There is considered to be a ‘global epidemic’ of diabetes,1 and it is predicted that the prevalence of diabetes will increase in New Zealand, especially in Maori and Pacific Islanders.2 Considerable importance is attached to the reliability of mortality statistics, collated by the New Zealand Health Information Service (NZHIS) [http://www.nzhis.govt.nz], derived predominantly from death certification, although with support from other sources.3

The reliability of death certification in NZ, however, has recently been brought into question following our recent publication of a retrospective review of 600 death certificates (during 1999) from the Mortality Review Database in Christchurch Hospitals.4 We found that 104 cases (17%) had previously documented diabetes, of which only 47 (45%) had diabetes recorded on either the death certificate or the coroner’s report.4 Furthermore, diabetes was recognised in the NZ Health Information Service coding in only 45 (43%) of cases despite the fact that hospital clinical coding was able to identify 94 (90%) cases.4 A previous NZ study showed that diabetes was missed from 36% of death certificates when present.5

In addition, we found that there were 159 (32%) of those people not known to have had previous diabetes that showed a highest random plasma glucose ≥11.1 mmol/L (range up to 34 mmol/L, median value 13.7) in their clinical records. Of these 159, 33 had two or more glucose elevations ≥11.1 mmol/L documented in their clinical records.

Documented diabetes is thus under-reported on more than 50% of death certificates and not compensated by NZHIS coding. Deficiencies in diabetes death certification could be addressed by better training of doctors and medical students. More efficient linkage of NZHIS with hospital coders might also improve the recognition of diabetes in NZ. Notwithstanding the above, the real concern is that of the cohort we examined, only 7.8% would be recorded as having had diabetes when the actual figure is 43.8%.

Reliance on death certification will result in significantly distorted mortality statistics and under-recognition of diabetes as a health problem in NZ.