Analyzing patient-reported outcome data in oncology care

Electronic Patient-Reported Outcome Measures (ePROMs) are collected routinely at an oncology institution throughout patients’ treatment and follow-up. Clinicians from the institution's head and neck team are investigating whether head and neck cancer patients with varying levels of comorbidity experience different changes in their ability to perform activities of daily living before and after radiotherapy. Clinicians hypothesize that patients with more severe baseline comorbidity experience worse changes in symptoms and quality of life scores from pre- to post-radiotherapy compared to those with no or only mild comorbidity. The institutions’ ePROM questionnaire for patients with head and neck cancer includes multiple symptom questions, plus the EuroQol EQ-5D-5L. The latter compromises five dimensions including usual activities; scores are measured on a 1 to 5 scale, where 1 indicates patients have no problems doing their usual activities and 5 indicates patients are unable to do their usual activities [1]. The team is considering the most appropriate method to analyze the data, given the ordinal nature of the endpoint and the longitudinal structure of the dataset.