Randomised clinical trials are still the gold standard when it comes to the development of new drugs. There are, however, important limitations to trials. Such as, patients included in clinical trials are often not representative of the general population, which limits the applicability of trial results in clinical practice. In this article, the advantages and disadvantages of observational data are discussed. For example, observational data are generally more representative of the general population and can include large numbers of patients. However, there are important biases that should be considered when performing observational studies. Of these, so-called ‘confounding by indication’ is the most important form of bias, which means that reasons for certain treatment allocations are also associated with outcomes of treatment, which can disrupt the analyses. In summary, real-world data can add to clinical trials, but bias in these studies cannot be completely resolved. For this reason, clinical trials remain essential and should attempt to use less stringent inclusion criteria in order to improve the generalisability of their results.
(BELG J MED ONCOL 2023;17(1):11–4)