Estimates of DALE for 191 countries: methods and results.

Colin D Mathers Ritu Sadana Joshua A Salomon Christopher JL Murray Alan D Lopez

Global Programme on Evidence for Health Policy Working Paper No. 16

World Health Organization June 2000

1. Introduction The annual assessment of world health is a key component of the global public policy process to improve health levels and reduce health inequalities. Current estimates of death and disability in countries disaggregated by age, sex and cause, are useful for several public health purposes, ranging from the monitoring of new epidemics to progress in reducing old ones for which disease control programmes are in place in countries. To adequately describe health patterns in almost 200 countries according to age, sex and cause, a vast array of estimates need to be generated. It then becomes extremely difficult to ascertain the main findings of such a review unless the are summarised in some fashion. In this year´s World Health Report [1], the primary summary measure of population health used is Disability-Adjusted Life Expectancy, or DALE. DALE measures the equivalent number of years of life expected to be lived in full health, or healthy life expectancy. This technical paper provides details of the methods and data sources used to prepare the DALE estimates for the 191 member countries of WHO. In constructing the estimates, we sought to address some of the methodological challenges regarding comparability of the health status data collected [2]. The Global Burden of Disease project developed two summary measures, the Disability- Adjusted Life Year (DALY) and Disability-Adjusted Life Expectancy (DALE), to provide a comprehensive assessment of the global burden of disease and injury [3, 4], to inform global priority setting for health research [5], and to report on trends in population health across the world [1, 6]. Both these summary measures of population health (SMPH) combine information on the impact of premature death and of disability and other non-fatal health outcomes. The burden of disease methodology provides a way to link information at the population level on disease causes and occurrence to information on both short-term and long- term health outcomes, including impairments, functional limitations (disability), restrictions in participation in usual roles (handicap), and death. DALYs are a gap measure; they measure the gap between a population's actual health and some defined goal, while DALE belongs to the family of health expectancies, summarizing the expected number of years to be lived in what might be termed the equivalent of "full health". Both DALE and DALYs require a number of social value choices relating among other things, to the valuation of time spent in states of health worse than ideal health, the definition of an implied norm for population health, and the weighting of years of life lived at different ages. Murray and Lopez [7, 34] published disability-adjusted life expectancy (DALE) estimates for the eight regions of the world based on the estimates of severity-weighted disability prevalence developed for the non-fatal component of disease and injury burden. As a summary measure of the burden of disability from all causes in a population, DALE has two advantages over other summary measures. The first is that it is relatively easy to explain the concept of an equivalent “healthy” life expectancy to a non-technical audience. The increasing popularity of health expectancy indicators among policy makers has been documented (van de Water et al. 1996; Barendregt et al. 1998) [8, 9]. The second is that DALE is measured in units (expected years of life) that are meaningful to and within the common experience of non-technical audiences (unlike other indicators such as health gaps, mortality rates or incidence rates).

1 Improving the overall level of population health has been identified by WHO as one of the five intrinsic goals of health systems (see Section 1.3 below), and DALE provides the best available SMPH for measuring the overall level of health for populations in a way that is appropriately sensitive to probabilities of survival and death and to the prevalence and severity of health states among the population.

1.1 Background In the last two decades, considerable international effort has been put into the development of summary measures of population health (SMPH) that integrate information of mortality and non-fatal health outcomes and international policy interest in such indicators is increasing [10]. Efforts to develop summary measures of population health have a long history [11–17]. In the past decade, there has been a markedly increased interest in the development, calculation and use of summary measures. The concept of combining population health state prevalence data with mortality data in a lifetable to generate estimates of expected years of life in various health states (health expectancies) was first proposed in the 1960s [11, 13] and Disability-free Life Expectancy was calculated for a number of countries during the 1980s. An informal international research network, the Network on Health Expectancy (Réseau Espérance de Vie en Santé or REVES) was established in 1989 with objectives including the harmonisation of calculation methods and identification of the conditions necessary for comparison of health expectancy estimates, both across populations and over time [18–22]. During the 1990s, Disability-Free Life Expectancy (DFLE) and related measures were calculated for many countries [23–26]. In 1993, OECD included disability-free life expectancy among the health indicators reported in its health database [27] and by 1999 the number of countries for which some estimates of disability-free life were available had grown to 29 [28]. However, DFLE and related measures incorporate a dichotomous weighting scheme, i.e., that does not account for varying levels of severity (see Section 1.2 below for more detail on this). The threshold definition of disability, therefore, has a dramatic effect on the results [29]. Wilkins and Adams [30] suggested a more sensitive weighting scheme based on the severity of functional limitations, leading to the disability-adjusted life expectancy (DALE) approach. DALE are described in more detail in the following section. Another type of summary measure, Disability-Adjusted Life Years (DALYs) has been used in the Global Burden of Disease Study [7, 31–36] and in a number of National Burden of Disease Studies [37–46]. Reflecting this rising interest in the academic and policy communities, the United States’ Institute of Medicine convened a panel on summary measures and published a report that included recommendations to enhance public discussion of the ethical assumptions and value judgements, establish standards, and invest in education and training to promote use of summary measures. [10]. More recently, WHO convened a conference of experts across a of disciplines including descriptive , public health, health economics and philosophy and ethics to discuss issues around the conceptual, technical and ethical basis for summary measures of population health. A book addressing these issues based on the papers presented at Marrakech is in preparation [47].

Interest in summary measures relates to a range of potential uses. Murray, Salomon and Mathers [48] identified eight of these: 1) Comparing the health of one population to the health of another population.

2 2) Comparing the health of the same population at different points in time. 3) Identifying and quantifying overall health inequalities within populations. 4) Providing appropriate and balanced attention to the effects of non-fatal health outcomes on overall population health. 5) Informing debates on priorities for service delivery and planning. 6) Informing debates on priorities for research and development in the health sector. 7) Improving professional training curricula in public health. 8) Analyzing the benefits of health interventions for use in cost-effectiveness analyses. Broad interest and use of summary measures in the policy arena demonstrates the recognition of their value at the practical level for many of these purposes. The World Health Report 2000 has used DALE as a summary measure of the level of population health in member countries in order to provide a comparative assessment of levels of health (use 1 above), and as a component of the composite health goal performance measure (see section 1.3 below). Over time, successive reporting on DALE will provide evidence of progress towards achieving global goals for improving health (use 2 above).

1.2 Relationship of DALE to other SMPH The Global Burden of Disease Study (GBD) developed a new SMPH, the Disability-Adjusted Life Expectancy (DALY). For a review of the development of DALYs and recent advances in disease burden measurement, see Murray and Lopez [3, 49]. The GBD also used Disability- Adjusted Life Expectancy (DALE) as a simple summary for comparative purposes across populations. This section explains the relationship between these two summary measures, the comparative advantages of each, and why WHO is using both types of indicators in its annual assessment of the health situation of member countries and in the assessment of performance.

Health expectancies and health gaps On the basis of a simple survivorship curve, SMPH can be divided broadly into two families: health expectancies and health gaps. The bold curve in Figure 1 is an example of a survivorship curve S(x) for a hypothetical population. The survivorship curve indicates, for each age x along the x-axis, the proportion of an initial birth cohort that will remain alive at that age. The area under the survivorship function is divided into two components, A which is time lived in full health and B which is time lived at each age in a health state less than full health. The familiar measure of life expectancy at birth is simply equal to A+B (the total area under the survivorship curve. A health expectancy is generally of the form:

Health expectancy = A + f(B) (1) where f(.) is a function that weights time spent in B by the severity of the health states that B represents. When a set of health state valuations are used to weight time spent in health states worse than ideal health, the health expectancy is referred to as a health-adjusted or disability- adjusted life expectancy (DALE). Another type of health expectancy is exemplified by disability-free life expectancy in which time spent in any health state categorized as disabled

3 is assigned arbitrarily a weight of zero, and time spent in any state categorized as not disabled is assigned a weight of one (i.e., equivalent to full health).

4

Figure 1. Survivorship function for a population

Survivors (%) 100 90 80 C 70 B 60 50 A (Full health) 40 30 20 10 0 0 102030405060708090100 Age (years)

In contrast to health expectancies, health gaps quantify the difference between the actual health of a population and some stated norm or goal for population health. The health goal implied by Figure 1 is for everyone in the entire population to live in ideal health until the age indicated by the vertical line enclosing area C at the right1. In the specific example shown, the normative goal has been set as survival in full health until age 100. By selecting a normative goal for population health, the gap between this normative goal and current survival, area C, quantifies premature mortality. A health gap is generally of the form:

Health gap = C + g(B) (2) where g(.) is a function that weights time spent in B by the severity of the health states that B represents. Note that because health gaps measure a negative entity, namely the gap between current conditions and some established norm for the population, the weighting of time spent in B is on a reversed as compared to the weighting of time spent in B for a health expectancy. More precisely, full health is 1 in a health expectancy, whereas death or a state equivalent to death is 1 in a health gap. Because health gaps measure the distance between current health conditions and a population norm for health, they are clearly a normative measure. Years of life lost measures are all measures of a mortality gap, or the area between the survivorship function and some implied target survivorship function (area C in Figure 1). Mortality gap measures were first suggested by Dempsey (1947) [50] and potential years of life lost has been extensively used as a population health indicator since its first calculation by Romeder and McWhinnie [51]. Murray [3] and others [16, 52] have since proposed and calculated a variety of health gaps.

1 Figure 1 graphically illustrates the magnitude of both health expectancies and health gaps only when a population has a stable distribution with a zero population growth rate. In practice, health expectancies are not sensitive to differences in the age structure of different populations. Health gaps are usually reported in absolute terms so that health gaps are sensitive to variations in the age distribution of different populations although age independent forms of health gaps can be formulated.

5 Health state expectancies and disability-adjusted life expectancies We can categorise health expectancies into two main classes: those that use dichotomous health state weights and those that use health state valuations for an exhaustive set of health states. Examples of the first class include: Disability-free life expectancy: This health expectancy gives a weight of 1 to states of health with no disability (above an explicit or implicit threshold) and a weight of 0 to states of health with any level of disability above the threshold. Other examples of this type of health expectancy include active life expectancy, independent life expectancy and dementia-free life expectancy. Life expectancy with disability: This is an example of a health expectancy which gives 0 weight to all states of health apart from one specified state of less than full health (in this case, disability above a certain threshold of severity). If health state 3 in Figure 2 is ‘moderate disability’ then the area under the survival curve corresponding to health state 3 represents life expectancy with moderate disability. Other examples of this type of health expectancy include handicap expectancy, severe handicap expectancy and unhealthy life expectancy. Examples of the second type of indicator include: Health-adjusted life expectancies: These have been calculated for Canada and Australia using population survey data on the prevalence of disability at four levels of severity together with more or less arbitrary severity weights [53–55]. More recently, Canada has produced the first estimates of health-adjusted life expectancy based on population prevalence data for health states together with measured utility weights [56]. Disability-adjusted life expectancy: This was calculated for the Global Burden of Disease Study using disability weights reflecting social preferences for seven severity levels of disability [7]. DALE has also been calculated for Australia using prevalence data from the Australian Burden of Disease Study [46] and preference weights derived from the Global Burden of Disease Study and from a Dutch study using similar valuation methods [57]. Although health states form a continuum, in practice they are generally conceptualised and measured as a set of mutually exclusive and exhaustive discrete states ordered on one or more dimensions. If we enumerate health states using a discrete index h, then we can calculate disability-adjusted life expectancy as:

L   DALE x * wh (u) S h (u) du (3) h x where u represents age and the integral is over ages from x onwards. If the weight wh for state h is independent of age u, then

 L . / DALE  w  S u du  w HE x  h * h () /  h hx (4) h x 0 h where HEhx is the health state expectancy at age x for years lived in state h.

In terms of the four health states illustrated in Figure 2, if HE1,0 to HE4,0 are the health state expectancies at birth for each of the four states, and we give age-independent weights w2, w3, w4 (less than 1) to the three states of less than full health, then the disability-adjusted life expectancy at birth and total life expectancy at birth are given by:        DALE0 HE1,0 w2 HE2,0 w3 HE3,0 w4 HE4,0 (5)

6     LE0 HE1,0 HE2,0 HE3,0 HE4,0 (6)

Figure 2. Survivorship functions for four health states

Survivors (%) 100 90 Health state 2 80 70 Health state 3 60 50 Health state 4 40 30 20 10 0 0 102030405060708090100 Age (years)

Terminology In the mid-1990s, REVES developed a set of recommendations for terminology that was widely adopted [58]. With the development of health gaps measures in the 1990s, there has been some shift in the use of these terms, and health expectancy is now used to denote the general class of summary measures that relate to the area under the survival curve. We use the revised terminology proposed by Mathers [59]: Health expectancy (HE): Generic term for summary measures of population health that estimate the expectation of years of life lived in various health states. Health state expectancy: Generic term for health expectancies which measure the expectation of years lived in a single specified health state (eg. Disability-free). Disability-adjusted life expectancy (DALE): General term for health expectancies which estimate the expectation of equivalent years of good health based on an exhaustive set of health states and weights defined in terms of health state valuations. Health-adjusted life expectancy (HALE) is a synonym for DALE. Healthy life expectancy: Used as a simple synonym for DALE.

Critical appraisal of health expectancies Murray, Salomon and Mathers [48] proposed a set of desirable properties for evaluating summary measures of population health (SMPH) based on common sense notions of population health of the following type:

7 If two populations are identical in every way except that infant mortality is higher in one, then we expect that everybody would agree that the population with the lower infant mortality is healthier. They suggested a minimal set of desirable properties for summary measures that will be used to compare the health of populations: 1. If age-specific mortality decreases in any age-group, everything else being the same, then a summary measure should improve (i.e. a health gap should decrease and a health expectancy should increase)2. 2. If age-specific prevalence of some health state worse than ideal health increases, everything else being the same, a summary measure should get worse. 3. If age-specific incidence of some health state worse than ideal health increases, everything else being the same, a summary measure should get worse. 4. If age-specific remission for some health state worse than ideal health increases, everything else being the same, a summary measure should improve. 5. If the severity of a given health state worsens, everything else being the same, then a summary measure should get worse. Mathers [59] has assessed health expectancies against these five criteria. All health expectancies meet criterion 1. Health expectancies based on prevalence data (for example, those calculated using Sullivan’s method) meet criteria 1 and 2 but fail criteria 3 and 4 (until prevalence rates change to reflect the change in transition rates). Health expectancies based on transition rates (for example, those calculated using the multistate life table method) meet criteria 1, 3 and 4 but fail criterion 2. Disability-adjusted life expectancies (DALE) meet criterion 5, whereas health expectancies using dichotomous health state weights (eg. disability-free life expectancy) do not. Table 1 summarises these conclusions.

Table 1. SMPH criteria met by various forms of health expectancies

Health state expectancies Disability -adjusted life expectancies Dichotomous weights Polytomous weights (eg. DFLE) (eg. DALE) Prevalence-based measures 1, 2 1, 2, 5 Transition-rate based measures 1, 3, 4 1, 3, 4, 5

Health state expectancies such as DFLE give an implicit value of zero (equivalent to the valuation of death) for disability above a certain threshold, below this threshold the valuation is 1. This that the summary indicator is not sensitive to changes in the severity distribution of disability within a population (criterion 5). The overall DFLE value for a population is largely determined by the prevalence of the milder levels of disability and comparability between populations or over time is highly sensitive to the performance of the disability instrument in classifying people around the threshold. For this reason, Murray, Salomon and Mathers [48] concluded that health state expectancies are not appropriate for use

2 This criterion could be weakened to say that if age-specific mortality decreases in any age-group, everything else being the same, then a summary measure should improve or stay the same. The weaker version would allow for deaths beyond some critical age to leave a summary measure unchanged. Measures such as potential years of life lost would then fulfil the weak criterion.

8 as SMPH, and that DALE is the most appropriate form of health expectancy for use as an SMPH. Murray, Salomon and Mathers [48] proposed two other desirable attributes of summary measures that are to be used to inform policy discussions. These are not attributes based on arguments about whether a population is healthier than another but rather on practical considerations: 1. Summary measures should be comprehensible and feasible to calculate for many populations. Comprehensibility and complexity are different. Life expectancy at birth is a complex abstract measure but is easy to understand. Health expectancies are popular because they are also easily understood. 2. Summary measures should be linear aggregates of the summary measures calculated for any arbitrary partitioning of sub-groups. Many decision-makers, and very often the public, desire information that is characterized by this type of additive decomposition. In other words, they would like to be able to answer what fraction of the summary measure is related to health events in the poor, in the uninsured, in the elderly, in children, and so on. Additive decomposition is also often appealing for cause attribution. Most health expectancies satisfy the first attribute. However, they cannot be additively decomposed in respect of causes or population sub-groups. Disability-adjusted life expectancies are additively decomposable into health expectancies for specified levels of disability severity (see above). This form of decomposition may be useful in understanding which levels of disability severity are contributing most to changes in population health. Health state expectancies should thus be understood as a decomposition of a DALE summary measure than as SMPH in themselves. This interpretation is consistent with the usual ways in which families of health state expectancies are presented for a population [60, 61]. In general, health gaps can be decomposed into the contribution of various causes in a more intuitive and easily communicated fashion than health expectancies. DALYs are additive across causes to give the total health gap for a population. Disability-adjusted life expectancy and a health gap measure such as the DALY thus fulfill different needs for SMPH to summarise and report on trends and achievements in population health across countries. Wolfson [62] outlined a vision of a coherent and integrated statistical framework, with summary measures of population health status at the apex of a hierarchy of related measures, rather than a piecemeal set of unconnected measures. The macro measures at the apex of the system, such as DALE and DALYs, provide a broad population-based overview of trends and causes (Figure 3). DALE would be used for monitoring overall progress in improving the level and distribution of health, and DALYs would be used for quantifying the causes of health losses, for identifying the potential for health gain and for linking health interventions to changes in population health. As shown in Figure 3, such a system should include the capability to ‘drill down’ below the summary measures to component parts such as incidence rates, prevalence rates, severity distributions, case fatality rates, etc. It should also allow us to ‘drill down’ below whole of population level to examine inequalities in health and to estimate the impacts of a given intervention on various sub-groups.

9 Figure 3. The pyramid of population health measures

Health states and valuations

Prevalence Duration Incidence, remission Case fatality Mortality

Ethnic, etc. Structural Behaviour, environme Socioeconomic groups Physiological and pathophysiological Geographic Diseases Injuries Age and sex Impairments Disabilities

1.3 Role as a measure of health system goal performance The World Health Report [1] has carried out an assessment of the performance of health systems of member countries in achieving three main (intrinsic) goals for the health system: health, responsiveness and fairness in financing. WHO’s work on operationalizing the measurement of goal attainment is focussed on measuring these three goals as well as relating goal attainment to resource use in order to evaluate performance and [1, 63]. In operationalizing the assessment of level of health for member countries, WHO has chosen to use DALE as the summary measure for the reasons outlined above. In the World Health Report 2000, health inequality is assessed for member countries in terms of child mortality inequality. In future assessments, it is planned to move to the use of more comprehensive indicators based on inequality in DALE within the population [64–66]. A third related concept is efficiency or composite goal performance. Efficiency is how well we achieve the socially desired mix of the five components of the three goals compared to the available resources. DALE is one of the measures used in the development of the composite measure of health system goal attainment and in the analysis of health system performance. Composite goal performance and individual goal performance are discussed in more detail in the World Health Report [1] and in related technical papers.

