Mortality in young people admitted to hospital for diabetes: database study

Oxford record linkage study

Table A

Table B
 
 

Posted as supplied by authors
 

Oxford record linkage study

The Oxford record linkage study was a project of the former Oxford regional health authority which covered the counties of Oxfordshire, Berkshire, Northamptonshire and Buckinghamshire. The database comprised the routine collection of hospital admission statistics, collated at regional level, from 1963-99. The hospital admission statistics were those collected as Hospital Activity Analysis (now termed Hospital Episode Statistics) for all NHS hospitals in the areas covered. The areas were part of Oxfordshire and its constituent hospitals from 1963, Oxfordshire and West Berkshire and their hospitals from 1966, six health districts of the former Oxford regional health authority from 1975, and eight health districts from 1987. The data are anonymised and archived. (Reference: Goldacre MJ, Simmons H, Henderson J, Gill LE. Trends in episode based and person based rates of admission to hospital in the Oxford record linkage study area. BMJ 1988; 296: 583-5.)
 

Multiple admissions

Of the 4992 hospital admissions, 3942 people (79.0%) were admitted only once, 784 (15.7%) twice, 192 (3.8%) three times, and 74 (1.5%) on four or more occasions.

Calculations of standardised mortality ratios

Standardised mortality ratios were calculated by applying the age and sex specific mortality in the resident population covered by the database to the number of people admitted with diabetes in the corresponding age-sex strata. They were standardised for calendar year as well as for age group and sex; the age groups used were <1, 1-4, 5-9, 10-12, 13-14, 15-19, 20-24 and 25-29 years.

Assignment of cause of death

Deaths from natural causes and all causes were those coded as the underlying cause of death on death certificates; deaths from accidents and suicides were those recorded as coroners’ verdicts (“accidental” or “misadventure” verdicts for accidents, and “suicide” or “open verdict” for suicides). The standardised mortality ratios for accidents were for 1979-98 only (although suicides were coded consistently in our database from 1968, accidental death was not coded consistently until 1979).

Sex specific standardised mortality ratios

These SMRs are sex-specific: they represent mortality in males with diabetes relative to mortality in males in the general population of the same age, and mortality in females with diabetes relative to mortality in females in the general population of the same age. SMRs for males and females, relative to the combined male and female general population, are 8.5 (95% CI = 5.7 to 11.8) for ‘all deaths’ in males, and 8.5 (5.6 to 11.9) for ‘all deaths’ in females, 7.6 (1.4 to 18.6) for suicides in males, and 16.1 (5.8 to 31.6) for suicides in females. The larger sex differentials in SMRs when calculated for male and female mortality, separately, arise because mortality rates are substantially higher in the young male general population than in the young female general population. This is particularly so for suicide. The suicide rate in the general population aged 0-29 years, covered by the database, increased from 4.5 per 100<thin>000 population (95% CI = 3.8 to 5.2) in 1968-78 to 9.3 (8.4 to 10.2) per 100<thin>000 in 1989-96 for males, and from 2.1 (1.7 to 2.7) in 1968-78 to 2.6 (2.1 to 3.1) in 1989-96 for women.
 

Posted as supplied by authors
 
 

Table A Trends in annual person-based hospital admission rates with a primary diagnosis of diabetes mellitus among young males and females, aged 0-12, 13-19 and 20-29 years, in the Oxford record linkage study population, 1968-96
 

*Admission rates were standardised by the direct method, using the standard European population as the standard population.

These results show that there has been little appreciable trend over time in admissions for diabetes among young people in the Oxford region.
 
 

Posted as supplied by authors
 
 

Table B Standardised mortality ratios, and trends in standardised mortality ratios, at 3 years after hospital admission for diabetes mellitus in young people aged 0-29 years, 1968-96; both including and excluding elective hospital admissions
 

These results show that the exclusion of elective hospital admissions for diabetes mellitus made no substantial difference to the study findings.