Background: Current modes of presentation of mortality and hospitalization data, based on codes in Chapter 1 of the International Classification of Diseases (ICD), hide the true importance of infection and other microbial diseases (infections). Our aim was to produce a new aggregation of ICD-9 to estimate their true impact.
Methods: ICD-9 codes for infections, irrespective of ICD-9 chapter, were extracted to form the new aggregation. We reanalysed routine mortality (1979-1983 and 1989-1993) and hospital finished consultant episode data (FCEs) (1989-1993) in the area of the former Northern Regional Health Authority (population approximately 3 million) using the new aggregation of codes. This area is in the North East of England. Age- and sex-specific rates of death and FCE from infection were calculated using mid-1981 and mid-1991 population estimates.
Results: Using ICD-9 Chapter 1 codes, 839 (0.4 per cent) deaths in 1989-1993 were classified as due to infection, compared with 12655 (6.7 per cent) with the new aggregation. In 1979-1983, 10.3 per cent of all deaths were due to infection. The highest rates of infection death were in the oldest age groups and infants. FCEs for infections were 1.2 per cent using ICD-9 Chapter 1 codes, but 5.4 per cent using the new aggregation. The highest rates of infection FCEs were in infants. Most infection deaths and hospitalizations were for respiratory disorders. Using the new aggregation, infection moved from the 13th ranking cause of death to the third ranking cause; and from the 15th ranking cause of hospitalization to the 10th.
Conclusions: The usual method of presenting data coded by ICD-9 chapters greatly understates the contribution of infection to morbidity and mortality and may mislead policy makers assessing the priority to be given to infections.