I'm trying to detect clusters of Lyme disease that are not explained by an area's "greenness." SaTScan cannot use continuous data for covariates, so I've created NDVI quintiles. The categories show some correlation with LD rate but I'm unsure if doing this invalidates the analysis.