10 2. Previous approaches to the calculation of DALE A key step in the construction of a health expectancy or a health gap is comparing time lived in a health state worse than full health with time lived in full health (in health expectancies) and with time lost due to premature mortality, compared to some normative goal (in a health gap). Two sets of issues are common to both health expectancies and health gaps: the conceptual framework and measurement strategy to describe health states and the conceptual framework and measurement strategy to value time spent in health states. We first review general issues in measuring and valuing health states in Section 2.1. We then review the two main approaches which have been used for calculating DALE based on whether the health state prevalences are derived from population health/disability surveys (described in Section 2.2) or from a disease-specific approach based in a full burden of disease analysis (described in Section 2.3).

2.1 Measurement and valuation of health states The literature on both description and valuation of health states is vast and rapidly expanding [67–70]. Murray [3] provides a more detailed discussion with regards to the original GBD approach.

Defining and measuring health status Until recently, most health expectancies have been defined in terms of disability (functional limitations), or handicap (role limitations, dependence, restrictions in participation). In its early meetings, the Network on Health Expectancy (REVES) agreed that the WHO International Classification of Impairments, Disabilities, and Handicaps or ICIDH [71] should provide the conceptual framework for the development of health expectancy indicators based on impairment, disability and handicap states. The original ICIDH framework recognised four dimension - diseases or disorders, impairments, disabilities and handicaps. In the context of a health condition (disease or disorder), impairment corresponds to any loss or abnormality of psychological, physiological or anatomical structure or function; disability corresponds to any restriction or lack (resulting from an impairment) of ability to perform an activity in the manner or within the range considered normal for a human being; and handicap or social disadvantage for a given individual results from an impairment or a disability that limits or prevents the fulfilment of a normal role (depending on age, sex and social and cultural factors). Handicap is characterised by discordance between the activity and status of the individual and the expectations of his or her social environment [71]. The beta draft revision of the ICIDH-2 [72] replaces the concept of handicap by the concept of social participation and includes limitations in performing more complex activities (formerly handicaps) as types of activity (the concept replacing disability). Impairments are renamed functional abilities. The ICIDH-2 classifies domains of functioning rather than persons and is a classification scheme rather than a health state measurement instrument. It thus cannot be used directly to classify persons according to health state for constructing summary population measures of health. Robine and Jagger [25] have reviewed the ICIDH and other models of the disablement process and note that there is considerable confusion and disagreement over the boundaries between impairment and disability, and disability and handicap, particularly in relation to where functional limitations and complex activity restrictions fall.

11 A wide range of instruments have been developed in various languages to use individual responses to measure various dimensions or domains of health states. Some of the more widely used instruments are summarised in Table 2. Some instruments sacrifice the capacity to discriminate between health states by restricting the number of questions or items in the survey and restricting the number of response categories in order to increase measured reliability – for example, this is the strategy used in the Euroqol EQ5D, which includes five

Table 2. Health domains included in 12 generic health status measurement instruments.

Health Domains QWB McM SIP QLI NHP FSQ CP Duke SF-36 WHO EQ6D WHO QOL DASII (multi-dimensional profile) 1970 1976 1976 1981 1981 1986 1987 1990 1992 1996 1999 1999 Overall Well-Being 4 General Health 4 4 4 4 Perceived Health 4 Change in Health 4 Physical Health 4 4 4 Activities/roles 4 4 4 4 4 4 Work 4 4 Home 4 Recreation 4 Ambulation 4 Eating 4 Energy/vitality 4 4 Mobility/fitness 4 4 4 4 4 4 4 4 4 Pain/discomfort 4 4 4 4 4 Self Care 4 4 4 4 4 Sleep/Rest 4 4 Social Health 4 4 4 Activities/roles 4 4 4 4 Communication 4 4 4 4 4 Support 4 4 Mental Health 4 4 4 4 Activities/roles 4 Alertness 4 Anxiety/Depression 4 4 Cognition 4 4 Emotional status 4 4 4 4 Outlook 4 Self-esteem 4 Handicap/Participation 4 4 Environmental Context 4 Source: Sadana, 2000. QWB:Quality of Well-Being Scale, McM: McMaster Health Index, SIP: Sickness Impact Profile, QLI: Quality of Life Index, NHP: Nottingham Health Profile, FSQ: Functional Status , CP: COOP Charts for Primary Care Practice, Duke: Duke Health Profile; SF-36:Short-Form 36 Health Survey, EQ6D: EuroQol 6 domain Quality of Life Scale, WHOQOL: WHO Quality of Life Bref Field Trial Version, WHODAS II: WHO Disability Assessment Schedule. Souvce: [73]

12 domains with three level categories on each [74]. Other instruments such as SF-36 have many more items and more response categories per item. Increased discriminatory power often comes at the price of increased complexity, which may have important implications for valuation to time spent in a health state. A fundamental problem with current self-reported instruments is a lack of cross-cultural comparability (including comparisons of the same community over periods long enough that cultural health norms may have changed). This is not simply a question of language and the interpretation of the meaning of different categorical responses in different languages. The endpoints of scales for a given domain such as best or worst mobility may also have very different meanings across different cultures or across socio-economic groups within a society. This is discussed further in Section 2.2 and is examined in detail by Sadana et al [2] who have analysed over 60 representative health surveys. The challenge for developing a profile for standardising health state descriptions is to include all domains considered to be important in terms of societal health goals and in terms of health state valuation and to develop methods for measuring each of these domains for individuals in ways that maximise comparability across population groups. WHO is proposing as part of the Global Burden of Disease 2000 project to develop a standardized description of health states covering a broad range of health states for use in population surveys and health state valuation. The challenge in seeking a standardized description involves trade-offs between completeness of description and parsimony. Other desirable properties include cross-cultural validity, and usability by younger and older adults with widely varying education levels and cultural backgrounds.

Valuing health states In order to use time as a common currency for years of life lived in various states of health and for time lost due to premature mortality, we must numerically value time lived in non- fatal health states. The health state valuations (or disability weights) used in DALY and DALE calculations represent societal preferences for different health states. They range from 0 representing a state of good or ideal health (preferred to all other states) to 1 representing states equivalent to being dead. These weights do not represent the lived experience of any disability or health state, or imply any societal value of the person in a disability or health state. Rather they quantify societal preferences for health states in relation to the societal ‘ideal’ of good health. At the time that the GBD was underway, and even today, there is no body of empirical measurement of health state descriptions and valuations that can be used (a) to describe the average health state in multiple domains associated with different diseases, injuries and risk factors and (b) to value these average health states. As an effort to provide a practical interim solution to these major data deficiencies, the GBD used a multiple methods (ordinal , two forms of person trade-off, time trade-off and visual analogue) approach with small groups of public health professionals to measure values for approximately 20 indicator health states ranging from mild to severe [3]. A deliberative approach was used with small groups in order to ensure that the people involved understood and were aware of the implications of their choices. Final weights for conditions were based on the person trade-off estimates in order to reflect social rather than individual preferences for health states. Other health conditions were valued by ordinal against the indicator conditions. There is a growing consensus among health economists that health state preferences should reflect the preferences of the general population when they are to be used as part of a process of broad health policy assessment, priority setting or resource allocation [75–76]. Health

13 experts were used in the GBD valuation exercise for convenience reasons due to the practical difficulties in ensuring that lay persons fully understood the impact and severity distribution of the conditions being valued. The Disability Weights Project for Diseases in the Netherlands [57] attempted to address this problem by defining the distribution of health states associated with a health condition by using the modified EuroQol health profile to describe the health states. Few differences were seen in the average PTO preferences assigned by a lay panel (people with an academic background but no medical knowledge) compared with those of two panels of physicians. The Dutch study concluded that it makes little difference whether the valuation panel is composed of health care experts or lay people, as long as accurate functional health state descriptions are included in the specifications of the health problems being valued. Since the development of the original protocol for health state value measurement in the GBD, a series of convenience samples of international public health practitioners has been organized, and a number of modifications and refinements of the original protocol have been examined. In eleven different groups, valuations for 15 to 22 states – with a set of 14 states common to all exercises – have been measured using a multi-method approach with internal consistency checks and group discussions. The study locations have included the United States, Mexico, Brazil, the Maghreb countries (Morocco, Algeria and Tunisia), Japan, Australia, the Netherlands, and four multi-national groups of health care practitioners. Murray and Lopez [49] compared the health state valuations for each state across ten of these groups. Overall, the intraclass for the ten studies was 0.954, indicating that this measurement approach yields similar values in groups from very different communities. The work completed by Ustun and colleagues in 14 countries [77], which measures rank correlations for a set of 17 health conditions, provides further evidence that valuations of health states appear to be quite stable across diverse settings. As further large scale empirical studies are undertaken in different countries, it is likely that some important variations in average health state valuations will be found, particularly with respect to the contribution of selected domains such as sexual function or pain (see for example [78]). Nevertheless, it is unlikely that the magnitude of this variation will have major implications for summary measures of population health. More recently, Mahapatra [79] has carried out in India the first large scale population survey which has adapted the GBD methods to obtain health state valuations from the general population. The valuations from this survey indicate similar ranks for the states included in the GBD studies, with somewhat higher disability weights overall. The most likely explanation for the higher weights is the use of a visual analog scale to obtain the valuations, which has produced higher disability weights than other methods in previous empirical studies. Further work is underway on developing and refining instruments for in the general community, as well as on understanding how responses to different valuation questions relate to strength of preference for different health states. One of the main objectives for the ongoing WHO work on a standardized description of health states for use in population surveys is to facilitate reliable and valid measurements of valuations of time spent in health states in populations across the world. If large-scale empirical assessment in many different countries to inform health state valuations for the calculation of DALE for member states are to be achieved, instruments that are reliable and valid for populations with widely varying educational attainment need to be developed.

14 2.2 Estimating DALE from health survey data To date, few health expectancy calculations have been carried out based on health state profiles that address more than one domain of health. Health-adjusted life expectancies (HALE) have been calculated for Canada and Australia using population survey data on the prevalence of disability at four levels of severity together with more or less arbitrary severity weights [53–55]. More recently, Canada has produced the first estimates of health-adjusted life expectancy based on population prevalence data for health states together with measured utility weights [62]. This is the only example of a DALE which is based on a true multi- domain health status instrument together with measured population preference weights (using a standard gamble, non-deliberative approach). There are two main problems with the use of self-report health status survey data to estimate DALE: S the problems of comparability of self-reported health status or disability across populations and across time. S the estimation of disability weights for the corresponding health states. The first of these problems is illustrated by Australian estimates of disability-free and handicap-free life expectancy from 1981 to 1998 based on disability prevalence data from the population surveys of disability (Figure 4). The prevalence of handicap increased substantially between 1981 and 1988, from 9.4 to 13.7 per cent for males and from 8.7 to 12.2 per cent for females. It is highly likely that a substantial part of these increases is due to changes in community awareness and perceptions of handicaps, changes in income support programs, availability of aids, and increasing levels of diagnosis of some health problems [61]. In contrast, the prevalence of severe handicap remained largely unchanged over the period 1981 to 1993, then jumped substantially between 1993 and 1998. The latter change is thought to be at least partly due to changes in , although the actual questions used were largely unchanged [80].

Figure 4: Trends in disability-free life expectancy (DFLE), handicap-free life expectancy (HFLE), severe handicap-free life expectancy (SHFLE) and total life expectancy (LE), by sex, Australia 1981 to 1998

Males Females 80.00 80 .00 LE

LE SHFLE 70.00 70 .00

SHFLE HFLE

60.00 HFLE 60 .00 DFLE Expectancy at birth (years) birth at Expectancy Expectancy at birth (years) birth at Expectancy

DFLE 50.00 50 .00 1980 1985 1990 1995 2000 1980 1985 1990 1995 2000 Year Year

15

There are similar problems in cross-national comparability of self-report health status data. During the last two decades, REVES and international agencies (including WHO, OECD and Eurostat) have put considerable effort into promoting the development and use of standardized health status and disability instruments in order to improve the cross-national comparability of population health data. As a result of these efforts, there are now a number of multi-country surveys that have used strictly comparable instruments and survey methods. Analyses of these surveys have shown that substantial problems with comparability of self- report health data remain [2]. This is illustrated in Figure 5, which shows results from the second wave of the EC Health Panel Survey for 13 European countries [81]. This figure shows the distribution of perceived health status (on a five point scale running from very good to very bad) for people aged less than 60 who reported no chronic conditions and no disability. Among this “healthy” group, there are very substantial variations in the prevalence of both good and poor health states, which are unlikely to reflect real differences in population health. Uncritical use of such self-report data would result in up to two-fold variations in the average prevalence of (severity-weighted) disability across European countries.

Figure 5: Distribution of perceived health status among people aged 0-59 years who have no chronic conditions, are not hampered in daily activities and have not cut down activities due to health problems, EC Health Panel Survey, Wave 2, 1995

Austria

Portugal

Spain

Greece

Italy Very good Ireland Good Fair UK Bad France Very bad

Luxembourg

Belgium

Netherlands

Denmark

Germany

0 0.2 0.4 0.6 0.8 1

16 Sadana et al [2] document in more detail how the problems in the comparability of self-report health data relate not only to differences in survey design and methods, but much more fundamentally to unmeasured differences in expectations and norms for health. Recent analyses of surveys containing both self-report and objective measurements of health status have documented systematic biases in self-report data according to age, sex, socioeconomic disadvantage, and other measures of social disadvantage within populations [2]. There are related problems in estimating disability weights for health states measured in self- report health surveys. Even where these surveys collect information on severity of disability, severity is not generally measured in a form which can be easily translated into disability weights reflecting health state preferences. For the foreseeable future, this means that summary measures of population health for comparative purposes must make use of survey results on self-reported health status instruments with great care, and then only if supported by many other condition-specific epidemiological datasets. In order to decompose summary measures into component causes (diseases, injuries or risk factors), such condition-specific data sets will also be needed. This is discussed further in the following section.

2.3 The disease-specific approach Burden of disease analysis uses a disease-specific approach to estimate the disability and loss of healthy years of life associated with a comprehensive and exhaustive set of health conditions. In particular, DALYs are calculated as the sum of years of life lost due to mortality (YLL) and equivalent years of healthy life “lost” due to disability (YLD). YLD for a particular health condition (disease or injury) are calculated by estimating the number of new cases (incidence) of the condition occuring in the time period of interest. For each new case, the number of years of healthy life lost is obtained by multiplying the average duration of the condition (to remission or death) by a severity weight that quantifies the equivalent loss of healthy years of life due to living with the health condition [3]. Burden of disease analysis involves making YLD estimates for a comprehensive set of at least 100 disease and injury categories involving analysis of many more disease stages, severity levels and sequelae. For some conditions, numbers of incident cases are available directly from disease registers or epidemiological studies but for most conditions, only prevalence data are available. In these cases, a software program called DISMOD© is used to model incidence and duration from estimates of prevalence, remission, case fatality and background mortality [3]. Many different sources of information are used to calculate YLD. An iterative process and extensive consultation with relevant experts is required to ensure consistency of epidemiological estimates. Murray and Lopez [7] presented estimates of DFLE and DALE for each region of the world using Sullivan’s method and the severity-weighted prevalence of disability derived from the YLD estimates in the Global Burden of Disease Study. For these calculations, severity- weighted disability estimates were not discounted or age weighted. Murray and Lopez calculated disability prevalences for seven disability classes with an adjustment to allow for independent co-disability between different disability classes. The expected years of healthy life lost ranged from 8% in the Established Market Economies (life expectancy at birth of around 77 years) to 15% in sub-Saharan Africa (life expectancy at birth of around 50 years). The Australian Burden of Disease Study [46] also estimated DALE for Australia using Sullivan’s method. Rather than estimate prevalence of seven disability classes, this study estimated undiscounted prevalence YLD per 1,000 population as a direct measure of severity-

17 weighted disability prevalence and adjusted for comorbidity between disease and injury causal groups rather than for co-disability. Total DALE at birth were 68.7 years for males and 73.6 years for females in Australia for 1996, similar to the values for the EME estimated in the GBD. It was estimated that approximately 9% of total life expectancy at birth was lost due to disability for both males and females in Australia, again similar to the 8% lost in the EME. The disease-specific approach to the calculation of DALE used in the GBD and the Australian studies has a number of advantages over the health survey approach: S it ensures consistency with the health gap measure (DALYs) of the burden of disease S it ensures inclusion of all causes of disability (including those resulting in forms of disability poorly reported in health surveys eg. substance abuse, intellectual disability) S it avoids problems of self-report biases. However, there are currently two major limitations with this approach: S problems with comorbidities, and S the data demands for calculating YLD for a comprehensive set of conditions. Comorbidity refers to the not uncommon situation where a person has two or more health problems that result in disability (either dependently or independently of each other). It makes little sense to simply add the independently determined disability weights for conditions that are found to coexist as this can lead to the illogical possibility of having a combined weight of more than one (i.e. more disabling than death), particularly in the case of two heavily weighted conditions. Both the GBD and the Australian studies made adjustments for comorbidity assuming that conditions occurred independently (ie. the probability of having 2 conditions was the product of the average probabilities for having each condition) and adjusted the disability weights for comorbid conditions assuming a multiplicative model [7, 46]. Further work is needed to determine whether such simple comorbidity models are adequate for characterising the burden of disease and the distribution of disability by severity in a population. Substantial effort will be required to improve on the estimation of the prevalence of non-independent comorbidity for future iterations of the GBD. Mathers et al [46] have compared disability and handicap prevalence data derived from a national population survey in Australia with weighted disability prevalences derived from the estimates of total YLD in the Australian Burden of Disease project. The total prevalent YLD per 100 population can be thought of as a severity-weighted disability prevalence measured as a percentage of the population of that age. The disability survey data were used to estimate weighted disability prevalence rates (%) by age and sex for the Australian population in 1998. Weights for six disability and handicap severity levels were chosen to line up as closely as possible with appropriate preference weight ranges for the Dutch disability weights defined in terms of EuroQol health state descriptions. Results for males and females combined are shown in Figure 6 and compared with the prevalence YLD from the Australian Burden of Disease study. YLD associated with short-term conditions lasting less than six months (such as colds and flu) have been excluded, since the survey definition of disability included only chronic disability lasting six months or more. YLD associated with anxiety disorders and mild to moderate (but not severe) depression have also been excluded, since the majority of disability associated with these conditions is unlikely to have been captured by the ABS Disability Survey. The YLD-based prevalence estimates correspond quite closely to the survey-based prevalence estimate at younger and middle ages and at ages 75 and over. For ages in the range 55–74

18 years, the YLD-based prevalence is significantly higher than the survey-based prevalence. This may reflect the impact of chronic diseases prevalent at these ages that are not being picked up by the Disability Survey screening questions. The Australian Burden of Disease Study made some attempts to ensure consistency between the overall population prevalence derived from survey data for some particular impairments (such as intellectual impairment, renal failure, amputations) and for the total prevalence of these sequelae added across relevant conditions for which YLD were estimated in the burden of disease study. In the following Section, we describe a combined approach which uses available population health survey data together with the disease-specific approach to develop consistent estimates of the prevalence of disability for the calculation of DALE.

