Evaluation of public housing lead risk assessment data

A unique data set from lead risk assessments performed on 67 public housing developments from across the United States was made available for analyzes . The data set includes results of lead analysis from 5906 dust wipes and f rom 1222 soil samples. A total of 487 dwelling units in these developments, as well as associated common areas, were sampled, all by the same team of inspectors. The number of dwelling units within a development that were sam pled reflected the guidelines then in force, the 1990 Interim HUD Guideline s, rather than those specified in the 1995 Guidelines. Median dust lead loa dings for floors, 151 mug m(-2) (14 mug ft(-2)), and window sills, 936 mug m(-2) (87 mug ft(-2)), were much less than former HUD limits of 1076 mug m( -2) (100 mug ft(-2)) and 5380 mug m(-2) (500 mug ft(-2)), respectively and are only about one-third of the recently established limits of 431 mug m(-2 ) (40 mug ft(-2)) and 2690 mug m(-2) (250 mug ft(-2)). In contrast, the med ian lead loading for window troughs, 8560 mug m(-2) (795 mug ft(-2)), was a lmost identical to the HUD clearance limit of 8610 mug m(-2) (800 mug ft(-2 )). There was a strong positive correlation between floor and window trough lead loading values for samples from the same dwelling units and those fro m common areas of the housing developments. Door threshold samples, which m ay reflect conditions exterior to the dwelling unit, were collected from 53 dwelling units. Median lead loading levels of these samples were more than ten times higher than those in floor samples from the same dwelling units, were about the same as window sill samples and about one-half of levels in window trough samples. Composite sample results, simulated by averaging re sults from four samples within a dwelling unit, revealed that in order to h ave the same rate of excedence of standards, the composite standards would have to be reduced, for example, from the single sample value of 1076 mug m (-2) (100 mug ft(-2)) to 527 mug m(-2) (49 mug ft(-2)) for floor samples an d from the single sample value of 8610 mug m(-2) (800 mug ft(-2)) to 5160 m ug m(-2) (479 mug ft(-2)) for window troughs. For this public housing data set, the portion of the units in developments containing more than 225 unit s which exceeded the established limit for window samples was the same when using either the full data set or a random one-half of the data set. This suggests that, for this data set, the number of dwelling units sampled was excessive . Thus, the required increase in the number of dwelling units to be sampled specified in the 1995 Guidelines for developments with more than 225 dwelling units, may not have been necessary if this data set is repres entative of public housing developments in the United States.