AN AUDIT OF BEHÇET'S SYNDROME RESEARCH: A 10-YEAR SURVEY

F. Esen, E.K. Schimmel, H. Yazici, Y. YaziciMedical Student, Istanbul University, Cerrahpasa Medical Faculty, Istanbul, Turkey University of Mississippi Medical Center, Jackson, MS, United States Division of Rheumatology, Department of Medicine, Istanbul University, Cerrahpasa Medical Faculty, Istanbul, Turkey NYU Hospital for Joint Diseases, New York University School of Medicine, New York, NY, United StatesObjectives: We had the impression and some preliminary data suggesting that the use of diseased control groups and proper use of power calculations were neglected in manuscripts published even in widely read rheumatology and medicine journals We aimed to formally survey these and other methodological shortcomings in published manuscripts on one specific topic, Behçet's syndrome (BS), within the last decade. We reason that recognizing such methodological shortcomings will eventually lead to better quality clinical and basic science manuscripts. Methods: The articles published in the 15 highest-impact-factor rheumatology, ophthalmology, dermatology and general medicine journals between January 1999 and January 2009 were searched for original papers on BS. The study designs (study types and time element), control groups, gender ratio and mean age of patients, the use of power calculations and reporting of negative results were specifically tabulated. Results: Most of the studies on BS were cross-sectional (83%). Prospective longitudinal studies were few (7%). In a considerable proportion of papers, some basic demographic data was missing (21%). Power calculations were rare (3%) and were not even considered in clinical hypothesis testing. Diseased control groups were present in slightly over one half of clinical and laboratory original research while, as we had noted earlier, only 9% of genetic association studies included diseased controls. Only 12% of all manuscripts concerned mainly negative outcomes. Conclusion: A considerable number of the research articles on BS with methodological weaknesses in study design appeared in our better journals. The generalizibitiy of what we observed in BS to other research topics needs to be formally studied. References: 1. Esen F, Celik A, Yazici H. The use of diseased control groups in genetic association studies. Clin Exp Rheum 2009; 27(Supl.53): S4-5 2. Yazici Y, Adler NM, Yazici H. Most tumor necrosis factor inhibitor trials in rheumatology are undeservedly called `efficacy and safety' trials: a survey of power considerations. Rheumatology 2008; 47:1054-1057 Disclosure of Interest: None declaredCitation: Annals of the Rheumatic Diseases, volume 69, supplement 3, year 2010, page 475Session: Epidemiology, health services and outcome research (Poster Presentations )