OBJECTIVES: Cancer patients, at the time of diagnosis, often possess multiple comorbid conditions that may or may not contribute to their ability to initiate and continue treatments, manage adverse events and achieve desired outcomes. Several instruments have been developed using real-world data from claims or health records to quantify comorbid conditions and their impact on costs and clinical outcomes. In retrospective, observational real-world research studies where access to medical history may be incomplete, comorbidity assessment that leverages the patient’s current medication profile may be of value.

METHODS: A comprehensive search of PubMed was undertaken to identify studies of comorbidity assessment among patients with cancer treated in real-world settings published within the last 10 years. A keyword search using terms associated with cancer, real-world settings and comorbidity assessment instruments was executed, yielding 41 potential publications. Titles and abstracts of these publications were reviewed to identify 30 studies that were considered for this evaluation.

RESULTS: Most studies captured comorbidities using standardized instruments, particularly the Charlson Comorbidity Index. Other approaches included targeted comorbidity searches, patient-reported information, as well as linkage of electronic health records and claims data. None of the included publications described use of patients’ medication histories to derive chronic conditions.

CONCLUSIONS: Assessment of comorbidity is important in cancer research, as it may influence treatment selection, confound patient outcomes and preclude clinical trial eligibility. In the real-world setting, patients’ medication history may be readily available and there is an opportunity to explore how this information could be used to measure comorbidities through validated approaches. Such techniques would need to consider how to differentiate off-label, acute and non-specific use among cancer populations. Nonetheless, a valid and reliable comorbidity assessment derived from baseline medication profiles could provide meaningful insights for cancer populations.