For this report, the committee reviewed the published medical literature to determine whether nutrition-based therapies have a beneficial impact in preventing or treating diseases among Medicare beneficiaries (i.e., predominantly older-aged persons). The fundamental question was one of efficacy: For the disease or condition under study, does manipulation of diet in older persons have beneficial effects? To answer this question, the committee reviewed the published medical literature with a focus on original research and systematic reviews. To identify relevant literature, systematic searches of online databases were conducted and key references were elicited from committee members, other experts (invited speakers and consultants), and the reference lists of publications.
Early in the course of this study, the committee developed and agreed upon the literature search strategy. The initial focus was a comprehensive search of relevant online databases. The online database Medline was selected because it offers an effective and efficient means of searching the medical literature. Medline was accessed through Ovid, a commercial database vendor, and through Grateful Med of the National Library of Medicine (NLM). The search strategy was individualized by topic area or disease state. Searches were limited to human subjects, age ≥65 and/or the entire population, the English language, and 1976–present. Searching also incorporated standardized terminology found in NLM’s Medical Subject Headings (MeSH) and the MeSH tree structures.
All retrieved citations were reviewed to determine whether the citation was relevant to this report and if relevant, whether to obtain the full
paper. All citations were entered into the study’s bibliographic database, which at the conclusion of the study contained approximately 850 references to abstracts, journal articles, books, congressional reports, and conference proceedings.
The evidence reviewed was heterogeneous (i.e., the types of nutrition-based therapies varied as did the types of outcomes reported). The latter included both “clinical” outcome variables (e.g., number of hip fractures, hospital admissions) and “intermediate” outcome variables (e.g., bone mineral density, blood pressure). The vast majority of studies report the effects of nutrition-based therapy on intermediate outcomes. The committee also evaluated the full spectrum of nutrition-based therapies. For example, for common outpatient conditions such as hypertension, the therapies included behavioral interventions that involved counseling, feeding studies that controlled food intake, and trials of dietary supplements. Multidisciplinary interventions were also evaluated as long as there was a nutritional component to the program. The type of individuals who conducted the study interventions and thus provided the nutrition therapy was not uniform: whereas a registered dietitian provided the nutrition-based therapy in most instances, there were several studies in which other personnel (e.g., nurses) administered the intervention. The setting for interventions varied as well (e.g., outpatient versus home-based, group versus individual counseling).
Although research conducted among the elderly was of primary interest, evidence from studies in non-elderly population groups, as well as evidence from studies that covered a broad age range of individuals—both non-elderly and elderly—was also considered. In many instances, fewer studies were conducted in elderly population groups than in the non-elderly. In this setting, the committee considered studies in non-elderly populations as relevant to its deliberations as long as the evidence appeared consistent across the adult lifespan.
In preparing this report, the committee evaluated and then classified the type and quality of available evidence for each of the diseases or conditions under study. Several classification schemes are available (NIH, 1998; USPSTF, 1995; Yusuf et al., 1998). For this report, the committee adapted the approach used by the National Institutes of Health (NIH)
task force on the identification, evaluation and treatment of overweight and obesity (NIH, 1998). The only substantive difference between the approach in this report and that of the NIH task force is in the committee’s classification of systematic reviews. In addition to meta-analyses, other types of systematic reviews were considered as long as the review involved a comprehensive search of the medical literature. A separate category, “systematic review,” was also created—rather than having to combine these reports with randomized trials. The classification scheme, ordered from lowest to highest quality of evidence, is summarized below:
Consensus reports and guidelines. In many instances, expert panels convened by professional societies or government agencies evaluated and synthesized the available evidence. In addition to describing the available data, these reports have the advantage of proposing clinical management strategies, which integrate nutrition-based therapy in the overall management of patients.
Observational studies. In contrast to consensus reports and guidelines, this category pertains to original research and includes both observational studies (i.e., case series, case-control studies, cohort studies) and nonrandomized trials of nutrition-based therapies. Because individuals were not randomly allocated to the treatment under study, the potential for bias exists. Hence, this grade of evidence is less persuasive than randomized, controlled trials.
Systematic reviews. In recent years, systematic approaches have been used to identify, evaluate, and synthesize evidence. This report includes meta-analyses of clinical trials as well as other systematic reviews as long as these reviews performed a comprehensive search of the medical literature. Systematic reviews are considered more complete and less subject to bias than expert opinion and previous types of reviews.
Some evidence from randomized, controlled trials. For several conditions, clinical trials were conducted; however, the number or size of the trials was small and/or the data were somewhat inconsistent.
Extensive evidence from randomized, controlled trials. This category applies to those nutrition-based therapies for which there are several clinical trials documenting a clinically meaningful health benefit from the therapy. Furthermore, the data from these trials had to be generally consistent. Overall, this is the strongest category of evidence supporting nutrition-based therapy.
At its initial meeting, the committee identified common medical conditions for which nutrition-based therapy was generally recommended,
as well as other conditions for which the evidence appeared much less persuasive. The committee then assigned working groups to review and summarize the published medical literature. The committee also sought expert advice from people who were not committee members, primarily through presentations at subsequent committee meetings. At several points, the working groups presented their assessments to the entire committee. These presentations included an overall assessment of the efficacy of nutrition-based therapy in the elderly, as well as practical aspects of implementation (e.g. Who is qualified to implement the therapy—a highly skilled professional or a variety of general health professionals? What is the appropriate setting? What number of contacts is required; and so forth). The committee then considered the costs of nutrition-based therapy and identified potential cost offsets. Methods pertaining to the cost analyses are presented separately in chapter 14.
NIH (National Institutes of Health). 1998. Clinical guidelines on the identification, evaluation, and treatment of overweight and obesity in adults—The evidence report. Obes Res 2:51S–209S.
USPSTF (U.S. Preventive Services Task Force). 1995. Guide to Clinical Preventive Services, 2nd ed. Report of the U.S. Preventive Services Task Force. Washington, D.C.: U.S. Department of Health and Human Services, Office of Public Health, Office of Health Promotion and Disease Prevention.
Yusuf S, Cairns JA, Cann AJ, Fallen EL, Gerch BJ. 1998. Evidence-Based Cardiology. London: BMJ Publishing Group.