The value of ischaemic heart disease vital statistics since 1968.

Although individual cases of IHD (Ischaemic Heart Disease) are frequently misclassified, the overall reliability of IHD vital statistics in industrialized countries is better than claimed. Random classification errors produce a correct mean value, whereas systematic misclassifications produce an opposite change in one or more different causes of death, since total number of deaths is correct in the age group and countries under consideration. Therefore it is important to compare a certified cause of death to several others, including total mortality. The comparison of death rates should be done in many countries, over several years, for each sex and between sexes. Since the 8th revision (1968) the comparability between years has improved. The mean mortality rates over the available years, the slopes calculated over the years and the % change of IHD in men and in women were correlated with different causes of death in men and in women, totalling 111 comparisons. Some of these (cancer of the rectum, colon, prostate and breast, diabetes and stroke) are related with nutrition. In total 71 out of the 111 correlations were significant, 28 of which with p less than 0.001. Criteria have been defined for checking the reliability of the slopes of IHD mortality with time since 1968. The trends in IHD mortality were compared with observed changes in nutrition. These nutritional changes, especially those of saturated fat, are generally consistent with the observed rates or time trends of the diseases under consideration. In conclusion, although classification errors occur they are not of a level of magnitude that makes valid conclusions impossible.