Since 1995, researchers at the SRP at Boston University (BU) have used Geographic Information System (GIS) data and increasingly sophisticated statistical methods to examine the geographical distribution of disease, which can provide important clues to the origins of the disease. These environmental epidemiological studies are complicated by factors including small sample size for case populations; the fact that disease registries only contain data on where people lived at the time of diagnosis, which may not be the time of exposure; and methodological problems of understanding, estimating and evaluating potential biases.