Reporting treatment outcomes in observational data: A fine balance.

Analyses of observational data have been gaining momentum in the evaluation of ever increasing spectrum of disease modifying therapies for multiple sclerosis. While high cost-effectiveness and generalisability represent their main advantages, these studies are also burdened with high risk of bias that may lead to erroneous conclusions. In this viewpoint, we highlight the key role of rigorous and transparent statistical methodology in the studies of observational data and encourage its thorough editorial scrutiny.