Introduction: Predictive indices of hemodynamics may be modelled from the acquisition of continuous physiologic data, but to date are not related to a specific outcome. The Inadequate Oxygen Delivery Index (IDO2) - a patient specific, FDA 510(k) cleared algorithm (Etiometry, Boston, MA) is continuously computed from multiple physiologic and laboratory values collected by the T3 Data Aggregation and Visualization platform. A lower IDO2 is correlated with higher lactate and lower mixed venous oxygen levels. The goal of this study was to evaluate the association between implementation of IDO2 index in clinical practice and ICU length of stay (LOS).
Hypothesis: The initiation of IDO2 will be associated with a decrease in LOS.
Methods: In this retrospective cohort analysis of neonates admitted to Boston Children’s (BCH) and The Hospital for Sick Children (HSC) following cardiac surgery, patients were divided into pre IDO2 (Aug 2015-Mar 2016, BCH; May 2015-Mar 2016, HSC) vs post IDO2 (Aug 2016-Mar 2017, BCH; May 2016-Mar 2017, HSC). IDO2, age, physician login, and hours of T3 platform use were collected. At BCH, STAT score, weight, gender, and prematurity were also collected. Generalized linear regression models risk adjusted for age and location (multicenter) and STAT score, weight, and prematurity (BCH only) were created to compare LOS pre vs post IDO2.
Results: Of the 342 patients admitted pre IDO2 (BCH n=119; HSC n=223), and 331 patients admitted post IDO2 (BCH n=134; HSC n=197), there was no difference in age, STAT score, weight, or prematurity (p>0.05). The mean hours of T3 platform use increased post IDO2 (BCH 17 vs. 32 hrs/wk, 85%; HSC 19 vs. 46 hrs/wk, 145%), as did the mean number of physician logins (BCH 26 vs 58, 64%; HSC 29 vs 71, 145%). In the multicenter analysis, there was a relative decline in LOS of 0.75 (0.60-0.94), p=0.01 post IDO2. In the BCH model (further adjusted for STAT score, age, prematurity, and gender) there was a relative decline in LOS of 0.64 (0.45-0.91), p=0.01 post IDO2.
Conclusion: The introduction of a patient specific physiologic index resulted in a significant relative reduction in LOS in a large multicenter cohort of neonates. Further analysis is ongoing to risk adjust the entire cohort and explore the impact of IDO2 on clinical decision making.