Predictive Algorithm to Improve Financial Toxicity Monitoring in Patients with Cancer

Financial hardship, which is affected by socioeconomic and clinical factors such as health insurance, employment, social support, income, cancer type and stage, and treatment plans, impacts approximately 25% of patients with cancer. The negative effects of cancer-related financial hardship on patients, known as “financial toxicity,” can lead to increased anxiety and depression, treatment nonadherence, and even early mortality.

One proven solution to identify and prevent financial toxicity in patients with cancer is oncology financial advocacy (OFA), which assesses patients’ financial needs and connects them to hospital, community, and government relief programs. OFA, however, is currently underutilized due to lack of awareness, understaffing, and difficulty identifying at-risk patients. Additionally, present screening tools to assess financial toxicity, like the Comprehensive Score for Financial Toxicity (COST), rely on patient-reported experiences and have not been validated for assessing diverse patients.

To better address the issue of identifying patients at risk for financial toxicity, PC3I Faculty Meredith Doherty, PhD, LCSW; Ravi Parikh, MD, MPP, FACP; and Tamara Cadet, PhD, LICSW, MPH are developing a predictive algorithm to monitor toxicity in patients with cancer. The tool utilizes machine learning to access real-time electronic health record (EHR) data and financial records to identify at-risk patients and refer them to on-site financial advocates. After comparing various predictive models, the best performing algorithm will be validated with a group of 100 patients who received chemotherapy in the previous 4 weeks. By efficiently and accurately identifying patients at-risk for financial toxicity, this project will reduce negative effects of financial hardship and improve health outcomes.

Pending successful implementation, this tool will result in earlier and greater use of financial advocacy services, therefore promoting improved access to cancer services. Additionally, because financial toxicity disproportionately impacts people of color and patients who are low-income, a successful intervention will improve health equity at Penn Medicine. Findings from this pilot study will be used to develop an R01 application that responds to the NCI Notice of Special Interest: Addressing Cancer-Related Financial Hardship to Improve Patient Outcomes (NOT-CA-22- 045) to develop and test a machine-learning-generated electronic health record-based “nudge,” creating default referral pathways for referring at-risk patients to financial advocacy or social work services.

Leonard Davis Institute (LDI); National Institute on Aging (NIA) K23: Mentored Patient-Oriented Research Career Development Awards