Estimating disability-adjusted life years

This page is a knowledge crystal of subtype assessment. The page identifier is Op_en3227
Moderator:Jouni (see all)
Upload data {{#opasnet_base_link:Op_en3227}}

Main message:
Question: What is Disability Adjusted Life Years? Answer: The Disability Adjusted Life Years (DALY) has emerged in the international health policy lexicon as a new measure of "burden of Disease". The conceptual framework for DALY is described and justified in a recent paper "Quantifying the burden of disease: the technical basis for disability-adjusted life years"

Disability-adjusted life years (DALYs) are a method for combining different health impacts such as mortality and morbidity into a single common metric. The DALY is one of the most commonly used integrated health measures and was first introduced by Murray and Lopez (1996) in collaboration with World Health Organization and the Worldbank in an attempt to introduce morbidity in mortality-based health discussions. In effect, the DALY integrates many dimensions of public health impact e.g. the number of persons affected by a particular agent or event, the severity and duration of any health effects.

This document describes briefly how Disability Adjusted Life Years (DALYs) can be used in a Health Impact Assessment (HIA). After a short introduction about integrated health measures, the calculation steps are described. Together with Excel sheets made for this purpose (“DALY worksheets Intarese”), this information can be used as a basis for the calculation of environmental DALYs.

The proponents of DALY's use the metric for atleast two separate exercise

1. Positive exercise of measuring the burden of disease.

2. The normative exercise for resource allocation.

The burden of disease is simply measured as the sum of DALYs attributable to premature mortality and morbidity. for resource allocation, Murrey suggested that DALYs be used "in conjunction with the literature on cost effectiveness of health intervention" so as "using estimates of the burden of disease in determining health resource allocations". In using DALY for this purpose, the object is to minimize and aggregate DALYs subject to a given budget.

Murrey states "{T}he he intended use of an indicator of the burden of disease is critical to its design. at least 4 objectives are very important

• to aid in sitting health service (both curative and preventive) priorities.

• to aid in sitting health research priorities.

• to aid in identifying disadvantaged group and targeting of health intervention

• to to provide a comparable measure of output for intervention, program and sector evaluation and planning.

Not everyone appreciate the ethical dimension of heath status indicator. Nevertheless, the first two objectives listed for measuring burden of disease could influence the allocation of resource among individuals, clearly establishing an ethical dimension to the construction of an indicator of the burden of disease.

Integrated health measures

Health effects of environmental factors can vary considerably with regard to their severity, duration and magnitude. This makes it difficult to compare different (environmental) health effects. An integrated health measure, using the same denominator for all health effects, can help with interpretation and comparison of health problems and policies. They quantify and summarize (environment-related) health effects and can be used for:

• Comparative evaluation of environmental health impacts (“how bad is it?”)
• Evaluation of the effectiveness of environmental policies (largest reduction of disease burden)
• Estimation of the accumulation of exposures to environmental factors (for example in urban areas)
• Communication of health risks

An example of an integrated health measure is the DALY (Disability Adjusted Life Years). DALYs combine information on quality and quantity of life. They give an indication of the (potential) number of healthy life years lost due to premature mortality or morbidity. In these calculations, morbidity is weighted for the severity of the disorder.

Calculation

DALY in case of recovery:

Peffect  * (Prec* YLDrec * Sev)

DALY in case of death

Peffect * [Pdie *(YLDdie * Sev) + L.E. - C.A. - YLDdie]

DALY if 'diseased' to natural life expectancy (chronic disease)

Peffect * (1 - Prec - Pdie) * (L.E. - C.A.) * Sev

Where:

• Peffect = Probability of the considered effect (dose-response relationship)
• Prec = Probability to recover from a disease
• Pdie = Probability to die from a disease
• YLDrec = Years Lived with Disability in case of full recovery
• YLDdie = Years Lived with Disability in case of death
• Sev = Severity of the effect (0-1)
• L.E. = Life expectancy
• C.A. = Current Age

Calculating DALYs

DALYs can be calculated using the Excel sheets provided for the Intarese project (pilot version available only).

