Predictive Risk Model for Early Post-Treatment Acute Care Use in Adolescent and Young Adult Patients With Cancer.

Abstract

Adolescent and young adult (AYA) patients with cancer (ages 15-39 years at diagnosis) continue to have health care needs after completing treatment. Risk models to predict acute care events (ACEs) in this population may lead to earlier identification of high-risk populations. This study aimed to develop and validate a clinical risk model to predict ACEs in AYA patients with cancer during the early post-treatment period. Using the University of North Carolina Lineberger Cancer Information and Population Health Resource, we identified AYAs diagnosed between 2006 and 2018 who were 2-5 years after diagnosis. The primary outcome was any ACE, defined as either hospitalization or emergency department (ED) visit. Patients were randomly assigned to development (70%) and validation (30%) cohorts. Logistic regression models were developed using stepwise inclusion of predictive variables. Model performance was evaluated using sensitivity, specificity, positive predictive value (PPV), and AUC. The study cohort included 7,393 patients (development = 5,276, validation = 2,217) with an average follow-up of 1.9 years. The most common cancers were breast (17%), thyroid (14%), and gynecologic (10%). Patients with ACEs (n = 3,572, 48%) were more frequently female, Black, and publicly insured. Defining high risk as the top 20% of scores, in the validation cohort, the selected model achieved an AUC of 0.76, a specificity of 0.94, a sensitivity of 0.34, and a PPV of 0.84. This is the first validated risk model to predict ACEs in AYA patients with cancer during the early post-treatment period. The model, designed for seamless electronic health record integration, enables early identification of high-risk patients, presenting opportunities for targeted interventions to reduce acute care use. Further validation across different health care systems is planned and will expand clinical applicability.