RESUMO
OBJECTIVE: To test the impression of an increased cancer incidence on the island of Pellworm (in the far North of Germany) and to illustrate the feasibility of a general practice-based approach in epidemiological research. DESIGN: Cancer incidence on Pellworm was prospectively registered in the only general practice on the island from 1986 to 1992. Age-standardized rates and expected rates were calculated on the basis of the Saarland cancer registry, the only registry in Germany. Standardized incidence ratios and 95% confidence intervals (CI) for Poisson-distributed events were also calculated. The cancer data were summed up over a seven-year period. SETTING: The only general practice on the island of Pellworm, a North Frisian marshland-island. SUBJECTS: The total practice population between 1986 and 1992 (N = 1172) RESULTS: The crude annual cancer incidence rate for Pellworm was, according to the impression, unexpectedly high: 634/100 000 for men and 502/100 000 for women. After age standardization, however, this increased rate of cancer incidence was even lower than in the Saarland (ratios: 0.86 for men and 0.95 for women). Only the incidence of neoplasms of the lymphatic and haematopoietic tissue in men exceeded the limits of statistical likelihood (ratio: 3.21; 95% CI: 1.17-7.10). CONCLUSION: The impression of an increased rate of cancer on Pellworm could not be validated. The overall incidence rate was even lower than expected. Only the rate of leukaemia/lymphoma (men) was significantly higher. Reasons for this result could not be detected by a descriptive approach. General practice is a suitable place for studies in cancer epidemiology, especially in such ideal circumstances as a clearly limited area and complete and reliable documentation.