engleski

Genetic association studies: the quality of study design and data analysis

Background. Genetic association studies suffer from some serious shortcomings in the study design and data analysis. Our aim was to evaluate the methodological quality of genetic association studies published in various scientific journals during 2003– 2005. Methods. English-language articles on genetic association studies published during 2003-2005 were identified in MEDLINE using the following search strategy: (“ Factor V” OR “ prothrombin” OR “ methylenetetrahydrofolate reductase” OR “ plasminogen activator inhibitor 1” ) AND “ polymorphism” . Only articles available as full text were included. Articles were evaluated by two independent reviewers using the checklist of 28 items. Each item was given a score and total score was calculated for each article. Maximum total score was 13. Results. 115 articles were identified, out of which only 57 were genetic association studies. Average total score was 7.59&plusmn ; ; ; ; 1.96. There was not a single article with a maximum total score. Total article score did not correlate with the journal impact factor (r=0.3971 ; 95% Ci 0.1547-0.5944 ; P=0.002). Total score did not differ across years (P=0.624). Only 3 studies provided power calculation. In 33/57 studies gene-dose effect was not presented, while other well known established risk factors were not reported in 32/57 studies. Hardy-Weinberg equilibrium was calculated only in 30/57 studies. Conclusions. Overall methodological quality of genetic association studies published in various scientific journals during the studied period is moderate to low and it does not depend on the quality of the journal. Guidelines for reporting genetic association studies are needed to the advantage of researchers, reviewers and editors.

Genetic association studies; genetic polymorphism

nije evidentirano