Analysing geographic variation in cardiovascular mortality: methods and results.

One common type of epidemiological study is to compare disease-specific mortality rates in different areas. In evaluating the geographic association between mortality and other characteristics multiple regression techniques may be used. This paper presents one example concerning British geographic variation in cardiovascular mortality as an illustration of the methods and problems of statistical analysis and interpretation for such studies. In particular the relevance of weighted regression and spatial patterns is considered." The negative association between cardiovascular mortality and water hardness is explored. A series of comments is also included at the end of the paper (pp. 329-41). (EXCERPT)