Sociodemographic differences in pain assessment and management in cancer care across two academic centers.

113 Background: Pain management is essential to quality cancer care. Prior studies suggest gaps in pain reporting and management, though variability across regions and institutions is less described. This study examines pain reporting, surveillance, and management by sociodemographic characteristics in patients with cancer at two demographically distinct academic cancer centers. Methods: This retrospective cohort study examines electronic health records of patients receiving radiotherapy between 2013-2022/21 at two hospital systems: A, B. Linear and logistic regression analyses assessed inpatient pain level reporting, frequency of pain scores, and receipt of opioid and IV analgesics across sex, language preference, and race at each institution. Results: 8797 patients were treated at A and 3270 at B. Most were male (A 54.8%, B 50.6%). Patient-reported race differed by site (White, A 59.6%, B 71.4%; Black, A 6.5%, B 16%; Asian, A 18.6%, B 6%). English language preference was common (ELP; A 86.6%, B 87.2%). Patients were younger at A (median 56 years [IQR, 38-66]) than B (64 [56-71]). Median pain scores were higher at B (4 [IQR 0-6]) than A (2 [0-5]). Women had higher pain scores than men (A β 0.32, p < 0.001; B 0.34; p < 0.001). ELP patients reported greater pain compared to non-ELP (NELP) patients (A β 0.78, p < 0.001; B 0.84; p < 0.001). Compared to White patients, Black (A β 0.88, p < 0.001; B 0.32; p < 0.001) and Native American (A β 0.59, p < 0.001; B 1.06, p < 0.001) patients reported greater pain and Asian patients reported lower pain (A β -0.72, p < 0.001; B -1.03; p < 0.001). Time between pain assessment was similar across sites (A 2.08 hours [1.00-3.79], B 2.64 [1.23–5.59], longer for NELP patients (A β 0.02, p < 0.001; B 0.02, p < 0.001), but highly variable by sex and race across institutions. Controlling for pain score, treatment varied by demographic group. On a per admission basis, women received fewer opioids (A OR 0.79, p < 0.001; B 0.84, p = 0.002), but variable IV medication (A 0.94 p = 0.02; B 0.98, p = 0.65). NELP was associated with fewer opioids at A (0.76, p < 0.001) but not B (1.10, p = 0.27). NELP was associated with less IV medication at B (0.87, p = 0.04) but not at A (0.98, p = 0.51). Opioid receipt and IV medication also differed among racial groups between the institutions. Conclusions: This study highlights sociodemographic differences in pain reporting and management as well as variability across institutions, suggesting regional, cultural, and institutional differences.