Integration of Early Physiological Responses Predicts Later Illness Severity in Preterm Infants

Physiological parameters routinely and noninvasively collected in the first 3 hours of life can accurately predict morbidity in premature infants. Not All Preemies Are Alike Premature babies can be full of surprises. Although smaller and more premature babies generally experience more complications, the hospital course of individual infants can vary greatly. Preemies born the same size and at the same gestational age can have vastly different outcomes, ranging from death to healthy survival with minimal medical problems. Ideally, the infants who are likely to do well could stay in local hospitals where they are born, whereas their high-risk counterparts would be transferred to specialty referral centers for more aggressive treatment. Distinguishing these groups of patients has been the Holy Grail of neonatology for some time, however. Ranging from the old classic, the Apgar score, to the newest inventions such as SNAP, SNAPPE, and CRIB scores, these many different prediction methods attest to the difficulty of the task. Now, Saria et al. have developed a way to take advantage of the cardiorespiratory monitors that are ubiquitous in the neonatal intensive care unit and use routinely collected data to predict infants’ clinical outcomes more accurately than can be achieved with any of the scoring systems in use today. After infants are born prematurely, they are usually attached to a cardiorespiratory monitor within minutes of their delivery. The monitors continuously display and store each baby’s vital sign data, including heart rate, respiratory rate, and oxygen saturation. This continuous stream of vital sign data continues as each infant transfers from the delivery room to the neonatal intensive care unit, and then until the patient is discharged home, or longer as necessary. Saria et al. have found that physiologic data derived from routine monitoring in the first 3 hours of life can predict future outcomes. The authors used heart rate and respiratory rate, as well as variability in these parameters, and oxygen saturation and time of hypoxia in a computational model that was able to predict the patients’ risk of future morbidity. The model proved particularly accurate in predicting the risk of high morbidity due to infections and cardiopulmonary complications, even when these were not diagnosed until days or weeks later. PhysiScore, the new method developed by Saria et al. for assessing the prognosis of premature infants, is an important development given its improved specificity and sensitivity over preexisting scoring techniques. Moreover, it relies on readily accessible noninvasive data that are already routinely collected on all infants, and can be quickly calculated by computer as early as 3 hours into the infant’s life. PhysiScore is a timely and necessary invention and has the potential to optimize medical management for most premature infants. Physiological data are routinely recorded in intensive care, but their use for rapid assessment of illness severity or long-term morbidity prediction has been limited. We developed a physiological assessment score for preterm newborns, akin to an electronic Apgar score, based on standard signals recorded noninvasively on admission to a neonatal intensive care unit. We were able to accurately and reliably estimate the probability of an individual preterm infant’s risk of severe morbidity on the basis of noninvasive measurements. This prediction algorithm was developed with electronically captured physiological time series data from the first 3 hours of life in preterm infants (≤34 weeks gestation, birth weight ≤2000 g). Extraction and integration of the data with state-of-the-art machine learning methods produced a probability score for illness severity, the PhysiScore. PhysiScore was validated on 138 infants with the leave-one-out method to prospectively identify infants at risk of short- and long-term morbidity. PhysiScore provided higher accuracy prediction of overall morbidity (86% sensitive at 96% specificity) than other neonatal scoring systems, including the standard Apgar score. PhysiScore was particularly accurate at identifying infants with high morbidity related to specific complications (infection: 90% at 100%; cardiopulmonary: 96% at 100%). Physiological parameters, particularly short-term variability in respiratory and heart rates, contributed more to morbidity prediction than invasive laboratory studies. Our flexible methodology of individual risk prediction based on automated, rapid, noninvasive measurements can be easily applied to a range of prediction tasks to improve patient care and resource allocation.