Predicting general practitioner utilisation at a small area level across Western Australia.

Equitable delivery of GP services is a key goal in universal healthcare systems. In Australia, information to evaluate equitable delivery is limited, especially at finer geographic scales, leaving an information void that needs to be filled to inform, prioritise and target interventions. To fill this void, GP utilisation was estimated by combining responses on GP utilisation from a national survey differentiated by demographic and area-based socioeconomic and remoteness characteristics with similar characteristics represented geographically at a fine scale. These estimates were then compared to actual GP utilisation to evaluate their predictive reliability. Comparable estimates were found in the greater metropolitan area, with 76% of areas having estimated GP utilisation within ±10% of actual utilisation. Larger discrepancies were found as areas became remoter, with 84% of areas reporting estimated utilisation that was higher than actual utilisation. Comparing the geographic differences between estimated and actual utilisation allowed us to examine the reliability of our methodology. Given the identified limitations, a proxy for GP utilisation at a small area level can be created, a dataset that is not currently published at this geography. This approach has the potential to be applied Australia-wide, providing another valuable tool to evaluate the equitable delivery of primary health care nationally.