Observational analytical studies, especially retrospective case-control studies, are considered methodologically suspect by many investigators. In general, there are a number of limitations to the value of the results of case-control studies, and in assessing the validity of included cases and controls, a first question would be: To what extent would further follow-up alter the estimate of the association between a postulated risk factor and studied disease? In case of cohort studies, there is no bias of misclassification of individuals into exposed and unexposed. Because of their prospective nature, cohort studies are stronger than case-control studies when well executed but they also are more expensive. Moreover, cohort studies are unsuitable for uncommon diseases or diseases with low incidence in the population, and usually involve a large number of people.
In theory, well designed randomised controlled clinical trials (RCTs) avoid most of pitfalls of observational studies and categorized as level II or even level I evidence in case of large multicentric RCTs. It is surprising that nearly all systematic reviews, including Cochrane Collaboration reviews, concentrate on clinical trials. Little attention has been withdrawn towards the importance of pooling observational studies as cohort, case-control and ecological studies. Combining RCTs may provide more evidence; however combining data from observational studies is sometimes desirable, especially in studying the treatment of a chronic disease. In addition, identification of chronic diseases risk factors is currently the most crucial problem in the epidemiology of non-infectious diseases and calls for pooling the results of dispersed observational studies.
Key words: Meta-Analysis, Observational Studies, Randomised Controlled Trials. Article Language: Turkish English
|
