5 Factors That May Predict Which Newborns Need More Follow-Up Care

Cleveland Clinic researchers identified five variables associated with increased pediatric office utilization in the first month after discharge, bringing them one step closer to developing a neonatal risk prediction tool.

The American Academy of Pediatrics recommends that infants have two pediatrician visits in their first month of life—a newborn visit and a 1-month visit. Although most newborns do well after discharge, some require repeated pediatric visits, emergency care or hospital readmission. Currently, clinicians have no reliable way to identify these infants before they leave the hospital.

Neonatologist Anirudha Das, MD, MPH, and his Cleveland Clinic Children’s colleagues have been developing a risk-scoring system integrated into infants’ electronic health records that could predict the likelihood of healthcare utilization in babies born slightly preterm at 35 to 36 weeks, allowing pediatric providers to shift resources to higher-risk infants.

The findings from a recent analysis were presented at the Pediatric Academic Societies.

The work builds on a 2020 pilot study led by Dr. Das that quantified parental questions or concerns among two groups of NICU infants. The findings, he explains, were somewhat paradoxical. Parents of infants with advanced higher gestational age and weight, but a shorter NICU stay, had similar questions or concerns to parents with more premature infants, who had greater medical complexity and a longer NICU stay.

“That prompted us to explore some factors that might predict why they would go to the emergency department [ED] or get readmitted or have more pediatrician visits than expected,” says Dr. Das. This launched a four-year quest to collect data from infant-mother dyads within the institution and develop a risk-scoring system to identify those who might require additional predischarge resources.

Identifying predictors of high office visit utilization

Using data from 21,061 infants born between 2016 and 2021, the team examined three outcomes: readmission or at least one ED visit, four or more office visits in the first month after discharge, and a composite outcome. Overall, 4.5% of infants were readmitted or had an ED visit, 21% had four or more office visits, and 25.5% met the composite outcome.

The dataset was randomly split for variable selection and model validation, with cross-validated c-statistics used to assess performance. Models for hospitalization/ED use and the composite outcome did not perform well enough for inclusion, but the office-visit model identified several predictive variables:

• Maximum transcutaneous bilirubin
• Gestational age
• Percent weight loss after birth
• Maternal BMI at delivery
• Maximum heart rate

“Together, these variables reflect newborn maturity, physiologic adaptation after birth and maternal health—factors that may influence how much medical attention an infant requires after discharge,” he explains. The researchers weren’t surprised by the findings, but another question looms larger: to what extent do these affect outcomes?

“Gestational age alone, especially if it’s late preterm, may not predict whether an infant will be readmitted,” Dr. Das says. “However, gestational age coupled with one or more other variables could increase risk considerably.”

Making progress on the long-term plan

The risk-scoring system will use the composite variables and, at a predetermined cutoff value, enable a provider to predict a higher likelihood of healthcare utilization.

The risk prediction project was initially funded with a Caregiver Catalyst Grant in 2022. Today, the outcomes data, including nearly 100,000 electronic medical records of infants born within the Cleveland Clinic health system. Dr. Das says the team is completing additional statistical analyses and plans to publish those findings.

“The scoring system will inform pediatric providers about which babies are more likely to need more support,” says Dr. Das. “We can focus on those babies and identify upfront how to prevent readmission, ED visits and excessive office visits.”

Although the model is being developed with Cleveland Clinic Children’s data, the broader goal is to offer a framework other institutions can adapt using their own patient data.

“Our long-term goal is to move from reacting to problems after discharge to preventing them before they occur," Dr. Das says. "If we can identify the babies most likely to need additional support, we can deliver that support proactively.”