Investigation of spacial clustering of rare diseases: childhood malignancies in North Humberside.

STUDY OBJECTIVE--The aims of the study were (1) to test for uniformity of distribution of childhood leukaemias and other malignancies; and (2) to consider the aetiological implications of unusual distributions. DESIGN--A test for spacial clustering was applied using a method which allows for unequal distribution of the population at risk and avoids using census data to provide population denominators. When clustering was identified, four possible aetiological links which had already been suggested to the Leukaemia Research Fund Centre were examined in a local area. SETTING--The study was carried out in the Yorkshire Health Region in the north of England. PATIENTS--144 children under 15 years of age with a diagnosis of malignant disease known to the Yorkshire Regional Childhood Tumour Registry between 1974 and 1986 were included in the analysis. Of these 53 had leukaemias and nine had lymphomas. MEASUREMENTS AND RESULTS--Significant localised clustering was found in North Humberside, though not in the whole of the Yorkshire Health Region. A number of clustered cases were identified, some of whom were in a post code sector, Hull 10, to the west of Kingston-upon-Hull, about which concern had been expressed since 1985. There was however no evidence that disease clustering was confined to this area. Four previously suggested hypotheses about causation in this particular area were examined but the results were negative or inconclusive. CONCLUSIONS--The identification of spacial clustering must be seen as only the first step in a series of investigations; it can only rarely lead to aetiological conclusions by itself, but it can motivate and target other investigations.