801 - Real time Neonatal Oxygenation Compliance analyzed by Etiometry: Tracking Risks for Retinopathy of Prematurity and Bronchopulmonary Dysplasia in Preterm infants

Publication Number: 801.6380

Sofia I. Perazzo, Childrens National Hospital, North Bethesda, MD, United States; Adam Tomczak, Etiometry Inc, Boston, MA, United States; Lauralee Davis, Children's National Hospital, Washington, DC, United States; Lamia Soghier, George Washington University School of Medicine and Health Sciences, Washington, DC, United States

Background: Preterm infants are at risk of Retinopathy of Prematurity (ROP) and Bronchopulmonary Dysplasia (BPD) due to prolonged mechanical ventilation and oxygen supplementation during their Neonatal Intensive Care Unit (NICU) stay. To minimize risks, staff maintain oxygen saturation from 90% to 95% to avoid hypoxia and hyperoxia. Etiometry® is a real-time data management software that collects, analyzes, visualizes, and archives pulse oximetry data. Clinicians can potentially use this data for assessment of day-to-day compliance with set parameters.

Objective: To determine compliance with oxygen saturation parameters (time in range) for neonates with ROP and BPD compared to controls to use as baseline data for quality improvement.

Design/Methods: Oxygen saturation data for neonates < 37 weeks gestational age (GA) at birth was automatically uploaded from the bedside monitor to Etiometry® every 5 seconds. Ventilation mode (invasive vs. non-invasive vs. room air), duration of oxygen therapy, ROP, and BPD diagnoses were extracted from the medical record. Patients were divided into 3 groups ( < 28, 28- < 34, and 3 4- < 37 weeks GA). We calculated time in range and FiO2 adjustments resulting in 5 compliance states (Table 1). We calculated total compliance time and divided groups into high and low compliance groups (above and below median) (Figure 2). Neonates with ROP and/or BPD were compared to controls using multivariate analysis. Relationships between GA and intervenable events were analyzed using Spearman’s rank correlation coefficient.

Results: Data from 879 unique patients (45,931days) over 50 months were analyzed. Total compliance rate for all groups was 34.8% with 7.2% intervenable hypoxic and 36.9% intervenable hyperoxic (Figure 1). Intervenable events increased with lower GA (p < 0.001) (Figure 3). In the low compliance group, neonates < 28 weeks GA had 1.83 odds , and 28 - < 34 weeks GA had 1.91 odds of developing ROP compared to the high compliance group (Figure 2). No statistically significant difference was seen in the 34- < 37 weeks GA. Similarly, the low compliance group in < 28 weeks had twice the odds and 28- < 34 weeks GA 4.74 odds of developing BPD compared to the high compliance group (Figure 2).

Conclusion(s): Intervention for hypoxic events occurred more frequently than hyperoxic events. Ventilation mode and oxygen therapy can now directly be captured by Etiometry allowing us to develop real-time pathways to alert clinicians to degree of non-compliance to decrease FiO2 use. Interventions can then be developed to improve compliance for quality improvement.

Table 1.
Group A <28 weeks, Group B =28- <34 weeks, Group C 34- <37. Normoxic = % of time between 90-95% saturation, Intervenable hypoxic states= % of time that the saturation is below 90% where FiO2 was less than 100%, Non-intervenable hypoxic state= % time that saturation is below 90% where FiO2 is at 100%, Intervenable hyperoxic state= % time that saturation is above 95% where FiO2 was above 21%, Non- intervenable hyperoxic state= % time that saturation is above 95% where FiO2 was 21%.

Comparison between High and Low fraction compliance group for ROP and BPD in each gestational age group.
Group A= < 28 weeks, Group B = 28- <34 weeks, Group C= 34- <37 weeks.

Figure 3: Correlation between intervenable events and Gestational Age
Figure shows the correlation between GA and number of intervenable events.