Assessing the association of sociodemographic and site factors with patient experience.

396 Background: Patient experience (PEx) is part of the IHI’s quadruple aim. Better PEx has been associated with better performance on other health system metrics. Developing an optimal PEx improvement strategy requires understanding the extent to which patient socio-demographic and site/encounter factors explain the performance gap. Methods: We identified PEx surveys submitted for adult outpatient medical oncology encounters between 1/1/2024 and 12/31/2024 at a Boston-based comprehensive cancer center and its 7 community sites. A subset of 8 PEx measures were selected for analysis based on prior performance and investigator consensus. Multivariable logistic regression models were used to identify predictors of top box scores for each measure; covariates included in typical PEx analyses were added sequentially to a base model: 1 st ) age, sex, education, insurance, and social vulnerability index (SVI); 2 nd ) race/ethnicity; and 3 rd ) encounter site, encounter type (consult, established), time since first encounter, and number of prior PEx surveys submitted. Model performance was assessed using the area under the receiver operating characteristic curve (AUROC). Results: There were 35,031 eligible PEx responses (18% response rate): mean age 68 (SD 11), 55% female, 25% less than college education, 10% non-White, 49% commercial insurance, 43% from community sites. All 8 measures demonstrated an opportunity for improvement (i.e., top box score <90%; range 68%-86%). The Table shows the AUROC for 3 sequential models for each measure. All models demonstrated low accuracy (i.e., AUROC <0.65). Race/ethnicity was significantly associated with lower patient experience for all measures except wayfinding (P<.05), but did not meaningfully enhance discrimination (i.e., AUROC change >0.05) for any measure. Site/encounter factors modestly improved model performance only for wayfinding (AUROC change 0.082). Conclusions: For the 8 selected PEx measures, improving Institutional performance may require targeting system-wide processes and culture, whereas addressing equity-related gaps may require distinct focused interventions. Future research should seek to identify covariates that more strongly discriminate measures of PEx. PEx Measure Base model (age, sex, education, insurance, SVI) Base model + race/ethnicity Base model + race/ethnicity + site + encounter factors Kept informed about delays 0.544 0.547 0.560 Wayfinding around the facility 0.542 0.546 0.628 Timeliness of appointment scheduling 0.563 0.566 0.606 Ease of reaching office staff on the telephone 0.545 0.546 0.566 Extent to which appointments began on time 0.538 0.541 0.556 Timeliness of nursing staff in responding to a message regarding a clinical concern 0.539 0.558 0.586 Availability of information about supportive resources and services during care/treatment 0.548 0.555 0.570 Education about how to care for yourself at home 0.555 0.559 0.583