Assessing Treatment-Related Toxicity Using Administrative Data, Patient-Reported Outcomes, or Physician-Graded Toxicity: Where Is the Truth?

The accurate determination of the frequency and severity of treatment-related complications is vital to informing patients and clinicians in their decision-making process. In published studies, complications are assed via administrative data, patient-reported outcomes, and physician-graded toxicity, each with their strengths and limitations. Administrative data provide a vast, accessible history of patient data, but are limited in the ability to accurately capture diagnosis and causality, and are subject to differing interpretations of billing codes. Patient-reported outcomes provide direct and nuanced descriptions of both symptoms and bother; but are by definition subjective, affected by nonrespondents, and results (scores) are often difficult to interpret for patients and clinicians alike. Physician-graded toxicity is a relatively more objective measure, but relies on both clinicians fully assessing all relevant symptoms and patients accurately reporting them to the clinician. Understanding these strengths and limitations will help clinicians become more informed readers of the published literature.