Improving the evidence-base for access to primary health care in Canterbury: a panel study

Objective: Despite many reforms and initiatives, inequities in access to primary health care remain. However, the concept of 'access' and its measurement is complex. This paper aims to provide estimates of general practice visit frequencies for 'attenders' (those who seek consultation) and the proportion of 'non-attenders' (those who never seek consultation) of primary health care services. Methods: A panel study of people enrolled within a large primary health care organisation of affiliated general practices. Standard and zero-inflated regression models were assessed. Results: 980,918 visits were made by 388,424 people, averaging 2.64 visits/person/year. The zero-inflated negative binomial model was superior, and significant age, gender and ethnic differences were observed in attender and non-attender profiles. More Asian (21.0%), Pacific (19.6%) and Maori (17.1%) people were non-attenders than European/Other (9.0%) people. Among attenders, males, Asian and Pacific people, and young to middle-aged adults, generally had relatively lower visit rates. Conclusions: Interpretation of utilisation data must be made with caution because of two distinct characteristics: the differential rates of non-attenders and the highly dispersed distribution of attenders. Implications: Improved understanding of differential non-attender rates and attender visit distributions by demographic factors needs to be considered when addressing improved access to general practice services.