Background: Information on non-fatal health outcomes of disease and injury has been largely neglected in health planning because of the conceptual and definitional complexity of measuring morbidity and disability in populations. One of our major objectives was to quantify disability for inclusion in health policy debates. We analysed these health outcomes in terms of disability-free life expectancy (DFLE) and disability-adjusted life expectancy (DALE).
Methods: Published and unpublished data were systematically reviewed to estimate the incidence, prevalence, and duration of 483 disabling sequelae of 107 diseases and injuries. To ensure internal consistency of these estimates, a software programme (DISMOD) was applied many times until consistent parameters were identified. The severity of disability, on a scale of 0 (perfect health) to 1 (death), was measured in a deliberate manner by the person-trade-off method. Spearman's and Pearson's correlation coefficients were used to measure disability weights among groups. Prevalence of seven classes of disability was back-calculated from the distribution of each disabling sequela across disabilities. Prevalence for each class of disability for different age-sex groups was used to calculate seven forms of DFLE and DALE based on Sullivan's method.
Findings: Prevalence of most disability classes is highest in sub-Saharan Africa and lowest in established market economies. Low-severity disabilities (class I and class II) are the most common. The expectation at birth of class I disability ranges from 6.5 years in established market economies to 14.7 years in sub-Saharan Africa, and for class II disabilities, from 8.5-18.4 years. DFLE varies significantly among regions: DFLE for class I disabilities at birth ranges from 9.9 years in sub-Saharan Africa to 47.7 years in established market economies for females and DFLE for class V disabilities ranges from 43.4 years for men in sub-Saharan Africa to 74.8 years for women in established market economies. The proportion of expected life span at birth lived with disability adjusted for severity, varies from about 8% in established market economies to 15% in sub-Saharan Africa, with little difference between men and women. In high-income regions, nearly 90% of expected disability is due to non-communicable diseases and most of the remainder to injuries. In poorer regions, almost half of expected disability is due to communicable diseases and injuries.
Interpretation: The higher proportion of lifespan spent disabled in high-mortality populations is consistent with the compression of morbidity hypothesis. The threshold definition of disability used substantially affects the results of DFLE, DALE, which incorporates severity weights for disabilities, is a useful summary measure of the burden of disability and mortality.
PIP: Published and unpublished data were reviewed to estimate the incidence, prevalence, and duration of 483 disabling sequelae of 107 diseases and injuries in an attempt to quantify disability for inclusion in health policy debates. The DisMod computer program was applied many times until consistent parameters were identified. The severity of disability was measured by the person-trade-off method, disability weights were measured across groups, and the prevalence of seven classes of disability was back-calculated from the distribution of each disabling sequelae across disabilities. The prevalence for each class of disability for different age-sex groups was used to calculate seven forms of disability-free life expectancy (DFLE) and disability-adjusted life expectancy (DALE). The prevalence of most disability classes is highest in sub-Saharan Africa and lowest in established market economies, with low-severity disabilities being the most common. DFLE varies significantly among regions. In high-income regions, almost 90% of expected disability is due to non-communicable diseases and most of the rest to injuries. However, in poorer regions, almost half of expected disability is due to communicable diseases and injuries. The higher proportion of life span spent disabled in high-mortality populations is consistent with the compression of morbidity hypothesis.