The uncertainty surrounding expected outcomes at periviable gestation leads to several major challenges. First, clinicians may be unsure of how to counsel families. Second, the lack of clarity makes families more anxious and causes trauma. Third, it is difficult for both clinicians and families to make the most informed decisions for the neonate. This is important because making a decision to resuscitate when there are very poor chances for a good outcome could lead to a futile attempt at resuscitation leading to death, or potentially a survivor that has severe neurodevelopmental disability. On the other hand, making a misinformed decision to not resuscitate and proceed to comfort care when there is a good chance of survival without disability could be even more tragic. We will develop and test a modern, comprehensive predictive model for outcomes at periviable gestation using an existing infrastructure for data collection and implementation, the California Perinatal Quality Care Collaborative (CPQCC). This population-based network of neonatal intensive care units includes both academic and community units, which means that results will be generalizable. CPQCC already has an existing data infrastructure that includes maternal and neonatal data, including follow-up data at 2 years of age, giving an opportunity to study outcomes that do not exist in similar networks. The setting of the CPQCC allows for a unique opportunity to both improve on current prediction tools, and to implement and evaluate the prediction tool in a real-world setting.
In Aim 1, we will build a predictive model for outcomes in periviable gestation using the most up-to-date data possible using a broad population-based cohort. This model will be used to build an on-line estimator that will be used by 20 hospitals across California.
In Aim 2, we will evaluate how current practice across ~140 California neonatal intensive care units align with prognostic estimates from the models built in Aim 1. In this Aim, we will evaluate whether certain patient level factors and hospital level factors appear to fall outside the norms of typical practice in relationship to prognosis, for therapies provided to the mother prior to birth, and the infant after birth.
In Aim 3, we will implement usage of the estimator across California neonatal intensive care units in waves of 20 hospitals each over a 1 year period. We will then compare if and how practices change for periviable gestation infants.
In Aim 4, we will conduct a cost-effectiveness analysis of implementing this estimator in clinical practice. This research will fill several gaps in our knowledge of the use of prediction models for periviable birth, particularly the gap in our understanding of how using an estimator in practice may influence and improve clinical decisions and outcomes.
When preterm birth occurs at the most extreme youngest gestation, the hope of neurologically intact survival must be weighed against the very severe trauma that infants and families may undergo in efforts to resuscitate and care for an infant that may ultimately have a very short life or severe life-long disability. Clinicians lack the ability to provide the most accurate and up-to-date information to the family so that they can come to the best decision for the infant, as current prediction models rely on older data, are not tailored to the individual institution where the birth is occurring, and use limited information to make predictions; furthermore, there is a gap in knowledge in the following: how the use of prediction tools might impact counseling and clinical practice; whether tools can make any difference in outcomes such as survival without morbidity and length of stay before death; and the cost impact and cost-effectiveness of using such tools. We propose to expand the knowledge base on prediction models for periviable gestation in an existing network of neonatal intensive care units in California ? involving 140 hospitals from which to obtain data?by implementing the prediction model with established collaborative quality improvement methods, evaluating the impact of implementation on practices and outcomes, and examining the cost impact of implementation in clinical practice.
