Handling of statistical heterogeneity in meta-analyses of observational studies on treatment comparative effectiveness

For some diseases, several treatments are available and it is therefore necessary to compare their effectiveness. When several studies have been published on such comparison, it is becoming common practice to combine the results through a method called meta-analysis. Combining results is however challenging owing to differences in populations and methodologies between studies. Not to consider such difference may invalidate the results.

Definitions of schizophrenia relapse as a study outcome and impact on the comparative effectiveness of oral versus injectable antipsychotics in the real-world setting

Schizophrenia is a devastating disease for which there is no cure. Available treatments aim at minimizing the symptoms. However, patients often stop taking their tablets, which leads to a worsening of their symptoms, referred to as relapse. In order to avoid the problems of discontinuation, some treatments are administered through an injection. Several studies have been conducted to compare the benefits of tablets versus injectable treatments on the risk of relapse. However, to combine those results is a challenge because relapse is not defined the same way across those studies.