614 - Pandemic Allocation of Ventilators Model Penalizes Infants with Bronchopulmonary Dysplasia

Poster Number: 614
Publication Number: 614.306

Anupama Sundaram, UH Rainbow Babies & Children's Hospital, Strongsville, OH, United States; Jonathan M. Fanaroff, UH Rainbow Babies & Children's Hospital, Cleveland, OH, United States; Deanne Wilson-Costello, Rainbow Babies and Children's Hospital, Cleveland, OH, United States; Melissa Alberts, UH Rainbow Babies & Children's Hospital, COOPER CITY, FL, United States; Naini Shiswawala, Case Western Reserve University School of Medicine, Cleveland, OH, United States; Rita M. Ryan, Case Western Reserve University, Cleveland, OH, United States

Background: In response to the COVID-19 pandemic, to appropriately distribute limited resources, institutions developed ventilator allocation models. In one proposed “real-life” planned, but never implemented, model, neonates compete with adults for ventilators using a scoring system based on predicting survival. Additional points were given to comorbid conditions increasing 1 and 5-year (y) mortality. In adults, for example, 4 points were given for advanced cirrhosis, with mortality of 71.3% based on literature, and for neonates with severe bronchopulmonary dysplasia (sBPD); 2 points for moderate BPD (mBPD). We hypothesized that this allocation model overestimates mortality in neonates with m/s BPD and would inappropriately penalize these infants.

Objective: To evaluate if the current pandemic ventilator allocation protocol correctly predicts long-term mortality in infants with m/s BPD.

Design/Methods: A retrospective chart review was performed on all infants born ≥22 weeks and < 1500 grams admitted to Rainbow Babies and Children’s Hospital in 2015 to identify babies with m-sBPD (NIH 2001 definition). Main outcomes were 1 and 5y mortality. By design, follow-up was ascertained only for infants with later visits to our hospital system. Categorical variables analyzed with chi-square testing; continuous variables analyzed by Wilcoxon rank sum test. Sample size calculation based on 1y survival stated a need of at least 15 patients with known mortality outcomes to show a significant difference between BPD-related mortality and adult co-morbid condition mortality.

Results: In 2015, 28 infants were diagnosed with m-s BPD. Four infants had mBPD; 24 had sBPD. All infants (100%) with mBPD survived to 5y. For infants with sBPD, by 1y, 2 died (8%), 22 survived (92%). By 5y, 3 died (12.5%) and at least 13 survived (54%). We did not have 5y outcome on 8 infants (33%).Conclusion(s): Infants with m/s BPD had lower than predicted, by the allocation model, 1 and 5y mortality, suggesting the points assigned in the model are too high for these conditions. Newly published BPD data evaluating mortality prior to discharge was 2% and 4.8% for m-s BPD. With our sBPD 1y mortality at 8% and mBPD 5y mortality at 0% we believe this model would have over-penalized these babies. Limitations include small sample size and limited outcome data. Further investigation is ongoing to address limitations. We hope to use our findings to inform future pandemic triage protocols. Local mortality rates should be used when available for a ventilator allocation protocol.
Figure 1: Outcome of subjects according to disease severity
Figure 2 – Ventilator allocation protocol for the COVID-19 pandemic