Remarks about systematic reviews of diagnostic tests

Abstract

From the start of the second half of the 20 th century, the evolution of medicine was seen to be accelerating, and the new millennium began with a fast pace. Doctors and other health professionals cannot survive unless they are up to date with the constant changes imposed by technology , especially in big cities. Over the last few decades, the avalanche of new drugs and modern methods of treatment has encouraged doctors to look for faster ways to follow these changes, as well as ways to verify the true efficacy of these new interventions. All of these changes led to evidence-based medicine , which has subsequently become known as evidence-based healthcare, when other health-care professionals are included in this. Evidence consists of nothing more than the results of assessments through scientific studies with reproducible methodological quality (for all the data described in the work). This is possible in intervention studies on high-prevalence diseases because they present large numbers of participants. However, in cases of low-prevalence diseases, the results give rise to doubts. This has led to the use of systematic reviews, since these are able to fulfill the need to evaluate intervention studies. However, this is only possible when studies are designed to have similar objectives and interventions and their methodological quality can be assessed. In a systematic review, it is desirable and often feasible to obtain the sum of statistical data from several studies, which is called meta-analysis. More recently, the same need has arisen in the field of diagnostics. The evolution of diagnostic equipment and the emergence of new laboratory kits with promises of faster, more accurate and less invasive diagnostic methods has been widely broadcasted in the media. This directly influences patients' opinions, and affects the people responsible for doing the examinations and governments. Patients obviously want to have access to the best examination that there is, while professionals and governments want to make sure that these new tests really are superior to the existing ones, so that the possibly high financial investment can be justified. Therefore, systematic reviews on diagnostic accuracy studies are considered to be of great relevance. Given the lack of consensus on the most appropriate way of conducting the systematic review method, the Cochrane Collaboration, a pioneer in implementing systematic reviews on intervention studies, decided to disseminate and encourage the development of systematic reviews on accuracy studies by creating a section aimed only at reviews on diagnostic accuracy studies. This new section forms part of the RevMan (Review Manager) software, which the Cochrane Collaboration maintains in order to guide the elaboration of reviews and enable production of meta-analysis whenever possible. This brilliant initiative has caught the attention of many researchers and has encouraged them to work in this field, but just as in any other movement in its initial phase, several unresolved issues still hinder the work. The tools for assessing the quality of individual diagnostic accuracy studies are different from those applied in intervention studies. This has generated great confusion among researchers and among editorial boards evaluating such studies. There are several published systematic reviews on accuracy studies that are full of inappropriate terms and were methodologically designed as if they were intervention studies. Furthermore, there is a lack of significant accuracy values, such as sensitivity, specificity and predictive values.