Association of patient-reported symptoms with telehealth use and symptom resolution in oncology.

563Background: Oncology practices continue to explore how best to incorporate synchronous telehealth visits within service offerings. We evaluated the association of patient-reported symptoms in offering such services to patients, patient acceptance of the offering, and symptom resolution post visit compared to in-person care. Methods: An observational cohort of adults aged 21+ treated for a malignancy 6 months-3 years prior to the scheduled telehealth or in-person oncology visit April 2022-April 2023 at one of two academic medical centers (one in North Carolina, one in Michigan) or their community affiliates was created by joining pre- and post-visit telephone survey data with electronic health record (EHR) data and geo-coded community characteristics. Black adults and those scheduled for telehealth visits were over sampled. Surveys inquired about telehealth visit offerings and three physical (fatigue, pain and insomnia) and three psychological (anxiety, sadness and discouraged) symptoms using the PRO-CTCAE. We used Chi-square tests to compare proportions of patients reporting each symptom by being offered a telehealth visit and accepting that offer. To adjust for non-random assignment of visit type, we used propensity score (PS)-weighted average changes in symptom composite scores pre-/post-visit to assess differences in symptom resolution by visit type. Results: N=773 patients completed a pre-visit survey. Mean age was 63.3 years (SD=13.2), 67.1% female, 40.0% Black, 78.3% some college or more, and 28.7% reported fair/poor health. Physical (fatigue 79.8%; pain 74.0%; insomnia 65.6%) and/or psychological (anxiety 64.5%; sadness 63.8%; discouraged 36.6%) symptoms were common. 18.6% of patients reported being offered a telehealth visit, with those reporting anxiety and sadness significantly less likely to be offered (16.4% and 16.2%, respectively, p<0.05). Among those offered a telehealth visit, 52.1% reported accepting the offer with the proportion not differing by symptom presence (p>0.30). N=610 patients completed the post-visit survey and had EHR visit data available (n=496 in-person and n=114 telehealth). While telehealth visit use varied by patient socio-demographic and clinical characteristics (e.g., gender, race, travel distance and cancer diagnosis) after PS weighting and sample trimming (n=477) balance in most observed characteristics was achieved. Across both physical and psychological symptoms, the weighted changes in composite scores from pre- to post-visit were negligible and statically not significant regardless of visit type (95% CI: -0.05, 0.01 for in-person; 95% CI: -0.12, 0.05 for telehealth, p>0.29). Conclusions: Although telehealth access within oncology varies, thoughtfully offered and accepted telehealth visits appear to be non-inferior to in-person visits for near-term symptom resolution.