Figure 6: Comparison of severity-weighted prevalence of disability from 1998 ABS Disability Survey with prevalence YLD (per cent), by age, 1996

Severity-weighted prevalence (%) 30

1998 ABS Disability Survey 25 YLD for chronic conditions

20

15

10

5

0 0 20406080 Age (years)

19 3. Methods

3.1 Overview of approach As discussed above, ideally we would use survey data for disability caused by specific conditions to estimate YLD for those conditions, in a way that ensured consistency between epidemiological estimates of disease incidence, prevalence, case fatality, progression to disabling sequelae, and severity distributions of resulting health states. Because of the severe problems found with the comparability of self-reported health status data in population surveys [2], we have developed an analytical approach for this first analysis of DALE for all WHO member countries, which combines the condition-specific approach based in burden of disease analysis with the use of available representative population health survey data on the population distribution of health states. In brief, it involves the following steps: 1. The development for each country of age-sex specific weighted disability prevalence estimates based on burden of disease analyses at country level which build on condition- specific epidemiological information to the maximum extent possible. 2. The construction of latent health factor scores from representative population health surveys. 3. The estimation of weighted disability prevalence from these latent health factor scores using the disability estimates from step 1 as prior estimates. The rescaling of the factor scores to improve comparability of survey data and adjust for self-report biases is based on estimating a parsimonious set of self-report bias parameters which provide best fit between factor scores and prior disability estimates . 4. The use of Sullivan’s method to calculate DALE from posterior disability estimates plus country life tables. Where health survey data is not available, DALE are calculated using the prior disability estimates from the country-level epidemiological analyses. The advantages and limitations of this approach are discussed further in Section 5, which also outlines the approach which will be taken over the next year to improve the estimation of DALE for WHO member countries.

3.2 Life tables and cause of death distributions for countries As a first step towards the estimation of DALE for WHO Member States, it is crucial to develop for each country the best possible assessment for 1999 of overall mortality levels by age and sex and the corresponding life table. New life tables were developed for all 191 WHO Member States starting with a systematic review of all available evidence from surveys, , sample registration systems, population laboratories and vital registration on levels and trends of child and adult mortality [82]. To aid in the cause of death and burden of disease analysis used as inputs to the DALE estimation process, the six WHO regions of the world were divided into 5 mortality strata on the basis of their level of child (5q0) and adult male mortality (45q15). The matrix defined by the six WHO Regions and the 5 mortality strata leads to 14 subregions, since not every mortality stratum is represented in every Region. These subregions are defined in Table 3.

20

Table 3. WHO Regions and mortality subregions used in epidemiological analyses and reporting for the World Health Report 2000.

Population Region Mortality subregion WHO Member States (millions) AFRO D High child Algeria, Angola, Benin, Burkina Faso, Cameroon, Cape Verde, Chad, Comoros, 286 High adult Equatorial Guinea, Gabon, Gambia, Ghana, Guinea, Guinea-Bissau, Liberia, Madagascar, Mali, Mauritania, Mauritius, Niger, Nigeria, Sao Tome and Principe, Senegal, Seychelles, Sierra Leone, Togo AFRO E High child Botswana, Burundi, Central African Republic, Congo, Côte d'Ivoire, Democratic 330 Very high adult Republic of the Congo, Eritrea, Ethiopia, Kenya, Lesotho, Malawi, Mozambique, Namibia, Rwanda, South Africa, Swaziland, Uganda, United Republic of Tanzania, Zambia, Zimbabwe AMRO A Very low child Canada, Cuba, United States of America 318 Low adult AMRO B Low child Antigua and Barbuda, Argentina, Bahamas, Barbados, Belize, Brazil, Chile, 425 Low adult Colombia, Costa Rica, Dominica, Dominican Republic, El Salvador, Grenada, Guyana, Honduras, Jamaica, Mexico, Panama, Paraguay, Saint Kitts and Nevis, Saint Lucia, Saint Vincent and the Grenadines, Suriname, Trinidad and Tobago, Uruguay, Venezuela (Bolivarian Republic of) AMRO D High child Bolivia, Ecuador, Guatemala, Haiti, Nicaragua, Peru 70 High adult EMRO B Low child Bahrain, Cyprus, Iran (Islamic Republic of), Jordan, Kuwait, Lebanon, Libyan 137 Low adult Arab Jamahiriya, Oman, Qatar, Saudi Arabia, Syrian Arab Republic, Tunisia, United Arab Emirates EMRO D High child Afghanistan, Azerbaijan, Djibouti, Egypt, Iraq, Morocco, Pakistan, Somalia, 348 High adult Sudan, Yemen EURO A Very low child Andorra, Austria, Belgium, Croatia, Czech Republic, Denmark, Finland, France, 410 Low adult Germany, Greece, Iceland, Ireland, Israel, Italy, Luxembourg, Malta, Monaco, Netherlands, Norway, Portugal, San Marino, Slovenia, Spain, Sweden, Switzerland, United Kingdom EURO B Low child Albania, Armenia, Bosnia and Herzegovina, Bulgaria, Georgia, Kyrgyzstan, 215 Low adult Poland, Romania, Slovakia, Tajikistan, The Former Yugoslav Republic of Macedonia, Turkey, Turkmenistan, Uzbekistan, Yugoslavia

EURO C Low child Belarus, Estonia, Hungary, Kazakhstan, Latvia, Lithuania, Republic of Moldova, 246 High adult Russian Federation, Ukraine SEARO B Low child Indonesia, Sri Lanka, Thailand 289 Low adult SEARO D High child Bangladesh, Bhutan, Democratic People's Republic of Korea, India, Maldives, 1,219 High adult Myanmar, Nepal WPRO A Very low child Australia, Brunei Darussalam, Japan, New Zealand, Singapore 153 Low adult WPRO B Low child Cambodia, China, Cook Islands, Fiji, Kiribati, Lao People's Democratic 1,521 Low adult Republic, Malaysia, Marshall Islands, Micronesia (Federated States of), Mongolia, Nauru, Niue, Palau, Papua New Guinea, Philippines, Republic of Korea, Samoa, Solomon Islands, Tonga, Tuvalu, Vanuatu, Viet Nam World 5,968

Because of increasing heterogeneity of patterns of adult and child mortality, WHO has developed a system of two-parameter logit life tables for each of the 14 mortality subregions used in the World Health Report [83]. This system of model life tables was used extensively in the development of life tables for each Member State and in projecting life tables to 1999 when the most recent data available were from earlier years. Details on the data and methods used for each country are given by Lopez et al [82]. In countries with a substantial HIV

21 epidemic, separate estimates were made of the numbers and distributions of deaths due to HIV/AIDS and these deaths incorporated into the life table estimates [84]. The detailed distribution of causes of death by age and sex were estimated for each WHO Member State based on data from national vital registration systems that capture 16.7 million deaths annually. In addition, information from sample registration systems, population laboratories and epidemiological analyses of specific conditions were used to produce better estimates of the cause of death patterns. Cause of death patterns were carefully analysed to take into account incomplete coverage of vital registration in countries and the likely differences in cause of death patterns that would be expected in the low coverage areas of countries with incomplete data. Techniques developed in the Global Burden of Disease Study to undertake this analysis were further developed using a more extensive database and more robust modelling techniques [85]. Special attention was paid to problems of misattribution or miscoding of causes of death in cardiovascular disease, cancer injuries and general ill-defined categories. Deaths coded to ill- defined cardiovascular categories were reclassified using a correction algorithm described by Lozano et al [86]. A complete age-period-cohort model of cancer survival in each region was used to identify cancer sites with significant undercoding of mortality in order to reclassify cancer deaths coded to ill-defined categories [87].

3.3 Prior estimates for countries The disease-specific approach described in Section 2.3 has been used to develop the best possible initial (prior) estimates of weighted disability prevalence by age and sex for all 191 WHO member countries. These estimates are based on preliminary burden of disease analyses at country level which build on condition-specific epidemiological information to the maximum extent possible. This Section describes in more detail how they were developed. Step 1. As part of its annual assessment of world health in the World Health Report, WHO is updating and revising its estimates of disease burden for the 14 mortality subregions of the world. This involves carrying out detailed and comprehensive reviews of the incidence, prevalence, duration, and case fatality in all the regions of the world for each of 109 major disease and injury causes of mortality and disability by age group and sex (see Table 4). The ongoing revisions to the Global Burden of Disease analysis at WHO draw on a wide range of data sources. Various methods have been developed to reconcile often fragmented and partial estimates of epidemiological parameters that are available from different studies. A specific software tool, DisMod, is developed to ensure that the results of these assessments are internally consistent, and in particular, are consistent with cause of death distributions [7]. For a review of the development of DALYs and recent advances in disease burden measurement, see Murray and Lopez [7, 49]. Annex Table 4 of the World Health Report (2000) summarises the disease burden estimates for the 14 mortality subregions of the world in 1999. The ten leading causes of DALYs for each of these regions are also shown in this report (Table 9).

22 Table 4. Disease and injury categories used for regional burden of disease analyses for the World Health Report 2000. I. Communicable, Maternal & Perinatal II. Noncommunicable (continued) A. Infectious and Parasitic A. Malignant Neoplasms (continued) 1. Tuberculosis 14. Bladder 2. STD's excluding HIV 15. Lymphoma a. Syphilis 16. Leukemia b. Chlamydia 17. Other Cancers c. Gonorrhea B. Other Neoplasms d. Other STD's C. Diabetes Mellitus 3. HIV D. Nutritional/Endocrine 4. Diarrhoeal Diseases E. Neuro-psychiatric 5. Childhood Cluster 1. Major Affective Disorder a. Pertussis 2. Bipolar Affective Disorder b. Polio 3. Psychoses c. Gonorrhea 4. Epilepsy d. Measles 5. Alcohol Dependence e. Tetanus 6. Alzheimer's and other de 6. Meningitis 7. Parkinson's Disease 7. Hepatitis 8. Multiple Sclerosis 8. Malaria 9. Drug Dependence 9. Tropical Cluster 10. PTSD a. Trypanosomiasis 11. Obsessive Compulsive b. Chagas' Disease 12. Panic disorder c. Schistosomiasis 11. Other Neuro-psychiatric d. Leishmaniasis F. Sense Organ e. Lymphatic Filariasis 1. Glaucoma f. Onchocerciasis 2. Cataracts 10. Leprosy 3. Other Sense Organ 11. Dengue G. Cardiovascular 12. Japanese Encephaliti 1. Rheumatic Heart Disease 13. Trachoma 2. Ischemic Heart Disease 14. Intestinal Nematodes 3. Cerebrovascular Disease a. Ascaris 4. Inflammatory Cardiac b. Trichuris 5. Other c. Hookworm H. Respiratory 15. Other Infectious 1. COPD B. Respiratory Infections 2. Asthma 1. ALRI 3. Other Respiratory 2. AURI I. Digestive 3. Otitis Media 1. Peptic Ulcer Disease C. Maternal Conditions 2. Cirrhosis of the Liver 1. Hemorrhage 3. Appendicitis 2. Sepsis 3. Other Digestive 3. Hypertensive disorders of pregnancy J. Genito-Urinary 5. Obstructed Labor 1. Nephritis/Nephrosis 6. Abortion 2. Benign Prostatic Hypertension 7. Other Maternal 3. Other Genito-Urinary D. Perinatal Conditions K. Skin Disease E. Nutritional L. Musculo-Skeletal 1. Protein-Energy malnutrition 1. Rheumatoid Arthritis 2. Iodine Deficiency 2. Osteoarthritis 3. Vitamin A Deficiency 3. Other Musculo-Skeletal 4. Anemia M. Congential Abnormalities 5. Other Nutritional N. Oral Health II. Noncommunicable 1. Dental Caries 2. Periodontal Disease A. Malignant Neoplasms 3. Edentulism 1. Mouth and Oropharynx 4. Other Oral Health 2. Esophagus III. Injuries 3. Stomach A. Unintentional 4. Colon/Rectum 1. Motor Vehicle Accidents 5. Liver 2. Poisoning 6. Pancreas 3. Falls 7. Trachea/Bronchus/Lung 4. Fires 8. Melanoma and other Skin 5. Drowning 9. Breast 6. Other Unintentional Injuries 10. Cervix B. Intentional 11. Corpus Uteri 1. Self-inflicted 12. Ovary 2. Homicide and Violence 13. Prostate 3. War

23 Step 2. As described in Section 3.2, WHO has prepared estimates of numbers of deaths for each of its 191 Member States according to sex, age group (0, 1-4, then 5-year age groups to 85+) and 130 disease and injury causes (covering all causes of disease and injury). These estimates were used to calculate YLL by sex, age group and detailed cause for each Member State. Step 3. This country-level mortality data (Step 2), some country level epidemiological data and regional burden of disease estimates (Step 1) were then used to develop country-level estimates for YLD and total DALYs by sex, 5 year age group, and detailed cause as follows. For specific disease and injury causes where mortality is responsible for a significant proportion of the total burden (YLD/YLL ratio less than 5), regional estimates of YLD/YLL ratios by age and sex together with country-level estimates of YLL were used to estimate country-level YLD. This process ensures that country-specific knowledge on the epidemiology of the disease (as reflected in the country-level mortality estimates of that disease) is used to adjust the regional-level patterns of disability due to that cause. For specific disease and injury causes where mortality is not responsible for a significant proportion of the total burden (YLD/YLL ratio is 5 or higher), regional estimates of YLD rates per 1,000 population by age and sex were used together with country-level population distribution estimates and estimates of health expenditure per capita to make first estimates of the resulting YLD for each country. For some diseases, notably cancers, major depression and chronic respiratory conditions, available country-specific epidemiological estimates were also examined. In order to estimate disability prevalence at population level, it is also necessary to estimate the YLD associated with residual categories of disease and injury such as ‘Other chronic respiratory diseases’ or ‘Other malignant neoplasms’. We followed the procedure developed by the Global Burden of Disease Study [7, page 211] to estimate YLD for all of these residual categories. Step 4. For each member country, we then used the incidence YLD by age, sex and detailed cause (Step 3) to estimate undiscounted and un-age-weighted prevalence YLD by 5 year age group, sex and detailed cause. The method for conversion of incidence YLD to prevalence YLD used was dependent on the average duration of condition as follows: Short duration (<5 years): Prevalent YLD are equal to incident YLD Moderate duration (5 years to 50% of remaining life expectancy): We assume incident YLD are evenly distributed across the age interval a to a+L, where a is average age of onset and L is average duration. Long duration (50% or more of remaining life expectancy): We construct a life table for years lived with condition using the country life table and proportionately increasing mortality rates at all ages to match remaining life expectancy to the average duration of condition. We then use the Lx (years lived) column of the resulting life table to distribute incident YLD across age groups. Step 5. Adjustment for comorbidity. As discussed in Section 2.3, the total prevalent YLD per 100 population can be thought of as a severity-weighted disability prevalence measured as a percentage of the population of that age. However, summation over all conditions of the prevalence YLD calculated in Step 4 would result in overestimation of disability prevalence because of comorbidity between conditions. We correct for independent comorbidity between major condition groups (these approximately correspond to the Chapters of the International Classification of Diseases) as follows:

24     PYLDs,x 1 1- PYLDs,x,g (7) g where PYLDs,x,g is the prevalence YLD per 100 population for sex s, age x and cause g. The resulting PYLD per 100 population for sex s, age x gives the severity-weighted prevalence of disability by age and sex.

3.4 of health surveys Population representative data obtained through national sample surveys which include an assessment of general health status and physical and cognitive disability, have been collected and critically reviewed in an accompanying Discussion Paper [2]. Table 5 lists these surveys and summarises some of their characteristics. In order to estimate the prevalence of disability (non-fatal health) by five year age groups and sex at the country level, Sadana et al. [2] outline an analytic approach which addresses some of the methodological challenges regarding the comparability of the health status data collected. These include differences in the range and depth of questions and response scales [88] as well as differences in the interpretation of responses given different norms and expectations for health by age, sex or other sub- population groups. After conducting several validity and reliability checks, the analysis confirmed a latent dimension of disability that is common across population survey data and estimated the level of disability for each country and sex separately. As shown in Figure 7, the cumulative distribution of disability prevalence by severity is approximately exponential according to the detailed analyses carried out for the Global Burden of Disease study [7]. The distributions of latent health factor scores derived from the analysis of country health surveys were also generally exponential (see Figure 7 right-hand graph). The distribution of disability by severity level (or disability weight) can thus be approximately described by a two parameter exponential distribution as follows:   x d(x)  e  (8)  where x is the disability weight (severity) measured on a scale where 1 represents good health and 0 represents a state equivalent to death. The of this distribution is: d   P  (9) The parameter is readily interpreted as the proportion of the population with disability (with non-zero disability weight) and  as the average disability weight among the people with disability. Figure 8 compares for selected countries the mean values by age and sex for the latent health factor scores (expressed as a per cent on a scale where 0 represents good health and 100% the worst possible health state) with the prior mean values of severity-weighted disability from the condition-specific analyses. There are some countries where the latent health factor scores correspond reasonably well with the prior disability estimates (eg. Both sexes in Ireland, males in the USA). However, for many countries there are substantial differences in the latent health factor scores for males and females (eg. Portugal, USA) or in age trends (eg. Brazil). For many developing countries, very few people report disability in population surveys, resulting in quite low mean values for the latent health factor scores compared to the prior disability estimates. The graph for India in Figure 8 illustrates this situation.

25 Figure 7: Left: cumulative prevalence of disability by severity weight. Left: from GBD 1990. Right: for males in four European countries (using disability weight based on latent health factor scores unadjusted)

5-14 1.000 100 0.200 0.300 0.400 0.500 0.600 0.700 0.800 0.900 1.000

Germany UK Italy Portugal

0.100 60+

10 revalence p 45-59

0.010 prevalenceCumulative (%) Cumulative Cumulative

15-44

0-4 1 5-14 0.001 80 60 40 20 0 1 – Disability weight Disability w eight

As documented in detail by Sadana et al [2], there are country-specific biases in self-reported health data by sex and age that confound cross-national comparisons of health status or disability. Yet these data do appear to contain information on the health of the population. The task is to improve the comparison of health status across surveys and countries. Approaches that have attempted to map similar questions from surveys across countries have not been successful in achieving this aim, due to the substantial problems with self-report differences between cultures [2]. In order to obtain strict comparability of health status across surveys, the latent health factor derived from different surveys should map similar health states in different populations to the same score. Even if there were no self-report biases of the type documented above, there is still a question of whether the factor analyses of different surveys would result in a common underlying factor which measures health states on the same ordinal scale. For example, it is possible that surveys which contain questions that discriminate mild states of less than ideal health better may spread out such states more on the latent factor scale of 0 to 1 than other surveys. It is clear that for the estimation of DALE on an internationally comparable basis, if we are to use self-reported health survey information on population health status, we need to adjust the self-report data for differences in survey content and self-reporting biases in a way that will produce comparable estimates of severity-weighted disability prevalence.

26 Table 5. Sixty-four household interview surveys from 48 countries: survey characteristics by region

Region Age Sample Size Country Year Range M F Type of Survey AFRO D Ghana 1987 0+ 6833 7162 Integrated Household Survey (LSMS type) AFRO E Côte d'Ivoire 1988 0+ 4933 5174 Integrated Household Survey (LSMS type) South Africa 1994 0+ 21239 22703 Integrated Household Survey (LSMS type) United Republic of Tanzania 1995 14+ 2210 2068 Sub-national health survey AMRO A United States of America * 88-94 17+ 9401 10649 National health survey (NHANES III) United States of America 1994 18+ 35914 41523 National health survey (NHIS-D Phase I) AMRO B Brazil 96-97 4+ 9410 9999 Integrated Household Survey (LSMS type) Guyana 1996 0+ 3869 3924 Integrated Household Survey (LSMS type) Jamaica 1996 0+ 3422 3510 Integrated Household Survey (LSMS type) Panama 1997 0+ 10838 10599 Integrated Household Survey (LSMS type) Paraguay 1996 4+ 5543 5925 Integrated Household Survey (LSMS type) AMRO D Peru 1991 0+ 5795 6051 Integrated Household Survey (LSMS type) Peru * 1994 0+ 9402 9883 Integrated Household Survey (LSMS type) EMRO B Bahrain 1991 60+ 448 351 WHO Collaborating Study on Ageing Jordan 1991 60+ 545 652 WHO Collaborating Study on Ageing Tunisia 1991 60+ 658 578 WHO Collaborating Study on Ageing EMRO D Egypt 1991 60+ 710 470 WHO Collaborating Study on Ageing Morocco 90-91 0+ 9444 10128 Integrated Household Survey (LSMS type) Pakistan * 90-94 0+ 10039 9792 National health survey (NHSP) Pakistan 1991 0+ 18731 17340 Integrated Household Survey (LSMS type) EURO B Bulgaria 1995 0+ 3350 3577 Integrated Household Survey (LSMS type) Kyrgyzstan 1993 14+ 2617 3030 Integrated Household Survey (LSMS type) EURO C Russian Federation 1998 14+ 3750 4916 Longitudinal Integrated Household Survey EURO A Austria 1995 15+ 3567 3874 Longitudinal Integrated Household Survey (European Community) Belgium 1994 15+ 3872 4249 Longitudinal Integrated Household Survey 1995* 15+ 3666 4066 (European Community) Denmark 1994a 15+ 2855 3048 Longitudinal Integrated Household Survey 1995* 15+ 2680 2824 (European Community) Denmark 1994b 15+ 2699 2913 National health survey (DHMS: SF-36) France 1994 15+ 6839 7494 Longitudinal Integrated Household Survey 1995* 15+ 6368 6936 (European Community) Germany 1994 15+ 4150 4366 Longitudinal Integrated Household Survey 1995* 15+ 3885 4073 (European Community)

(continued)

27 Table 5 (continued). Sixty-four household interview surveys from 48 countries: survey characteristics by region

Region Age Sample Size Country Year Range M F Type of Survey EURO A Greece 1994 15+ 5904 6589 Longitudinal Integrated Household Survey 1995* 15+ 5878 6396 (European Community) Ireland 1994 15+ 4922 4982 Longitudinal Integrated Household Survey 1995* 15+ 4263 4268 (European Community) Italy 1994 15+ 8660 9071 Longitudinal Integrated Household Survey 1995* 15+ 8704 9079 (European Community) Luxembourg 1994 15+ 990 1056 Longitudinal Integrated Household Survey 1995* 15+ 957 1011 (European Community) Netherlands 1994 15+ 4457 4950 Longitudinal Integrated Household Survey 1995* 15+ 4299 4852 (European Community) Portugal 1994 15+ 5556 6065 Longitudinal Integrated Household Survey 1995* 15+ 5691 6167 (European Community) Spain 1994 15+ 8625 9285 Longitudinal Integrated Household Survey 1995* 15+ 7837 8443 (European Community) United Kingdom 1994 15+ 4986 5531 Longitudinal Integrated Household Survey 1995* 15+ 3995 4396 (European Community) SEARO B Indonesia 1995 0+ Total 9901 National health survey (SKRT) Indonesia * 93-94 0+ 5509 5103 Longitudinal Integrated Household Survey Indonesia 1990 60+ 568 634 WHO Collaborating Study on Ageing Sri Lanka 1990 60+ 638 562 WHO Collaborating Study on Ageing Thailand 1990 60+ 598 601 WHO Collaborating Study on Ageing SEARO D Bangladesh 1996 15+ 5266 6311 Sub-national Integrated Household Survey DPR of Korea 1990 60+ 585 596 WHO Collaborating Study on Ageing India 95-96 0+ 32353 30561 National Integrated Household Survey 6 2 (Round 57) Myanmar 1990 60+ 511 710 WHO Collaborating Study on Ageing Nepal 1994 0+ 9263 9592 Integrated Household Survey (LSMS type) WPRO B China 1993 0+ 7836 7846 Longitudinal Integrated Household Survey Fiji 1986 60+ 360 321 WHO Collaborating Study on Ageing Malaysia 1986 60+ 389 589 WHO Collaborating Study on Ageing Philippines 1986 60+ 326 491 WHO Collaborating Study on Ageing Republic of Korea 1986 60+ 348 593 WHO Collaborating Study on Ageing

*If more than one survey from a country, survey with * is used as input to DALE estimations.

28

Figure 8. Comparison of latent health factor scores and prior disability prevalences from condition- specific analyses, selected countries, 1999

Males: Latent health factor score Males prior disability estimates

Females: Latent health factor score Females prior disability estimates

United Kingdom Ireland 40 40

30 30

20 20

Mean value (%) 10 Mean value (%) 10

0 0 15 25 35 45 55 65 75 15 25 35 45 55 65 75 Age group Age group

Portugal Br azil 40 40

30 30

20 20 Mean value (%) Mean value (%) 10 10

0 0 15 25 35 45 55 65 75 15 25 35 45 55 65 75 Age group Age group

USA India 40 40

30 30

20 20 Mean value (%) Mean value (%) value Mean 10 10

0 0 15 25 35 45 55 65 75 15 25 35 45 55 65 75 Age group Age group

29

Figure 9. Estimation of true health status from self-reported health status in representative national population surveys

Survey item 1 Country/culture

Perceived/ Survey item 2 Sex reported health status Survey item 3

Age . . . Socioeconomic status

True Survey item n health status Mortality

Healt determinants

Health/welfare system

This estimation problem is illustrated in Figure 9. Determinants of “bias” in self-report health data such as age, sex, socioeconomic status and other population-specific (cultural/environmental) factors are also determinants of true health status. We made a number of attempts to develop models to estimate true health status from observed health status by making use of the self-reported health status of reference groups (such as young, high socioeconomic status people) who could be assumed to have good true health status. However, we found that the patterns of bias between social groups and across countries were so heterogeneous that such “bootstrapping” calibration procedures using internal health patterns within survey populations were not able to produce reliably comparable data across surveys. We thus concluded that the inter-country biases in self-report health survey data could only be adequately adjusted if external calibrators (of true health or of health determinants) were used to make appropriate adjustments to self-report data.

3.5 Posterior estimates In order to maximise comparability of the disability dimensions and to ensure that the resulting scores appropriately reflected health state preferences, we have developed numerical models for estimating and adjusting for age, sex and cross-country bias in reporting of health states. These use prior estimates of disability distributions based on burden of disease analyses at regional and country level (Section 3.3) in order to estimate a parsimonious set of parameters to adjust for differences in survey content, in self-report responses and for the mapping of the resulting latent health factor to health state preference weights. This procedure

30 enables us to incorporate all data from household surveys into our estimates of average levels of severity-weighted disability for countries. We briefly describe the model and the estimation procedure here. If f > denotes the latent health factor score derived from health survey data using confirmatory factor analysis (Sadana et al 2000) (2), and f denotes the true underlying health state (in terms of a disability weight ranging from 0 to 1), then we describe the relationship between these two factors in terms of three parameters as follows (see also Figure 10):  >f .  1  P / if f > ?  1 1  / 2 2 0 f  (10) >   1 if f 2

Figure 10. Adjustment of latent health factor score for reporting biases and disability severity

100 100 <1 80 >1 80

60 60

40 40 1

True disability weight 20  True disability weight 20 1

0 2 0 0 50 100  0 502 100 Latent health factor score Latent health factor score

 The first parameter 1 specifies the disability weight for the worst health state observed in the  health survey, the second parameter 2 specifies the latent health factor score above which all health states are equivalent to good health f  1(see Figure 10). This latter parameter allows for health surveys which include questions which discriminate between people in good health, thus resulting in a range of factor scores which all equate to good health. This ceiling effect is discussed in more detail by Sadana et al (2000) (2). The third parameter,  , specifies a power transformation which allows the latent health factor score to decrease with increasing severity faster or slower than the true disability weight as illustrated in Figure 10. We used this model to rescale latent health factor scores for all the available health surveys in a region as follows. Prior values for average disability severity were derived from the condition-specific country-level analyses (described in Section 3.3) by sex for age groups 0-4, 5-14, 15-24, 25-34, 35-44, 45-54, 55-64, 65-74, 75+. Nonlinear optimization methods were   then used to estimate values for parameters 2 and which minimized the difference between mean disability weight and the mean value (across all age groups combined) for the  rescaled factor score f . The parameter 1 was set at 0.1 in all models. Sensitivity analyses

31 found that the value assumed for this parameter in a range from around 0.0 to 0.2 did not significantly affect the mean values for the rescaled factor scores.   We aimed to estimate common values 2 and for all countries in a region to the extent possible, ie. to simultaneously fit equation (10) to data for a number of countries. We found that this was rarely possible due to the heterogeneity of the self-report biases across countries,  and in some instances, we also needed to allow 2 to differ for ages 65 and over, to allow for age-related variations in reporting behaviour. Health survey data for a number of countries was restricted to ages 60 and above (see Table 5). For these countries, the model parameters were estimated by fitting the model to prior disability distributions for four 5-year age groups (60-64, 65-69, 70-74, 75+). Figure 11 compares the average values of the rescaled health factor (the posterior disability prevalence) with the latent health factor scores by age and sex for selected countries. In some countries, the posterior estimates are quite similar to the latent health factor scores (eg. Ireland, Greece, males in the USA). However, for many countries the fitted bias parameters result in quite substantial changes in the posterior estimates (eg. United Kingdom, Brazil, females in Portugal). Figure 12 compares the posterior (adjusted) disability prevalence estimates for seven age groups (15-24, through to 75+) to the prior disability prevalence estimates (from the condition-specific analyses) for 13 European countries. For these countries, the latent health factor scores were derived from a single common health survey using the same instrument in all 13 countries (see Table 5). Figure 13 similarly compares posterior and prior disability prevalence estimates for nine age groups in 4 African countries. Figure 14 illustrates the differences between the posterior disability prevalence estimates and the (unadjusted) latent health factor scores derived from the initial factor analyses of the country health surveys. The first 6 graphs are for selected countries in Europe and the Americas. The health survey data results in moderate changes in the level and age trends of average weighted disability prevalence for the European countries. For the USA, prior and posterior estimates are quite close, apart from for the oldest male age group. For Brazil, the use of the health survey data results in reductions in the prior disability estimates across all ages for both sexes. The second six graphs (Figure 14 continued) are for selected countries in Africa, Asia and the Western Pacific region. The health survey data resulted in increases above estimated prior levels of disability at younger ages in both Ghana and Côte d’Ivoire. The analysis of health survey data resulted in a reduction in estimated prior levels of disability at older ages in the Republic of Korea (South Korea) and increases in the estimated prior levels of disability at older ages in the Democratic People’s Republic of Korea (North Korea). Health survey data were only available for ages 60 and over for Thailand and Myanmar. Analysis of this data resulted in increases in prior estimates of disability for the older old (ages 75 and over) in both countries. Using the methods outlined above, we obtained estimates of weighted disability prevalence by sex and age for all WHO Member States. For countries where we were not able to obtain unit record data for appropriate household surveys, we used the estimated prevalence of disability by five year age groups and sex calculated by age, sex and cause at country level based on regional estimates of the burden of disease for 1999 together with country-specific epidemiological information. Figure 15 shows the resulting distributions of weighted disability prevalence by age and sex for the 14 WHO mortality subregions in 1999. These prevalences have been calculated by

32 summing disability prevalences weighted by population numbers across Member States in each subregion.

33

Figure 11. Posterior (adjusted health survey) disability prevalence versus latent health factor score, selected countries, 1999

Males: Latent health factor score Males posterior estimates

Females: Latent health factor score Females posterior estimates

United Kingdom Ireland 40 40

30 30

20 20 Average (%) Average (%) 10 10

0 0 15 25 35 45 55 65 75 15 25 35 45 55 65 75 Age group Age group

Gree ce Portugal 40 40

30 30

20 20 Average (%) Average (%) 10 10

0 0 15 25 35 45 55 65 75 15 25 35 45 55 65 75 Age group Age group

USA Br az il 35 35

30 30

25 25

20 20

15 15

Average (%) 10 Average (%) 10

5 5

0 0 15 25 35 45 55 65 75 15 25 35 45 55 65 75 Age groups Age groups

34 Figure 12. Comparison of posterior (adjusted) disability prevalence versus prior disability prevalence for seven age groups by sex, 13 European countries, 1995

35

30 ) 25

20

15

10 Modelled disability(%

5

0 0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 Prior disability (%)

Figure 13. Comparison of posterior (adjusted) disability prevalence versus prior disability prevalence, for nine age groups by sex, 4 African countries countries

70

60

50

40

30

20 Modelled disability (%) disability Modelled

10

0 0.0 10.0 20.0 30.0 40.0 50.0 60.0 Prior disability (%)

35

Figure 14. Posterior (adjusted health survey) disability prevalence versus prior disability prevalence, selected countries, 1999

Males: Prior disability estimates Males posterior estimates

Females: Prior disability estimates Females posterior estimates

Germ any Ne the r lands 35 35

30 30 25 25

20 20 15 15

10 10

Wgt. disability prev. disability (%)Wgt. 5 prev. disability (%)Wgt. 5

0 0 15 25 35 45 55 65 75 15 25 35 45 55 65 75 Age groups Age groups

United Kingdom Italy 35 35

30 30 25 25

20 20 15 15

10 10

Wgt. disability prev. disability (%)Wgt. 5 prev. disability (%)Wgt. 5

0 0 15 25 35 45 55 65 75 15 25 35 45 55 65 75 Age groups Age groups

USA Br az il 35 35

30 30

25 25

20 20

15 15

10 10

Wgt. disability prev. (%) 5 Wgt. disability prev. (%) 5

0 0 15 25 35 45 55 65 75 15 25 35 45 55 65 75 Age groups Age groups

36 Figure 14 (continued). Posterior (adjusted health survey) disability prevalence versus prior disability prevalence, selected countries, 1999

Males: Prior disability estimates Males posterior estimates

Females: Prior disability estimates Females posterior estimates

Ghana Côte d'Ivoire 40 40 35 35 30 30 25 25 20 20 15 15 10 10

Wgt. disability prev. (%) 5 Wgt. disability prev. (%) 5 0 0 0 5 15 25 35 45 55 65 0 5 15 25 35 45 55 65 Age groups Age groups

Republic of Korea Democratic People's Republic of Korea 50 50

40 40

30 30

20 20

10 10 Wgt. disability prev. (%) Wgt. disability prev. (%)

0 0 35 40 45 50 55 60 65 70 75 35 40 45 50 55 60 65 70 75 Age groups Age groups

Thailand Myanmar 50 50

40 40

30 30

20 20

10 10 Wgt. disability prev. (%) Wgt. disability prev. (%)

0 0 35 40 45 50 55 60 65 70 75 35 40 45 50 55 60 65 70 75 Age groups Age groups

37 3.6 Calculation of DALE Sullivan’s method was used to compute DALE for each Member State from the country life table and the severity-weighted prevalence estimates. Sullivan's method involves using the observed prevalence of disability at each age in the current population (at a given point of time) to divide the hypothetical years of life lived by a period life table cohort at different ages into years with and without disability. The method is illustrated in Table 6.

Table 6. Illustration of Sullivan's method for the calculation of disability-free life expectancy.

Ordinary life table Disability Years Years LED DFLE Survivors Years Expectation prevalence with without LE with Disability (%) Age lx lived Lx of life ex disability disability disability -free LE 0 100000 496210 74.98 4.5 22130 474080 16.60 58.38 5 99134 495425 70.63 9.6 47506 447919 16.52 54.11 10 99045 495018 65.69 8.6 42568 452450 16.05 49.64 15 98940 493916 60.76 5.7 28100 465816 15.64 45.12 20 98572 491448 55.98 7.6 37433 454015 15.41 40.56 25 97997 488469 51.29 8.5 41623 446846 15.12 36.17 30 97383 485285 46.60 10.6 51280 434005 14.79 31.81 35 96722 481816 41.90 12.2 59013 422803 14.36 27.54 40 95988 477781 37.20 14.3 68247 409534 13.86 23.34 45 95079 472220 32.53 17.9 84507 387713 13.27 19.26 50 93701 463324 27.97 23.5 108766 354558 12.57 15.40 55 91452 448652 23.59 30.9 138780 309872 11.68 11.90 60 87702 424469 19.48 41.6 176738 247731 10.60 8.88 65 81656 386806 15.73 44.0 170265 216541 9.22 6.50 70 72512 332217 12.38 58.3 193526 138691 8.04 4.34 75 59796 259645 9.45 59.6 154714 104931 6.51 2.94 80 43550 173081 7.02 73.2 126672 46409 5.39 1.63 85 25802 132424 5.13 81.5 107916 24508 4.18 0.95 Notes: First four columns are from a standard life table for a population. lx is the number of survivors at age x in the hypothetical life table cohort. Lx is the number of years of life lived by the life table cohort between ages x and x+5. prevx is the prevalence of disability between ages x and x+5 in the population Years lived with disability YDx = Lx * prevx, Years lived without disability YWDx = Lx * (1-prevx) DFLEx = Sum of years lived without disability for ages x and above, divided by lx DLEx = Sum of years lived with disability for ages x and above, divided by lx

DALE can be calculated using the same method as illustrated in Table 7 where disability prevalence is replaced by severity-weighted disability prevalence.

38 Using standard notation for the country life table parameters, we calculated DALE at age x as follows:

Dx Severity-weighted prevalence of disability between ages x and x+5

YDx = Lx * Dx Equivalent years of healthy life lost due to disability between ages x and x+5

YWDx = Lx * (1- Dx) Equivalent years of healthy life lived between ages x and x+5 Lx is the total years lived by the life table population between ages x and x+5 DALE at age x is the sum of YWDi from i = x to w (the last open-ended age interval in the life table) divided by lx (survivors at age x):

 w . / (11) DALEx = YWDi /lx ix 0  w . / DLEx = YDi /lx = LEx - DALEx (12) ix 0 DLEx,the equivalent years of healthy life lost due to disability, is the sum of YDi from i = x to w divided by lx (survivors at age x).

3.7 Uncertainty analysis Uncertainty intervals have been estimated for life expectancies and other life table parameters for WHO member countries as described by Salomon and Murray [89]. To capture the uncertainty due to , indirect estimation techniques and projections, a total of 1000 life tables was developed for each Member State in order to quantify the uncertainty distribution of key life table parameters. In countries with a substantial HIV epidemic, recent estimates of the level and uncertainty range of the magnitude of HIV/AIDS deaths by age and sex have been incorporated into the life table uncertainty analysis. The degree of uncertainty in country-level weighted disability prevalences has also been estimated for each country. This is mainly determined by levels of uncertainty in (a) epidemiological estimates for prevalence, incidence and/or severity of disability associated with specific conditions, (b) estimation of prevalence YLD from incidence YLD, and (c) the approximate nature of adjustments for comorbidity. For all these reasons, the uncertainty distributions across different ages are likely to be highly correlated for children, for adults, and at older ages. To be conservative in our estimation of uncertainty, we assumed 100% correlation between uncertainty at each age within broad age ranges 0-14, 15-29, 30-44, 45-59, 60-69 and 70+ (so that for a given sample of the disability prevalence distribution, it is high at all ages or low at all ages within one of these ranges). The uncertainty distributions of the DALE estimates for each Member State were quantified by developing a total of 1000 DALE life tables for each Member State which simultaneously sampled the uncertainty in the life tables and the disability prevalences. The techniques used

39 for estimating uncertainty in life expectancies, DALE and in country rankings are discussed further in Salomon and Murray [89]. The uncertainty ranges for these quantities given in Annex Table B and shown in graphs and tables in Section 4 give the 10th and 90th percentile of the relevant uncertainty distributions. The ranges thus define 80% uncertainty intervals around the estimates. Rank uncertainty is not only a function of the uncertainty of the DALE measurement for each country, but also the uncertainty of the measurement of adjacent countries in ranking table.

Figure 15. Weighted disability prevalence (%), by age, sex, and WHO mortality sub-region, 1999

50% Males Age 0-4 ) 40% 5-14 15-44 45-59 30% 60+

20%

Weighted disability prevalence (% prevalence disability Weighted 10%

0% WprA EurA AmrA EmrB WprB EurB EurC SearB AmrB EmrD SearD AmrD AfrD AfrE

40 50% Females Age

0-4 ) 40% 5-14 15-44 45-59 30% 60+

20%

Weighted disability prevalence (% prevalence disability Weighted 10%

0% WprA EurA AmrA EmrB WprB EurB EurC SearB AmrB EmrD SearD AmrD AfrD AfrE

41 4. Results Using the methods outlined in the previous Section, we have estimated healthy life expectancy (DALE) for males and females in the 191 Member States of WHO for 1999, as well as for 14 mortality sub-regions of the world, the 6 WHO regions and for the total global population. These estimates of healthy life expectancy are based on country-specific estimates of mortality, cause of death patterns, epidemiological analyses and health survey data where available. We describe the results in this Section. Estimates and uncertainty intervals for DALE at age 0 and 60 are given in full in Annex Table A for each Member State.

4.1 DALE for WHO regions and the world in 1999 Country-level estimates for mortality and disability were aggregated to estimate life expectancy (LE) and healthy life expectancy (DALE) for each of the six WHO Regions and for the world (Table 7). Regional healthy life expectancies at birth in 1999 ranged from a low of 37 years for African males to a high of almost 70 years for females in the low mortality countries of mainly Western Europe. This is an almost 2-fold difference in healthy life expectancy between major regional populations of the world. Regional healthy life expectancies at age 60 in 1999 ranged from a low of 8.4 years for African males to a high of around 22 years for females in Europe and North America. The difference between DALE and total life expectancy is DLE (expected years “lost” due to disability), shown in Figure 16 as the light shaded areas. The equivalent healthy years “lost”

Table 7. Life expectancy (LE), healthy life expectancy (DALE), and years lost to disability as per cent of total LE (DLE%), at birth and at age 60, by sex and total, WHO regions and world, 1999

Persons Males Females WHO DALE LE DLE% DALE LE DLE% DALE LE DLE% Region (years) (years) (%) (years) (years) (%) (years) (years) (%) At birth AFRO 37.5 46.3 18.9 37.3 45.6 18.1 37.8 47.0 19.7 AMRO 65.2 72.7 10.4 62.3 69.5 10.4 68.1 76.0 10.3 EMRO 54.5 61.9 12.0 54.4 61.2 11.1 54.6 62.6 12.8 EURO 66.5 72.9 8.8 63.1 69.1 8.6 69.8 76.7 9.0 SEARO 53.9 61.2 11.9 53.3 60.2 11.5 54.4 62.1 12.4 WPRO 63.3 70.4 10.2 61.9 68.5 9.7 64.7 72.4 10.7 World 56.8 64.5 11.9 55.8 62.5 10.7 57.8 66.4 12.9 At age 60 AFRO 9.0 14.9 39.5 8.4 14.4 42.0 9.6 15.3 37.1 AMRO 16.0 20.5 21.7 14.5 18.6 22.3 17.5 22.3 21.3 EMRO 10.6 14.5 27.2 10.6 14.4 26.6 10.6 14.6 27.7 EURO 15.8 19.6 19.4 14.0 17.5 19.7 17.6 21.8 19.2 SEARO 12.1 16.1 24.7 11.5 15.4 25.3 12.7 16.8 24.1 WPRO 13.7 18.0 23.9 12.6 16.5 23.7 14.9 19.6 24.0 World 13.5 18.0 24.7 12.6 16.6 23.6 14.4 19.4 25.6

42 due to disability range from 18.9% (of total life expectancy at birth) in Africa to 8.8% in the European region. The equivalent healthy years “lost” due to disability at age 60 are a higher percentage of remaining life expectancy, due to the higher prevalence of disability at older ages. These range from around 40% in sub-Saharan Africa to around 20% in developed countries.

Figure 16. Disability-adjusted life expectancy (DALE), healthy years lost due to disability (DLE) and life expectancy (LE), at birth for total populations, WHO regions, 1999

80

75 DLE LE DALE 70

65 Global LE

60 Global DALE 55

50 Expectancy (years) 45

40

35

30 AFRO SEARO EMRO WPRO AMRO EURO WHO Region

4.2 DALE for the 14 mortality subregions of the world in 1999 When DALE is calculated for the 14 mortality subregions of the world, the range is even greater (Table 8). Subregional healthy life expectancies at birth in 1999 ranged from a low of 35 years for males in the very high mortality subregion of Africa to a high of almost 77 years for females in the low mortality countries of the Western Pacific region (these include Japan, Australia, New Zealand and Singapore). Figure 17 compares DALE and total life expectancy at birth across the 14 mortality subregions in 1999. Figure 18 similarly compares DALE and total life expectancy at age 60 across the 14 mortality subregions in 1999. The very low health expectancies of the African countries in both subregions D and E reflects the high burden of HIV/AIDS, malaria, other communicable, maternal, perinatal and nutritional conditions, and injuries. Figure 19 summarises the relationship between life expectancy and DALE for the 14 mortality subregions of the world, for both men and women. The gap between LE and DALE ranges from between 8 and 9 years for the subregions in Africa to between 5 and 6 years for females in developed countries. Despite the fact that people live longer in the richer, more

43

Table 8. Life expectancy (LE), healthy life expectancy (DALE), and years lost to disability as per cent of total LE (DLE%), at birth and at age 60, by sex and total, by mortality subregion, 1999

Persons Males Females WHO DALE LE DLE% DALE LE DLE% DALE LE DLE% Region (years) (years) (%) (years) (years) (%) (years) (years) (%) At birth AFRO D 40.3 49.4 18.5 40.0 48.5 17.6 40.6 50.4 19.4 AFRO E 35.4 43.9 19.3 35.3 43.3 18.6 35.6 44.5 20.0 AMRO A 70.4 76.9 8.4 67.9 74.0 8.3 73.0 79.8 8.5 AMRO B 62.7 71.0 11.7 59.5 67.5 11.9 65.9 74.5 11.6 AMRO D 55.9 64.1 12.7 54.5 62.3 12.6 57.4 65.8 12.8 EMRO B 61.0 67.7 10.0 61.7 67.3 8.3 60.2 68.2 11.6 EMRO D 52.5 60.1 12.8 52.0 59.3 12.3 52.9 61.0 13.3 EURO A 71.8 77.7 7.6 69.0 74.5 7.4 74.7 81.0 7.7 EURO B 62.8 69.8 10.0 60.7 67.0 9.4 64.9 72.6 10.6 EURO C 61.6 68.4 9.9 56.6 62.9 10.0 66.6 73.9 9.8 SEARO B 59.8 67.8 11.8 58.5 66.1 11.6 61.1 69.4 11.9 SEARO D 52.8 60.0 12.0 52.4 59.2 11.5 53.1 60.7 12.5 WPRO A 74.4 80.6 7.7 71.7 77.3 7.2 77.1 83.9 8.2 WPRO B 61.9 69.3 10.6 60.8 67.5 10.0 63.1 71.0 11.1 World 56.8 64.5 11.9 55.8 62.5 10.7 57.8 66.4 12.9 At age 60 AFRO D 9.6 15.6 38.3 9.1 15.1 40.1 10.2 16.0 36.6 AFRO E 8.5 14.2 40.6 7.7 13.8 43.9 9.2 14.7 37.5 AMRO A 17.1 21.5 20.2 15.4 19.5 20.8 18.8 23.4 19.6 AMRO B 15.1 19.7 23.4 13.7 17.9 23.7 16.4 21.4 23.2 AMRO D 11.4 15.8 28.0 11.1 15.3 27.4 11.7 16.4 28.5 EMRO B 11.0 14.5 23.7 11.4 14.5 21.1 10.6 14.5 26.4 EMRO D 10.4 14.6 28.7 10.2 14.4 29.2 10.5 14.7 28.2 EURO A 17.7 21.4 17.2 15.8 19.1 17.3 19.6 23.7 17.2 EURO B 14.5 18.3 20.8 13.3 16.7 20.5 15.7 19.9 21.0 EURO C 13.2 17.2 23.3 11.0 14.7 25.2 15.3 19.7 21.9 SEARO B 15.3 20.1 23.9 15.3 20.1 23.9 15.4 20.2 23.9 SEARO D 11.5 15.4 24.9 10.9 14.6 25.7 12.2 16.1 24.1 WPRO A 19.6 23.7 17.2 17.6 21.1 16.5 21.6 26.3 17.7 WPRO B 12.8 17.1 25.5 11.8 15.8 25.3 13.7 18.4 25.6 World 13.5 18.0 24.7 12.6 16.6 23.6 14.4 19.4 25.6

44 Figure 17. Disability-adjusted life expectancy (DALE) and life expectancy (LE), at birth for total populations, by mortality sub-region, 1999

WPR A EUR A AMR A EUR B AMR B WPR B EUR C EMR B SEAR B AMR D SEAR D LE EMR D DALE AFR D AFR E

30 40 50 60 70 80 Expectation (Years)

Figure 18. Disability-adjusted life expectancy (DALE) and life expectancy (LE), at age 60 total populations, by mortality sub-region, 1999

WPR A EUR A AMR A SEAR B AMR B EUR B EUR C WPR B SEAR D AMR D EMR B LE EMR D DALE AFR D AFR E

0102030 Expectation (Years)

45 Figure 19. Life expectancy and disability-adjusted life expectancy at birth and at age 60, for males and females, by WHO mortality subregion, 1999. The dotted line represents a situation of no time lived with disability, so that life expectancy and DALE coincide.

80 20

75 Male Male Female 70 Female

65 15 60

55

50 10 45 DALE at birth (years) birth at DALE DALE at age 60 (years)60 age at DALE 40

35 5 30 10 15 20 25 40 45 50 55 60 65 70 75 80 85 Life expectancy at birth (years) Life expectancy at age 60 (years)

developed countries, and have greater opportunity to acquire non-fatal disabilities in older age, disability has a greater absolute (and relative) impact on healthy life expectancy in poorer countries. Separating life expectancy into equivalent years of good health and years of lost good health thus widens rather than narrows the difference in health status between the rich and the poor countries. The relative contributions of diseases and injuries to variations in DALE are best summarised in terms of the loss of healthy life measured in DALYs. The World Health Report provides detailed estimates of DALYs for over 100 disease and injury categories for the 14 mortality subregions. The leading causes of DALYs worldwide are shown in Table 9 for males and females separately. While the rankings are broadly similar for the two sexes, there are important differences. Thus while perinatal conditions, HIV/AIDS and lower respiratory infections are the three leading causes of DALYs their relative importance differs slightly for males and females. More importantly, depression is the fourth leading cause of disease burden for females but ranks ninth for males. Maternal conditions are the seventh leading cause for females, causing almost 4% of their global disease burden in 1999. Road traffic accidents are a leading cause of overall disease and injury burden for males (3.9%) but not for females. In parts of South Asia, Eastern Europe and the Western Pacific, 20% or more of the entire disease and injury burden is due to injuries alone. There are marked contrast in epidemiological patterns between rich and poor regions of the world (Table 9). Thus in the more developed countries, the share of disease burden due to communicable, maternal, perinatal and nutritional conditions is typically around 5%, compared with 70-75% in Africa. Specifically, the leading causes of disease burden in Africa in 1999 were HIV/AIDS (20.0%). Malaria (10.0%) and acute lower respiratory infections (8.4%), compared with ischaemic heart disease, depression, alcohol dependence and stroke in the industrialized countries (Table 9).

46 Table 9. Top 10 causes of loss of healthy life expectancy (in DALYs) for the 14 mortality sub-regions, 1999

GLOBAL DALYs (000) % 1 Acute lower respiratory infections 96 682 6.7 2 HIV/AIDS 89 819 6.2 3 Perinatal conditions 89 508 6.2 4 Diarrhoeal diseases 72 063 5.0 5 Unipolar major depression 59 030 4.1 6 Ischaemic heart disease 58 981 4.1 7 Cerebrovascular disease 49 856 3.5 8 Malaria 44 998 3.1 9 Road traffic accidents 39 573 2.8 10 COPDa 38 1562.7 All causes 1 438 154 100

AFRO D DALYs (000) % AFRO E DALYs (000) % 1 Malaria 18 600 11.7 HIV/AIDS 58 671 27.3 2 HIV/AIDS 15 778 10.0 Malaria 18 238 8.5 3 Acute lower respiratory infections 14 858 9.4 Acute lower respiratory infections 16 871 7.8 4 Perinatal conditions 12 351 7.8 Diarrhoeal diseases 12 454 5.8 5 Diarrhoeal diseases 11 867 7.5 Perinatal conditions 11 746 5.5 6 Measles 8 762 5.5 Measles 8 701 4.0 7 Maternal conditions 4 954 3.1 Maternal conditions 7 571 3.5 8 Congenital abnormalities 3 180 2.0 Tuberculosis 5 563 2.6 9 Tuberculosis 3 158 2.0 Congenital abnormalities 3 257 1.5 10 Road traffic accidents 3 011 1.9 Road traffic accidents 3 207 1.5 All causes 158 439 100 All causes 214 921 100

AMRO A DALYs (000) % AMRO B DALYs (000) % 1 Ischaemic heart disease 3 298 8.5 Unipolar major depression 4 240 6.0 2 Unipolar major depression 2 623 6.8 Alcohol dependence 3 762 5.3 3 Alcohol dependence 1 863 4.8 Perinatal conditions 3 707 5.2 4 Diabetes mellitus 1 628 4.2 Nutritional/endocrine disorders 3 386 4.8 5 Road traffic accidents 1 612 4.2 Homicide and violence 3 000 4.2 6 Trachea/bronchus/lung 1 509 3.9 Ischaemic heart disease 2 915 4.1 7 Cerebrovascular disease 1 410 3.7 Road traffic accidents 2 901 4.1 8 COPDa 1 267 3.3 Acute lower respiratory infections 2 315 3.3 9 Alzheimer and other dementias 1 141 3.0 Osteoarthritis 2 224 3.1 10 Osteoarthritis 1 087 2.8 Cerebrovascular disease 2 208 3.1 All causes 38 627 100 All causes 70 969 100

AMRO D DALYs (000) % EMRO B DALYs (000) % 1 Perinatal conditions 1 283 7.8 Ischaemic heart disease 1 484 7.1 2 Acute lower respiratory infections 1 054 6.4 Unipolar major depression 1 312 6.3 3 Diarrhoeal diseases 770 4.7 Perinatal conditions 1 134 5.4 4 Alcohol dependence 656 4.0 Cerebrovascular disease 1 041 5.0 5 Unipolar major depression 646 3.9 Diarrhoeal diseases 977 4.7 6 Diabetes mellitus 625 3.8 Acute lower respiratory infections 921 4.4 7 HIV/AIDS 621 3.8 Road traffic accidents 881 4.2 8 Nutritional/endocrine disorders 588 3.6 Maternal conditions 704 3.4 9 Tuberculosis 471 2.9 Anaemias 607 2.9 10 Cirrhosis of the liver 453 2.8 Nutritional/endocrine disorders 492 2.4 All causes 16 346 100 All causes 20 895 100 (continued)

47 Table 9 (continued). Top 10 causes of loss of healthy life expectancy for the 14 mortality sub-regions, 1999

EMRO D DALYs (000) % EURO A DALYs (000) % 1 Perinatal conditions 10 621 10.4 Ischaemic heart disease 4 757 9.7 2 Acute lower respiratory infections 9 625 9.5 Unipolar major depression 3 376 6.9 3 Diarrhoeal diseases 9 146 9.0 Cerebrovascular disease 2 857 5.8 4 Congenital abnormalities 5 446 5.4 Alcohol dependence 2 562 5.2 5 Ischaemic heart disease 3 588 3.5 Trachea/bronchus/lung 1 774 3.6 6 Unipolar major depression 3 227 3.2 Alzheimer and other dementias 1 737 3.5 7 Measles 3 020 3.0 Osteoarthritis 1 508 3.1 8 Malaria 2 727 2.7 COPDa 1 441 2.9 9 Road traffic accidents 2 298 2.3 Road traffic accidents 1 423 2.9 10 Cerebrovascular disease 2 277 2.2 Diabetes mellitus 1 176 2.4 101 688 48 999

EURO B DALYs (000) % EURO C DALYs (000) % 1 Ischaemic heart disease 2 986 8.2 Ischaemic heart disease 7 113 14.0 2 Cerebrovascular disease 2 727 7.5 Cerebrovascular disease 5 584 11.0 3 Unipolar major depression 1 964 5.4 Unipolar major depression 2 297 4.5 4 Acute lower respiratory infections 1 435 3.9 Self-inflicted 2 200 4.3 5 Road traffic accidents 1 355 3.7 Road traffic accidents 1 589 3.1 6 Perinatal conditions 1 269 3.5 Osteoarthritis 1 481 2.9 7 Alcohol dependence 1 192 3.3 Poisoning 1 396 2.7 8 Osteoarthritis 1 151 3.2 COPDa 1 323 2.6 9 Diabetes mellitus 876 2.4 Homicide and violence 1 240 2.4 10 Diarrhoeal diseases 875 2.4 Alcohol dependence 1 140 2.2 36 484 50 868

SEARO B DALYs (000) % SEARO D DALYs (000) % 1 Perinatal conditions 6 362 11.2 Acute lower respiratory infections 33 746 9.5 2 Road traffic accidents 4 377 7.7 Diarrhoeal diseases 28 960 8.1 3 Tuberculosis 3 453 6.1 Perinatal conditions 26 353 7.4 4 Unipolar major depression 3 104 5.5 Ischaemic heart disease 20 133 5.7 5 Acute lower respiratory infections 2 904 5.1 Falls 13 742 3.9 6 Cerebrovascular disease 2 041 3.6 Unipolar major depression 12 165 3.4 7 Anaemias 1 749 3.1 Congenital abnormalities 10 889 3.1 8 Ischaemic heart disease 1 732 3.1 Tuberculosis 10 648 3.0 9 Falls 1 515 2.7 Cerebrovascular disease 8 639 2.4 10 Self-inflicted 1 207 2.1 Anaemias 8 462 2.4 56 604 355 876

WPRO A DALYs (000) % WPRO B DALYs (000) % 1 Unipolar major depression 1 255 8.2 COPDa 22 593 9.0 2 Cerebrovascular disease 1 160 7.6 Unipolar major depression 17 132 6.8 3 Ischaemic heart disease 721 4.7 Cerebrovascular disease 15 982 6.3 4 Alcohol dependence 714 4.7 Perinatal conditions 12 914 5.1 5 Osteoarthritis 599 3.9 Acute lower respiratory infections 10 798 4.3 6 Alzheimer and other dementias 580 3.8 Road traffic accidents 8 615 3.4 7 Self-inflicted 513 3.4 Self-inflicted 8 407 3.3 8 Stomach 462 3.0 Anaemias 7 764 3.1 9 Trachea/bronchus/lung 448 2.9 Falls 7 316 2.9 10 Road traffic accidents 414 2.7 Congenital abnormalities 6 955 2.8 15 235 252 204

(a) Chronic obstructive pulmonary disease (chronic bronchitis and emphysema)

48

4.3 Global DALE and disability severity distribution, 1999 Overall, for the entire population of the world, average life expectancy at birth in 1999 was 64.5 years, an increase of almost 6 years over the last two decades. Global healthy life expectancy at birth is 56.8 years, 7.7 years lower than total life expectancy at birth (see Table 7). At the global level, female healthy life expectancy is 57.8 years, 2.0 years higher than male healthy life expectancy at 55.8 years. Figure 20 shows global estimates of DALE, DLE and LE for 1999 at fifteen year age intervals. Expectancies at ages greater than zero are plotted as bars based on the relevant age, so that the total height of the bar represents expected age at death. Expected age at death rises with achieved age, since the person has already survived the risk of death at earlier ages. Although the primary focus of the DALE analyses for the World Health Report 2000 has been on estimating severity-weighted disability prevalence and disability-adjusted life expectancy, we have also made an estimate of the global pattern of disability prevalence in terms of the seven disability severity classes used in the Global Burden of Disease 1990 Study [7, 34]. Figure 21 shows the estimated global distribution of disability by severity, in terms of the corresponding survival curves for the global population. As discussed in Section 2.3, total global healthy life expectancy (DALE) is the total area under this curve, weighted by the complement (one minus) the average disability weight for each area.

Figure 20. Global disability-adjusted life expectancy (DALE), global healthy years lost due to disability (DLE) and global life expectancy (LE), by age, 1999

Years lost due to disability (DLE) Health expectancy (DALE)

90 90 Males Females 80 80

70 70

60 60

50 50

40 40

30 30 Expectation (years) Expectation (years) Expectation 20 20

10 10

0 0 0 1530456075 0 1530456075 Age (years) Age (years)

49 Figure 21. Estimated global survival curves for seven classes of disability severity, 1999. Disability severity classes and average weights as defined in Murray and Lopez (1996), average expected years lived in each disability class shown in brackets.

100

90 VII (1.2)

80 VI (3.3) V (2.1) 70 III (6.5) IV (3.4) 60

50 II (15.9)

40

Survivors (%) Survivors I (13.8) 30

20 No disability (18.1 years) 10

0 0 20406080100 Age (years)

4.4 DALE estimates and rankings for WHO Member States, 1999 Japan leads the world with an average healthy life expectancy of 74.5 years at birth in 1999 (Table 10). Female healthy life expectancy in Japan was 77.2 years for females and 71.9 years for males in 1999. After Japan, in second and third places, are Australia (73.2 years) and France (73.1 years), followed by a number of other industrialized countries of Western Europe. Note however, that there is a considerable range of uncertainty in the ranks for countries other than Japan, with an 80% uncertainty range of around 6 to 10 ranks for many countries. Canada is in twelfth place (72.0 years) with an uncertainty range of 8–14 in ranking and the USA in 24th place (70.0 years with a ranking range of 22–27). The bottom ten countries are all African countries with health expectancies in 1999 as low as 25.9 years in Sierra Leone.

Table 10. DALE at birth (years), top 10 and bottom 10 countries, 1999

Rank Top 10 countries Uncertainty in ranka Bottom 10 countries Uncertainty in ranka 1 Japan 74.5 1 1 Sierra Leone 25.9 191 2 Australia 73.2 2 - 10 2 Niger 29.1 189 - 190 3 France 73.1 2 - 7 3 Malawi 29.4 188 - 190 4 Sweden 73.0 2 - 8 4 Zambia 30.3 188 - 189 5 Spain 72.8 3 - 11 5 Botswana 32.3 185 - 187 6 Italy 72.7 3 - 10 6 Uganda 32.7 183 - 187 7 Greece 72.5 5 - 11 7 Rwanda 32.8 181 - 187 8 Switzerland 72.5 4 - 11 8 Zimbabwe 32.9 181 - 187 9 Monaco 72.4 2 - 15 9 Mali 33.1 181 - 186 10 Andorra 72.3 6 - 15 10 Ethiopia 33.5 180 - 185

(a) 10th percentile and 90th percentile of rank for DALE based on estimated uncertainty in DALE for countries.

50

Figure 22. Disability-adjusted life expectancy at birth, 1999

Healthy Life Expectancy Espérance de vie en santé Esperanza de vida saludable

High expectancy / Espérance élevée / Alta esperanza 71.0 - 74.6 67.0 - 70.9 64.0 - 66.9 62.0 - 63.9 60.0 - 61.9 55.0 - 59.9 45.0 - 54.9 35.0 - 44.9 25.9 - 34.9 Low expectancy / Espérance faible / Baja esperanza No Data / Pas de données / No hay datos Measure: Disability adjusted life expectancy at birth, both sexes, estimates for 1997 Mesure: Espérance de vie à la naissance corrigée de l'incapacité, population totale, estimations pour 1997 Medida: Esperanza de vida al nacer ajustada por incapacidad, ambos sexos, estimaciones para 1997

The boundaries and names shown and the designations used on this map do not imply the expression of any opinion whatsoever on the part of the World Health Organization concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. Dotted lines on maps represent approximate border lines for which there may not yet be full agreement. WHO 2000. All rights reserved

52 The worldwide pattern of health expectancies in 1999 is shown in Figure 22, highlighting the enormous variation between developing countries and developed countries, as well as between the lower and higher mortality regions of Europe. Japan leads the world with an average healthy life expectancy of 74.5 years at birth in 1999. Female healthy life expectancy in Japan was 77.2 years for females and 71.9 years for males in 1999. Total life expectancies were 77.6 and 84.3 for males and females. The high DALE reflects the very high life expectancy (low mortality) together with similar levels of disability to other low mortality countries. Australia ranks 2nd in the world with a DALE of 73.2. Note that the uncertainty ranges for DALE estimates are quite large: the 80% range for Australia is 72.7–74.1, which translates to an uncertainty range for its rank of 2–10. Uncertainty ranges for overall estimates of DALE at birth are plotted for the 191 WHO Member States in Figure 23 against mean DALE. Figure 24 shows a similar plot of the uncertainty in DALE ranks for Member States. Estimated healthy life expectancy for New Zealand is 69.2 years, 4 years lower than Australia. The life expectancy difference is lower at 2.8 years. New Zealanders lived longer than Australians until the 1970s, but during the 1980s New Zealanders fell behind Australians [90]. The mortality gap is compounded by a higher level of disability in New Zealand, reflecting higher rates of cardiovascular diseases, diabetes and injuries and the higher proportion of Indigenous people in the population. The USA ranks 24th with a DALE of 70.0, compared to Canada (12th with 72.0) and Cuba (33rd at 68.4). Other countries with reasonably high healthy life expectancies in the Americas include Dominica (69.8 years), Chile (68.6 years), Uruguay (67.0 years), and Argentina and Costa Rica at 66.7 years. Brazil is split, with a high healthy life expectancy in its southern half, and a lower one in the north. The total average is a relatively low 59.1 years, at 55.2 for males and 62.9 for females. China has a healthy life expectancy above the global average, at 62.3 years, 63.3 years for women and 61.2 for men. Other countries in the Asian region have lower DALE. Improving health in Viet Nam has resulted in a healthy life expectancy of 58.2 years, while Thailand has not improved significantly over the past decade, though it is still ahead of Viet Nam at 60.2 years. Healthy life expectancy in Myanmar is just 52 years, substantially behind its Southeast Asian neighbors. In Russia, healthy life expectancy is 66.4 for females, 3 years below the European average, but just 56.1 years for males, 7 years below the European average. This is one of the widest sex gaps in the world and reflects the sharp increase in adult male mortality in the early 1990s (see Section 4.5 below). Similar rates exist for other countries of the former Soviet Union. The bottom 10 countries for DALE are all in sub-Saharan Africa, where the HIV-AIDS epidemic is rampant (Table 10). Life expectancy in several countries in southern Africa has been reduced 15-20 years in comparison to life expectancy without HIV. Other African countries have lost 5-10 years of life expectancy because of HIV [84]. AIDS is now the leading cause of death in Sub-Saharan Africa, far surpassing the traditional deadly diseases of malaria, tuberculosis, pneumonia and diarrheal disease. AIDS killed 2.2 million Africans in 1999, versus 300,000 AIDS deaths 10 years previously. Figures 25 and 26 show worldwide patterns in DALE at birth for males and females respectively. Annex Table A provides details of the estimates for each Member State.

53 Figure 23.Uncertainty in DALE, versus mean DALE at birth, 191 Member States, 1999

Australia San M arino Finland Portugal Croatia Panama Bosnia and Azerbaijan Colombia Belarus Honduras Jordan Morocco Tuvalu Turkmenistan Democratic People's Comoros Cô te d'Ivo ire Swaziland Somalia Ethiopia Sierra Leone 20 30 40 50 60 70 80 80% uncertainty interval for DALE at birth (years)

Figure 24.Uncertainty in DALE ranking, versus mean rank, 191 Member States, 1999

Japan San M arino Malta New Zealand Armenia Republic of Korea Bahrain Paraguay China Russian Federation Jordan Brazil M arshall Islands M ongolia Yemen Senegal Chad Lesotho Liberia Sierra Leone 0 20406080100120140160180200 80% uncertainty interval for DALE rank

54 Figure 25. Disability-adjusted life expectancy at birth, males, 1999

(Years) 25.8 - 33.7 33.8 - 40.1 40.2 - 47.1 47.2 - 53.3 53.4 - 57.3 57.4 - 60.6 60.7 - 64.1 64.2 - 67.4 67.5 - 71.9 No Data

The boundaries and names shown and the designations used on this map do not imply the expression of any opinion whatsoever on the part of the World Health Organization concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. Dotted lines on maps represent approximate border lines for which there may not yet be full agreement. WHO 2000. All rights reserved

55 Figure 26. Disability-adjusted life expectancy at birth, females, 1999

(Years) 26.0 - 34.5 34.6 - 41.3 41.4 - 49.7 49.8 - 56.7 56.8 - 61.1 61.2 - 64.8 64.9 - 68.4 68.5 - 72.6 72.7 - 77.2 No Data

The boundaries and names shown and the designations used on this map do not imply the expression of any opinion whatsoever on the part of the World Health Organization concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. Dotted lines on maps represent approximate border lines for which there may not yet be full agreement. WHO 2000. All rights reserved

56 4.5 Expectation of years lived with disability There is a reasonably close correlation across countries between total life expectancy and DALE, both at birth (Figure 27) and at age 60 (Figure 28). The relationship between DALE and LE is very similar for males and females. However, when we examine the relationship between DLE (=LE – DALE) and LE (Figures 29 and 30), it is apparent that DLE at birth declines as LE increases. Taking into account the uncertainty intervals for both LE and DALE, these declines differ significantly from zero at the p=0.10 level for both males and females with regression slopes of –0.053 and –0.035 respectively. The difference between the male and female slopes is also statistically significant. The corresponding regression slopes at age 60 are –0.043 and +0.034 for males and females respectively. Only the female trend differs significantly from zero, but the male and female trends also differ significantly at age 60. This suggests that, for females but not males, there is some evidence of an expansion of morbidity at older ages, probably reflecting the greater life expectancies of older females than males together with the increasing prevalence of disability with age. At least cross-sectionally across countries in 1999, we estimate that there is an absolute compression of morbidity overall: DLE at birth decreases slightly in absolute terms (number of years) as life expectancy increases. This translates into a substantial decline in DLE as a proportion of total life expectancy as the latter increases (Figures 31 and 32). DLE represents the equivalent healthy years of life lost through living with disability resulting from diseases and injuries. Health expectancies are lower than life expectancies by amounts ranging from around 9 years in Africa to 6 years in countries such as Japan and Australia. These lost healthy years range from 20% of total life expectancy at birth in sub-Saharan Africa down to 10% for the low mortality countries in the Western Pacific region, primarily Japan, Australia, New Zealand and Singapore (Figures 33 and 34). Countries with longer life expectancy also have fewer lost years of healthy life due to disability. Higher levels of mortality are generally accompanied by more disability. Table 11 shows the top ten and bottom ten countries in the world in terms of DLE/LE%— the equivalent “lost” years of healthy life due to years lived with disability as a percent of total life expectancy at birth.

Table 11. DLE/LE% at birth, top 10 and bottom 10 countries, 1999

Rank Top 10 countries DLE/LE% Bottom 10 countries DLE/LE% 1 Greece 7.0 1 Niger 25 2 United Kingdom 7.1 2 Sierra Leone 25 3 Austria 7.4 3 Mali 22 4 Spain 7.5 4 Uganda 22 5 Italy 7.7 5 Malawi 22 6 Netherlands 7.7 6 Liberia 22 7 France 7.8 7 Rwanda 21 8 Japan 7.9 8 Zambia 21 9 Australia 8.0 9 Madagascar 21 10 Belgium 8.0 10 Burkina Faso 21

57 Figure 27. Disability-adjusted life expectancy (DALE) by total life expectancy at birth, by sex, 191 countries, 1999

85

Males 75 Females

65

55 DALE (years) DALE 45

35

25 30 35 40 45 50 55 60 65 70 75 80 85 Life expectancy (years)

Figure 28. Disability-adjusted life expectancy (DALE) by total life expectancy at age 60, by sex, 191 countries, 1999

25

Males 20 Females

15

DALE (years) 10

5

0 0 5 10 15 20 25 30 Life expectancy (years)

58 Figure 29. Expected years lost due to disability (DLE) by total life expectancy at birth, by sex, 191 countries, 1999

14

12

10

8

6 DLE at birth (years) birth at DLE

4 Males Females 2 Male Female 0 30 35 40 45 50 55 60 65 70 75 80 85 Life expectancy (years)

Figure 30. Expected years lost due to disability (DLE) by total life expectancy at age 60, by sex, 191 countries, 1999

10 Males 9 Females 8 Males Females 7

6

5

4 DLEat (years) age 60 3

2

1

0 0 5 10 15 20 25 30 Life expectancy (years)

59 Figure 31. Per cent of total life expectancy at birth lost due to disability, by sex, 191 countries, 1999

30

Males 25 Females

20

15

10 DLE/LE at birth (%) birth at DLE/LE

5

0 30 35 40 45 50 55 60 65 70 75 80 85 Life expectancy (years)

Figure 32. Per cent of total life expectancy at age 60 lost due to disability, by sex, 191 countries, 1999

60

55

50

45

40

35

30

25

20 DLE/LE at age 60 (%) DLE/LE 15

10 Males Females 5

0 0 5 10 15 20 25 30 Life expectancy (years)

60 Figure 33. Percentage of total life expectancy at birth lost due to disability, males, 1999

(Years) 6.7 - 7.9 8.0 - 9.0 9.1 - 10.4 10.5 - 11.6 11.7 - 13.5 13.6 - 15.4 15.5 - 17.4 17.5 - 19.2 19.3 - 24.3 No Data

The boundaries and names shown and the designations used on this map do not imply the expression of any opinion whatsoever on the part of the World Health Organization concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. Dotted lines on maps represent approximate border lines for which there may not yet be full agreement. WHO 2000. All rights reserved

61 Figure 34. Percentage of total life expectancy at birth lost due to disability, females, 1999

(Years) 7.4 - 8.7 8.8 - 9.9 10.0 - 11.0 11.1 - 12.5 12.6 - 14.2 14.3 - 16.2 16.3 - 18.8 18.9 - 22.0 22.1 - 26.7 No Data

The boundaries and names shown and the designations used on this map do not imply the expression of any opinion whatsoever on the part of the World Health Organization concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. Dotted lines on maps represent approximate border lines for which there may not yet be full agreement. WHO 2000. All rights reserved

62 4.5 Male-female differences in healthy life expectancy Health expectancies at birth are higher for females than males in most regions in the world (Figure 35). In 1999, female-male difference was highest at around 10 years or even higher in many countries of the former Soviet Union (including the Russian Federation, Latvia, Estonia, Belarus and Kazakhstan) and lowest (males 1 to 2 years higher than females) in countries in the Eastern Mediterranean region (including Turkey, Algeria, Iran and Saudi Arabia). For other countries of the world, the female-male difference in healthy life expectancy generally increases as average life expectancy increases (Figures 36).

