Abstract
Background Most healthcare data have a spatial dimension. For investigators preparing a study, displaying the target population sizes or certain healthcare utilization in different geographic areas in a map will be very visual for investigators to know how the disease is spread and how much healthcare resources are used. For programmers who manage the database, presenting the geographic distribution of patient demographics, enrollment and utilization in a map could be the first step to check whether the data volume is normal and complete. When healthcare data are presented to audience, an informative map is worth a thousand words! However, map creation usually involves specialized software and is time-consuming. In SAS 9.2, creating highly-customized maps becomes surprisingly easy using GMAP procedure. This presentation/poster introduces several SAS macros that can transfer frequency tables of variables in VDW into colors and symbols on maps.
Methods Map creation in SAS/GRAPH involves two steps. The first step is preparing three datasets: a map data set which contains the information needed to draw map boundaries; a response data set which contains the information that will be displayed on the map; and an annotation dataset which customizes the map by adding text, symbols, contours, images and html links to the map. The second step is using the PROC GMAP procedure with the three datasets. SAS macros introduced here emphasize on the use of annotation. Annotation instructs SAS/GRAPH on how to customize graphs as if you were drawing with a pencil on a paper. For example, polygon can be drawn to emphasize the contour of certain areas so that audience’s attention can be led to certain areas; legend, made by the annotation of text and symbols, can be added to wherever you want in the map and with format that are not in standard SAS legend templates; images can be overlaid with transparent background so that the map is not obscured by the background the of image.
Results These SAS macros for mapping VDW data have the advantages of creating highly customized maps by reusable code without extra cost of specialized mapping software or training.
