Longitudinal risk of physical functional impairment and health care utilization in cancer survivors within 3 years of diagnosis.

11081 Background: Despite known physical functional (PF) impairment in cancer survivors, patterns of its development and relationships to healthcare utilization remain poorly characterized. We examined PF impairment trajectories among outpatient cancer survivors in a large Midwestern academic health care system during the first three years after a new cancer diagnosis. We hypothesized that PF impairment would be common, vary by extent of disease, cancer care phase, with differing patterns of rehabilitation services (RS) referral, emergency department (ED) and cancer center urgent care (CCUC) visits. Methods: We included 2,254 adults with a cancer diagnosis, treated in medical oncology clinics, and completed 2 or more Patient Reported Outcome Measurement Information System – Physical Function (PROMIS-PF) surveys within a 3-year period. A PROMIS-PF T-score < 40 represents moderate – severe impairment. We used group-based trajectory modeling (GBTM) to identify patterns of impairment likelihood (0.7 threshold for accurate fit) in the 3 years after cancer diagnosis. Trajectory models identified the demographic profile, cancer type, stage, and treatment intent (intent to cure, non-intent to cure, other) most associated with each trajectory. We assessed trajectory associations with rates of RS referrals, and visits to ED and CCUC during the study period with multiple logistic regression (p<.05). Results: On average, participants were newly diagnosed within 11 months (s.d. 10.92) before enrollment in the study. We determined 4 PF impairment trajectory patterns during the first 3 years after a cancer diagnosis: (1) low probability of impairment which remained steady over time (“Low Steady”, 64% participants); (2) decreasing probability (“Decreasing”, 13%); (3) increasing probability (“Increasing”, 9%); (4) high probability of PF impairment that remained high (“High Steady”, 14%). Patients in trajectories 2, 3, and 4 were significantly more likely to have ED and/or CCUC visits and receive RS referrals than those in trajectory 1, although differences by cancer type, stage, and Beacon module between groups 2 – 4 were mixed. Conclusions: Over one-third of cancer survivors experienced substantial impairment and 14% had persistent, severe impairment with higher rates of RS referrals, ED and CCUC visits. These patterns had few variations by tumor type and cancer care phase. Our data suggest that patients in Trajectories 2 – 4 should be targeted for early and frequent monitoring with PROMIS-PF and referral for rehabilitation services to address impairments –ideally within the first six months after cancer diagnosis. PROMIS-PF effectively identified impairment variation across patients and time. Future studies will explore whether screening and triaging for appropriate cancer rehabilitation referral and intervention improve impairment trajectories.