The formula of a DALY is as follows:

• Number of people with environment-related morbidity or mortality x
• Severity factor (0 = healthy, 1 = death) x
• Duration (YLL for mortality)

A more detailed formula is given below :

DALY = AB * D * S

AB = AR * P * F

AR = (RR’-1)/RR’

RR’ = ((RR-1) * C) + 1

where:

• AB: Attributable Burden; the number of people in a certain health state as a result of exposure to the (environmental) factor that is being analyzed, not corrected for comorbidity.
• D: Duration of the health state; for morbidity, prevalence numbers have been used and therefore duration is one year (except for hospital visits, for which the mean duration of the specific hospital visit has been used). For mortality, the duration of time lost due to premature mortality is calculated using standard expected years of life lost with model life tables.
• S: Severity; the reduction in capacity due to morbidity is measured using severity weights. A weight factor, varying from 0 (healthy) to 1 (death), is determined by experts (clinicians, researchers, etc).
• AR: Attributive Risk; risk of getting a specific disease as a result of exposure to a certain (environmental) factor.
• P: Base prevalence for morbidity; number of deaths for mortality
• F: Fraction of the population exposed to the (environmental) factor under investigation (for air pollution, this fraction is set to 1, meaning that everybody is exposed to a certain degree)
• RR’: Adjusted Relative Risk
• RR: Relative Risk
• C: Concentration of the environmental factor, expressed in the unit of the Relative Risk

Discounting

DALY calculations can also include discounting factors. In discounting, future years of healthy life lived are valued less than present years (discounting normally 3%), or years lived by people in a certain age group (productive ages) are valued more than years lived by the very old and young. Ethical questions can be raised with regard to the use of these factors, and they are currently not included in the calculation sheets. They might become optional in newer versions of DALY calculation tools. More information and templates can be found at WHO health info.

Data

Number of people

The number of people with environment-related morbidity or mortality can be calculated using baseline incidence or prevalence of a disease, population exposure, and a proper exposure response function. It is important the definition and units of the environmental factor and the definition of the related health outcome match exactly with the definitions used in the exposure response function.

The input data can be found using the help of SP2 and WP1.3.

Severity factors

Severity weights (or disability weights) give an indication of the reduction in capacity due to the specific disease. A weight factor, varying from 0 (healthy) to 1 (death), is determined by experts (clinicians, researchers, etc). An overview of severity weights that have been collected in various studies can be found in an Australian report (appendix 1).

If severity weights for the selected health outcomes are not available in this overview, or not judged suitable, they can be derived using expert judgments. A helpful tool is the EuroQol (5D+) model. This is a model which evaluates health states based on six health dimensions: mobility, self-care, daily activities, pain or discomfort, anxiety or depression and cognitive functions.

If deriving new severity weights using expert panels is too time-consuming, it is sometimes possible to use existing severity weights for similar conditions, using expert judgment.

Duration

The duration of a health effect describes the number of healthy life years lost.

For morbidity, this is the time someone has the specified disease condition. This duration can be set to one year if prevalence data are used (assuming that prevalence approximately equals incidence multiplied by duration, and thereby assuming a steady-state equation where the rates are not changing). If incidence data are used, an estimation of the duration of a certain health state should be based on literature research, hospital registries or expert judgments.

For mortality caused by those environmental factors that are completely responsible for death (such as traffic is completely responsible for traffic accident mortality), the mean life expectancy minus mean age of death can be used as the number of years of life lost. The YLL are thus very dependent on the age group of the people that are affected and the remaining life expectancy they have. If age-specific information is available, this should be used to derive the value for duration. For national estimates, values based on national statistics should be used. For international calculations, or calculations that compare various countries, standard (European) values should be used. These can be derived from national statistics offices or Eurostat. If gender-specific health effect estimates are available, gender-specific duration estimates should be used. Life table analysis (not included in this file) can help to identify YLL. Some templates (excel sheets) are provided at WHO health info.

