Comparing the Ability of Physician-Reported Versus Patient-Reported Performance Status to Predict Survival in a Population-Based Cohort of Newly Diagnosed Cancer Patients

Abstract

Aims: Performance status is an important prognostic tool in cancer. In oncology, the Eastern Cooperative Oncology Group (ECOG) measure is commonly used. Patient-reported functional status (PRFS) is an emerging method that allows patients to provide an estimate of their function; however, there is limited information about its prognostic significance. The aim of this study was to compare the predictive validity of functional status as reported by patients and physicians in relation to the observed survival after a new cancer diagnosis.

Materials and methods: This was a retrospective, population-based study using observational data of newly diagnosed patients in Ontario, Canada. We included patients who had both PRFS and ECOG recorded on the same day during an outpatient cancer clinic visit between March 2013 and March 2018. The dataset was randomly divided into 60% training and 40% validation cohorts. One-year survival was estimated by modelling clinical characteristics with PRFS, with ECOG, and alone.

Results: In total, 13 045 patients met the inclusion criteria. Covariates were similar at baseline for both training and validation datasets. PRFS and ECOG scores were statistically significant predictors of overall survival. Higher PRFS and ECOG scores were both associated with inferior survival, hazard ratio = 1.71 (P < 0.0001) and hazard ratio = 1.90 (P < 0.0001), respectively. Models that included either PRFS or ECOG scores outperformed the model with clinical characteristics only. C statistics were 0.836, 0.839 and 0.811, respectively.

Conclusions: PRFS adds to survival modelling and is equally predictive as the ECOG scale. PRFS may be used instead of ECOG in clinical or research settings for survival estimation.

Keywords: Cancer; ECOG; patient-reported outcome; performance status; prognosis; survival.