Research design and methods of quantitative synthesis of medical evidence.

Abstract

To review the scientific principles, strengths, and limitations of research designs and methods of quantitative synthesis of medical evidence. We used MEDLINE to perform a systematic search for literature using the keywords research design, epidemiology, and biometry. Journals searched included six major journals in obstetrics and gynecology and three in general medicine. These sources were supplemented with texts and reviews from the general medical literature. We reviewed the publications identified by our search and evaluated critically the relevant reports. We summarized objectives and scientific guidelines for the common research methodologies and outlined their advantages and disadvantages. The standard of clinical research design is the randomized controlled trial (RCT), which, if performed with sufficient methodologic rigor, is least likely to have serious biases. Cohort, case-control, and cross-sectional studies are common observational studies used in reproductive health; such observational studies are more susceptible to biases that can distort the researcher's results and conclusions. Descriptive studies such as case series and case reports are often interesting as clinical vignettes but have limited scientific merit. Methods for quantitative synthesis of medical evidence, including meta-analysis, decision analysis, and cost-effectiveness analysis are being used with increased frequency in the reproductive health literature to summarize medical evidence. Various research methods have their own inherent advantages and disadvantages. An understanding of the scientific principles of these methods will enable the clinician to evaluate medical evidence critically.