A MEDLINE Survey Study on Reporting of Design Effect and Degree of Compliance with STROBE

Abstract

Objective: The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) Statement is a checklist and has been developed to improve the reporting ofobservational studies. The design effect for observational studies is an important factor in samplesize determinations however is not considered under STROBE. In this study, it is aimed to evaluate compliance with STROBE in the articles reviewed in MEDLINE and highlight the necessity of thinking out design effect. Material and Methods: A database search has been performedon the observational studies n=342, that are free full text, written in English and published inMEDLINE. The frequency of compliance with STROBE was examined and an auxiliary reviewwith respect to the design effect was done by one reviewer. Results: 342 articles evaluated,224 (65.5%) were cross-sectional and 118 (34.5%) cohort studies. The best compliance (above 90%) was for items 1(b), 2 and 14(a) while the worst fit (under 20%) was for items 12(d),13(c) and 16(c). In n=6 (1.8%) cross-sectional studies design effect was reported. Conclusion:In terms of the quality of the reporting of observational trials, current observational studies inMEDLINE could still benefit from increased reporting of methodologic details including the reporting design effect, sample size determinations, methods taking account of sampling strategyfor cross-sectional studies, power analyses and consideration of potential bias. For items undermethods section (items 4–12) compliance is lower than other items. Discarding design effect hasa potential of source of bias. Authors should report the design effect and can also be considerwithin the items 4 or 9 or 10.