Are geographic variations in death rates for the cardiovascular diseases real?

Abstract

No doubt the assignment of causes of death is often inaccurate, and any practicing physician can probably recall a large number of instances in which the cause of death was improperly entered on the death certificate. The magnitude of such inaccuracies undoubtedly varies somewhat from state to state and these inaccuracies may very well account entirely for small differences in death rates for cardiovascular diseases or coronary heart disease. It seems unlikely, however, that they could account for all of the large differences in death rates shown here. Evidence presented may be summarized as follows: 1. 1. The contrasts in death rates for cardiovascular diseases for some states, such as South Carolina and North Dakota, are sufficiently great to make it seem certain that some real differences exist. It seems likely that real geographic variations in death rates for cardiovascular disease and coronary heart disease of considerable magnitude exist within the United States. 2. 2. The quality and quantity of physicians available do not appear to have an appreciable effect upon cardiovascular death rates, and only a small percentage of the variation in coronary heart disease death rates appears to be related to this factor. 3. 3. Population and total death counts appear to be fairly accurate. 4. 4. Autopsies and other studies indicate that, in general, errors in cause-of-death entries on death certificates tend to be compensatory, so that for major groupings such as cardiovascular diseases or coronary heart disease net errors in the counts are possibly less than 7 per cent.