A feasibility study using geographic access to general practices and routinely collected data in public health and health services research

Background: Advances in geographic information systems (GIS) and increased availability of routinely collected data have the potential to contribute to public health and health services research. The aim of this feasibility study was to explore the use of GIS to measure access to general practices and its relationship to selected antenatal behaviours. Methods: We obtained the Perinatal Data Collection, 2004-2008, for South Western Sydney Local Health District from the New South Wales Ministry of Health. Using a GIS, we generated circular and polygon-based 1 km and 2 km road network buffers from a maternal residential address and then calculated the number of general practices within those buffers (density measures). Distances to general practices from maternal residences were also calculated in the GIS as another measure of geographic access to general practices (proximity measures). We used generalised estimating equations logistic regression to examine associations between access to general practices, and smoking during pregnancy and late first antenatal visit. Results: The numbers of general practices within circular buffers were significantly greater than within polygon-based road network buffers. The average distance to the five closest general practices from a mother's residence was 2.3 km +/- 3.0 km (median 1.4 km, range 0.05 km to 40.0 km). For two of the five measures of access, a higher number of general practices was associated with increased risk for late first antenatal visit. However, the associations were weak, with odds ratios very close to unity. Choice of metrics of geographic access to general practices did not alter the associations with smoking during pregnancy or late first antenatal visit. Conclusion: GIS can be used to quantify measures of geographic access to health services for public health and health services research. Geographic access to general practices was generally not associated with smoking during pregnancy or late first antenatal visit. GIS and routinely collected health data can be used to answer novel questions in public health and health services research.