Area (2008) 40.3, 354–364 Area Vol. 40 No. 3, pp. 354–364, 2008 ISSN 0004-0894 © The Authors. Journal compilation © Royal Geographical Society (with The Institute of British Geographers) 2008 Blackwell Publishing Ltd Examining access to GP services in rural Ireland using microsimulation analysis Karyn Morrissey*, Graham Clarke**, Dimitris Ballas † , Stephen Hynes* and Cathal O’Donoghue* *Rural Economic Research Centre, Teagasc, Athenry, Republic of Ireland Email: karyn.morrissey@teagasc.ie **University of Leeds, School of Geography, Leeds LS2 9JT † University of Sheffield, School of Geography, Sheffield S10 2TN Revised manuscript received 6 June 2008 This paper analyses the effect accessibility has on General Practitioner (GP) utilisation rates at the sub-national level for Ireland. Specifically, the aim of this paper is to estimate whether there is an urban–rural differential in GP utilisation rates. We do this by simulating micro-level healthcare data. Using this synthetic data, simple logit models are employed to estimate the likelihood that individuals in different jurisdictions will attend a GP surgery. These individual logit estimates are then inputted into a spatial interaction model to highlight areas with low GP accessibility given their health status. The policy implications of these results are discussed in relation to both the healthcare literature and current Irish healthcare policy. Key words: accessibility, spatial interaction models, spatial microsimulation, rural healthcare Introduction The main aim of most publicly provided healthcare systems is to increase the health status of its residents. Secondly, there is a need to ensure that all residents have equitable access to these services. As such, the importance of identifying areas or individuals with high (or unmet) healthcare demand becomes paramount to public health. The difficulty in measuring ‘healthcare need’ has long been realised (Barnett et al. 2002). Traditionally, standardised mortality ratios and morbidity rates have been used as indicators of healthcare need. Using these measures as proxies for healthcare demand has consistently shown that urban areas have much higher levels of ill-health (Congdon 1995; Phillimore and Reading 1992). However, mortality and morbidity rates are only proxies for healthcare need and as such their usefulness as indicators is much debated (Barnett et al. 2002). Given the limitations of such indicators as proxies for healthcare demand and the well-known link between socio-economic characteristics and health status (Haynes and Gale 1999), recent studies have moved towards using deprivation indicators as a proxy for health outcomes (Hayward et al. 1997; Barnett et al. 2001). Deprivation indicators have been successfully developed and used to estimate the link between healthcare need and deprivation in urban areas (Townsend et al. 1988; Carstairs and Morris 1989). However, research tends to reveal a weak link between deprivation and poor health in rural areas (Phillimore and Reading 1992). A number of key papers (Barnett et al. 2001; Haynes and Gale 1999) have suggested that this weak link between rural health status and deprivation may not be because rural areas are healthier places to live but rather is a result of an estimation bias created by using aggregate data. Rural areas generally have a larger geographical area and much smaller popu- lations compared to their urban counterparts. As the number of individuals commuting increases, rural areas also tend to be more socially heterogeneous