interventions for adults with diabetes mellitus, hypertension, or osteoarthritis. Self-management helped reduce hemoglobin A1c and blood pressure levels in diabetes and hypertension, respectively, but had minimal effect on pain and function in patients with arthritis. The authors could not identify any self-management program characteristics that predicted successful outcomes. Cautions The authors found evidence of possible publication bias. Implications Self-management programs may improve some outcomes in patients with some chronic diseases, but how to design an optimal program is not yet clear. The Editors Chronic diseases are conditions that are usually incurable. Although often not immediately life-threatening, they place substantial burdens on the health, economic status, and quality of life of individuals, families, and communities (1). In 1995, 79% of noninstitutionalized persons who were 70 years of age or older reported having at least 1 of 7 of the most common chronic conditions affecting this age group: arthritis, hypertension, heart disease, diabetes mellitus, respiratory disease, stroke, and cancer (1). Of these 7 conditions, arthritis is most prevalent, affecting more than 47% of individuals 65 years of age and older (2). Hypertension affects 41% of this population, and 31% of this group has some form of heart disease (of which ischemic heart disease and a history of myocardial infarction are major components). Diabetes mellitus affects approximately 10% of persons 65 years of age and older and increases the risk for other chronic conditions, including ischemic heart disease, renal disease, and visual impairment (2). Enthusiasm is growing for the role of self-management programs in controlling and preventing chronic disease complications (3-5). Despite this enthusiasm, experts do not agree on the definition of what constitutes a chronic disease self-management program, which elements of self-management programs are essential regardless of the clinical condition, or which elements are important for specific conditions. Several recent reviews on chronic disease self-management interventions have been published, including 2 Cochrane collaborations (6-13). Almost all have been disease-specific. One Cochrane review (12) concluded that there was insufficient evidence to assess the benefit of dietary treatment for type 2 diabetes mellitus programs, but exercise programs led to improved hemoglobin A1c values. A second Cochrane review of self-management for hypertension (11) used unpooled results to conclude that a reduction in the frequency of medication dosage increased adherence. There was not, however, consistent evidence of decreased blood pressure. Almost all previous reviews have been disease-specific or addressed specific intervention components within specific disease conditions (14-17). Two recent reviews assessed self-management programs across conditions. The first review provided a qualitative evaluation of self-management interventions across 3 conditions: type 2 diabetes mellitus, arthritis, and asthma (18). This review, which presented an overall optimistic assessment of self-management interventions, did not, however, include a quantitative synthesis of the data, nor did it address the issue of publication bias. The second review quantitatively assessed 71 trials (both randomized and nonrandomized) that included a self-management education program for patients with asthma, arthritis, diabetes mellitus, hypertension, and miscellaneous other conditions. Meta-analysis found statistically significant benefits for some outcomes within conditions. The authors could not detect meaningful differences in the effectiveness of the programs because of the varying intervention characteristics, such as the use of a formal syllabus, the type of program facilitator, the number of program sessions in which patients participated, and the duration of the program (19). In our review, we sought to quantitatively assess chronic disease self-management programs for older adults within and across disease conditions. We used empirical data from the literature to address 2 research questions: First, do chronic disease self-management programs result in improved disease-related outcomes for specific chronic diseases of high prevalence in older adults? Second, if self-management interventions are effective, are there specific components that are most responsible for the effect, within or across disease conditions? To address these questions, we focused on evaluating the effect of self-management programs for the 3 chronic conditions that have been most commonly studied in controlled trials of older adults: osteoarthritis, diabetes mellitus, and hypertension. Methods Conceptual Model Because there is no accepted definition of what constitutes a chronic disease self-management program, we used an intentionally broad definition to avoid prematurely excluding relevant studies. On the basis of a conceptual framework derived from the clinical literature and from discussions with social scientists with expertise in self-management, we defined chronic disease self-management as a systematic intervention that is targeted toward patients with chronic disease. The intervention should help them actively participate in either or both of the following: self-monitoring (of symptoms or of physiologic processes) or decision making (managing the disease or its impact through self-monitoring). We attempted to understand the characteristics particular to chronic disease self-management programs that may be most responsible for their effectiveness. On the basis of the literature and expert opinion, we postulated 5 hypotheses regarding the effectiveness of chronic disease self-management programs that feature the following characteristics: Tailoring. Patients who receive interventions tailored to their specific needs and circumstances are likely to derive more benefit than those receiving interventions that are generic. roup setting. Patients are more likely to benefit from interventions received within a group setting that includes others affected by the same condition than from an intervention provided in some other setting. Feedback. Patients are more likely to derive benefit from a cycle of intervention followed by some form of individual review with the provider of the intervention than from interventions where no such review exists. Psychological emphasis. Patients are more likely to derive benefit from a psychological intervention than from interventions where there is no psychological emphasis. Medical care. Patients who receive interventions directly from their medical providers (physicians or primary care providers) are more likely to derive benefit than those who receive interventions from nonmedical providers. Outcome Measures From the literature, we identified outcomes of interest to include the following: clinical outcomes, such as pain and function for osteoarthritis; measures that have strong links to clinical outcomes, such as hemoglobin A1c levels, fasting blood glucose levels, and patient weight for diabetes and blood pressure for hypertension; and intermediate outcomes, such as knowledge, feeling of self-efficacy, and health behaviors that are postulated to be related to clinical outcomes. Databases for Literature Search We used several databases and published documents to identify existing research and potentially relevant evidence for this report. For our primary source of citation information from 1980 until 1995, we used An Indexed Bibliography on Self-Management for People with Chronic Disease (20), published by the Center for Advancement of Health in association with the Group Health Cooperative of Puget Sound; we obtained any studies not listed in the bibliography (including those published later than 1995) by searching MEDLINE, PsycINFO, and CINAHL. We also used the Cochrane Library (its database of systematic reviews and the central register of controlled trials); the Assessment of Self-Care Manuals, published by the Evidence-based Practice Center at the Oregon Health Sciences University (21); and 77 other previously completed reviews relevant to this project. We retrieved all relevant documents referenced in these publications, and we updated our search in September 2004. Each review discussed at least 1 intervention aimed at chronic disease self-management. We also searched the Health Care Quality Improvement Projects database, maintained by the U.S. Centers for Medicare & Medicaid Services. This database contains reports known as narrative project documents, each of which describes an individual research project conducted by a Medicare Peer Review Organization; most projects in this database are not published elsewhere. Each report includes the project's background, aims, quality indicators, collaborators, sampling methods, interventions, measurement, and results. A complete description of our literature search has been reported elsewhere (22). Article Selection and Data Abstraction Two trained physician reviewers, working independently, conducted the article selection, quality assessment, and data abstraction; disagreements were resolved by consensus or third-party adjudication. Articles were not masked. We included all randomized trials that assessed the effects of an intervention or interventions relative to either a group that received usual care or a control group among the elderly and for our 3 conditions. Most studies compared their intervention with usual care or with a control intervention designed to account for the added attention received in the intervention (such as attending classes on vehicle safety instead of attending classes on self-management). Because our analysis was funded by the Centers for M