Defining regions for locality health care planning: a multidimensional approach

The increasing significance of the role of the general practitioner (GP) in the British National Health Service, evolving from a provider to purchaser and now a key player in the organisation of Primary Care Groups, suggests the need for GPs to possess more and more information about their registered population. GP catchment areas, though an essential basis for providing GPs with important information such as levels of accessibility to surgery, are rarely clearly or accurately defined. Previous approaches towards the definition of GP catchments have been confined to single regionalisation methods, such as mean distance measures, and are prone to problems of either overestimating or underestimating medical service areas. This problem is compounded by a lack of acknowledgement that the application of contrasting catchment methodologies to a common service population has the potential to yield vastly different results which can have serious implications for health care planning and resource allocation. The lack of sophistication in the definition of medical service areas calls for a new methodology to be considered. In this paper, attention is given to the adaptation of multidimensional regional analytical techniques developed outside the health domain and applied in a Regional Health Authority in Northern Ireland. The technique involves the creation of a Synthetic Data Matrix (SDM) which compares patient to GP flow (affiliation) information aggregated at the Census Enumeration District level across a number of catchment areas created using different methodologies. The SDM is then analysed using a modified version of the European Regionalisation Algorithm to create an optimal set of non-overlapping regions according to pre-defined population size and self-containment criteria. The results, a set of compact, robust and highly self-contained catchments, are extremely encouraging. The paper considers the future potential use of such a methodology for health care planning and highlights areas for further research in this field.