A composite neonatal adverse outcome indicator using population-based data: an update

Introduction Severe morbidity rates in neonates can be estimated using diagnosis and procedure coding in linked routinely collected retrospective data as a cost-effective way to monitor quality and safety of perinatal services. Coding changes necessitate an update to the previously published composite neonatal adverse outcome indicator for identifying infants with severe or medically significant morbidity. Objectives To update the neonatal adverse outcome indicator for identifying neonates with severe or medically significant morbidity, and to investigate the validity of the updated indicator. Methods We audited diagnosis and procedure codes and used expert clinician input to update the components of the indicator. We used linked birth, hospital and death data for neonates born alive at 24 weeks or more in New South Wales, Australia (2002-2014) to describe the incidence of neonatal morbidity and assess the validity of the updated indicator. Results The updated indicator included 28 diagnostic and procedure components. In our population of 1,194,681 live births, 5.44% neonates had some form of morbidity. The rate of morbidity was greater for higher-risk pregnancies and was lowest for those born at 39-40 weeks' gestation. Incidence increased over the study period for overall neonatal morbidity, and for individual components: intravenous infusion, respiratory diagnoses, and non-invasive ventilation. Severe or medically significant neonatal morbidity was associated with double the risk of hospital readmission and 10 times the risk of death within the first year of life. Conclusion The updated composite indicator has maintained concurrent and predictive validity and is a standardised, economic way to measure neonatal morbidity when using population-based data. Changes within individual components should be considered when examining longitudinal data.