Heather Siefkes
University of California-Davis, CA·Medicine & Clinical Research·Early-career
Congenital heart disease is the most common birth defect. Its most serious forms, called critical congenital heart disease (CCHD), can be missed by both prenatal ultrasound and the newborn exam.
- Fields
Medicine & Clinical Research
Biomedical Engineering
Computer Science
- Topics
critical congenital heart disease
coarctation of the aorta
newborn screening
- Grants
- 1 known grant, through Feb 2027
What they work on
- Externally validate a machine learning algorithm that combines pulse oximetry features, including oxygen saturation, perfusion index, and pulse delay, for detecting critical congenital heart disease in hospitalized newborns.
- Test a repeat-screen approach for newborns who fail the initial machine learning screen, mimicking the current oxygen-saturation standard, with the goal of keeping the false positive rate below 1%.
- Develop a dynamic machine learning model that updates each newborn's CCHD prediction with pulse oximetry readings taken after 48 hours of age, including at outpatient visits, to improve detection of coarctation of the aorta.
- Measure how long it takes to obtain artifact-free pulse oximetry waveforms from the right hand and foot of newborns, to inform how the screening could be put into practice.
More about the research
Newborns are now screened with pulse oximetry, a clip-on sensor that measures blood oxygen saturation. But this test misses about half of babies with CCHD who show no symptoms, nearly 900 newborns in the U.S. each year. Many of the missed babies have blockages in blood flow to the body, such as coarctation of the aorta (CoA), a narrowing of the main artery. These defects do not lower oxygen levels.
Why it matters
A screening algorithm that adds blood-flow measurements to standard oxygen readings could catch hundreds of babies with critical heart defects each year who are now missed, especially those with coarctation of the aorta. Earlier detection could prevent delayed diagnosis, poor outcomes, and deaths. The approach uses pulse oximetry data that newborn nurseries and clinics already collect.
Known grants
From NIH and NSF research grants. Labs often have other funding we can't see.
Closest labs
Karen Fairchild
University of Virginia, VA·Medicine & Clinical Research
Refine and expand algorithms that predict sepsis and necrotizing enterocolitis (NEC) in very low birth weight infants, applying time series analysis and machine learning to continuous heart rate and pulse oximetry oxygen data from four NICUs.
Shabnam Peyvandi
University of California-San Francisco, CA·Medicine & Clinical Research
Test whether lower brain tissue oxygen before birth leads to immature brain development in fetuses with transposition of the great arteries, using fetal neuroimaging to find a prenatal marker of poor brain maturation.
Elliot Botvinick
University of California-Irvine, CA·Biomedical Engineering
Design and build a minimally invasive, ultra-thin flexible sensor placed just under the skin that continuously measures glucose, lactate, oxygen, and pH — analytes that report tissue oxygen delivery and metabolic stress.
The school
University of California-Davis: 49 labs on PI Pages, $1.0B a year on research, 647 research doctorates a year.
Reach out
A start for your email. Fill in the brackets.