A Novel Approach to ICU Survivor Care: A Population Health Quality Improvement Project.

Objectives: Deliver a novel interdisciplinary care process for ICU survivor care and their primary family caregivers, and assess mortality, readmission rates, and economic impact compared with usual care.

Design: Population health quality improvement comparative study with retrospective data analysis.

Setting: A single tertiary care rural hospital with medical/surgical, neuroscience, trauma, and cardiac ICUs.

Patients: ICU survivors.

Interventions: Reorganization of existing post discharge health care delivery resources to form an ICU survivor clinic care process and compare this new process to post discharge usual care process.

Measurements And Main Results: Demographic data, Acute Physiology and Chronic Health Evaluation IV scores, and Charlson Comorbidity Index scores were extracted from the electronic health record. Additional data was extracted from the care manager database. Economic data were extracted from the Geisinger Health Plan database and analyzed by a health economist. During 13-month period analyzed, patients in the ICU survivor care had reduced mortality compared with usual care, as determined by the Kaplan-Meier method (ICU survivor care 0.89 vs usual care 0.71; log-rank p = 0.0108) and risk-adjusted stabilized inverse probability of treatment weighting (hazard ratio, 0.157; 95% CI, 0.058-0.427). Readmission for ICU survivor care versus usual care: at 30 days (10.4% vs 26.3%; stabilized inverse probability of treatment weighting hazard ratio, 0.539; 95% CI, 0.224-1.297) and at 60 days (16.7% vs 34.7%; stabilized inverse probability of treatment weighting hazard ratio, 0.525; 95% CI, 0.240-1.145). Financial data analysis indicates estimated annual cost savings to Geisinger Health Plan ranges from $247,052 to $424,846 during the time period analyzed.

Conclusions: Our ICU survivor care process results in decreased mortality and a net annual cost savings to the insurer compared with usual care processes. There was no statistically significant difference in readmission rates.