Aug. 26, 2026 — For many children, anemia develops quietly. The earliest stages often go unnoticed, with few symptoms to signal that the body has begun to decline. By the time fatigue, irritability or difficulty concentrating become apparent, the condition may already be affecting a child’s growth and development.

That challenge led University of Florida pediatricians Maria N. Kelly, M.D., and Molly Posa, M.D., to explore how clinicians could better identify children at risk for anemia, including those who may not show obvious symptoms.

Their early work examined the relationship between pediatric anemia and social determinants of health, finding that children with adverse social risk factors were more likely to have anemia.

Those findings helped lay the groundwork for a collaboration with the UF College of Medicine’s Quality and Patient Safety Initiative, or QPSi, to explore another question: Could health data help clinicians identify children at higher risk who might otherwise go undetected?

Using that data from approximately 29,000 children in Florida, the team developed a clinical decision-support tool that combines patient information, geographic trends and social determinants of health to improve pediatric anemia screening and help clinicians recognize children who may otherwise go undetected.

For the QPSi team, pediatric anemia presented an opportunity to improve care.

After examining health data from children across Florida, they found that nearly one-quarter showed signs of anemia, with rates varying considerably depending on where they lived.

“Pediatric anemia is common, its consequences are largely preventable if caught early and it’s relatively inexpensive and effective to treat once identified,” said David Cody Hall, Ph.D., a data scientist with QPSi. “The combination of high prevalence, real disparity and relative ease of treatment made it an excellent candidate for targeted data-driven screening intervention.”

Kelly said one of the biggest challenges in identifying pediatric anemia is that it often develops without obvious warning signs.

“Children with anemia look the same as children without,” she said. “You can’t always tell by doing an exam.”

Kelly and Posa, her co-principal investigator, have been working alongside the QPSi team to test how incorporating these factors into the clinical screening workflow could help identify additional children who may benefit from anemia assessment.

Traditionally, pediatric anemia screening criteria have relied on a child’s age and known risk factors, such as eating habits and family history.

Current guidance from the American Academy of Pediatrics recommends universal anemia screening around 9 to 12 months of age. After infancy, screening is left to a provider’s clinical judgment based on individual risk factors.

To better support providers in making those judgments, the QPSi team explored whether patterns within existing patient and community data could help identify children who may benefit from additional screening.

“We analyzed thousands of patients to identify relationships between hemoglobin results, patient age, geographic location, and multiple vulnerability measures,” Hall said.

The team examined several factors, including healthcare access, transportation, housing, education, neighborhood characteristics and economic conditions, as well as patient age and ZIP code.

Their analysis identified statistically significant pediatric anemia hotspots across North and Central Florida, providing the foundation for a rules-based decision-support model that can be incorporated into clinical workflows.

Rather than replacing provider expertise, Hall said the model is intended to complement clinical judgment by highlighting children whose combination of risk factors may unexpectedly warrant additional screening.

Posa said the model can also help clinicians identify potential risk factors that may not be immediately apparent during routine care.

“It’s really helpful because the whole point of it is making us realize there are other probable risk factors for anemia that maybe we’re not recognizing regularly,” Posa said.

The project also takes into account the role that social determinants of health play in pediatric care.

“Families experiencing food insecurity may struggle to provide foods rich in iron, making nutritional deficiencies more likely,” said Jefferey Miedema, M.B.A., a senior clinical project manager with QPSi.

But social factors such as food insecurity do not explain every pattern the team has identified. Researchers have also observed anemia in areas where children are not experiencing food insecurity, raising new questions about what other factors may be contributing to their risk.

Transportation barriers can also make it difficult to attend routine appointments or complete laboratory testing. This in turn might limit a patient’s access to preventive healthcare, which may delay diagnosis until symptoms become more severe.

By integrating these factors into the model, the QPSi team hopes to provide clinicians with additional context that supports more informed screening decisions while helping reduce disparities in pediatric anemia treatment across Florida.

For Miedema, the pediatric anemia project reflects QPSi’s broader mission of using data to improve healthcare quality and patient safety.

“This work serves as a practical example of using technology to close a known care gap,” he said. “It’s the same approach QPSi applies across other quality and patient safety initiatives.”

Although the project is still in its early stages, Miedema said, the model now screens thousands of pediatric patients each month for their risk of anemia.

Some physicians have reported screening children they otherwise may not have tested, creating an opportunity to identify cases of anemia that could otherwise go unrecognized. As the members of the QPSi team continue evaluating the model, they hope it will become another practical resource clinicians can use to support earlier diagnosis and improve outcomes for children throughout Florida.

“If we can identify children at risk in a way like this, I think it will allow us to improve care delivery to patients who need it the most,” Kelly said.

Turning the model’s predictions into better patient care requires another important step: putting that information into the hands of clinicians. At UF Health Pediatrics – Tioga, clinic manager Carol Ellis and nurse Cristy Brannin have helped integrate the model into day-to-day care.

Each week, the QPSi team provides the clinic with a report identifying patients who may be at higher risk for anemia. Ellis and Brannin review those reports and pend hemoglobin screening orders directly in patients’ charts, alerting pediatricians to the identified risk and giving them an opportunity to discuss screening with families during the visit.

Ellis and Brannin have also taken the initiative to track which patients are screened because of the model. The QPSi team plans to use those records to help evaluate patient outcomes and the model’s impact as the project advances.

“Carol and Cristy have been essential to getting this model out of the data and into actual patient care,” Miedema said. “Their initiative is what turns a risk score into a conversation between a provider and a family.”