Childhood Cancer Incidence Surveillance
Surveillance analyses can be performed on routinely collected childhood cancer incidence data. Such analyses are best performed at the census tract level. Above the tract level, environmental hazards are not uniform; below the tract level, the data are too sparse.
The traditional method of evaluating risk is calculation of rates. Even at the census tract level, this is unsatisfactory for childhood cancer beause the data are too sparse.
Grouping of tracts is undesirable because geographic detail is lost. Furthermore, one must make an arbitrary choice about how to group the tracts. Such an arbitrary choice can bias the results of the analysis.
A superior method is to adjust the tract boundaries so that every tract has an area proportional to its population. Now the population density is uniform over the whole map. Such a map is called an area cartogram, and the transformation that produces the cartogram is called a DEMP (Density Equalizing Map Projection). See "Cartograms and Density Equalizing Map Projections (DEMP)," dempcalc.com.