A Rules-Based Algorithm to Prioritize Poor Prognosis Cancer Patients in Need of Advance Care Planning

Accurate understanding of the prognosis of an advanced cancer patient can lead to decreased aggressive care at the end of life and earlier hospice enrollment.

Our goal was to determine the association between high-risk clinical events identified by a simple, rules-based algorithm and decreased overall survival, to target poor prognosis cancer patients who would urgently benefit from advanced care planning.

A retrospective analysis was performed on outpatient oncology patients with an index visit from April 1, 2015, through June 30, 2015. We examined a three-month window for "high-risk events," defined as (1) change in chemotherapy, (2) emergency department (ED) visit, and (3) hospitalization. Patients were followed until January 31, 2017.

A total of 219 patients receiving palliative chemotherapy at the University of Chicago Medicine with a prognosis of ≤12 months were included.

The main outcome was overall survival, and each "high-risk event" was treated as a time-varying covariate in a Cox proportional hazards regression model to calculate a hazard ratio (HR) of death.

A change in chemotherapy regimen, ED visit, hospitalization, and at least one high-risk event occurred in 54% (118/219), 10% (22/219), 26% (57/219), and 67% (146/219) of patients, respectively. The adjusted HR of death for patients with a high-risk event was 1.72 (95% confidence interval [CI] 1.19-2.46, p = 0.003), with hospitalization reaching significance (HR 2.74, 95% CI 1.84-4.09, p < 0.001).

The rules-based algorithm identified those with the greatest risk of death among a poor prognosis patient group. Implementation of this algorithm in the electronic health record can identify patients with increased urgency to address goals of care.