IMPORTANCE: Discharge directly home from PICU is increasingly described, but its prevalence and predictors are not known. OBJECTIVES: To describe the prevalence and predictors of discharge directly home from the PICU. DESIGN: Cohort of admissions between January 1, 2015, and June 30, 2021. SETTING: Multicenter study of 142 North American PICUs participating in Virtual Pediatric Systems dataset. PARTICIPANTS: Hospitalized children younger than 18 years admitted to a participating PICU who were discharged home either directly or after transfer to the acute care ward. We excluded admissions with unclear discharge disposition, and repeat admissions within a year. MAIN OUTCOMES AND MEASURES: Demographic, PICU admission, and unit characteristics were measured according to discharge disposition. A multilevel adjusted logistic regression model, clustered by patient and center, was used to evaluate predictors of discharge directly home. RESULTS: The cohort included 612,471 admissions (339,818 [55%] males), of which 141,427 (23.1%) were discharged directly home from the PICU. Across 142 sites, the proportion of those discharged directly home ranged from 3% to 100%. In adjusted models, predictors associated with discharge home included prior discharge home (odds ratio [OR], 30.77
95% CI, 27.92-33.92), age 2-5 years (OR, 2.35
95% CI, 2.08-2.66), surgical admission (OR, 2.68
95% CI, 1.34-5.33), and respiratory (OR, 2.08
95% CI, 1.64-2.63) and cardiovascular (OR, 1.56
95% CI, 1.18-2.06) complex chronic conditions. Large center size was associated with a lower likelihood of discharge home (OR, 0.58
95% CI, 0.49-0.69). CONCLUSIONS AND RELEVANCE: About one in four children are discharged directly home after a PICU admission. While the practice is common, there is a large variability between centers. Selected patients, including those with acute surgical admission and those with complex chronic respiratory and cardiovascular conditions, are more likely to go directly home from the PICU.