Automated patient selection and care coaches to increase advance care planning for patients with cancer.

Background: Advance care planning and serious illness conversations can help clinicians understand patients' values and preferences. Data are limited on how to increase the number of these conversations and what their effects are on care patterns. We hypothesized that using a machine learning survival model to select patients for serious illness conversations, along with trained care coaches to conduct the conversations, would increase uptake in patients with cancer at high risk of short-term mortality.

Methods: We conducted a cluster-randomized, stepped-wedge study on the physician level. Oncologists entered the intervention condition in a random order over 6 months. Adult patients with metastatic cancer were included. Patients with a less than 2-year computer-predicted survival and no prognosis documentation were classified as high priority for serious illness conversations. In the intervention condition, clinicians received automated weekly emails highlighting high-priority patients and were asked to document prognoses for them. Care coaches contacted these patients to conduct the remainder of the conversation. The primary endpoint was the proportion of visits with prognosis documentation within 14 days.

Results: We included 6372 visits with 1825 patients in the primary analysis. The proportion of visits with prognosis documentation within 14 days was higher in the intervention condition than in the control condition: 2.9% vs 1.1% (adjusted odds ratio = 4.3, P < .001). The proportion of visits with advance care planning documentation was also higher in the intervention condition: 7.7% vs 1.8% (adjusted odds ratio = 14.2, P < .001). For high-priority visits, the advance care planning documentation rate in intervention visits was 24.2% and in control visits was 4.0%.

Conclusion: The intervention increased documented conversations, with contributions by both clinicians and care coaches.