We tested the hypothesis that a relationship between ill health and deprivation exists for patients with diabetes, distinct from that experienced by the non-diabetic population. Age standardized admission and appointment rates and proportion of total activity for patients with and without diabetes were determined by electoral ward and correlated with the Townsend index of social deprivation for the health district of South Glamorgan (population 408,000). Both diabetic (r = 0.78, p < 0.001) and non-diabetic (r = 0.74, p < 0.001) in-patient admissions were positively correlated with social deprivation. This relationship also existed for attended out-patient appointments (r = 0.67, p < 0.001 and r = 0.45, p < 0.01, respectively). The proportion of diabetic to non-diabetic admissions by ward also showed a positive correlation for in-patients (r = 0.47, p < 0.001). This remained true for IDDM (r = 0.23, not significant) and NIDDM (r = 0.62, p < 0.001) diabetes, for admissions for coronary heart disease (r = 0.50, p < 0.001) and cerebrovascular disease (r = 0.29, p < 0.05), elective admissions (r = 0.30, p > 0.05), and emergency admissions (0.46, p < 0.001). Our results suggest that secondary care utilization is positively correlated with social deprivation and that this relationship is stronger in the diabetic population. This may be due to different prevalence rates or increased complications requiring hospital treatment in different social circumstances. Further research is required to examine these factors more closely.