BE-A-PAL: Behavioral economics and automated analytics to improve palliative care among patients with advanced cancer.

Abstract

268 Background: Patients with advanced solid malignancies often experience high symptom burden and aggressive end-of-life care. Early specialist palliative care (PC) can improve symptom management, mood, and quality of life. However, many patients do not receive PC before they die, with clinician inertia and difficulty identifying high-risk patients being barriers to initiating PC referrals. Methods: In a 2-arm pragmatic randomized clinical trial among patients with stage 3 or 4 lung or non-colorectal gastrointestinal malignancies, oncology clinics were randomly assigned to intervention or usual practice. In the intervention arm, oncologists received automated algorithm-based default notifications prompting specialty PC referral. The control arm received usual care, in which oncologists could refer to PC at their discretion. The algorithm was adapted from NCCN guidelines, identified high-risk patients based on prognosis or symptom/psychosocial burden, and was integrated into the EHR. Oncologists were alerted weekly about eligible patients and had the option to opt-out of PC referral. If there was no response or agreement, a PC coordinator introduced and offered a PC visit to the patient. The primary outcome was a completed PC visit within three months of identification among high-risk patients. We report descriptive results from an interim analysis of the trial. Results: 15 practices (7 clinics and 32 physicians in intervention; 8 clinics and 31 physicians in control) and 567 patients (299 intervention; 268 control) were randomized. Rates of completed early PC visits were 46.4% (139/299) the intervention arm and 11.2% (30/268) in the control arm. The opt-out rate in the intervention arm was 10.7%. Among clinicians who did not opt-out, 79% of their patients agreed to an initial PC visit. Adjusted analyses and demographic data will be completed and available by September 2023. Conclusions: Algorithm-based default PC referrals meaningfully increased utilization of specialty PC within a community oncology practice. Targeting high-risk patients using guideline-based risk stratification prevented overwhelming PC capacity. Further analyses will assess the impact of the intervention on end-of-life care, acute care utilization, patient-reported outcomes, and explore clinician perspectives. Clinical trial information: NCT05590962 .