Tweetable abstract Recent RWD studies suggest that the performance of immunotherapy in NSCLC is not as good as that seen in RCTs. However, analyses using RWD (and their interpretation) require careful appraisal and quantification of possible sources of bias.