Study to Collect Data for Neonatal Abstinence Syndrome (NAS) and Evaluate the Automated Data Collection Process

Overview

Substance abuse during pregnancy is on the rise through both prescribed and illicit use of controlled substances, which has increased neonatal abstinence syndrome (NAS). The prevalence of opioid use during pregnancy has increased by 333% from 2013 to 2014 and continues to rise. Approximately 1 in 3 women were prescribed opioids during pregnancy from 2008 to 2012. In the US, NAS was diagnosed every 25 minutes in 2014. By 2019, it became every 15 minutes. Although there are medication-based interventions for the treatment of NAS, used in up to 80% of opioid-exposed infants, these treatments carry risks of toxicity and drug interactions. Despite the steep medical costs and the risks of treatment, current tools to assess the severity of NAS are subjective and suffer from examiner bias, resulting in poorer clinical outcomes, such as longer lengths of stay in the Neonatal Intensive Care Unit (NICU), for these babies. Studies have shown that continuous vital sign monitoring improves outcomes and decreases the length of stay in general practice. Preliminary machine learning models have been able to predict pharmacological treatment for Neonatal Opioid Withdrawal Syndrome (NOWS). This project will collect physiological and behavioral data of NAS patients to develop an AI algorithm and establish the advantages of continuous monitoring in NAS. The AI algorithm, processed by machine learning, will help predict NAS symptoms, automate scoring, and provide healthcare personnel with predictive analytics to guide suggested treatments.

The current diagnostic and assessment framework for NAS heavily relies on subjective methods, primarily the Finnegan Neonatal Abstinence Score (FNAS). FNAS helps providers evaluate pharmacological and non-pharmacological treatments and monitor the progress of infants with NAS. The Eat Sleep Console (ESC) approach has been implemented in some hospitals to emphasize non-pharmacological interventions as the primary method of managing and treating NAS. Despite the high prevalence of NAS and the significant resources allocated to its management, the healthcare system continues to grapple with an unmet clinical need for standardized diagnostic and treatment protocols. The reliance on subjective assessments contributes to this challenge, as FNASS and ESC introduce variability in care that can affect outcomes. Developing objective, reliable tools for assessing NAS severity and guiding treatment decisions remains a critical need in neonatal care, promising to enhance the efficiency and effectiveness of interventions for these vulnerable patients. Recent studies have underscored the lack of consistency in diagnosing and treating NAS, revealing a broad spectrum of practices across different pediatric healthcare settings. This problematic inconsistency leads to varied patient outcomes and a lack of clarity on best practices. This multicenter study will collect data that will be used to develop an AI-based tool that can automate scoring with predictive analytics. Additionally, the investigators aim to establish the advantages of continuous monitoring in NAS that should lead to decreased length of stay in the NICU and improved patient outcomes.