The Apparent Discrepancy Between Social Inequality in Disability-Free and Quality-Adjusted Life Expectancy

To examine how using binary rather than continuous adjustments for health-related quality of life (HRQOL) influences the estimated magnitude of social inequality in lifetime health and to explore the implications for equity weighting in distributional cost-effectiveness analysis.

We used 2018 Health Survey for England and UK Office for National Statistics mortality and population data to compare health distributions in quality-adjusted life expectancy at birth (EuroQol 5-Dimension 5-Level instrument [EQ-5D-5L], 2018 value set), disability-free life expectancy (limiting long-term illness), and healthy life expectancy (self-assessed health), across neighborhood deprivation quintiles. Indirect equity weights were derived using Atkinson inequality aversion parameters, and implied mean HRQOL scores were cross-tabulated by age and deprivation to understand discrepancies.

Continuous adjustment yielded substantially smaller inequality gaps between the most and least deprived quintile groups than binary adjustments: quality-adjusted life expectancy gap 11.25, disability-free life expectancy gap 16.00, healthy life expectancy gap 18.53. Equity weights based on binary adjustments were substantially higher, especially at higher levels of health inequality aversion (4 times higher at Atkinson parameter 10). Social group HRQOL differences reflected both increased prevalence of morbidity and increased HRQOL burden of morbidity. However, the morbidity prevalence differential between groups was larger than the HRQOL differential.

Disability-free and healthy life expectancy yield larger estimated magnitudes of health inequality than quality-adjusted life expectancy. This is due to their binary scaling, which implicitly assigns a HRQOL value of 0 (as bad as death) to disability or non-good health. A simple adjustment to narrow both ends of the binary scale renders the approaches more comparable.