Prospectively Identifying People for Goals of Care Conversations (QI620)

Abstract

•Apply the methods used to estimate the impact within attendees' own organizations.•Compare the methods used to those currently employed at attendees' individual organizations. Our aim was to develop a method to prospectively target a population that is likely to benefit from goals of care discussions by first conducting a retrospective analysis of those dying in a large urban hospital system following recurrent admissions and escalating healthcare utilization. Across the hospital system, 1,906 people died in 2018; 253 experienced ≥2 admissions within 6 months of death. Most were over age 62 with 87 different primary diagnoses. Sixty-six percent presented from a nonhealthcare location. Palliative care consultation at any point in time was 70% among the group. However, 66% were within the last 6 months of life and 36% of patients only received palliative care during the final hospitalization. Forty of 55 people who were enrolled in hospice also had palliative care consultation. The average length of stay for all hospice enrollees was 34 days and the median was 3 days. The average number of days between palliative care consultation and death was 32 with a median of 10 days. Thirteen percent of patients had a document uploaded as an advance directive. Thirty-four percent received ICU care during the final hospital admission for an average of 5 days. The total cost of care for the final hospitalizations alone was $6.3M. These data informed a care model that prospectively identifies people at risk for in-hospital death to prioritize goals of care discussions. The model uses age, recurrent admissions, and location prior to presentation as criteria to alert admitting hospital staff. A fourth variable, ICU admission, was added to contrast differences in volume. Applied retrospectively to 2018, the model yielded 4,779 unique patients and 1,221 if ICU admission was added. In order for healthcare organizations to optimize care delivery it is advantageous to anticipate required clinical skills that can contribute to quality outcomes and to allocate resources to their development.