ACO Spending and Utilization Among Medicare Patients at the End of Life: an Observational Study

Abstract

Background: End-of-life (EOL) costs constitute a substantial portion of healthcare spending in the USA and have been increasing. ACOs may offer an opportunity to improve quality and curtail EOL spending.

Objective: To examine whether practices that became ACOs altered spending and utilization at the EOL.

Design: Retrospective analysis of Medicare claims.

Patients: We assigned patients who died in 2012 and 2015 to an ACO or non-ACO practice. Practices that converted to ACOs in 2013 or 2014 were matched to non-ACOs in the same region. A total of 23,643 ACO patients were matched to 23,643 non-ACO patients.

Main measures: Using a difference-in-differences model, we examined changes in EOL spending and care utilization after ACO implementation.

Key results: The introduction of ACOs did not significantly impact overall spending for patients in the last 6 months of life (difference-in-difference (DID) = $192, 95%CI -$841 to $1125, P = 0.72). Changes in spending did not differ between ACO and non-ACO patients across spending categories (inpatient, outpatient, physician services, skilled nursing, home health, hospice). No differences were seen between ACO and non-ACO patients in rates of ED visits, inpatient admissions, ICU admission, mean healthy days at home, and mean hospice days at 180 and 30 days prior to death. However, non-ACO patients had a significantly greater increase in hospice utilization compared to ACO patients at 180 days (DID P-value = 0.02) and 30 days (DID P-value = 0.01) prior to death.

Conclusions: With the exception of hospice care utilization, spending and utilization were not different between ACOs and non-ACO patients at the EOL. Longer follow-up may be necessary to evaluate the impact of ACOs on EOL spending and care.

Keywords: ACO; End-of-life care; Healthcare spending; Healthcare utilization; Medicare.