For environmental conditions that only accelerate death in people that are already diseased, only a percentage of the actual Year of Life Lost (YLL) can be attributed to the environmental factor. This estimate of the duration should then be based on literature research or expert judgments. Also here, it is important to take into account which age group/ gender is affected.

Measuring the Burden of Disease:Implication of the DALY framework

What is Burden

The DALY approach measure the burden of disease through reduction in human function (Murray, 1994, p.438). The multiple dimensions of human function are mapped onto a undimensional scale between 0 (perfect health) and 1 (death) along which six discrete disability classes are distinguished. Human function is represented by ability to perform certain activities of daily living, such as learning, working, feeding and clothing oneself. The space in which ill-health is assessed is limitation in these actitivites rather than, for example, that of pain or suffering which would be the relevant categories in a utility based framework. Another space for assessment might be reduction well-being, a notion that is broader than utility and is captured by general capability to function - including physical functioning. Of course, there will be utility or well-being consequences associated with reduction in human function, but these are not the basis for the DALY metric.

An often-cited advantage of DALYs and similar composite indicators such as QALYs, is that they allow fatal and non-fatal health outcomes to be combined into a single indicator. A necessary condition for a finite scale which has perfect health (or quality of life) at one end and death at the other is that the values of all health states, including death, be bounded. In the DALY scale death differs from disability merely by reducing human function to nought. While having an indicator that combines states of imperfect health with death is clearly convenient, there is an obvious information loss in reducing death to simply another health state. Some will argue that the two events are incommensurable, and that a lexical priority attaches to life over death. At any rate, this suggests that information about mortality and morbidity should be presented seperately - even if the trade-offs were conceded between the two events.

DALYs attempt to measure the burden of disease in a somewhat narrow sense. As discussed, they represent the qyantity of ill-health experienced by individuals through functional limitation and premature death. The burden that is measured does not reflect individuals' differential ability to cope with their functional limitation. Moreover, burdens which fall on family, friends and society at large (e.g. the economic cost of illness) are not included. Only in the use of unequal age-weights does there appear to be an attempt to capture the indirect health burden of illness. We return to the rationale for an ethical implications of unequal age-weighting

DALYs use standardized maximum life expectancies (80 years for men and 82.5 years for women) which are considerably higher than the levels of life expectancy currently achieved in developing countries. Using these standardized life expectancies either in measuring global burden of disease or in cost-effectiveness analysis implicitly assumes that health interventions alone are capable of achieving an increase in life expectancy to those higher levels. It is clear that many non-health circumstances will also need to change for life expectancy to rise to the level used in the DALY calculations. These interventions would have to address the socio-economic determinants of health. They would include raising incomes, increasing female education, improving water supply and sanitation conditions, improving workplace safety, and reducing accidents and violence. Hence the burden that is measured by DALYs is the burden of disease and underdevelopment and not that disease alone.

Standard expectation of life and gender gap

To calculate the DALYs from morbidity and premature mortality, a standard expectation of life at birth of 82.5 years is chosen for women and of 80 years for men. This gap is considerably smaller than the observed gender gap in life expentancy in low mortality populations, for example, Japan which has a gender gap of some 6 years. However, the gender gap of 2.5 years is argued to correspond purely to the biological difference in survival potential between males and females (Murray, 1994, p.434), factoring out the effects on life expectancy of males' greater exposure to social and other risk factors. It is nonetheless, an arbitary choice.

The assumed gender gap in life expectancy may have important implications for the estimation of the disease burden of women relative to that of men. World Bank (1993, p. 28) estimates that [F]emales have about a 10 percent lower disease burden per 1,000 per population than males for the world as a whole. The smaller the gender group, ceteris paribus, the smaller will be the female contribution to the burden of disease relative to the male contribution. If the true biological gap happens to be greater than 2.5 years, then the calculation in Murray et al. (1994) and World Bank (1993) will understate the burden of disease of females relative to that of males.

