Oncologist perspectives on the acceptability, appropriateness, and feasibility of the Cancer and Aging Research Group (CARG) chemotherapy toxicity prediction tool for older adults

Introduction: The clinical uptake of validated chemotherapy toxicity predictor tools for older adults with cancer remains low. In this qualitative study, we sought to evaluate oncologist perspectives on the acceptability, appropriateness, and feasibility of the Cancer and Aging Research Group (CARG) chemotherapy toxicity predictor tool. Materials and Methods: We conducted semi-structured qualitative interviews with 18 medical oncologists in the M Health Fairview system to understand barriers to CARG tool use and implementation solutions. A trained researcher conducted interviews, and two coders analyzed interview transcripts to identify themes. Using an implementation science framework, we categorized oncologist perspectives into the outcomes of acceptability, appropriateness, and feasibility. Results: We identified four themes: (1) current methods for assessing chemotoxicity risk, (2) acceptability - perceptions of the CARG tool, (3) appropriateness - perceptions of the CARG tool in practice, and (4) appropriateness - integration of the CARG tool into oncologist workflow. Participants highlighted the relevance of the CARG questions but noted that certain treatment regimens required additional information (e.g., cardiac function or pre-existing neuropathy). They also noted that the topline results lack nuance and are difficult to interpret, with concern about the tool keeping up with the rapid pace of oncology advances. They pointed out that the tool was not applicable for every patient, especially newer treatments, and questioned the benefit over standard of care. However, they emphasized that a trusted colleague who could be a champion could aid buy-in, and a workflow priority was a seamless integration into the electronic health record. Discussion: Practicing academic and community-based medical oncologists noted several implementation considerations for the CARG tool. These data have implications for health systems and policymakers who wish to implement chemotoxicity predictor tools into routine practice, and for researchers and learning health systems in designing and conducting pragmatic trials.