The present paper examines the relationship between deprivation and changing patterns of public hospital admissions in Christchurch, New Zealand, between 1992 and 1997, during a time of economic restructuring and rapid change in the health sector. The total set of admissions into Christchurch Hospital was geocoded according to the meshblock domicile of each patient. Domiciles were classified into 10 decile categories using the NZDep91 and NZDep96 measures of deprivation. Regression analysis was used to measure changes in the relationship between deprivation and different types of admissions. Differences between admission rates for people living in the most and least deprived areas increased over time, especially following the implementation of the 1993 health reforms. This was most marked for younger adults (ages = 25-44 years), day patients, and especially, acute day patients, ambulatory-care-sensitive admissions and re-admissions. The average length of stay also varied by deprivation and appeared to be an important cause of the increasingly high rate of re-admissions. On average, patients from more affluent areas are hospitalised longer than low-income patients, although the differences narrow over time. The results suggest that the widening social gap in hospitalisation rates is a result of the effects of poverty and problems of access to primary care. However, more research on different admission pathways and causes of admissions for different patients from different parts of the city is needed to confirm these observations.