ArticlesRegional patterns of disability-free life expectancy and disability-adjusted life expectancy: Global Burden of Disease Study
Introduction
In this second instalment of a four-part series on the findings of the Global Burden of Disease Study (GBD)1 we report the regional rates and patterns of disability by age, sex, and region, with various health expectancy measures for 107 diseases (see Lancet 1997; 349: 1269–76 for part 1; parts 3 and 4 follow in the next two issues). Disability-free life expectancy (DFLE) and disability-adjusted life expectancy (DALE)-a health-adjusted expectancy based on the GBD's disability severity weights-are used to describe regional differences in health expectancy. We also discuss the relevance of the cross-sectional pattern of DALE by region to the debate on the compression of morbidity hypothesis.
Health expectancies refer to life expectancy in various health states,2 and can be divided into indicators such as DFLE, in which the expected length of life lived without a given impairment or disability is calculated, or into health-adjusted life expectancy, which can be estimated by calculation of life expectancy for different health states with adjustment for severity weights. Both types of health expectancies may be useful ways to summarise the health status of the population. International comparisons of DFLE and other health expectancies have, however, been severely hampered by differences in calculation and definition.3 Some investigators have examined trends in health expectancies in only one country to try to reduce such discrepancies.4, 5, 6 Even interpretation of trends in DFLE has been confounded by changes in definition and method. When measurements are based on self-reported disability, trends in health expectancies may be affected by changes in the perception of illness, the willingness to take on the sick role, and the cost to the individual of missing work or school.7, 8, 9
Trends in life lived with disability that have accompanied the rise in life expectancy during this century have been subject to extensive debate.10, 11, 12 There are three types of theories about the changes in disability that go with longer life expectancy. Fries and colleagues13, 14 argue that with improvements in survival, the prevalence of disability will decrease and, therefore, the proportion of life lived with disability will also decrease. This theory is often called compression of morbidity. Conversely, other theories predict that the proportion of life lived with disability will increase as mortality declines. Gruenberg15 and Kramer16 suggest that as the length of survival of individuals with chronic disorders such as Down's syndrome increases, the prevalence of these disorders will also rise. Others17, 18, 19, 20 suggest that improved survival among frail individuals who have higher expected incidence rates of disability will lead to an increased prevalence of disability. A third, “mixed” theory predicts that the progression of chronic diseases to severe disability will be slowed by medical intervention, which will lead to a decline in the prevalence of severe disability, but a rise in the prevalence of mild disability;21 increasing life expectancy would also contribute to the latter. Available cross-sectional estimates of health expectancies and longitudinal analyses were not very useful in the investigation of these theories. For example, recent evidence from France suggests that a compression of morbidity is occurring, but similar studies in Australia have more ambiguous results.4 Several data sources suggest that the prevalence of disability in the USA is rising.10
Section snippets
Methods
We emphasised in the GBD examination of internal consistency of epidemiological estimates, and, therefore, that incidence, prevalence, case-fatality, and death rates for each disease or sequela were all compatible with each other. As discussed in more detail elsewhere,22 the efforts to ensure internal consistency included reviews of all available published and unpublished surveys or studies for each sequela and repeated estimation of rates specific for age and sex that were based on available
Results
Table 2 shows summarised prevalences of each of the seven classes of disability by age, sex, and region. For nearly every class of disability and every region, prevalence rises with age. The exception is female class III disability, for which prevalence reaches a peak in the 15-44 years age-group in the six developing regions. This pattern is largely due to a concentration of infertility caused by sexually transmitted diseases and maternal disorders. The rise in prevalence with age is much less
Discussion
The estimates of DFLE and DALE may be affected by two potential sources of bias. First, the approximation method used to estimate the prevalence of disability from the residual categories of diseases may be inaccurate. If the set of diseases and injuries that have been estimated are representative of the relation between disability and mortality for all conditions, then the estimates may not be biased. One factor that may have compromised representativeness is that idiopathic disabilities (for
References (29)
Cost utility analysis: what should be measured
Soc Sci Med
(1994)Methods for quality adjustment in life years
Soc Sci Med
(1992)- et al.
Variability among methods to assess patients' well-being and consequent effect on a cost-effective analysis
J Clin Epidemiol
(1992) - et al.
An empirical evaluation of two assessment methods for utility measurement for life years
Socio-Econ Plann Sci
(1987) - et al.
Global health statistics: a compendium of incidence, prevalence and mortality estimates for over 200 conditions
(1996) - et al.
Health expectancy indicators: a review of the work of REVES to date
- et al.
World atlas of health expectancy calculations
- et al.
Changes in life expectancy and disability-free life expectancy in the United States
Popul Dev Rev
(1989) - et al.
Estimates of change in chronic disability and institutional incidence and prevalence rates in the US elderly population from the 1982, 1984 and 1989 National Long Term Care Survey
J Gerontol
(1993)