The Lead Safe API uses characteristics about the patient to determine if they are at a higher risk for lead poisoning. This determination is based on a machine learning algorithm that considers many factors that directly and indirectly correlate with future lead poisoning.
The model behind Lead Safe API, originally developed by the Data Science for Social Good Program at the University of Chicago, includes clinical data, census data, open data on buildings, and data belonging to the CDPH Lead Program. Because all of this data is collected and stored separately, applying it can be difficult. One challenge was turning a large amount of information into a comprehensive, yet comprehensible, data set. Another is providing this data to physicians in a format they can use. This project led to a streamlined method of sharing data across sectors of health by using an API, which will also be made available to other healthcare providers in the future.
“Decades of data on lead prevalence exist, but data silos impede its potential impact on addressing disparities that are perpetuated due to social determinants of health,” says Nivedita Mohanty, M.D., AllianceChicago’s Chief Research Officer and Director of Evidence Based Practice.
CDPH worked with University of Chicago data scientists to develop and deploy a predictive algorithm to identify lead risk. AllianceChicago provided leadership on the design and implementation. AllianceChicago is also conducting the initial pilot within their health centers to make this information available to clinicians who work directly with families living in at-risk neighborhoods.