Determinants of direct discharge home from critical care units: a population-based cohort analysis
Martin CM, Lam M, Allen B, Richard L, Lau V, Ball IM, Wunsch H, Fowler RA, Scales DC. Crit Care Med. 2020; 48(4):475-83. Epub 2019 Dec 23. DOI: https://doi.org/10.1097/CCM.0000000000004178
Objective — To describe trends and patient and system factors associated with direct discharge from critical care to home in a large health system.
Design — Population-based cohort study of direct discharge to home rates annually over 10 years. We used a multivariable, multilevel random-effects regression model to analyze current factors associated with direct discharge home in a subcohort from the most recent 2 years.
Setting — One hundred seventy-four ICUs in 101 hospitals in Ontario.
Patients — All patients discharged from an ICU between April 1, 2007, and March 31, 2017.
Interventions — None.
Measurements and Main Results — Overall, 237,200 patients (21.1%) were discharged directly home from an ICU. The rate of direct discharge to home increased from 18.6% in 2007 to 23.1% in 2017 (annual increase of 1.02; 95% CI, 1.02-1.03). There were marked variations in rates of direct discharge to home across all critical care units. For medical and surgical units, the median odds ratio was 1.76 (95% CI, 1.59-1.92). In these units, direct discharge to home was associated with younger age (odds ratio, 0.36; 95% CI, 0.34-0.39 for age 80-105 vs age 18-39), fewer comorbidities (odds ratio, 1.74; 95% CI, 1.63-1.85 for Charlson comorbidity index of 0 vs 2), diagnoses of overdose/poisoning (odds ratio, 1.35; 95% CI, 1.23-1.47) and diabetic complications (odds ratio, 1.35; 95% CI, 1.2-1.51), and admission after a same-day procedure (odds ratio, 2.82; 95% CI, 2.46-3.23 compared with emergency department). ICU occupancy was inversely associated with direct discharge to home with an odds ratio of 0.88 (95% CI, 0.87-0.88) for each 10% increase.
Conclusions — High rates of direct discharge to home with evidence of significant practice variation combined with identifiable patient characteristics suggest that further evaluation of this increasingly common transition in care is warranted.