Leveraging patient-reported outcomes for enhanced symptom management among thoracic radiation patients.

381 Background: Patient reported outcomes (PROs) can help guide patient-provider discussions and generate high alert values (HAVs) that prompt clinical response. We analyzed PRO outcomes among Thoracic Radiation Oncology (TRO) patients to identify predictors of HAVs and provider response. Methods: We retrospectively analyzed PRO responses among TRO patients seen at our institution from 2021-2024. Patients received surveys comprised of 13 symptom items from the PRO-CTCAE Measurement System, 72 hours prior to each consult, follow up, and weekly on-treatment visit (OTV). HAVs (prespecified high severity/frequency PRO responses) sent alerts to the care team via Epic Inbasket. Demographic (age, gender, race/ethnicity, marital status) and clinical (cancer type, metastatic disease [mets], smoking status, alcohol use, psychiatric diagnoses/treatment) data were collected. Multivariable logistic regression models (MVA) were used to identify variables associated with 1) patients reporting a HAV and 2) documented provider response to HAVs within 7 days. Results: Among 2973 patients who completed at least one survey, 52% were male with median age 68 (IQR 61-75). Most common cancer types included lung (64%) and esophagus (13%). Within this cohort, 40% of patients generated at least one HAV. Common HAVs included pain (49%), dyspnea (17%), anxiety (14%), and sadness (9%). Younger age (OR 1.02/yr, p<0.01 ), former or current heavy smoking (OR 1.29, p<0.01; OR 1.70, p<0.01; ref never smoker), divorced/separated/widowed status (OR 1.32, p=<0.01; ref married/significant other), psychiatric diagnosis or treatment (OR 1.39, p=0.01; OR 1.79, p<0.01), mesothelioma/trachea/thymus primary (OR 1.94, p<0.01; ref lung), and bone mets (OR 2.03, p<0.01) were associated with greater reporting of HAVs. Patients with lung mets were less likely to generate HAVs (OR 0.72, p=0.02). Providers responded to pain HAVs 79% of the time, while response rates for anxiety and depression HAVs were <25%. On MVA, providers were more likely to respond to HAVs reported prior to OTVs and for patients with bone mets (Table). In comparison to alerts related to pain, providers were less likely to respond to HAVs for dyspnea, anxiety, sadness, or palpitations. Both HAV risk and provider response did not differ by gender, race/ethnicity, or presence of other types of mets. Conclusions: 40% of patients completing PROs generated HAVs. Providers frequently acknowledged and documented interventions for pain-related HAVs, but infrequently documented interventions for mental health-related HAVs. This may indicate room for improvement in how psychosocial care is integrated into oncology care. Significant Variables Provider response (%) OR p-value Visit type Consult 54 Ref OTV 68 1.63 0.010 HAV type Pain 79 Ref Dyspnea 69 0.69 0.046 Anxiety 21 0.08 <0.001 Sadness 14 0.04 <0.001 Palpitations 7 0.02 <0.001 Bone mets No 57 Ref Yes 74 1.58 0.021