Preventing Childhood Lead Poisoning
Targeting Proactive Inspections for Lead Hazards
The Challenge
Lead poisoning imposes lifelong health and economic costs on hundreds of thousands of people every year in the United States. A ban on leaded consumer products in the United States was not enacted until the late 1970s (Needleman, 1998). To this day, lead in paint remains a significant hazard to children in particular. Exposure to lead is associated with premature birth, edema, herniation, atrophy, and white-matter degeneration (Cleveland, et al, 2008; Bellinger, 2008). Elevated blood lead levels are associated with lower IQ in children as well as with poorer achievement on reading and math standardized tests in the third grade (Evens et al, 2015). Lead also correlates with crime rates (e.g. Stretesky and Lynch, 2004). Lead-related child health issues conservatively bear a price tag of over $40 billion annually (Landrigan, et al, 2002).
Existing Approach
The current approach to identifying and remediating lead hazards is to test all children for elevated blood lead levels and then investigate and remediate the indicated homes. This helps prevent exposure for future residents but comes too late for poisoned children, who disproportionately come from low socioeconomic status and minority groups. In addition, health professionals around the country devote an enormous and concerted effort to the problem of lead exposure, yet many face a shortage of resources. For example, at current levels of funding and staffing, it would take the Chicago Department of Public Health (CDPH) 76 years and $98 million to simply inspect the city’s 197,157 old buildings, let alone remediate them. The only hope of making a significant impact with the available budget is to use it more efficiently.
Our Solution
The Center for Data Science and Public Policy (DSaPP) is using machine learning to create a real-time, actionable system for health professionals to predict and remediate lead poisoning hazards before children are poisoned and suffer lifelong health and development consequences. This integrated and innovative system will ensure that resources are used most efficiently and ultimately mean healthier children. This prevention stance will serve to better level the playing field by remediating a core determiner of health for low-income and at-risk children nationwide.
What We’ve Done
In conjunction with CDPH, DSaPP has built an empirical model to predict the risk of a child being poisoned so that an intervention can take place before that happens (Potash, et al, 2015). Using two decades of blood lead level tests, home lead inspections, property value assessments, and census data, the model, when fully built and integrated technologically, will allow inspectors to prioritize houses on a long list of potential hazards and identify children who are at the highest risk. By building statistical models that predict exposure based on evidence such as the age of a child’s home, the history of children’s exposure at that address, and economic conditions of the child’s neighborhood, CDPH and DSaPP have the potential to link high-risk children and pregnant women to inspection and lead-based paint mitigation before any harm is done.
What We’re Doing
- Validating the model. Our DSaPP team has trained the existing empirical model to optimize for predictive accuracy. To fine-tune and test this approach and model, CDPH will field test the model by providing a dynamic testing regime in which we will choose addresses for city officials to inspect, receive feedback and results from CDPH inspectors, then update the model and continue to iterate and improve. We’re choosing residential units from three categories — high risk, medium risk, and low risk — to validate that the model gives accurate predicted probabilities for each category. CDPH believes it can conduct 400 of these inspections to test and improve the model with real-time input and responsive feedback.
- Implementing the model at the point of care. We received a Data Across Sectors for Health (DASH) grant from RWJF. Through this grant, we are implementing the model on City of Chicago servers and building an API that Alliance of Chicago healthcare providers will use to provide doctors with point-of-care risk scores.
Are You Our Next Partner?
We’d like to partner with more organizations that have the data, staff, resources, and willingness necessary to adapt and implement the model. To do this project, you will need at least three years of the following data:
- Individual blood lead level
- Test date
- Date of birth
- Home address
- Building lead inspections
- Remediation compliance
- Year home was constructed
- Birth certificates, WIC records, EMRs, or some other data source that identifies where babies live within three months of birth
You will get more from the model if you also provide the following:
- Name
- Demographic information (sex, race, whether the child is an immigrant, etc.)
- Type of test (venous / capillary)
- Laboratory that conducted the test
- Interior hazard found
- Exterior hazard found
- Building improvement history
- Building condition (e.g. sound/unsound)
We have also developed a specialized Data Maturity Framework that helps you assess your organizational and data readiness for this project. Our experience suggests you need to fall within the colored sections for success:
You can find our standard legal agreements here and notes on dumping your database and setting up a computational platform here.
Ready to Contact Us?
If you think you fit, please let us know.