Selecting a lead hazard control strategy based on dust lead loading and housing condition: I. Methods and results.

A methodology was developed to classify housing conditions and interior dust lead loadings, using them to predict the relative effectiveness of different lead-based paint hazard control interventions. A companion article in this issue describes how the methodology can be applied. Data from the National Evaluation of the HUD Lead Hazard Control Grant Program, which covered more than 2800 homes in 11 U.S. states, were used. Half these homes (1417) met the study's inclusion criteria. Interior interventions ranged from professional cleaning with spot painting to lead abatement on windows, and enclosure, encapsulation, or removal of other leaded building components. Modeling was used to develop a visual Housing Assessment Tool (HAT), which was then used to predict relative intervention effectiveness for a range of intervention intensities and baseline floor and windowsill dust lead loadings in occupied dwellings. More than 117,000 potential HATs were considered. To be deemed successful, potential HATs were required to meet these criteria: (1) the effect of interior strategy had to differ for HAT ratings of good vs. poor building condition and/or baseline dust lead loadings; (2) the HAT rating had to be a predictor of one year post-intervention loadings; (3) interior intervention strategy had to be a predictor of one-year loadings; (4) higher baseline loadings could not be associated with lower one-year loadings; and (5) neither exterior work nor site/soil work could result in higher predicted one-year loadings for either HAT rating. Of the 1299 HATs that met these criteria, one was selected because it had the most significant differences between strategy intensities when floors and sills were considered together. For the selected HAT, site/soil work was a predictor of one-year loadings for floors (p = 0.009) but not for sills (p = 0.424). Hazard control work on the building exterior was a predictor of both sill and floor one-year loadings (p = 0.004 and p < 0.001, respectively). Regardless of the type of interior intervention strategy, interior work was a predictor of both floor and sill one-year loadings (each p < or = 0.001).