Asthma morbidity in England and Wales appears to have increased in recent decades, despite advances in therapy, and this is widely attributed to increasing asthma prevalence. This increase has not, however, been fully reflected by mortality trends, and in children and young adults, there have been no clear changes. In adults aged Š45 yrs (in whom >85% of current asthma deaths are recorded), mean annual mortality doubled between the mid1970s and the early 1990s in both sexes [1,2]. The rate of change and the degree of annual fluctuation increased with age, and the apparent increase was most evident in those aged Š65 yrs.The cause of the marked increase in recorded asthma deaths in this age group is uncertain. If correct, it could reflect improved diagnostic recognition, an increase in asthma prevalence, an increase in disease severity or an adverse effect of medication. The 1979 and 1984 changes in international coding practice (International Classification of Diseases (ICD) 9, implementation of rule 3) artificially increased the mortality rate in those aged <45 yrs and >75 yrs, respectively, but this provides insufficient explanation for the observed trends [2][3][4].Alternatively, the increase may be largely artefactual. The accuracy of death certification is known to decline with advancing age, and there is concern that recent trends, especially in the elderly, may be attributable to diagnostic transfer [5,6]. An audit investigation in the period 1980-1989 in one district of the Northern Health Region of England suggested that the majority of certified "asthma deaths" had occurred in elderly smokers who had neither died from asthma nor suffered from it [7]. Other studies have suggested that asthma death certification for the whole population may overestimate the number of true asthma deaths by 13-47%, with the degree of inaccuracy in certification rising in the elderly to 39-80% [8][9][10].The recorded asthma mortality rates of 3.94 and 3.64 per 100,000 for the Northern Health Region (population 3.07 million) for the years 1991 and 1992, respectively, closely reflect the national figures for England and Wales of 3.67 and 3.48 (total population 51 million). Therefore, experience within the northern region as a whole is likely to be representative of the national picture. The aim of this study was, consequently, to estimate the magnitude of any inaccuracy in death certification for asthma within this region, and hence, to assess whether the apparent increase in asthma mortality in the elderly could have occurred, at least partly, as a result of diagnostic transfer. Methods SubjectsPermission was obtained from 13 of the 16 local Health Authorities for a review of death certificates for the years , 1991. D.W.E.C. Reid, V.J. Hendrick, T.C. Aitken, W.T. Berrill, S.C. Stenton, D.J. Hendrick. ©ERS Journals Ltd 1998. ABSTRACT: Asthma mortality appeared to increase two-fold in the UK from the mid-1970s to the early 1990s, but there is evidence of inaccuracy in asthma death certification and so a region-wide investigat...