While DALYs take account of higher female life expectancy in calculating years lost to premature mortality, the valuation of those years can be sharply reduced by age-weightning and discounting. As an illustration Table 1 shows the estimate of time lost, and of its value, from the death of a female and a male infant, respectively. The female advantage in years lost of 3% is reduced by age-weighting to 1.5% and is further reduced by discounting to 0.3% for the calculation of DALYs.

Age-weighting and the value of time lived at different ages

Age-weighting assigns a different value to time lived at different ages. Thus, in the construction of a DALY, a year lived at age 2 counts for only 20% of a year lived at age 25 where the age-weighting function is at maximum, while that lived at age 70 counts 46% of the maximum. In a human capital framework, age-weighting might be justified in terms of the differential productivity of an individual at different stages of his life cycle. This approach allows one to impute a money value to life and to disability according to the respective (discounted) income streams foregone. Although it provides a consistent justification for age-weighting (and for discounting), valuing people's lives in terms of a money metric, through their instrumental worth in production, is hard to defend ethically. Murray (1994) himself explicitly rejects the human capital approach, arguing that it inadequately reflects human welfare (p.435). What, then, is the basis for assigning different relative values to years of life lost at different ages?

Murray (1994) views unequal age weights as an attempt to capture different social roles at different ages, arguing that [H]igher weights for a year of time at a particular age does not mean that the time lived at that age is per se more important to the individual, but that because of social roles the social value of that time may be greater (p.435). He claims that social roles vary with age because the young, and often the elderly, depend on the rest of the society for physical, emotional and financial support (p.434). How different roles and changing leves of dependency with age (p.434) are supposed to affect the burden of illness to the individual is far from clear to us. We take it that unequal age weights do no constitute a differential intrinsic valuation of years lived at different ages. Rather, there appears to be an instrumental justification for valuing the time of people in middle age-groups more highly than that of the young or elderly. Presumably ill-health among the middle age-groups also has an indirect effect on the health of the young and elderly because the latter depend on the former for care.

However, if age-weighting is supposed to reflect an instrumental valuation of people's time, even if solely in terms of its health impact, then a host of other instrumentalities with health impacts will need to be incorporated. From the viewpoint of preventing ill-health the social value of time will clearly differ for different occupation groups in the population. For instance, doctors' and nurses' time could be argued to be more valuable than that of other professions. More indirectly, the time of people who have a greater capacity to contribute, through taxation, to the size of the health budget should be valued more highly. However, a person's occupation or tax bracket are not part of the information set used to calculate DALYs and nor are other (for example, social and economic) factors which directly and indirectly influence individual's health.

Murray (1994) apparently believes that [U]nequal age weights [also] has broad intuitive appeal, and goes on to state that informal polling of tuberculosis programme managers by the author in an annual training course has revealed that everyone polled believes that the time lived in the middle age groups should be weighted as more important than the extremes (p.435). But what precise question were his group of programme managers asked? Did it concern an intrinsic valuation of time lived at different ages or an instrumental valuation? How do we know that it is not reflecting their view of income levels and productivity through the life cycle? Were they made aware of the implications of age-weighting for resource allocation? It is not obvious to us that the author has properly solicited from his programme managers their value judgements concerning age-weighting per se.

It is also not at all clear that the programme managers were provided with (adequate) information about other adjustments made in the DALY formula to life years lived. It is possible, for instance, that, they had in mind different functional capacities at different ages, in other words, a higher level of functioning in the middle age-groups comapred with either end. But reduction in human function will be captured separately and independently through Murray's disability weights. Even in function and age were correlated (and followed the shape of the age-weighting function), applying age weights on top of disability weights would amount to penalizing reduced functional capacity twice over.

Disability weights

In the DALY framework, the effects of illness are captured through six disability classes which assign increasing weights associated with the extent of loss of physical functioning. Murray (1994, p.439) states that weights for the six classes have chosen by a group of independent experts. As in the case of age weights, the meaning attached to the different weighting of health states depends in an important way on the precise question that was asked of these experts. Their responses would also depend on understanding the use to which such estimates would be put.

