Does routine hospital discharge data accurately reflect true myocardial infarction cases according to WHO criteria in patients under 65?
Routine hospital ICD discharge codes for acute myocardial infarction have limited positive predictive value (67%) compared to standardized WHO criteria, suggesting they should not be used alone to monitor coronary heart disease morbidity trends.
Data from a 1983 Auckland coronary heart disease register applying current World Health Organization criteria have been used to validate routine hospital discharge data. The register contained 905 patients under 65 years admitted to hospital and 858 of these patients were matched with hospital discharge records. Of the registered definite myocardial infarction cases 86% received the International Classification of Diseases code 410 (acute myocardial infarction); 9% of these cases received a code 411-414 (other forms of coronary heart disease or angina) and 5% received other codes. Only 405 of the 604 cases (67%) coded 410 in the hospital discharge data were true definite myocardial infarctions according to the World Health Organization criteria. The routine hospital International Classification of Diseases data do not provide diagnostic groups sufficiently close to World Health Organization categories for them to be used alone to monitor trends in coronary heart disease morbidity rates.
Beaglehole et al. (Sun,) studied this question.