Unlocking Large Data Sets To Drive Population Health Gains in Pediatrics

A Look Into the Work of Nicoleta Serban and Its Population Health Impact for Pediatrics

Technologies, policies and therapies closely related to population health take data analysis and its application to high levels of volume and, hopefully, positive impact.

While not tangible like innovations that result in a new physical device, drug, app, diagnostic or imaging tool, health analytics based on large data sets can address unmet needs at the population level.

Nicoleta Serban is Peterson Professor in the H. Milton Stewart School of Industrial and Systems Engineering at Georgia Tech. “Her research interests on Health Analytics include large data representation and analysis with a focus on processing patient-level health information into data features dictated by various considerations, such as data-generation process and data sparsity; machine learning and statistical modeling to acquire knowledge from a compilation of health-related datasets with a focus on geographic and temporal variations; and integration of statistical learning into informed decision making in healthcare delivery and into managing the complexity of the healthcare system.”

What does this mean for pediatric care innovation and provision?

“Looking back ten years or so to when my work at Georgia Tech became closely associated with Children’s Healthcare of Atlanta (CHOA), I was among a few researchers attempting to make sense of very large healthcare data sets, including Medicaid claims data. Our work then, as now, seeks to improve access and delivery of care for pediatric populations in ways that could make an easily understandable, and have short and long-term impact,” Dr. Serban told us.

“Research directions identified by partners like CHOA and the Institute for People and Technology at Georgia Tech (IPaT) brought together the network and funding that supported research we knew was promising in its potential to inform health policy impacting children’s health.”

Dr. Serban’s prolific research and publications extend into multiple health areas of concern at the population level. For the sake of brevity, we share two of those with evidence of impact; preventive dental care and interdisciplinary interventions for treatment of pediatric feeding disorders.

From Dr. Serban et. al. in the 2017 publication Disparities in Preventive Dental Care Among Children in Georgia, “Across communities, high-income children had better access to preventive dental care than low-income children had. Identifying tracts with disparities in access could result in more efficient allocation of public health dental resources.

In the same year, the Georgia legislation authorized Georgia dentists to allow their dental hygienists to provide preventive services in safety net settings, such as long-term care facilities and school-based clinics without their supervision.

“We were able to demonstrate that access to simple preventive care in rural communities could have a meaningful, positive impact in the lives of those kids and their families while alleviating the state’s overall financial burden associated with healthcare spending at the same time.”

Another of Dr. Serban’s studies demonstrated healthcare cost savings for treating children with feeding disorders when they have access to in-patient interdisciplinary care versus the common gastrointestinal tube used for feeding.

Developed over the course of more than a decade, the Marcus Autism Center and Children’s Healthcare of Atlanta developed an interdisciplinary evidence-based Feeding Program. The program’s “intensive multidisciplinary treatment model combines daily behavioral intervention and parent training with nutrition therapy, oral-motor therapy and medical oversight. The retrospective study evaluating the program, “Intensive Multidisciplinary Intervention for Young Children with Feeding Tube Dependence and Chronic Food Refusal: An Electronic Health Record Review,” was published in the Journal of Pediatrics. The analysis included 81 patients ages 10 months to 19 years who relied exclusively on a feeding tube for most nutrition and whose previous treatments were unsuccessful. They received four meals per day over five days a week for two months or more. At discharge, their oral nutrition intake increased by 70% and food refusal decreased 68%. At follow-up, 72% of patients were completely weaned from their feeding tube.”

Dr. Serban et. als.’ “Economic Analysis of Intensive Multidisciplinary Interventions for Treating Medicaid-Insured Children with Pediatric Feeding Disorders” found, “this model of care not only holds benefit in terms of improving overall quality of life but also brings significant expenditure savings in the short and long term.”

Private pay insurance companies instated a CPT code for such interdisciplinary care for children with feeding disorders in the same year. Medical CPT codes are critical for claims processing. Private pay insurance companies can now pay per diem using the CPT code (other services). Legislation through the state to cover the intensive model is the ultimate goal allowing Medicaid to use the CPT code.

“With this treatment and code, thousands of children will be able to avoid growth delays, decreased cognitive ability or compromised immune function, not to mention the financial savings to the healthcare system as a whole,” Dr. Serban said.

Using Data Analytics to Create Real Change

Nicoleta Serban

“Data analysis has to be informed by collaboration with clinicians, policy makers, and researchers,” Dr. Serban said. “Concurrent data, modeling, along with the vision and willingness to go the extra mile creates real change. That is what makes the most out of our work and its potential.

“Looking back, especially regarding the Medicaid claims data, it was probably better we didn’t know exactly how big of a challenge we were up against. We knew the work was important and just kept going. Because of that collective commitment, we now have several researchers working with these large data sets to generate the evidence and insights we need, to improve the lives of kids and families in significant numbers.”

Looking ahead Dr. Serban points to the potential of the Medicaid claims data dating back to 2005, demonstrating empirical evidence of the impact programs have on outcomes as kids become adults.

“If a child received mental health treatment, what does the data show in their quality of life ten years on? We will be able to see that evidence, that result, much more clearly given that we have access to data from back 2005. Thanks to the commitment of the Institute including IPaT, we continue learning from large data sets in order to improve the lives of kids and families close to home and across the country.”

Programs nascent and newly in-flight include implications of Medicaid expansion in states for expecting mothers, as well uncovering opportunities for improvement in the care delivery of children living with diabetes and  or children diagnosed with mental health conditions.

“As W. Edwards Deming said, ‘In God we trust. All others bring data.’ That’s my mantra as well and my favorite thing to share with others when speaking about our work,” Dr. Serban told us. “I believe innovation, regardless of its specific discipline in health, reaches its highest potential when it centers on the three effect framework, efficiency, effectiveness and equity. That ethos guides our work, every single day.”

Georgia Tech Pediatric Technologies thanks Dr. Serban for helping us share her story here.