A more appropriate measure of burden must take account of the way in which individual and social resources can compensate for the level of disability experienced. The individual's actual loss of functioning will depend on both his uncompensated disability state and the factors which affect his capability to cope with that disability (given his circumstances). Compensated disability weights would depend inter alia on the individual's income (for example, whether he can employ somebody to prepare his meals and provide other assistance with his activities of daily living), and on the provision of local service facilitate his daily activities. The latter might include designated parking, transport services for the disabled, etc. Compensated disability weights would come closer to reflecting the true burden of disability as experienced by the individual. The DALY approach does not distinguish between the quantity of ill-health and the burden associated with it.

A final question about the construction of disability weights relates to the manipulations necessary to restrict the maximum disability weight for an individual to 1. In particular, although DALYs are aggregated accross individuals, problems caused by co-morbidity (an individual experiencing multiple illnesses) are not adequately dealt with in the framework, and can lead to an over estimation of the total disease burden.

As they stand Murray's disability categories do not distinguish functionings associated with illness and those associated with age (but no illness). For example, the most severe disability class (Class 6) involves disability states in which an individual needs assistance with activities of daily living such as eating, personal hygiene or toilet use (Murray, 1994, Table 2, p.438). Infants are not capable of feeding themselves: does this imply they are disabled? Do they by virtue of the functional limitations of their stage of development contribute to the burden of disease? If disability weights are to be useful and consistent, they should be defined so as to avoid confounding age with disability.

Time preference and the discounting of future life

In the DALY formula, future year of life lived are valued less than present years. With the recommended 3% discount rate this implies that one life saved today will be worth more than five lives in 55 years Discounting future lives in this way would justify many forms of environmental degradation today which benefits the present generation at the expense of future generations.For example, the benefit today of economic activities which emits green house gases at present rates could well outweigh the harm to future generation if future lives are valued at only one-fifth of present lives.

We can see no justification for an estimation of the time lost to illness or death which depends on when the illness or the calculation occurs. Suppose a person experiences an illness today and another person, identical in respects experiences an illness of exactly the same description next year. Discounting amounts to concluding that the quantity of the (same) illness is lower in the latter case. This does not accord with intuition or even with common use of language.

As in the case of age-weighting, a logicall consistent defence of discounting could be provided if the human capital approach to valuing life were adopted. Life would then be reducible to a monetary value, and discounting it justified because of the oppurtunity cost of money. But Murray (1994) eschews this framework yet invokes economic cost-benefit arguments to defend "social time preference" (p.440).

Because life cannot be reduced to money, the usual arguments for discounting money do not apply to discounting DALYs. Yet Murray (1994) fails to distinguish between discounting DALYs (or utility) and discounting money (or consumption). Hence the usual cost-benefit reasons presented by him for discounting future consumption (money) (e.g. by appeal to the marginal utility of consumption failing with expected future growth of consumption) are irrelevant to discounting utility or DALYs. Moreover, it is difficult to see how pure time preference in the discounting future utility or future DALYs can be justified.

The only defensible argument for treating future periods differently rests on the possibility that the world might end. A construction which could accomodate uncertainty is the minimization of expected (in the statistical sense) undiscounted DALYs. Under this objective function lives in each period were weighted by the probability that the world will exist in that period. Note that a 3% discount rate implies a 50% chance that the world will end in 69.7 years. How many people would be willing to take an odds-on chance that the world will end within their, or their children's, lifetime? We reckon that the discount rate implied by the probability of the world ending within the planning horizon for DALY calculations is infinitesimally small. For any practical purpose, the assumption of a zero discount rate is likely to command more assent than even a very small one.

