Objective.Data suggest that the use of disease control groups and proper use of power calculations were neglected in published reports. We surveyed these and other methodological shortcomings in reports published within the last decade about one specific topic, Behcet’s syndrome. We reason that recognizing such methodological shortcomings will lead to better quality clinical and basic science articles.Methods.Articles published in the 15 highest impact factor journals on rheumatology, ophthalmology, dermatology, and general medicine between January 1999 and January 2009 were searched for original reports on Behcet’s syndrome. Study designs (study types and time element), control groups, demographic data, use of power calculations, and reporting of negative results were specifically tabulated.Results.Most studies on Behcet’s syndrome were cross-sectional (83%). Prospective longitudinal studies were few (7%). In a considerable proportion of papers (21%), some basic demographic data were missing. Power calculations were rare (3%) even in randomized controlled trials and were not considered at all in clinical hypothesis-testing. Disease control groups were present in slightly over half of clinical and laboratory original research, while just 13% of genetic association studies included disease controls. Only 12% of all reports concerned mainly negative outcomes.Conclusion.A considerable number of the published research articles have methodological weaknesses. The generalizability of what we observed in Behcet’s syndrome to other research topics needs to be formally studied.
Power Calculations for ZIP and ZINB Models
