A suggestion for quality assessment in systematic reviews of observational studies in nutritional epidemiology

OBJECTIVES It is important to control the quality level of the observational studies in conducting meta-analyses. The Newcastle-Ottawa Scale (NOS) is a representative tool used for this purpose. We investigated the relationship between high-quality (HQ) defined using NOS and the results of subgroup analysis according to study design. METHODS We selected systematic review studies with meta-analysis which performed a quality evaluation on observational studies of diet and cancer by NOS. HQ determinations and the distribution of study designs were examined. Subgroup analyses according to quality level as defined by the NOS were also extracted. Equivalence was evaluated based on the summary effect size (sES) and 95% confidence intervals computed in the subgroup analysis. RESULTS The meta-analysis results of the HQ and cohort groups were identical. The overall sES, which was obtained by combining the sES when equivalence was observed between the cohort and case-control groups, also showed equivalence. CONCLUSIONS The results of this study suggest that it is more reasonable to control for quality level by performing subgroup analysis according to study design rather than by using HQ based on the NOS quality assessment tool.