In a testing lab, tiny blood samples sit in neat rows, each one representing a newborn baby. Most of those babies are healthy. But hidden among them could be a child with cystic fibrosis, a serious genetic disease that can cause life-threatening lung and digestive issues if not treated early.
Newborn screening for cystic fibrosis, or CF, is “a state-level public health initiative that is universal across the U.S. and many other nations,” said Doug Bish, Professor of Operations Management at Culverhouse. “Our goal is to improve it.”
Doug and Ebru Bish, also a Professor of Operations Management, recently published a paper, “Improving Cystic Fibrosis Screening Through a Novel Testing Design,” in Production and Operations Management. Their research introduces a data-driven framework for a more accurate and efficient CF newborn screening protocol, specifically designed to navigate the complex economic and logistical constraints faced by individual state health departments.
The challenge begins with cost. “Because comprehensive genetic testing is both high-cost and capacity-constrained, newborn screening must follow a tiered approach. Each state starts its CF protocol with an inexpensive biochemical test that measures immunoreactive trypsinogen (IRT) levels,” Doug explained.
Today, all babies receive this low-cost IRT test. Only babies with elevated IRT levels (based on a certain cutoff) move on to second-tier genetic testing, which can look for many of the over 1,000 known genetic variants that can cause CF. The problem is, the IRT test has its limits. “When used as the first-tier test in the screening protocol, the IRT biomarker is not perfectly reliable on its own, and risks missing CF cases,” Ebru noted.
“A state’s clinical infrastructure and budget typically only allow for comprehensive genetic testing for only about 5% of their newborn population, or less, which creates a bottleneck in the screening process,” Ebru added. “Our research provides a data-driven framework to ensure that this 5% is comprised of the highest-risk infants. By optimizing the selection criteria for the second tier of testing, we can maximize the detection of CF cases within the specific laboratory and budgetary limits of any given state.”
Their innovative solution? “Adding a small-panel low-cost genetic test to the first tier of the protocol, which targets a few high-prevalence variants,” Doug explained.
About 70% of CF cases are linked to just one or two common variants. “While a limited genetic panel lacks the breadth to serve as a standalone screening tool, it becomes incredibly powerful when integrated with the existing IRT biochemical test,” Doug noted. “By combining these two distinct data points, the protein levels from the blood and the presence or absence of common genetic variants, we can significantly increase the positive predictive value. This means we can more accurately identify which infants should undergo comprehensive genetic testing, reducing both false-positives and missed diagnoses.”
Using data from states like New York and California, they tested how different combinations of IRT cutoffs and limited genetic panels would perform. One thing surprised Ebru along the way. “I was struck by the lack of standardization in CF screening processes across the country,” Ebru noted. “Even for the standard IRT test, cutoff points and screening protocols vary from state to state. Currently, many of these ‘rules’ lack a clear, data-driven justification.”
Ebru acknowledges that because states differ in their demographics and laboratory capacities, a single national standard might not be the answer. “States have unique resource constraints, so they don’t necessarily need identical systems, but they do need a rigorous framework to guide their policy choices,” she explained.
“Our research introduces a decision support system that allows policymakers to intentionally balance screening efficiency with accuracy. We are moving away from arbitrary rules toward an optimized approach where the choice of a cutoff and screening protocol are backed by mathematical evidence,” Doug added.
At its core, this research leverages data-driven optimization to solve a key public health challenge: making smarter, data-driven decisions within the reality of limited medical resources. By optimizing the CF screening process, the study ensures that more infants with cystic fibrosis are identified at the earliest possible stage, providing a vital window for medical intervention when treatment can make the most difference in a child’s life.