Hart, C., Ecob, R & Davey Smith. G. People, places and coronary heart disease risk factors: a multilevel analysis of the Scottish Heart Health Study Archive. Social Science and Medicine. Vol.45,6, 893-902.

Geographical variation in four risk factors for Coronary Heart Disease (diastolic blood pressure, cholesterol, alcohol consumption and smoking) were examined, for men and  women separately using data from the Scottish Heart Health Study, a large dataset set up in part to investigate geographical variation in Coronary Heart Disease in Scotland, between 1984 and 1996. Multilevel analysis found evidence of significant between district variation in three (diastolic blood pressure, cholesterol, smoking) out of the four risk factors after controlling for socio-economic and other variables at the individual level. Results suggest that place may have a role in the distribution of coronary heart disease risk.

Ecob, R, Robertson, C and Watt. G. Has Regional variation in mortality rates declined since 1931, and in all age groups, in Britain? A reanalysis using formal statistical modelling. J Epidemiology and Community Health. 1997, Vol 541,5,502-509.

Raymond Illsley and Julian LeGrand were responsible for collating and analysing a dataset of mortality by age and sex (from Census data and Death Registration data around Census years) and region in Great Britain from 1931 to 1981, and interestingly pointed to a reduction in variance between regions in mortality in younger but not older age groups (Illsley and LeGrand, 1993). We extended this dataset to add in data for 1991 and analysed this using both age-period-cohort models and bootstrap simulation tests (the latter to test for changes in the regional variance in log mortality rates over the sixty year period). Differences (reductions) in regional variance were confirmed for the younger age groups (only). Models for changes in regional variances based on period effects provided better description of observed variances than those based on cohort effects. For both sexes the data was well fitted by a model which assumed up to a quadratic effect of period, varying according to age, with extra dummy variables to account for excess mortality in the war years. In younger, but not older, cohorts there was evidence of a rise in the variance between regions between 1981 and 1991.