Machine learning models to predict and benchmark PICU length of stay with application to children with critical bronchiolitis.

OBJECTIVE: To create models for prediction and benchmarking of pediatric intensive care unit (PICU) length of stay (LOS) for patients with critical bronchiolitis. HYPOTHESIS: We hypothesize that machine learning models applied to an administrative database will be able to accurately predict and benchmark the PICU LOS for critical bronchiolitis. DESIGN: Retrospective cohort study. PATIENTS: All patients less than 24-month-old admitted to the PICU with a diagnosis of bronchiolitis in the Pediatric Health Information Systems (PHIS) Database from 2016 to 2019. METHODOLOGY: Two random forest models were developed to predict the PICU LOS. Model 1 was developed for benchmarking using all data available in the PHIS database for the hospitalization. Model 2 was developed for prediction using only data available on hospital admission. Models were evaluated using R2 values, mean standard error (MSE), and the observed to expected ratio (O/E), which is the total observed LOS divided by the total predicted LOS from the model. RESULTS: The models were trained on 13,838 patients admitted from 2016 to 2018 and validated on 5254 patients admitted in 2019. While Model 1 had superior R2 (0.51 vs. 0.10) and (MSE) (0.21 vs. 0.37) values compared to Model 2, the O/E ratios were similar (1.18 vs. 1.20). Institutional median O/E (LOS) ratio was 1.01 (IQR 0.90-1.09) with wide variability present between institutions. CONCLUSIONS: Machine learning models developed using an administrative database were able to predict and benchmark the length of PICU stay for patients with critical bronchiolitis.

Characteristics of Children Who Deteriorate After Transport and Associated Preadmission Factors.

OBJECTIVE: The incidence of deterioration and associated characteristics are largely unknown for children transported for admission from referring emergency departments (EDs) to general inpatient units. This study describes this population and identifies associated preadmission characteristics. METHODS: This single-center cohort study included children ≤ 18 years old transferred from an ED and directly admitted to general inpatient units from 2016 to 2019. Deterioration was defined as 1 or more of the following occurring within 24 hours of admission: rapid response team activation, transfer to the intensive care unit (ICU), or cardiac or respiratory arrest. ICU transfer was the secondary outcome. Logistic regression was performed. RESULTS: One thousand nine hundred eighty-eight patients were included; the median age was 4.2 years, 53.9% were male, and 44.1% had respiratory diagnoses. Deterioration occurred in 135 (6.8%) children overall and in 10.1% of children with respiratory complaints. Deterioration was associated with ≥ 2 complex chronic conditions (adjusted odds ratio [aOR] = 2.09; 95% confidence interval [CI], 1.04-4.19) and a longer stabilization time (per 10 minutes) (aOR = 1.17; 95% CI, 1.01-1.36). ICU transfer was associated with ≥ 2 complex chronic conditions (aOR = 2.33; 95% CI, 1.13-4.80), supplemental oxygen via nasal cannula (aOR = 2.13; 95% CI, 1.18-3.85), and nebulizer treatment (aOR = 2.77; 95% CI, 1.21-6.35). CONCLUSION: Deterioration was experienced by 7% of children admitted to a general unit, with the majority having respiratory complaints. Transport teams should consider the potential for increased risk of deterioration among children with respiratory disease, multiple complex chronic conditions, and a nasal cannula or nebulizer therapy. The clinical significance of marginally longer stabilization times is unclear and warrants further investigation.

Resources and Costs Associated With Repeated Admissions to PICUs.

OBJECTIVE: To determine the costs and hospital resource use from all PICU patients readmitted with a PICU stay within 12 months of hospital index discharge. DESIGN: Cross-sectional, retrospective cohort study using Pediatric Health Information System. SETTING: Fifty-two tertiary children's hospitals. SUBJECTS: Pediatric patients under 18 years old admitted to the PICU from January 1, 2016, to December 31, 2017. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: Patient characteristics and costs of care were compared between those with readmission requiring PICU care and those with only a single PICU admission per annum. In this 2-year cohort, there were 239,157 index PICU patients of which 36,970 (15.5%) were readmitted and required PICU care during the 12 months following index admission. The total hospital cost for all index admissions and readmissions was $17.3 billion, of which 21.5% ($3.71 billion) were incurred during a readmission stay involving care in the PICU; of the 3,459,079 hospital days, 20.3% (702,200) were readmission days including those where PICU care was required. Of the readmitted patients, 11,703 (30.0%) received only PICU care, accounting for $662 million in costs and 110,215 PICU days. Although 43.6% of all costs were associated with patients who required readmission, these patients only accounted for 15.5% of the index patients and 28% of index hospitalization expenditures. More patients in the readmitted group had chronic complex conditions at index discharge compared with those not readmitted (83.9% vs 54.9%; p < 0.001). Compared with those discharged directly to home without home healthcare, patients discharged to a skilled nursing facility had 18% lower odds of readmission (odds ratio 0.82 [95% CI, 0.75-0.89]; p < 0.001) and those discharged home with home healthcare had 43% higher odds of readmission (odds ratio, 1.43 [95% CI, 1.36-1.51]; p < 0.001). CONCLUSIONS: Repeated admissions with PICU care resulted in significant direct medical costs and resource use for U.S. children's hospitals.