It appears to us that Murray's positive arguments in support of discounting are based largely on attempts to avoid some awkward implications of the use of undiscounted DALYs for cost-effective analysis. The first of these states that "if health benefits are not discounted, then we may conclude that 100% resources should be invested in any disease eradication plans with finite costs as this will eliminate infinite streams of [undiscounted] DALYs which will outweight all other health investments that do not result in eradication" (p.440). Quite apart from whether it is necessary to invest 100% of resources to eradicate diseases, we fail to see how this statement provides an argument for discounting DALYs. In the burden of disease framework it would seem a desirable outcome to eradicate a disease, for precisely the author's goal of minimizing aggregate ill-health.

Murray also invokes the so called "time paradox", arguing that if health benefits are not discounted then "one will always choose to put off investing in a health project until the future" because "...the budget could be invested and yield a positive return" (p.440). Whether or not any "time paradox" arises, if Murray's concern is that, without discounting, present health outcomes will be sacrificed in favour of future health outcomes (leading to an undesirable inequality between generations) then this concern for equity should be incorporated directly in a temporally natural way. One way of making the criterion sensitive to inequality is to express it as a strictly convex, additive function of undiscounted DALYS

Finally, discounting at 3% in the Murray-World Development Report 1993 framework implies that we should save the life of a 20-year-old person rather than an infant: more age-weighted and discounted DALYs are prevented in the former case. But does this accord with general intuition? It is the non-monotonic feature of Figs. 4 and 5 in Murray (1994, p.436 and p.441, respectively) and Box Fig. 1.3 in World Bank (1993, p.26) which jars with our basic intuitions. Discounting, which in itself is totally indefensible in the context of lives and life years, can be shown to compound the problems inherent in age-weighting. Together they comprise the most unappealing features of the DALY formula.

Sensitivity

Much is made in Murray et al. (1994) of the extensive sensitivity analysis undertaken on the global burden calculations to the various assumptions concerning unequal age weights, discount rate, and disability-class weights. Two points are relevant here. First, even though changing these parameters may result in small changes to the overall estimates, this does not constitute evidence that the approach is correct. Insensitivity to parameter changes can hardly validate a formula! This paper has raised various concerns about the ethical underpinnings of the DALY approach. These concerns are little affected by any lack of sensitivity of the overall calculations to particular assumptions.

Secondly, although the overall burden of disease calculation may not be very sensitive to changing crucial parameter values, this in no way indicates that the calculation for individual diseases is not highly sensitive to the underlying assumptions. Evidence from a recent study of the burden of trachomatous visual impairment (Evans and Ranson, 1995) suggests that at the level of specific diseases the calculations are indeed highly sensitive to several of the assumptions in the DALY framework, including the discount rate. It is possible that individual diseases are sensitive in compensating directions, resulting in relative lack of overall sensitivity.

Whose values?

There appear to be at least four distinct agents whose values are incorporated in the DALY-minimization exercise. First, there is a social planner who specifies the exercise (minimizing the burden of ill-health) and who determines the DALY function used to measure it. Secondly, there are a number of other agents whose values are incorporated into the DALY through the parameters of this function: for age-weighting, TB (tuberculosis) programme managers (perhaps qua individuals); for disability weights , a group of independent experts; for the discount rate, the authors of the World Bank Disease Control Priorities Study (Jamison et al., 1993). It is entirely arbitary to appeal to different agents' values for the different parameters without prior justification or reasoning. Furthermore, it has to be asked why the social planner's objective is to minimize DALYs if individuals themselves have diffferent objectives. And if compelling reasons can be provided for the social planner to override individual preferences and minimize DALYs, why should the social planner rely on individual values for choosing DALY parameters?

Even if it could be argued that individuals values should be incorporated in the choice of parameters, a precondition for doing so is that everyone should agree on the form of the DALY function - i.e. share a common definition of ill-health. Otherwise, the responses to questions asked in determining parameter values (for example, disability weights or age-weights) will depend on the individual's own conception of ill-health and on his understanding of the purpose for which the estimate is intended. When these differ among individuals the responses provided cannot be compared, let alone averaged.