Methods in Nutrition Research

Abstract

Many different research designs are available for investigating nutrition-related questions. Each one has strengths and weaknesses. Much nutrition-related research is carried out using epidemiological study designs. Many epidemiological studies use observational designs; these include prospective cohort studies, case-control studies, cross-sectional studies, population studies, and historical studies. Results from randomized controlled trials (RCTs) are internally valid and are usually more reliable for causal inference than those from observational epidemiological studies. However, many kinds of nutritional studies do not fit well into the RCT framework. A recent development is the use of systematic reviews and meta-analyses. First, the literature is reviewed according to strict criteria. Findings from the studies are then analyzed in order to make pooled risk estimates. These gain power by increasing sample size; however, variables and concepts will vary across studies. This strategy can be valuable in evaluating consistency and in considering the possibility of residual confounding when only observational studies are available. Experiments on animals are often used for nutrition-related research. However, animal models have many differences from the human situation. Much mechanistic research is carried out where researchers study the details of body functioning such as physiology and biochemistry. This can illuminate detailed cellular and molecular relationships, but integrating such findings with the biology of intact humans is problematic. There is evidence of conflict of interest in studies that potentially affect profit of the research funder, most typically in industry funding. This indicates a need for particular care in interpreting such studies. A sufficient pool of research funds, independent of the food industry, is highly desirable to support needed research. Research studies involving human subjects require approval from an ethics committee.