Telehealth and cancer care delivery: An updated real-world analysis of the impact of telehealth on patient access and treatment utilization at a comprehensive cancer center.

1644 Background: Telehealth in oncology (onc) has persisted due to its convenience and potential to improve equitable access to care. Data on telehealth’s impact on treatment (tx) access are needed to guide long-term telehealth policies. Methods: We identified adult patients (pts) who completed ≥1 medical, surgical, or radiation onc visit with an associated cancer diagnosis at our center from 2017-2019 (pre-telehealth) vs 2021-2023 (post-telehealth). Data from 2020 was excluded, given COVID irregularities. We selected 3 disease groups with varying telehealth use – breast (low), GI (medium), and GU (high) – allowing us to better control for external factors (eg COVID). We compared changes in visit distribution, sociodemographic, and tx patterns within and between disease groups pre- vs post-telehealth using Chi-square and ANOVA tests. Logistic regression analyses identified post-telehealth predictors of receiving tx, including the percentage (%) of in-person (IP) visits per pt. Results: We analyzed 109,200 encounters and 26,907 pts pre-telehealth vs 143,159 encounters and 50,168 pts post-telehealth. Pt volume increased in all groups post-telehealth (+99% breast, +55% GI, +104% GU). Post-telehealth, the % of video visits (VVs) increased in breast (1.5% to 28.8%), GI (2.0% to 55.2%), and GU (3.0% to 80.9%). In all groups, the % of pts seen from outside the SF Bay Area decreased (-7.5 breast, -1.4 GI, -6.1 GU). In breast and GU, the % of pts receiving cancer tx decreased (-5.2, -16.0, p<0.01), while the % of pts receiving GI tx was stable (-0.9). For infusion tx, predictors of tx receipt included metastatic disease (OR 3.9, 95% CI 3.6-4.2) and higher % of IP medical onc visits (OR 3.6, 95% CI 3.3-3.9); while living outside the Bay Area was negatively associated with infusion tx (OR 0.5, 95% CI 0.4-0.5). For surgery, the % of IP surgical onc visits (OR 0.7, 95% CI 0.6-0.7) and living outside the Bay Area (OR 0.8, 95% CI 0.7-0.9) were negatively associated with tx receipt. For radiation tx, metastatic disease (OR 2.8, 95% CI 2.5-3.1), but not living outside the Bay Area (OR 0.6, 95% CI 0.5-0.7), was associated with tx receipt. The % of IP radiation onc visits was not significant (OR 1.1, 95% CI 1.0-1.4). Conclusions: Among 3 disease groups with varying telehealth use, pt volume increased in all groups, while the % of pts receiving tx at our center decreased. This suggests that providers may be providing collaborative care while pts receive care locally. In regression analyses, higher % of IP medical onc visits was positively associated with receiving infusions, highlighting the importance of IP visits during active medical onc tx. However, this was not seen for surgery and radiation tx. For these time-limited interventions, pts seen virtually are still likely to access tx at our center. Further analyses are needed to identify pts and visit types best suited for telehealth.