To the Editor Cancer registries are useful for observational studies but are limited in terms of clinical characteristics and do not provide specific information on treatment selection or intermediate outcomes such as recurrence. Treatment assignment is not random but is based on a number of patient- and clinician-related factors. Because the propensity scores developed with administrative data rely on the availability of measured covariates associated with the treatment assignment, these scores do not necessarily balance patient characteristics and can potentially increase the imbalance in other clinically important covariates, leading to a substantial overestimation of treatment effects.1
Schernberg A, Haie-Meder C, Chargari C. Level of Evidence and Ethical Considerations for Locoregional Treatments in Metastatic Cervical Cancer. JAMA Oncol. 2019;5(4):574–575. doi:10.1001/jamaoncol.2018.5096
Customize your JAMA Network experience by selecting one or more topics from the list below.
Create a personal account or sign in to: