Sindy Li: "On DALY Disability Weights"

By Sindy Li, Fellow

To discuss this post, please comment here

Abstract

I explain the concept of disability weights (used in the DALY measure of the “Global Burden of Disease” studies), describe how they are estimated, and discuss a few issues with its measurement, with some implications in the area of mental health. [1]

Longer Abstract

Disability weights measure the severity of a disease or disability state. It converts a year lived in the disease or disability state to a year lost due to early death, and ranges from 0 (perfect health) to 1 (dead). It enters into the calculation of disability-adjusted life years, and hence is used for measuring disease burden in populations and prioritizing health interventions.

The latest version of the Global Burden of Disease study’s disability weights aim to capture only the impact of disease on an individual’s health. They are calculated from data from household surveys using pairwise comparison (of health states with lay descriptions) and person trade-off questions.

I have concerns over the accuracy of lay descriptions in the pairwise comparison survey questions, respondents’ ability to understand the implications of various health states, the reliability of the person trade-off method, and the exclusion of non-health aspects (while measures incorporating it, like QALY, exist).

Regarding mental illness, some studies have questioned the validity of current methods in estimating its disease burden, and argued that they are underestimated. Even though these studies directly concern survey methods used for estimating QALY, similar issues could affect the current methods for estimating DALY.

1. DALY (Skip if you are familiar with it)

When we measure the cost-effectiveness of an intervention that aims to improve human welfare (often but not limited to health), we often use the unit “$ per DALY averted”. What does it mean?

DALY is “disability-adjusted life years”. It is a measure of the disease burden in a population, and is expressed as the number of years lost due to ill-health, disability or early death. It was first introduced in the 1990 Global Burden of Disease (GBD) study (published in 1996). DALY could be about a specific disease or the overall disease burden in the population (i.e. across all causes of ill-health, disability or early death in the population). Even though it measures disease burden in the population, it is calculated at the individual level for the “average” person in the population (or a certain category, e.g. sex and age group, in the population) at a certain point in time. Quoting the WHO webpage that defines DALY, “the sum of these (individual-level) DALYs across the population, or the burden of disease, can be thought of as a measurement of the gap between current health status and an ideal health situation where the entire population lives to an advanced age, free of disease and disability.” To put it simply, “One DALY can be thought of as one lost year of ‘healthy’ life.”

2. Disability Weights

Disability weights: what are they?

One of the advantages of DALY is that, in addition to early death, it takes into account ill-health and disability caused by the disease under consideration. Disability weight is a measure used to quantify the severity of the latter two, and it enters into the calculation of DALY. It is used to convert time lived under the disease to time lost due to early death, and ranges from 0 (perfect health) to 1 (dead).

Where do disability weights come from? They are from survey responses using various methods (which I will describe below), and from the first GBD 1990 study to the most recent GBD 2015 study the methodology has changed quite a bit.

(Here we focus on disability weights used to calculate DALY since DALY is what organizations like GiveWell use to prioritize interventions. There are other disability weights, e.g. those used in calculating quality-adjusted life years, or QALY.)

What elements could/should be incorporated into disability weights?

I start with a conceptual discussion of elements that could be incorporated into considerations of the “goodness” of a certain health state (all were once incorporated in GBD disability weights, though some no longer are):

  1. Private value, health aspect only: the individual’s functional losses due to diseases, excluding non-health aspects.

  2. Private value, non-health aspects: other ways in which the disease affects the individual’s quality of life. An example mentioned by the WHO paper is stigma suffered by the individual due to the disease; it could include any other non-health aspect of the effect of the disease on quality of life, e.g. effect of income loss due to reduced productivity.
    (1 and 2 combined, i.e. the disease’s impact on the individual’s overall quality of life, seems close to the idea of QALY.)

  3. Externalities: impact of the individual’s disease on others, e.g. family members.

  4. Societal judgements: this concerns a social planner who needs allocate limited resources to various causes (e.g. how much of the fixed health care budget to allocate to people with various severities of various diseases). In principle, elements 1-3 (i.e. the disease's impact on the individual and others) necessarily enter this consideration and are the only factors that should (other than time discounting and age weighting, which we exclude from this discussion entirely); another potential factor, which is how we weigh one member of society versus another, should be independent of their health states. In practice, when people were given such (hypothetical) choices, other factors that may affect their decisions may include weighing individuals with different conditions differently due to moral judgement or discrimination etc. Also, when making these judgements people may not think about non-health aspects of private value if the question primes them to think about health outcomes only (e.g. in the GBD 2010 study), and given the way the question is phrased it may also be unlikely that people think about externalities.

Two questions are worth thinking about:

  • Which of the above elements should enter disability weights?

  • Which of the above elements can be captured by survey responses?

The answer to the first question depends on the purpose of disability weights. The WHO uses it to measure disease burdens in populations around the world, and to prioritize health interventions. At least for the second purpose, it seems that including elements 1-3 would be appropriate (as I mentioned, given 1-3, element 4 seems redundant, except for moral judgement or discrimination based on health states which is not what we want to take into account for this purpose anyway). However, as we will see, the GBD’s philosophy on what should be included has changed quite a bit over time and seems to have converged on focusing on the health impact on the individual.

(On the first purpose, it is unclear to me why one would ever care about the consequences of diseases on the health aspect only, excluding other aspects of well being, even when looking at disease burdens: any relevant intervention would affect all aspects, and even health professionals ultilatemly care about the overall well being of the population. The only argument I can see is if the health aspect can be measured much better. However, health professionals may have different opinions than economists. See point 5 in “Some questions on the latest method” below.)

What can we actually capture (measure)?

On the second question, there are a few methods commonly used in GBD and related studies to assess people’s valuation of different health states, and here I discuss which of the above elements each method captures.

  1. Time trade-off technique (TTO): respondents are asked to imagine someone like themselves in full health, and consider how many years of life they are willing to sacrifice in order to avoid some health condition for the rest of their life.

    • This may capture at least the health aspect and possibly also the non-health aspect of quality of life, although depending on the phrasing respondents may be primed to focus on the health aspect.

  2. Person trade-off technique (PTO): respondents are asked to think in the shoes of a decision maker who has to choose between policies that involve trade-off between severity of illness, the size of the health gain and the number of people helped.

    • This explicitly captures societal judgements. (In practice, respondents may not think of externalities, and may be primed to focus on the health aspect by the question.)

  3. Visual analogue scale (VAS): respondents are asked to consider the consequences of living with the disease for one year, and rate the severity on a scale from 0 (the worst imaginable health state) to 100 (the best imaginable health state).

    • As respondents are specifically primed to think in terms of healthiness, it probably captures only the health aspect of private value. (Although one could imagine it being adapted to be about overall quality of life.)

  4. Pairwise comparisons: respondents are presented with two hypothetical individuals in different health states (described briefly in lay language), and asked which person they regarded as healthier.

    • As respondents are specifically primed to think in terms of healthiness, it probably captures only the health aspect of private value. (Although one could imagine it being adapted to be about overall quality of life.)

Comparing different methods: a study on participants (health professionals and general public with an academic background) in five European countries (England, France, the Netherlands, Spain, and Sweden) finds “a reasonably high level of agreement on disability weights in Western European countries with VAS and TTO methods, which focus on individual preferences, but a lower level of agreement with the PTO method, which focuses more on societal values in resource allocation.” (Agreement here is captured by correlation, since the three measures are all quantitative but are from different questions.) This could be due to different considerations involved in individual- versus societal-level trade-offs, or could be because people are less good at thinking about societal-level trade-offs hence responses based on the PTO method are intrinsically less accurate.

Evolution of GBD’s philosophy on what goes into disability weights

What do disability weights developed by the GBD studies actually capture? In sum, the latest study aims to capture only the health aspect of the disease’s impact on the individual. Below is the evolution of GBD’s philosophy on what elements should be incorporated in disability weights (quotes from section 3 of “WHO methods and data sources for global burden of disease estimates 2000-2015”; highlights added by me):

“In the earliest version of the GBD 1990 study, the burden of disease was defined as loss of welfare/subjective well-being/quality of life (World Bank, 1993).

“Murray (1996) subsequently argued that the health state values should reflect societal judgements of the value of averting different diseases rather than individual judgments of the disutility of the diseases. As a result, the 1996 version of the GBD 1990 used two forms of the person-trade-off (PTO) method to assess social preferences for health states.

“During the period 2000-2008 in which WHO was carrying out updates of the GBD using the original disability weights, with some revisions and additions (Mathers, Lopez & Murray, 2006), the conceptual thinking behind the GBD made explicit the aspiration to quantify loss of health, rather than the social value of the loss of health, or of wellbeing (Murray & Acharya, 2002; Salomon et al, 2003)” … “Thus in the GBD terminology, the term disability is used broadly to refer to departures from optimal health in any of the important domains of health and disability weights should reflect the general population judgments about the ‘healthfulness’ of defined states, not any judgments of quality of life or the worth of persons or the social undesirability or stigma of health states.”

This is also where they currently stand, as the latest study uses pairwise comparisons and explicitly asks about healthfulness. As I argued above, I am not sure including only the health aspect of the disease’ impact on the individual (and excluding non-health effects on the individual’s quality of life or externalities on other people) in disability weights is the sensible thing to do when they are used for prioritizing health interventions, but it is possible that the WHO prefers this for methodological reasons (e.g. this can be measured more accurately, for reasons similar to that mentioned at the end of the previous subsection). However, the method from their latest studies still seem to incorporate societal values, at least on the health aspect (see description below on the GBD 2010 study).

Evolution of GBD’s methodology for estimating disability weights

  • GBD 1990:

    • Philosophy: aims to capture societal preferences.

    • Respondents: health professionals.

    • Method: PTO.

    • Process: invited health professionals from different cultural background to weigh the severity of a set of 22 indicator disabling conditions between 0 (perfect health) and 1 (equivalent to death), through a deliberative process in small groups (8 to 12 participants) for them to reach a consensus. Respondents were asked two versions of the person trade-off question, one about extending life for people in a given health state versus extending life for healthy people, the second about giving health back to people in a given health state versus extending life for healthy people. (The authors argue that “while such trade-offs may affront our perceptions about what is morally acceptable, they are practised implicitly throughout the world's health care systems. The philosophy of the GBD is that the more explicitly these preferences are set out, the more meaningfully they may be debated.”)

    • Criticisms:

      • Method: forcing consistency between the two PTO questions may not make sense (Arnesen and Nord, 1999).

      • Subjects: the use of judgements from health professionals rather than the general population, or those with the conditions. Since the PTO method captures both the disease’s impact on the individual and societal values, we need to ask 1) whether health professionals have an accurate understanding of the  disease’s impact on the individual; 2) whether their values represent society’s in these decisions.

      • Some have criticised the use of universal weights rather than weights that varied with social and cultural environment. This concern may not be relevant for later studies that focus on only the health aspect (that does not include concerns for externalities, stigmas etc.). Also, in the GBD 1990 study, the authors claim that “there is surprisingly wide agreement between cultures on what constitutes a severe or a mild disability”, and that “despite their diverse cultural backgrounds, the participants reached consensus on these weights.”

  • GBD 2010:

    • Philosophy: aims to quantify loss of health from diseases, rather than of wellbeing or social values.

    • Respondents: general population.

    • Method:

      • Household survey from 13,902 people in five countries (Bangladesh, Indonesia, Peru, Tanzania, and the USA).

        • Open-access web-based survey of 16,328 people.

        • Both surveys included pairwise comparison, where respondents were presented with two individuals in different health states described briefly in lay language (in 30 words or less), and asked which person they regarded as healthier. Probit regression analysis was used to translate paired comparison responses into estimates of health on a continuous scale.

        • The web survey included additional PTO questions (which they call “health equivalence questions”, but they explicitly ask respondents “which program would you say produced the greater overall population health benefit?”, hence excluding non-health aspects). Responses from these are needed to anchor the estimates from the analysis of paired comparisons onto a unique 0–1 disability weight scale.

        • Estimated disability weights for 220 health states.

    • Findings:

      • Concluded that they found strong evidence of highly consistent results across the samples from different cultural environments.

      • Found that the new disability weights are much higher for some health states (such as heroin addiction, acute low back pain) and much lower for a larger number of health states, including infertility (0.01, previously 0.18), moderate to profound hearing loss (0.02-0.03, previously 0.12-0.33), blindness (0.20, previously 0.60) and intellectual disability (for severe intellectual disability 0.126, previously 0.82).

    • Criticisms:

      • Philosophy: claims to be quantify loss of health from diseases, rather than of wellbeing or social values; but still asked PTO questions to some respondents and in fact these are necessary to obtain the disability weights. PTO is necessary for their method, but it seems incorrect to claim that societal values (at least on the health aspect) are not incorporated here.

      • Method: some argue that inadequate descriptions of the consequences of health conditions have led to surprisingly low disability weights. For instance, explicitly framing the discrete choice comparisons of sequelae in terms of “who is healthier” may lead to underestimate of the disability weight for blindness. There could be other ways in which the lay descriptions of these health states were inadequate.

  • GBD 2013:

    • Philosophy: aims to quantify loss of health from diseases, rather than of wellbeing or social values.

    • Respondents: general population.

    • Method: web-based surveys in 2013 of 30,660 respondents in four European countries (Hungary, Italy, the Netherlands, and Sweden) on 183 health states, of which 30 were revised descriptions and 18 were for new health states.

    • Findings:

      • The revised descriptions resulted in increased disability weights for intellectual disability and hearing loss, as well as changes in various other weights.

    • Note on drug use disorders (quoting from “WHO methods and data sources for global burden of disease estimates 2000-2015”):

      • “The GBD 2010 disability weights for drug dependence (cannabis, amphetamines, cocaine, heroin) were substantially higher than those used in previous versions of the GBD 2010.”

      • “Salomon et al (2012a) attributed the increase in disability weights for drug use disorders to the fact that the drug use lay descriptions attributed functional outcomes to particular causes (such as heroin use), which was deliberately avoided in most other lay descriptions.” … “Explicit reference to use of or addiction to illicit drugs as the cause of the functional outcomes described could have biased disability weights upwards by introducing wellbeing considerations beyond ‘loss or health’, or reflecting moral or social disapproval.”

      • “Since the YLDs are intended to quantify functional losses in a comparable way across sequelae, and to exclude non-health aspects, we previously used a set of disability weights for the drug dependence sequelae derived from lay descriptions in which the drug name is masked by being described as ‘use of medication’.”

      • “The GBD 2013 and 2015 studies addressed this issue by introducing a set of mild health states for drug and alcohol dependence and estimating severity distributions for alcohol and drug dependence.”

      • In sum, the WHO thinks that the disability weights should quantify functional losses in a comparable way across sequelae, and exclude non-health aspects. They used to achieve this by using a set of disability weights for the drug dependence sequelae derived from lay descriptions with masked drug names. Now they deal with it by eliciting responses on mild and severe health states for drug dependence. (This is relevant for mental illness discussed at the end.)

Some questions on the latest method

  1. Do the lay description accurately represent consequences of the health conditions, even after revisions? This is important as the responses and resulting disability weights have seem to been sensitive to the descriptions.

    • We may need to further improve the descriptions and examine the sensitivity of responses to descriptions.

  2. Relatedly, do the general public (especially those who have never been in the health conditions described) have accurate understandings of the consequences?

    • It may help to compare responses from people who have always been healthy to those from people who had experienced both states of full health and certain disabilities to understand the magnitude of the bias.

    • Overall, the direction of the bias is ambiguous: inaccurate descriptions may lead respondents to over- or under-estimate the badness of a disease, while failure of affective forecasting typically lead healthy respondents to overestimate its badness, at least for certain conditions.

  3. The method still needs PTO (in order to map answers onto a 0-to-1 scale). Given previous findings on the disagreement between results from PTO and from methods that only involve individual-level trade-off, how much should we trust the PTO results used here?

    • We need to understand whether the disagreement is due to different considerations entering individual- versus societal-level trade-offs, or intrinsic inaccuracy of PTO results. On the former, people may somehow value the lives of the disabled less than the lives of those with full health, leading to differences in thinking about the hypothetical scenarios where 1) one is trading off one’s own disability with years of living (in TTO method), 2) one is trading off the lives of others with disability and lives of others with full health (in PTO method). It is possible that this is less of an issue with pairwise comparison, as opposed to TTO, since for pairwise comparison one is comparing two other hypothetical individuals, not oneself (but there may still be a difference between “who is healthier” and “how would I allocate healthcare resources across the two type of people”).

  4. Finally, does it really make sense to only care about the health aspect?

    • My intuition is no, although

      • Maybe the WHO prefers to measure only the health aspect because it can be measured better. This does not seem plausible, as another approach to measuring the disease burden, QALY, is explicitly about quality of life incorporating non-health aspects. TTO is one method for estimating QALY, among others (VAS, standard gamble, EQ-5D etc.). They all seem to involve individual-level comparisons without societal judgements, which seems desirable due to the reason mentioned in 3. If non-health aspects could be important for well being, which they seem to for many conditions, maybe we should simply prefer using QALY over DALY (both for measuring disease burden and for prioritizing health interventions).

      • Maybe health professionals have reasons for focusing on the health aspect.

      • (I am not familiar with comparison of disability weights from DALY and QALY methods, issues with QALY methodology, or why the WHO prefers to use DALY rather than QALY in GBD.)

    • Externalities are also important to take into account, but they seem much harder to measure in surveys.


 

3. Implications for Mental Health

(I am discussing this because mental health is an area that we might end up donating to.)

Some reasons that survey responses may yield under-/unreliable estimates of the disability weights of mental illness

Some authors doubt the suitability of the TTO method (used in eliciting QALYs) for the severity of mental illness. They argue that “not only is it hard imagine mental illnesses without experiencing them”, “but people fail to account for the fact they will adapt to some conditions but not others” (failure of affective forecasting typically lead healthy respondents to overestimate the badness of certain diseases, because they do not take into adaptation; however, adaptation may be less applicable to mental illness). They advocate for the use of subjective well-being measures (i.e. life satisfaction and daily affect). Here is an example of the possible scale of underestimate under the TTO method: the TTO method finds that people would give up the same number of years in their remaining life to be cured of “some difficulty walking” and “moderate anxiety or depression”, i.e. they value the two health states the same way; in contrast, subjective well-being scores imply that “‘moderate anxiety or depression’ is associated with 10 times a greater loss to life satisfaction, and 18 times a greater loss to daily affect, than ‘some difficulty walking’ is”. Even though the disability weights used in DALY are estimated from responses to pairwise comparison and PTO questions, they could be underestimated for mental illness for the same reasons.

The Wikipedia article on DALY mentions some concerns over whether the disease burden surveys fully capture the impacts of mental illness, due to factors including ceiling effects. The studies cited find problems with survey methods including EQ-5D, SF-36 and 15D. EQ-5D is used to estimate QALY; I am not sure if any is involved in measuring DALY, but surveys for obtaining disability weights used in DALY might suffer from the same issues.

Finally, a direct comment on the disability weights used in DALY: depending on the wording of the “lay descriptions” in the pairwise comparisons used to measure disability weights, respondents may or may not include considerations of non-health aspects (e.g. stigma, social and economic exclusion) when answering questions on mental illness. These non-health aspects may be important ways in which how mental illness negatively affects an individual’s life, hence focusing on the health aspect and abstracting away from non-health aspects may significantly underestimate its severity. (This is only speculation; I do not have an idea of the importance of non-health aspects of mental illness, or how much respondents incorporate them.)

Other reasons why the disease burden of mental illness could be underestimated (not directly due to disability weights, but mentioned here for completeness)

A recent study argues that the global burden of mental illness is underestimated by more than a third under currently used approaches, due to five main causes: overlap between psychiatric and neurological disorders; the grouping of suicide and self-harm as a separate category; conflation of all chronic pain syndromes with musculoskeletal disorders; exclusion of personality disorders from disease burden calculations; and inadequate consideration of the contribution of severe mental illness to mortality from associated causes. Using published data and adjusting for these factors, they estimate that the global burden of mental illness accounts for 32.4% of years lived with disability (YLDs) and 13·0% of disability-adjusted life-years (DALYs), instead of the earlier estimates suggesting 21·2% of YLDs and 7·1% of DALYs.

Conclusion

(Same as the longer abstract)

Disability weights measure the severity of a disease or disability state. It converts a year lived in the disease or disability state to a year lost due to early death, and ranges from 0 (perfect health) to 1 (dead). It enters into the calculation of disability-adjusted life years, and hence is used for measuring disease burden in populations and prioritizing health interventions.

The latest version of the Global Burden of Disease study’s disability weights aim to capture only the impact of disease on an individual’s health. They are calculated from data from household surveys using pairwise comparison (of health states with lay descriptions) and person trade-off questions.

I have concerns over the accuracy of lay descriptions in the pairwise comparison survey questions, respondents’ ability to understand the implications of various health states, the reliability of the person trade-off method, and the exclusion of non-health aspects (while measures incorporating it, like QALY, exist).

Regarding mental illness, some studies have questioned the validity of current methods in estimating its disease burden, and argued that they are underestimated. Even though these studies directly concern survey methods used for estimating QALY, similar issues could affect the current methods for estimating DALY.

 

Notes


[1] Here is a shorter and fun read on the same topic covering a subset of my points: https://www.jefftk.com/p/disability-weights.
[2] Murray CJL, Lopez AD, editors (1996). The global burden of disease: a comprehensive assessment of mortality and disability from diseases, injuries and risk factors in 1990 and projected to 2020. Cambridge: Harvard University Press. [3] http://www.who.int/healthinfo/global_burden_disease/metrics_daly/en/
[4] It helps to look at the precise definition of DALY in Appendix A, but in short, the higher the disability weight of a certain disease (the closer it is to 1), the worse it is (i.e. one year living under the disease being closer to one year lost due to early death), and the higher the resulting YLD / DALY / disease burden.
[5] For summary see section 3 of “WHO methods and data sources for global burden of disease estimates 2000-2015” which can be found on http://www.who.int/healthinfo/global_burden_disease/data_sources_methods/en/
[6] According to the WHO paper cited above, “the GBD results for the year 2001 provided a framework for cost-effectiveness and priority setting analyses carried out for the Disease Control Priorities Project.”
[7] See Appendix B for details of how the VAS, PTO and TTO methods were administered in Schwarzinger et al. (2003). My description here of the pairwise comparison method is taken from Salomon et al. (2012).
[8] Schwarzinger M, Marlies EA Stouthard MEA, Burström K, Nord E (2003). Cross-national agreement on disability weights: the European Disability Weights Project. Population Health Metrics. 1:9.
[9] Murray CJL, Lopez AD, editors (1996). The global burden of disease: a comprehensive assessment of mortality and disability from diseases, injuries and risk factors in 1990 and projected to 2020. Cambridge: Harvard University Press.
[10] Their main result is from one meeting in Geneva, but robustness is checked by correlating disability weights obtained from this meeting with the pooled results from nine additional exercises with other participants using the same protocol.
[11] Salomon JA, Vos T, Hogan DR, Gagnon M, Naghavi M, Mokdad A et al (2012a). Common values in assessing health outcomes from disease and injury: disability weights measurement study for the Global Burden of Disease Study 2010. Lancet.380:2129–2143.
[12] Salomon JA, Vos T, Hogan DR, Gagnon M, Naghavi M, Mokdad A et al (2012b). Common values in assessing health outcomes from disease and injury [Supplementary appendix]. Lancet.
[13] Nord E (2013). Disability weights in the Global Burden of Disease 2010: unclear meaning and overstatement of international agreement. Health Policy. 111 (1): 99–104.
Taylor HR, Jonas JB, Keeffe J, Leasher J, Naidoo Kovin, Pesudovs K et al (2013). Disability weights for vision disorders in Global Burden of Disease Study. Lancet.381:23–24.
[14] Salomon JA, Haagsma JA, Davis A, Maertens de Noordhout C, Polinder S, Havelaar AH, Cassini A, Devleesschauwer B, Kretzschmar M, Speybroeck N, Murray CJL, Vos T (2015). Disability weights for the Global Burden of Disease 2013 study. Lancet.
[15] Halpern, Jodi; Robert M. Arnold (2008). "Affective Forecasting: An Unrecognized Challenge in Making Serious Health Decisions". Journal of General Internal Medicine. 23 (10): 1708–1712.
Loewenstein, G; PA Ubel; N Schwarz; D Smith (2004). "Misimagining the unimaginable: the disability paradox and health care decision making". Health Psychology. 24 (4): 57–62.
[16] Sarah J. Whitehead, Shehzad Ali; Health outcomes in economic evaluation: the QALY and utilities. Br Med Bull 2010; 96 (1): 5-21.
[17] Fujiwara, D., & Dolan, P. (2014). Valuing mental health. Policy, 4(2.1). Also see https://www.academia.edu/25088361/What_Should_A_Utilitarian_Billionaire_Do_To_Maximise_Happiness
[18] https://en.wikipedia.org/wiki/Disability-adjusted_life_year#Criticism
[19] Vigo, D., Thornicroft, G., & Atun, R. (2016). Estimating the true global burden of mental illness. The Lancet Psychiatry, 3(2), 171-178.
[20] http://www.who.int/healthinfo/global_burden_disease/metrics_daly/en/
[21] https://en.wikipedia.org/wiki/Prevalence
[22] See section 2 of “WHO methods and data sources for global burden of disease estimates 2000-2015”.
[23] https://en.wikipedia.org/wiki/Disability-adjusted_life_year

Appendix A: Detailed definition of DALY

To introduce the definition we focus on one specific disease and the associated DALY. Quoting the WHO webpage that defines DALY, “DALYs for a disease or health condition are calculated as the sum of the Years of Life Lost (YLL) due to premature mortality in the population and the Years Lost due to Disability (YLD) for people living with the health condition or its consequences.

DALY = YYL + YLD

The YLL basically correspond to the number of deaths multiplied by the standard life expectancy at the age at which death occurs. The basic formula for YLL” … “is the following for a given cause, age and sex:

YYL = N*L

where:

  • N = number of deaths

  • L = standard life expectancy at age of death in years”

And “YLD = P*DW

where:

  • P = number of prevalent cases

  • DW = disability weight”

Prevalence (of a disease in a particular population) is the proportion of the population affected by a disease, and captures both the incidence (a measure of new cases arising in a population over a given period) and the average duration of the disease.

Disability weight is “a weight factor that reflects the severity of the disease on a scale from 0 (perfect health) to 1 (dead)”.

Technically, all the variables above are indexed by (c,a,s,t) where c is cause, a is age, s is sex and t is year, except for:

  • L(s,a): a standard loss function specifying years of life lost for a death at age a for sex s;

  • DW(c,s,a): disability weight for cause c, age a and sex s.

For discussion of age weighting and time discounting (which the latest Global Burden of Disease studies have ceased using), and prevalence vs. incidence, see the Wikipedia article on DALY or section 2 of “WHO methods and data sources for global burden of disease estimates 2000-2015”.

Appendix B: Details of how the VAS, PTO and TTO methods were administered in Schwarzinger et al. (2003)

“In the self-administered VAS participants were asked to consider the consequences of living with the disease stage for one year. The disease stages were first ranked by decreasing severity, and then scored on a vertical thermometer graded from 0 (the worst imaginable health state) to 100 (the best imaginable health state) considering the consequences of living with the disease stage for one year. The best and the worst disease stages were scored first.

“In the PTO, panel participants played the role of decision-makers in their country prioritising between two preventive programmes. Several assumptions about the programmes were made explicit in the panel sessions:

- Prevention means the reduction of occurrence in two to four years; programmes are of the same costs and otherwise equal (e.g. age, sex, socio-economic status of groups);

- Both programmes include people of various ages;

- Loss of production for society and burden on family or caretakers were to be disregarded in decisions.

“The PTO session began with the following example: ‘Programme A prevents the occurrence of a rapidly fatal dis- ease in 100 people in your country in 2 to 4 years' time. The identity of these people is unknown. With the programme they will live in normal health for a normal life- time. Programme B prevents the occurrence of severe vision disorder in a number of people in your country in 2 to 4 years' time. The identity of these people is unknown. With the programme they will avoid the state and live in normal health for a normal life time.’

“In TTO, panel participants had to imagine someone like themselves in full health, and choose between living their remaining 10 years of life in the chronic disease stage or less time in full health.”

To discuss this post, please comment here