Machine learning risk stratification for high-risk infant follow-up of term and late preterm infants.

BACKGROUND: Term and late preterm infants are not routinely referred to high-risk infant follow-up programs at neonatal intensive care unit (NICU) discharge. We aimed to identify NICU factors associated with abnormal developmental screening and develop a risk-stratification model using machine learning for high-risk infant follow-up enrollment. METHODS: We performed a retrospective cohort study identifying abnormal developmental screening prior to 6 years of age in infants born ≥34 weeks gestation admitted to a level IV NICU. Five machine learning models using NICU predictors were developed by classification and regression tree (CART), random forest, gradient boosting TreeNet, multivariate adaptive regression splines (MARS), and regularized logistic regression analysis. Performance metrics included sensitivity, specificity, accuracy, precision, and area under the receiver operating curve (AUC). RESULTS: Within this cohort, 87% (1183/1355) received developmental screening, and 47% had abnormal results. Common NICU predictors across all models were oral (PO) feeding, follow-up appointments, and medications prescribed at NICU discharge. Each model resulted in an AUC > 0.7, specificity >70%, and sensitivity >60%. CONCLUSION: Stratification of developmental risk in term and late preterm infants is possible utilizing machine learning. Applying machine learning algorithms allows for targeted expansion of high-risk infant follow-up criteria. IMPACT: This study addresses the gap in knowledge of developmental outcomes of infants ≥34 weeks gestation requiring neonatal intensive care. Machine learning methodology can be used to stratify early childhood developmental risk for these term and late preterm infants. Applying the classification and regression tree (CART) algorithm described in the study allows for targeted expansion of high-risk infant follow-up enrollment to include those term and late preterm infants who may benefit most.

Quality improvement sustainability to decrease utilization drift for therapeutic hypothermia in the NICU.

OBJECTIVES: Therapeutic hypothermia (TH) is now standard of care for the neuroprotection of patients with moderate to severe hypoxic-ischemic encephalopathy (HIE). TH misuse results in increased medical complication rates and high health care resource utilization. Quality improvement (QI) methodology can address drift from clinical guidelines. Assessment of sustainability of any intervention over time is an integral part of the QI methodology. METHODS: Our prior QI intervention improved medical documentation using an electronic medical record-smart phrase (EMR-SP) and demonstrated special cause variation. This study serves as Epoch 3 and investigates the sustainability of our QI methods to decrease TH misuse. RESULTS: A total of 64 patients met the diagnostic criteria for HIE. Over the study period, 50 patients were treated with TH, and 33 cases (66%) used TH appropriately. The number of appropriate TH cases between cases of misuse increased to an average of 9 in Epoch 3 from 1.9 in Epoch 2. Of the 50 cases, 34 (68%) had EMR-SP documentation included. Length of stay and TH complication rates did not vary between cases of TH misuse and appropriate TH use. CONCLUSIONS: Our study confirmed a sustained decrease in TH misuse, despite inconsistent use of EMR-SP. We speculate that culture change involving increased awareness of guidelines through education may have contributed more to a lasting change.

HOPE and DREAM: A Two-Clinic NICU Follow-up Model.

OBJECTIVE: The natural extension of inpatient-focused neonatal neurocritical care (NNCC) programs is the evaluation of long-term neurodevelopmental outcomes in the same patient population. CLINICAL DESIGN: A dedicated and collaborative team of neonatologists, neonatal neurologists, neuropsychologists, neurosurgeons, physical medicine and rehabilitation physicians, and psychologists are necessary to provide personalized medicine, developmental assessments, and parental education for NNCC graduates. To achieve this goal, we devised a two-clinic follow-up model at Children's Wisconsin: HOPE (Healthy Outcomes Post-ICU Engagement) and DREAM: Developmentally Ready: Engagement for Achievement of Milestones) clinics. Those infants with significant neurologic diagnoses attend DREAM clinic, while all other high-risk neonatal intensive care unit (NICU) infants are seen in the HOPE clinic. CONCLUSION: These clinic models allow for a targeted approach to post-NICU care, which has improved family engagement and perceptions of value. KEY POINTS: · Infants with neurologic compromise are a specialized population with increasing survival.. · Interdisciplinary NICU follow-up brings together previously separated outpatient service lines.. · Our novel clinic model allows for specialized developmental assessments..

Quality improvement for reducing utilization drift in hypoxic-ischemic encephalopathy management.

OBJECTIVES: Therapeutic hypothermia is an effective neuroprotective intervention for infants with moderate or severe hypoxic-ischemic encephalopathy (HIE). With the introduction of new medical therapy comes a learning curve with regards to its proper implementation and understanding of eligibility guidelines. We hypothesized that variation in patient selection and lack of adherence to established protocols contributed to the utilization drift away from the original eligibility guidelines. METHODS: A retrospective cohort study was conducted including infants who received therapeutic hypothermia in the neonatal intensive care unit (NICU) for HIE to determine utilization drift. We then used QI methodology to address gaps in medical documentation that may lead to the conclusion that therapeutic hypothermia was inappropriately applied. RESULTS: We identified 54% of infants who received therapeutic hypothermia who did not meet the clinical, physiologic, and neurologic examination criteria for this intervention based on provider admission and discharge documentation within the electronic medical record (EMR). Review of the charts identified incomplete documentation in 71% of cases and led to the following interventions: 1) implementation of EMR smartphrases; 2) engagement of key stakeholders and education of faculty, residents, and neonatal nurse practitioners; and 3) performance measurement and sharing of data. We were able to improve both adherence to the therapeutic hypothermia guidelines and achieve 100% documentation of the modified Sarnat score. CONCLUSIONS: Incomplete documentation can lead to the assumption that therapeutic hypothermia was inappropriately applied when reviewing a patient's EMR. However, in actual clinical practice physicians follow the clinical guidelines but are not documenting their medical decision making completely. QI methodology addresses this gap in documentation, which will help determine the true utilization drift of therapeutic hypothermia in future studies.