Figure 35. Female - male difference in DALE Figure 36. Female - male difference in DALE at birth, versus total life expectancy, at birth, versus total life expectancy, mortality sub-regions, 1999 191 countries, 1999

15 15

EUR C 10 10

AMR B

5 5

0 0 EMR B Female - male difference in DALE at birth (years) Female - male difference in DALE at birth (years) birth at DALE in difference - male Female -5 -5 30 35 40 45 50 55 60 65 70 75 80 85 30 35 40 45 50 55 60 65 70 75 80 85 Life expectancy (years) Life expectancy (years)

Russia has one of the widest sex gaps in healthy life expectancy in the world: 66.4 years for females at birth but just 56.1 years for males (Figure 37). The most common explanation is the high incidence of male alcohol abuse, which led to high rates of accidents, violence and cardiovascular disease. From 1987 to 1994, the risk of premature death increased by 70% for Russian males. Since 1994, life expectancy has been improving for males. Similar rates exist for other major countries of the former Soviet Union. In Ukraine, female babies can expect to live an equivalent of 67.5 years of healthy life versus 58.5 years for male babies. In Belarus, 67.2 for female babies and 56.2 for male babies. There is generally a close correlation between the female-male difference in DALE at birth and the female-male difference in total life expectancy at birth (Figure 38). There are a number of countries with higher female than male life expectancies but higher male than female DALE. These fall in the bottom right quadrant in Figure 38 and include mainly countries in the Eastern Mediterranean region and North Africa.

63

64 Figure 37. Sex difference in disability adjusted life expectancy at birth, 1999

Years -2.1 - -0.9 -0.8 - 0.3 0.4 - 1.3 1.4 - 2.5 2.6 - 3.7 3.8 - 5.0 5.1 - 6.4 6.5 - 7.7 7.8 - 11 No Data

The boundaries and names shown and the designations used on this map do not imply the expression of any opinion whatsoever on the part of the World Health Organization concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. Dotted lines on maps represent approximate border lines for which there may not yet be full agreement. WHO 2000. All rights reserved

65 In North Africa and the Middle East, males and females have similar levels of healthy life expectancy, which is unusual. Also, the position of women in these societies is often not good, Less care is given to female children, and they have a higher risk for reproductive deaths than in other countries. In Saudi Arabia, the overall healthy life expectancy is 64.5 years -- 65.1 for male babies and 64.0 for female babies. In Bahrain, the overall healthy life expectancy is 64.4, but 63.9 for male babies and 64.9 for female babies; Qatar, 63.5 overall, and 64.2 for male babies, 62.8 for females; and Kuwait, 63.2 overall, with 63.0 for male babies and 63.4 for female babies.

Figure 38. Female - male difference in DALE at birth, versus female - male difference in DALE at birth, 191 countries, 1999

15

10

5

0 Female- difference male in DALEat birth(years) -5 -5 0 5 10 15 Female-male differences in life expectancy (years)

66 5. Discussion and conclusions This paper has described the methods used to produce first estimates of healthy life expectancy (DALE) for 191 countries in 1999. These estimates are based on a very substantial information and analytic base for mortality rates and life expectancies, cause of death distributions, internally consistent estimates of the incidence, prevalence and disability distributions for 109 disease and injury causes by age group, sex and region of the world, and an analysis of 60 representative health surveys across the world. As with any innovative approach, methods and data sources can and will be refined and improved. Careful scrutiny and use of the DALE results reported here will lead to progressively better estimates. A wide range of people from WHO programs, from countries and other agencies have already been involved and consulted in the development of these initial estimates of DALE for WHO member countries. It is anticipated that continuing and increasing collaboration and consultation will lead to progressive refinement and improvement of the initial estimates presented here. Over the next few years, WHO is planning a range of methodological and data collection exercises to improve and extend the information base underpinning the estimates of healthy life expectancies for Member States. These include: C Improved methods for calculating life tables for WHO member states and for estimating cause of death distributions, C Complete revision of the Global Burden of Disease Study for 14 regions of the world for the year 2000. To the extent possible, this will also provide country-level estimates of the incidence, prevalence and severity of specific health conditions causing significant levels of disability. It will involve widespread consultation with countries and with epidemiologistsand other disease experts. C Improving the comparability and reliability of health survey data. WHO is developing a standardized description of health states for use in population surveys and health state valuation. The resulting instrument is being piloted in 10 countries over the next year and will also be validated against objective measurements. The objective is to develop an instrument with the maximum cross-cultural validity, and usability by younger and older adults with widely varying education levels and cultural backgrounds, and to understand better the systematic determinants of differences between self-report and underlying true health. Another objective for the ongoing WHO population survey work is to facilitate reliable and valid measurements of valuations of time spent in health states in populations across the world. The aim is to obtain large scale empirical assessment in many different countries to inform health state valuations for the calculation of DALES for Member States. Future burden of disease analyses should ensure to the extent possible, that there is consistency between the condition-specific estimates for disabling sequelae and measured population levels of these impairments and disabilities. Given the limitations of self-report general disability data, both in terms of underreporting of certain types and causes of disability, and in terms of the problems of ensuring comparability of self-report data across populations, it is likely that the combination of condition-specific analyses with population- level data for specific disabilities and impairments will continue to be the analytic approach of choice for some considerable time. The use of population-level impairment and disability data

67 will also assist in developing improved methods for addressing comorbidity. This approach is also required for causal attribution, where it will be critical to link diseases, injuries and risk factors to one or several average health states. The new WHO framework for performance assessment also measures health inequality across individuals. Eventually it will use inequalities in DALE. For the World Health Report 2000, inequality has been measured by looking at differences in child survival, because that data is available now. As adult data are added, it is planned to estimate inequalities in DALE for Member States. In conclusion, we believe that healthy life expectancy (DALE) is the most appropriate single summary measure of population health for the measurement of level of health for Member States. Unlike other forms of health expectancy, it takes into account differences in the severity distribution of health states between populations. As WHO is concerned not just to increase the numbers of people in full health, but also to reduce the severity of disability for people in less than full health, DALE is the most appropriate summary measure for the comparison of level of health of populations. Improvements in the estimation of DALE for Member States will require further collaborative efforts between WHO and Member States to improve population health survey data, to improve the completeness of death registration systems, and to improve the estimation of burden of diseases and injuries at country level.

68 References

1. World Health Organization (2000). World Health Report 2000. Geneva, World Health Organization. 2. Sadana R. Mathers C, Lopez AD, Murray CJL (2000). Comparative analysis of more than 50 household surveys on health status. Geneva, World Health Organization, 2000 (GPE Discussion Paper No. 15). 3. Murray CJL (1996). Rethinking DALYs. In: Murray CJL. and Lopez A. eds. The Global Burden of Disease. Cambridge, Harvard University Press, 1996, chapter 1. 4. World Bank (1993). World Development Report 1993. Investing in health. New York: Oxford University Press for the World Bank. 5. World Health Organization (1996). Investing in health research and development. Report of the Ad Hoc Committee on Health Research Relating to Future Intervention Options. Geneva: World Health Organization, 1996. 6. World Health Organization (1999). World Health Report 1999. Geneva, World Health Organization. 7. Murray CJL, Lopez, AD, eds (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 (Global Burden of disease and Injury Series, Vol. 1). 8. Van de Water HP, Perenboom RJ, Boshuizen HC (1996). Policy relevance of the health expectancy indicator: an inventory of European Union countries. Health Policy 36(2):117-29 9. Barendregt JJ, Bonneux L, van der Maas, PJ (1998). Health expectancy: from a population health indicator to a tool for policy making. Journal of Aging & Health 10(2):242-258 10. Field MJ, Gold GM, eds. (1998). Summarizing Population Health: Directions for the Development and Application of Population Metrics. Institute of Medicine, Washington, D.C. National Academy Press. 11. Sanders BS (1964). Measuring community health levels. American Journal of Public Health, 54: 1063-1070. 12. Sullivan DF (1966). Conceptual problems in developing an index of health. US public Health Service Publication Series No. 1000. Vital and Health Statistics Series 2. No. 17. National Center for Health Statistics. 13. Sullivan DF (1971). A single index of mortality and morbidity. HSMHA Health Reports, 86: 347-354. 14. Piot M, Sundaresan TK (1967). A linear programme decision model for tuberculosis control. Progress report on the first test-runs. Geneva, World health Organization, 1967, WHO/TB/Tech.Information/67.55. 15. Fanshel S, Bush JW (1970). A health-status index and its application to health services outcomes. Operations research, 18(6):1021-1066. 16. Ghana Health Assessment Project Team (1981). A quantitative method of assessing the health impact of different diseases in less developed countries. International journal of epidemiology, 10(1):72-80. 17. Preston SH (1993). Health indices as a guide to health sector planning: a demographic critique. In Gribble JN, Preston SH, eds. The epidemiological transition. Policy and planning implications for developing countries. Washington D.C., National Academic Press, 1993.

69 18. Bone MR (1992). International efforts to measure health expectancy. Journal of Epidemiology and Community Health, vol.46, pp.555-558. 19. Mathers CD, Robine JM (1993). Health expectancy indicators: a review of the work of REVES to date.. In: J.M. Robine, C.D. Mathers, M.R. Bone and I. Romieu (eds), Calculation of health expectancies, harmonization, consensus achieved and future perspectives. Colloque INSERM Vol. 226, John Libbey Eurotext and Les Editions INSERM, France. 20. Robine JM, Blanchet M, Dowd EJ (eds.) (1992). Health Expectancy. OPCS, HMSO, London. 21. Robine JM, Mathers CD, Bone MR & Romieu I (eds) (1994). Calculation of health expectancies, harmonization, consensus achieved and future perspectives. Colloque INSERM Vol. 226, John Libbey Eurotext and Les Editions INSERM, France. 22. Mathers CD, McCallum J Robine JM (eds) (1994). Advances in health expectancies: proceedings of the 7th meeting of the international network on health expectancy (REVES), Australian Institute of Health and Welfare, Canberra, 1994. 23. Robine JM, Mathers C, Brouard N (1996). Trends and differentials in disability-free life expectancy: concepts, methods and findings. In: Caselli G, Lopez A (eds). Health and Mortality Among Elderly Populations. Clarendon Press, Oxford, pp182-201. 24. Crimmins, EM., Saito, Y, Ingengneri, D (1997). Trends in disability-free life expectancy in the United States, 1970-90. Population and development review, 23(3): 555-572. 25. Robine JM, Jagger C (1999). Developing Consistent Disability Measures to Address Public Policy Needs for Older Populations. Background Paper prepared for OECD Meeting on Implications of Disability for Ageing Populations: Monitoring Social Policy Challenges. Paris, 9-10 December 1999. 26. Saito Y, Crimmins EM, Hayward MD (1999). Health expectancy: an overview. NUPRI Research paper Series No. 67, Nihon University Population research Institute, Tokyo. 27. OECD (1993). OECD Health Systems: Facts and Trends 1960-1991. OECD, Paris. 28. OECD (1999). Eco-santé. OECD Health Database 1999. OECD, Paris. 29. Romieu I and Robine JM (1994). World atlas of health expectancy calculations. In Mathers C, McCallum J, Robine JM (eds.) Advances in health expectancies. Australian Institute of Health and Welfare, Canberra. 30. Wilkins R, Adams O (1983). Health expectancy in Canada, late 1970’s: demographic, regional and social dimensions. AJPH 73:1073-1080 31. Murray CJL, and Lopez A (1996b). Global Health Statistics. Cambridge, Harvard University Press (Global Burden of disease and Injury Series, Vol. 2). 32. Murray CJL, and Lopez A (1996c). Evidence-based health policy—lessons from the Global Burden of Disease Study. Science,, 274:740-743. 33. Murray CJL, and Lopez A (1997a). Mortality by cause for eight regions of the world: Global Burden of Disease Study. Lancet, 349: 1269-1276. 34. Murray CJL, and Lopez A (1997b). Regional patterns of disability-free life expectancy and disability-adjusted life expectancy: Global Burden of Disease Study. Lancet, 349: 1347-1352. 35. Murray CJL, and Lopez A (1997c). Global mortality, disability, and the contribution of risk factors: Global Burden of Disease Study. Lancet, 349: 1436-1442. 36. Murray CJL, and Lopez A (1997d). Alternative projections of mortality and disability by cause 1990-2020: Global Burden of Disease Study. Lancet, 349: 1498-1504. 37. Lozano R, Frenk J, Gonzalez MA (1994). El peso de la enfermedad en adultos mayores, Mexico 1994. Salud Publica de Mexico, 38: 419-429. 38. Lozano R, Murray CJL, Frenk J, and Bobadilla J (1995). Burden of Disease Assessment and Health System Reform: Results of a study in Mexico. Journal for International Development, 7(3): 555-564.

70 39. Fundación Mexicana Para La Salud (1995). Health and the Economy: Proposals for Progress in the Mexican Health System. Mexico, Salud. 40. República de Colombia Ministerio de Salud (1994). La Carga de la Enfermedad en Columbia. Bogota, Ministerio de Salud, 1994. 41. Ruwaard D. and Kramers PGN (1998). Public health status and forecasts. The Hague: National Institute of Public Health and Environmental Protection, 1998. 42. Bowie C, Beck S, Bevan G, Raftery J, Silvertion F, and Stevens J (1997). Estimating the burden of disease in an English region. Journal of Public Health Medicine, 19:87-92. 43. Concha M, Aguilera, X, Albala C, et al (1996). Estudio ecarga de enfermedad informe final. Estudio Prioridades de Inversio en salud Minsterio de Salud. 44. Murray CJL. Mahapatra P. Ashley R. Michaud C. George A. Horbon P. Akhavan D. et al. (1997). The Health Sector in Mauritius: Resource Use, Intervention Cost and Options for Efficiency Enhancement. Cambridge, Harvard Center for Population and Development Studies. 45. Murray CJL. Michaud CM. McKenna MT. and Marks JS (1998). U. S. Patterns of Mortality by County and Race: 1965-1994. Cambridge: Harvard Center for Population and Development Studies and Centers for Disease Control, 1998.; 46. Mathers C, Vos T, Stevenson C (1999). The burden of disease and injury in Australia. Australian Institute of Health and Welfare, Canberra: AIHW. 47. Murray CJL, Salomon J, Mathers CD, Lopez A, Lozano R (2000). Summary Measures of Population Health. Geneva, World Health Organization (in preparation). 48. Murray CJL, Salomon J, Mathers C (1999). A critical review of summary measures of population health. Geneva, World Health Organization, 2000 (GPE Discussion Paper No. 2). 49. Murray CJL, Lopez AD (2000). Progress and directions in refining the global burden of disease approach: response to Williams. Health EconomicsDem 9: 69-82. 50. Dempsey M (1947). Decline in tuberculosis: the death rate fails to tell the entire story. American review of tuberculosis, 56: 157-164. 51. Romeder JM. and McWhinnie JR (1977). Potential years of life lost between ages 1 and 70: an indicator of premature mortality for health planning. International Journal of Epidemiology 6:143-151. 52. Hyder AA, Rotllanat G, Morrow RH (1998). Measuring the Burden of Disease: Healthy Life- Years. American Journal of Public Health, 88(2):196-202. 53. Wilkins R, Adams OB (1992). Quality-adjusted life expectancy: weighting of expected years in each state of health. In Robine J-M, Blanchet M, Dowd JE (eds) Health expectancy, OPCS studies on medical and population subjects no. 54, HMSO, London, 1992. 54. Wilkins R, Chen J, Ng E (1994). Changes in health expectancy in Canada from 1986 to 1991. In: C.D. Mathers, J. McCallum, J.M. Robine (eds), Advances in health expectancies: proceedings of the 7th meeting of the international network on health expectancy (REVES), Australian Institute of Health and Welfare, Canberra 1994. 55. Mathers CD (1999a). Gains in health expectancy from the elimination of diseases among older people. Disability and Rehabilitation 21(5-6): 211-221. 56. Wolfson MC (1996). Health-adjusted life expectancy. Health Reports 8(1): 41-45. 57. Stouthard M, Essink-Bot M, Bonsel G, Barendregt J & Kramers P (1997). Disability weights for diseases in the Netherlands. Rotterdam: Department of Public Health, Erasmus University. 58. Mathers CD, Robine JM, Wilkins R (1994). Health expectancy indicators: recommendations for terminology.. In: C.D. Mathers, J. McCallum, J.M. Robine (eds), Advances in health expectancies: proceedings of the 7th meeting of the international network on health expectancy (REVES), Australian Institute of Health and Welfare, Canberra 1994.

71 59. Mathers CD (1999b). Health expectancies: an overview and critical appraisal. Invited paper. Global Conference on Summary Measures of Population Health, World Health Organisation, Marrakech, 6-9 December, 1999. 60. Robine JM (1994). Disability-free life expectancy trends in France, international comparison. In Mathers C, McCallum J, Robine JM (eds.) Advances in health expectancies. Australian Institute of Health and Welfare, Canberra. 61. Mathers CD (1996). Trends In Health Expectancies In Australia 1981-1993. Journal of the Australian Population Association 13(1): 1-16. 62. Wolfson MC (1998). Measuring health—visions and practicalities. Joint ECE-WHO Meeting on Health Statistics. Rome. 63. Murray CJL, Frenk J (1999). A WHO Framework for Health System Performance Assessment. Geneva, World Health Organization, 2000 (GPE Discussion Paper No.6). 64. Gakidou EE, Murray CJL, Frenk J (1999). A framework for measuring health inequality. Discussion Paper. Global Program on Evidence for Health Policy, WHO, Geneva. 65. Gakidou EE, King G (2000). Using an extended beta-binomial model to estimate the distribution of child mortality risk. Geneva, World Health Organization, 2000 (GPE Discussion Paper No. 18). 66. Gakidou EE. Murray CJL (2000). Estimates of the distribution of child survival in 191 countries. Geneva, World Health Organization, 2000 (GPE Discussion Paper No. 19). 67. McDowell I, Newell C (1996). Measuring health. A guide to rating scales and . Second Edition. Oxford: Oxford University Press. 68. Krabbe PFM, Essink-Bot M, Bonsel GJ (1996). The comparability and reliability of five health- state valuation methods. Social Science and Medicine 45:1641-1652. 69. Brazier J, Usherwood T, Harper R,Thomas K (1998). Deriving a preference-based single index from the UK SF-36 Health Survey. Journal of Clinical Epidemiology 51(11): 1115-1128. 70. Nord E (1992). Methods for quality adjustment of life years, Social Science and Medicine 1992; 34: 559-569. 71. World Health Organization (1980). International Classification of Impairments, Disabilities, and Handicaps. Geneva, World Health Organization, (Reprint 1993). 72. World Health Organization (1999). International Classification of Functioning and Disability - Beta-2 Draft, Short Version, WHO, Geneva, July 1999. 73. Sadana R 2000. Measuring Reproductive Health: Conceptual and Operational Approaches, Bulletin of the World Health Organization (forthcoming). 74. Van Agt HME. Essinck-Bot ML. Krabbe PFM. et al. (1994). Test-retest reliability of health state valuations collected with the EuroQol questionnaire. Social Science and Medicine 39:1537-1544. 75. Gold MR, Siegel JE, Weinstein MC, & Russell LB (1996). Cost-effectiveness in health and medicine. NY: Oxford University Press. 76. Ubel PA, Richardson J & Menzel P (1999). Societal value, the person trade-off, and the dilemma of whose values to measure for cost-effectiveness analysis. Health Economics (in press). 77. Ustun TB. Rehm J. Chatterji, S. Trotter R. Room R. Bickenbach J (1999). Are disability valuations universal? Multiple-informant ranking of the disabling effects of different health conditions in 14 countries, Lancet (in submission). 78. Sadana R 1998. A closer look at the WHO/World Bank Global Burden of Disease study's methodologies: how do poor women's values in a developing country compare with international public health experts? Paper presented at Eighth Annual Public Health Forum: Reforming Health Sectors, London School of Hygiene & Tropical Medicine, 21-24 April.

