Community networking as a means for identifying people with diabetes in a rural, predominantly bicultural community in New Zealand.

Aims. To assess the use of community networking to estimate the prevalence of diabetes in a predominantly New Zealand Maori and European community. Methods. A cross-sectional survey of people with known diabetes identified either through general practice or community networks (others with diabetes , public notices or public meetings) was undertaken. Ascertainment was comp ared using capture-recapture methods for two independent samples. Results. Overall ascertainment by community networking was greater for Maor i than Europeans (40 +/- 3% vs 15 +/- 2%, p < 0.001). Ascertainment using g eneral practice registers was comparable in the two ethnic groups (48 +/- 4 96 vs 55 +/- 6%, respectively). Women were more likely than men to be conta cted through community networking (odds ratio 1.47, 1.05-2.06). Conclusion. In closely knit communities, community networking provides an i ndependent source for estimating the prevalence of diabetes.