Impact of an ePRO-informed clinical decision support tool on clinical discussions during a follow-up visit for post-treatment head and neck cancer survivors: Results from the HN-STAR trial (WF-1805CD).

509 Background: ASCO guidelines for care of post-treatment head and neck cancer (HNC) survivors recommend comprehensive management of late effects, comorbidities, health behaviors, and cancer risks. We developed HN-STAR: a clinical decision support tool based on electronic patient-reported outcomes (ePROs) to improve implementation of HNC survivorship guidelines. Methods: We conducted a site-randomized controlled trial of oncology practices in the Wake Forest NCI Community Oncology Research Program (NCORP) Research Base (NCT04208490) comparing HN-STAR to usual care during a routine clinic visit. Eligible disease-free survivors completed treatment for HNC <2 years prior. In both arms, prior to the visit, survivors completed validated ePROs for 25 concerns (symptoms, well-being, and health behaviors). Concerns were characterized as absent or mild vs. burdensome. The HN-STAR tool presented survivor’s burdensome concerns and tailored recommendations for clinician review during the visit. Clinicians selected management plans in HN-STAR for each concern. Post-visit, all survivors reported which concerns were discussed at the clinic visit (regardless of whether the concern was reported pre-visit). Comparisons between groups were made by mixed models adjusted for correlation within practices. We hypothesized that survivors in the HN-STAR arm would report 1) more concerns discussed, 2) more of their burdensome concerns discussed, and 3) a greater proportion of their burdensome concerns discussed, compared to survivors receiving usual care. Results: 335 survivors (mean age 64 years, 23% Hispanic or Non-White, 26% non-metropolitan) at 26 practices reported a mean of 7.7 burdensome concerns (standard deviation [sd] 4.6) pre-visit, most commonly dry mouth (71%), altered taste (54%), neck or shoulder stiffness (48%), and fatigue (40%). In adjusted models, survivors in the HN-STAR arm had a greater proportion of burdensome concerns discussed at the visit than survivors in the usual care arm (60% vs. 44%, p=0.01) (Table). Conclusions: In a large, national, diverse, and symptomatic sample of HNC survivors at community oncology practices, using an ePRO-informed clinical decision support tool increased the likelihood that burdensome concerns were discussed in clinic. Alerting providers to burdensome concerns and presenting guideline-concordant management options has promise for improving provider-survivor interactions. Clinical trial information: NCT04208490 . HN-STARN=136 Usual careN=199 p-value # concerns discussed, median (interquartile range [IQR]) 7.0 (4.0, 10.0) 5.0 (2.0, 9.0) 0.25 # burdensome concerns discussed, median (IQR) 4.0 (2.0, 7.0) 3.0 (1.0, 5.0) 0.10 Proportion of burdensome concerns discussed, mean (sd) 0.60 (0.32) 0.44 (0.32) 0.01