Disability-Adjusted Life Year

A measure of disease burden combining years of life lost to premature death and years lived with disability, expressed as healthy years lost.

Source: Murray 1994

A disability-adjusted life year combines two components. Years of life lost measure premature death, taken as the difference between the age at death and a standard life expectancy at that age, drawn from a reference table applied to every population so the loss does not depend on where a person lived. Years lived with disability measure time in impaired health, calculated as the duration of a condition multiplied by a severity weight, on a scale from zero for full health to one for a state equivalent to death. The two are summed, so one disability-adjusted life year is one year of healthy life lost.

Source: Murray 1994

It is the standard unit of the global burden of disease estimates, which quantify how much health is lost to each condition in each country and how that has changed over time. Those estimates inform priority setting by identifying where burden is concentrated, support advocacy for neglected conditions, and provide the denominator for measures of how much is spent per unit of burden. Disability-adjusted life years are also used as the outcome measure in cost-effectiveness analyses conducted for low and middle income settings, where cost per disability-adjusted life year averted is the counterpart of cost per quality-adjusted life year gained.

Source: Murray 1994

A disability-adjusted life year counts health lost while a quality-adjusted life year counts health gained, so the measures run in opposite directions and their weighting scales are inverted. Disability weights are derived from comparative judgements about described states; quality weights are derived from trade-offs involving time or risk. Disability-adjusted life years apply a standard life expectancy uniformly, whereas quality-adjusted life years use survival modelled for the population concerned. The measures were built for different purposes, describing burden against informing individual funding decisions, and averting one disability-adjusted life year is not equivalent to gaining one quality-adjusted life year.

Source: Salomon et al. 2012

Two adjustments applied in the original formulation have been dropped. Age weighting counted years lived at different ages differently, rising through young adulthood and falling in later life, on the argument that people in those years support dependants. Discounting reduced the weight of years lost further in the future. Both were removed in the 2010 revision, on the grounds that they imported value judgements into a measure intended to describe burden rather than prescribe priorities. Their removal raised the recorded burden of conditions causing death in childhood, so estimates published either side of the change are not comparable.

Source: Salomon et al. 2012

That the disability weights rest on judgements by people who have not experienced the states described, and that those living with a condition rate it less severely. That applying a single standard life expectancy means a death at a given age counts identically everywhere, which is defended as equity and criticised for ignoring what was achievable locally. That aggregating very different kinds of loss into one figure conceals whether burden arises from many mild impairments or few severe ones. And that successive methodological revisions make time series difficult to interpret as change in health.

Source: Anand & Hanson 1997