Figure 2.

Accuracy of reversely identifying patient location from a hypothetical low-resolution patient map in Boston, Massachusetts. The accuracy of the reverse identification was determined by (A) the distance between the reversely identified and the original addresses and (B) the number of buildings in which the patient could reside, given the reversely geocoded address. The reversely geocoded location was on average within 28.9 meters (95% CI, 27.4–30.4) of the correct address. The mean number of buildings in which the patient might reside was 7.7 (95% CI, 7.0–8.3).