The generalizability paradox within palliative care clinical trials

We are increasingly recognizing that personalized advanced and chronic illness care requires meticulous assessment and management of supportive care needs across the entire disease trajectory. This requires drawing clinical decisions from a research evidence base that is presumably generalizable to a heterogeneous patient population, often with poor performance status, multi-morbidity, and a large symptom distress profile. As sometimes this is not the case, how do we improve evidence generation that can be consistently applied to all patients with advanced disease?