Using Machine Learning and Data to Predict Risk and Protect Families from Lead Exposure

Childhood lead exposure is a persistent public health issue linked to a range of developmental and neurological problems. Low-income and Black children are particularly vulnerable, as they are more likely to live in older homes (built before 1978) with Pb-based paint or near industrial sites that released Pb into the environment. Currently, state and local agencies rely on children’s blood lead levels to indicate whether neighborhoods and homes are lead free. This approach is reactive, slow, and potentially has grave impacts on the life course of impacted children. Simply put, by the time a child is tested, they may have already suffered irreversible harm.  

The PREDICT Healthy Homes Project is seeking to shift the current approach towards a proactive, preventive model that relies on risks factors rather than children to target neighborhoods and homes for lead exposure prevention.  

In this Roundtable Discussion, we will integrate a presentation and audience participation building on machine learning techniques, community engaged science and data on children’s total residential Pb exposures to identify high-risk areas and match interventions to household context. The session is intended for professionals at all career stages who are dedicated to ensuring healthy homes and improving public health.