Outcomes of an electronic patient-reported outcomes (ePRO)–based symptom management program (eSyM): A cluster randomized trial.

Abstract

11001 Background: Although ePROs have been shown to reduce resource utilization and improve outcomes among people with cancer, they have not been widely adopted. We conducted a pragmatic type II hybrid effectiveness-implementation cluster randomized stepped-wedge trial of an ePRO-based, EHR-integrated symptom management program (eSyM) across 6 health systems. Here, we report the primary effectiveness outcome comparing patients treated before (control/not exposed) versus after (intervention/exposed) eSyM deployment. Methods: Eligible patients were adults who started chemotherapy (CHEMO) or were discharged after surgery (SURG) for a suspected or confirmed GI, GYN, or thoracic cancer. The intervention included ePRO questionnaires based on PRO-CTCAE items, severe symptom alerts, self-management tip sheets, and communication support. Outcomes included having an emergency department (ED) visit or inpatient admission (INPT) within 30 and 90-days. Logistic regression models accounted for socio-demographic, clinical, calendar time, health system, and other factors. Secondary analyses stratified results by treatment and health system to assess for effect modification. Results: From Jan. 2018 to Feb. 2023, the control and intervention conditions accrued 21,112 and 18,830 patients, respectively (median age 62 vs. 65; female 68% vs. 63%). Patient enrollment by health system ranged from 3,961 to 14,560. In the intervention cohort, 51% of patients used eSyM to report symptoms. Crude 30-day event rates for the control and intervention cohorts were 5.4% vs. 6.2% for ED, and 8.5% vs. 9.1% for INPT. Accounting for other factors, there were no significant differences in ED or INPT at 30 (Table) or 90 days. Among SURG patients, there was significantly greater odds of ED, but not INPT, for the intervention vs. control cohort. Results varied by health system, with evidence of higher, similar, and lower odds for the intervention vs. control cohort. Conclusions: eSyM deployment did not significantly reduce ED or INPT events. Only half of exposed patients used eSyM to report symptoms. Since prior analyses found lower odds of acute care utilization among patients who reported symptoms via eSyM, implementation and engagement barriers may have substantially impacted effectiveness outcomes. Heterogeneity of effect by health system and treatment suggest that healthcare structures, processes, and baseline performance may influence the uptake and impact of ePRO-based symptom management systems. Clinical trial information: NCT03850912 . Events at 30 days OR for ED 95%CI OR for INPT 95% CI Overall 1.10 0.94-1.29 1.00 0.88-1.14 Stratified by treatment Chemo 0.93 0.72-1.20 0.85 0.70-1.05 Surg 1.23 1.01-1.49 1.10 0.93-1.30 Stratified by health system A 1.26 1.07-1.50 1.59 1.37-1.85 B 1.47 1.15-1.88 1.25 0.98-1.58 C 0.74 0.61- 0.90 0.71 0.61-0.82 D 0.90 0.68-1.20 0.92 0.75-1.12 E 0.92 0.70-1.23 1.09 0.87-1.38 F 1.19 0.96-1.47 0.80 0.68-0.95