72 79. Mahapatra P (2000). Ph.D. thesis. Harvard School of Public Health. Cambridge, Massachusetts (in preparation). 80. Australian Institute of Health and Welfare (AIHW) (1999). Australia’s Welfare 1999. Canberra: AIHW. 81. Eurostat (1999). European Community Household Panel Longitudinal Users’ Database. Waves 1, 2 and 3.Manual. European Commission. 82. Lopez AD, Salomon J. Ahmad O, Murray CJL (2000). Life tables for 191 countries: data, methods and results. Geneva, World Health Organization, 2000 (GPE discussion Paper No. 9). 83. Murray CJL, Lopez AD, Ahmad O, Salomon J (2000). WHO system of model life tables. Geneva. World Health Organization, 2000 (GPE Discussion Paper No. 8). 84. Salomon J. Gakidou EE, Murray CJL (2000). Methods for modelling the HIV/AIDS epidemic in sub-Saharan Africa. Geneva, World Health Organization, 2000 (GPE discussion Paper No. 3). 85. Salomon J. Murray CJL (2000). Compositional models for mortality by age, sex and cause. Geneva, World Health Organization, 2000 (GPE Discussion Paper No. 11). 86. Lozano R, Murray CJL, Lopez AD, Satoh T (2000). Miscoding and misclassification of ischaemic heart disease mortality. Geneva, World Health Organization, 2000 (GPE Discussion Paper No. 12). 87. Boschi-Pinto C. Murray CJL, Lopez AD, Lozano R (2000). Cancer survival by site for 14 regions of the world. Geneva, World Health Organization, 2000 (GPE Discussion Paper No. 13). 88. Rasmussen N, Gudex C and Christensen S (1999). Survey Data on Disability, Eurostat Working Papers, Population and Social Conditions 3/1999/E/n 20, European Commission, Luxembourg 89. Salomon J. Murray CJL (2000). Methods for life expectancy and disability-adjusted life expectancy uncertainty analysis. Geneva, World Health Organization, 2000 (GPE Discussion Paper No. 10). 90. Woodward A, Mathers C, Tobias M (2000). The health status of New Zealanders and Australians: different politics, different health? Paper prepared for the Health Inequalities Research Consortium Conference “Social origins of health and well-being: from the planetary to the molecular.” Canberra, July 2000.

73 ANNEX TABLE

74 Annex Table A. Disability-adjusted life expectancy (DALE) and life expectancy (LE) at birth and age 60, by sex, WHO Member States, 1999

Expected years Per cent of total Disability-adjusted life expectancy (DALE) in years Life expectancy lost to disability life expectancy at birth (years) at birth (DLEb) lost to disability Total Males Females Uncertainty population At Uncertainty At age Uncertainty Uncertainty At age Uncertainty a a a a Rank rangea Member State at birth birth range 60 range At birth range 60 range Males Females Males Females Males Females 1 1 Japan 74.5 71.9 71.6 - 72.3 17.5 17.3 - 18.1 77.2 76.9 - 78.0 21.6 21.3 - 22.4 77.6 84.3 5.7 7.1 7.3 8.4 2 2 - 10 Australia 73.2 70.8 70.5 - 71.3 16.8 16.6 - 17.3 75.5 75.2 - 76.2 20.2 19.9 - 20.9 76.8 82.2 6.0 6.7 7.8 8.1 3 2 - 7 France 73.1 69.3 69.0 - 69.7 16.8 16.5 - 17.4 76.9 76.5 - 77.8 21.7 21.4 - 22.7 74.9 83.6 5.6 6.7 7.5 8.0 4 2 - 8 Sweden 73.0 71.2 70.9 - 71.8 16.8 16.5 - 17.3 74.9 74.4 - 75.7 19.6 19.4 - 20.5 77.1 81.9 5.9 7.0 7.7 8.5 5 3 - 11 Spain 72.8 69.8 69.1 - 70.6 16.8 16.4 - 17.6 75.7 75.3 - 76.6 20.1 19.8 - 21.0 75.3 82.1 5.5 6.4 7.3 7.7 6 3 - 10 Italy 72.7 70.0 69.7 - 70.5 16.2 16.0 - 16.8 75.4 75.0 - 76.2 19.9 19.6 - 20.7 75.4 82.1 5.4 6.7 7.1 8.2 7 5 - 11 Greece 72.5 70.5 70.2 - 70.9 16.9 16.6 - 17.3 74.6 74.2 - 75.2 18.8 18.6 - 19.5 75.5 80.5 5.0 5.9 6.7 7.4 8 4 - 11 Switzerland 72.5 69.5 69.0 - 70.2 16.0 15.7 - 16.7 75.5 75.0 - 76.5 20.6 20.3 - 21.6 75.6 83.0 6.1 7.5 8.1 9.1 9 2 - 15 Monaco 72.4 68.5 67.5 - 69.6 16.4 15.9 - 17.2 76.3 75.6 - 77.3 21.5 21.1 - 22.5 74.7 83.6 6.2 7.3 8.3 8.7 10 6 - 15 Andorra 72.3 69.3 68.6 - 70.2 16.3 15.9 - 17.0 75.2 74.6 - 76.2 20.0 19.6 - 20.9 75.4 82.2 6.1 7.0 8.0 8.5 11 5 - 17 San Marino 72.3 69.5 68.6 - 70.5 15.7 15.3 - 16.5 75.0 74.4 - 76.0 19.6 19.2 - 20.5 75.3 82.0 5.8 7.0 7.7 8.6 12 8 - 14 Canada 72.0 70.0 69.7 - 70.5 16.0 15.8 - 16.6 74.0 73.6 - 74.9 18.9 18.6 - 19.8 76.2 81.9 6.2 7.8 8.1 9.6 13 8 - 15 Netherlands 72.0 69.6 69.3 - 70.1 15.4 15.3 - 16.0 74.4 74.0 - 75.3 19.7 19.4 - 20.6 75.0 81.1 5.4 6.7 7.2 8.2 14 11 - 17 United Kingdom 71.7 69.7 69.4 - 70.1 15.7 15.5 - 16.2 73.7 73.5 - 74.4 18.6 18.3 - 19.2 74.7 79.7 5.0 6.0 6.7 7.5 15 11 - 18 Norway 71.7 68.8 68.5 - 69.3 15.1 15.0 - 15.7 74.6 74.2 - 75.3 19.7 19.4 - 20.6 75.1 82.1 6.3 7.6 8.4 9.2 16 11 - 17 Belgium 71.6 68.7 68.4 - 69.2 15.8 15.6 - 16.4 74.6 74.2 - 75.3 19.6 19.3 - 20.4 74.5 81.3 5.8 6.7 7.8 8.2 17 12 - 18 Austria 71.6 68.8 68.4 - 69.4 15.2 15.0 - 15.8 74.4 74.1 - 75.1 18.7 18.4 - 19.4 74.4 80.4 5.6 6.0 7.5 7.4 18 15 - 20 Luxembourg 71.1 68.0 67.6 - 68.7 15.8 15.2 - 16.8 74.2 73.7 - 75.2 19.7 19.0 - 21.0 74.5 81.4 6.5 7.2 8.7 8.8 19 18 - 23 Iceland 70.8 69.2 68.6 - 70.1 14.9 14.2 - 15.9 72.3 71.7 - 73.4 17.0 16.4 - 18.3 76.1 80.4 6.8 8.1 9.0 10.0 20 19 - 24 Finland 70.5 67.2 66.9 - 67.7 14.5 14.2 - 15.0 73.7 73.4 - 74.4 18.5 18.3 - 19.3 73.4 80.7 6.2 7.0 8.4 8.6 21 18 - 24 Malta 70.5 68.4 67.9 - 69.2 14.8 14.5 - 15.6 72.5 72.0 - 73.4 17.3 17.0 - 18.2 75.7 80.8 7.3 8.3 9.6 10.3 22 20 - 24 Germany 70.4 67.4 67.1 - 67.9 14.3 14.1 - 14.9 73.5 73.2 - 74.1 18.5 18.2 - 19.1 73.7 80.1 6.3 6.6 8.6 8.3 23 19 - 25 Israel 70.4 69.2 68.9 - 69.7 15.6 15.3 - 16.3 71.6 71.2 - 72.4 16.9 16.7 - 17.8 76.2 79.9 7.1 8.3 9.3 10.4 24 22 - 27 United States of America 70.0 67.5 67.0 - 68.1 15.0 14.7 - 15.7 72.6 72.2 - 73.3 18.4 18.1 - 19.2 73.8 79.6 6.3 7.0 8.68.8 25 23 - 28 Cyprus 69.8 68.7 68.2 - 69.4 15.9 15.6 - 16.6 70.9 70.4 - 71.7 17.3 17.0 - 18.1 74.8 78.8 6.1 7.9 8.2 10.0 26 20 - 33 Dominica 69.8 67.2 66.2 - 68.2 15.0 14.3 - 15.6 72.3 71.0 - 73.4 17.9 17.2 - 18.7 74.1 80.3 6.8 8.0 9.2 10.0 27 25 - 30 Ireland 69.6 67.5 67.0 - 68.2 13.9 13.6 - 14.6 71.7 71.2 - 72.5 16.6 16.3 - 17.4 73.3 78.3 5.8 6.6 8.0 8.4 28 26 - 31 Denmark 69.4 67.2 66.8 - 67.9 14.2 13.9 - 14.8 71.5 71.2 - 72.2 17.2 16.9 - 18.0 72.9 78.1 5.7 6.6 7.9 8.4 29 26 - 31 Portugal 69.3 65.9 65.6 - 66.6 14.0 13.7 - 14.6 72.7 72.4 - 73.4 17.7 17.3 - 18.5 72.0 79.5 6.1 6.8 8.4 8.6 30 27 - 32 Singapore 69.3 67.4 66.9 - 68.2 14.4 14.1 - 15.2 71.2 70.7 - 72.2 16.8 16.5 - 17.8 75.1 80.8 7.7 9.6 10.2 11.8 31 26 - 31 New Zealand 69.2 67.1 66.8 - 67.6 14.4 14.1 - 15.0 71.2 70.8 - 72.0 17.0 16.8 - 17.9 74.0 79.4 6.8 8.1 9.2 10.2 32 30 - 34 Chile 68.6 66.0 65.2 - 67.0 14.3 13.6 - 15.3 71.3 70.9 - 72.2 17.8 17.3 - 18.8 73.4 79.9 7.4 8.6 10.1 10.8 33 31 - 36 Cuba 68.4 67.4 66.8 - 68.1 15.4 14.9 - 16.1 69.4 68.9 - 70.3 16.1 15.8 - 16.9 73.5 77.4 6.2 8.0 8.4 10.3 34 32 - 34 Slovenia 68.4 64.9 64.6 - 65.4 12.7 12.6 - 13.4 71.9 71.5 - 72.6 16.8 16.5 - 17.6 71.6 79.5 6.7 7.6 9.4 9.6 35 33 - 36 Czech Republic 68.0 65.2 64.9 - 65.7 12.7 12.6 - 13.2 70.8 70.5 - 71.5 16.4 16.2 - 17.1 71.5 78.3 6.3 7.5 8.8 9.5 36 34 - 45 Jamaica 67.3 66.8 65.5 - 68.0 18.9 18.1 - 19.7 67.9 66.5 - 69.3 18.2 17.3 - 19.1 75.2 77.4 8.4 9.5 11.2 12.3 37 36 - 46 Uruguay 67.0 64.1 63.1 - 65.0 15.3 14.8 - 15.8 69.9 68.8 - 71.0 18.3 17.6 - 19.0 69.4 77.2 6.4 7.9 9.1 10.2 38 36 - 42 Croatia 67.0 63.3 63.1 - 63.8 11.4 11.3 - 11.9 70.6 70.3 - 71.3 16.0 15.8 - 16.7 70.5 77.8 6.0 6.6 8.7 8.6 39 38 - 44 Argentina 66.7 63.8 63.5 - 64.3 14.7 14.4 - 15.3 69.6 69.2 - 70.3 18.1 17.8 - 19.0 70.6 77.8 6.8 8.2 9.6 10.6 40 37 - 45 Costa Rica 66.7 65.2 64.6 - 66.0 14.2 13.9 - 15.0 68.1 67.5 - 69.1 16.6 16.2 - 17.6 72.3 77.1 9.0 10.8 12.1 13.7

75 Annex Table A (continued). Disability-adjusted life expectancy (DALE) and life expectancy (LE) at birth and age 60, by sex, WHO Member States, 1999

Expected years Per cent of total Disability-adjusted life expectancy (DALE) in years Life expectancy lost to disability life expectancy at birth (years) at birth (DLE) lost to disability Total Males Females Uncertainty population At Uncertainty At age Uncertainty Uncertainty At age Uncertainty Rank range Member State at birth birth range 60 range At birth range 60 range Males Females Males Females Males Females 41 36 - 46 Armenia 66.7 65.0 64.4 - 65.9 14.5 14.2 - 15.5 68.3 67.6 - 69.3 15.5 15.1 - 16.5 74.2 78.9 7.3 8.8 10.1 11.4 42 39 - 45 Slovakia 66.6 63.5 63.2 - 64.0 12.7 12.6 - 13.1 69.7 69.4 - 70.3 16.0 15.9 - 16.7 68.9 76.7 5.4 7.0 7.8 9.1 43 36 - 54 Saint Vincent and the Grenadines 66.4 65.0 63.8 - 66.2 15.9 15.2 - 16.7 67.8 66.4 - 69.0 16.7 15.9 - 17.5 71.8 75.2 6.8 7.4 9.5 9.8 44 39 - 50 Georgia 66.3 63.1 62.2 - 64.0 13.8 13.5 - 14.6 69.4 68.7 - 70.3 16.6 16.2 - 17.4 69.4 76.7 6.3 7.3 9.1 9.5 45 41 - 48 Poland 66.2 62.3 61.6 - 63.0 12.5 12.1 - 13.1 70.1 69.7 - 70.7 16.6 16.4 - 17.3 67.9 76.6 5.6 6.5 8.2 8.5 46 39 - 51 Yugoslavia 66.1 64.2 63.1 - 65.3 15.1 14.4 - 15.7 68.1 66.9 - 69.2 17.5 16.8 - 18.1 71.8 76.4 7.6 8.2 10.6 10.8 47 40 - 53 Panama 66.0 64.9 63.6 - 66.1 17.3 16.4 - 18.1 67.2 65.9 - 68.5 17.4 16.5 - 18.3 72.6 75.8 7.8 8.6 10.7 11.4 48 39 - 61 Antigua and Barbuda 65.8 63.4 62.0 - 64.6 14.4 13.7 - 15.2 68.3 66.9 - 69.6 16.8 15.9 - 17.6 71.4 76.8 8.0 8.5 11.2 11.1 49 40 - 63 Grenada 65.5 62.4 61.1 - 63.6 14.1 13.5 - 14.8 68.5 67.2 - 69.7 16.9 16.2 - 17.7 69.1 75.9 6.7 7.4 9.7 9.7 50 46 - 58 United Arab Emirates 65.4 65.0 64.0 - 65.9 11.7 10.9 - 12.5 65.8 64.6 - 67.0 12.6 11.8 - 13.5 72.2 75.6 7.3 9.8 10.0 13.0 51 46 - 58 Republic of Korea 65.0 62.3 61.6 - 63.1 12.1 11.6 - 12.7 67.7 66.7 - 68.7 15.2 14.5 - 15.8 68.8 76.0 6.4 8.3 9.3 10.9 52 49 - 57 Venezuela, Bolivarian Rep. of 65.0 62.9 62.4 - 63.6 13.4 13.1 - 14.2 67.1 66.5 - 68.1 15.7 15.2 - 16.7 70.9 76.2 8.1 9.0 11.4 11.8 53 47 - 59 Barbados 65.0 62.4 61.2 - 63.8 14.5 13.8 - 15.8 67.6 66.9 - 68.7 16.6 15.8 - 17.9 72.7 77.8 10.3 10.2 14.2 13.1 54 45 - 67 Saint Lucia 65.0 62.4 61.1 - 63.6 14.1 13.5 - 14.8 67.6 66.4 - 68.7 15.8 15.2 - 16.5 68.9 74.9 6.5 7.3 9.4 9.8 55 48 - 58 Mexico 65.0 62.4 61.6 - 63.3 14.7 14.4 - 15.6 67.6 67.1 - 68.5 16.8 16.4 - 17.9 71.0 77.2 8.6 9.6 12.2 12.4 56 49 - 64 Bosnia and Herzegovina 64.9 63.4 62.3 - 64.5 13.3 12.7 - 14.1 66.4 65.2 - 67.5 15.3 14.5 - 15.9 71.2 75.0 7.9 8.6 11.0 11.5 57 52 - 62 Trinidad and Tobago 64.6 62.8 62.2 - 63.5 12.0 11.8 - 12.6 66.4 65.9 - 67.1 13.9 13.7 - 14.6 68.7 73.4 5.9 7.0 8.5 9.5 58 52 - 66 Saudi Arabia 64.5 65.1 64.3 - 65.9 12.7 12.1 - 13.4 64.0 62.6 - 65.2 12.8 12.2 - 13.5 71.0 72.6 5.8 8.7 8.2 12.0 59 51 - 69 Brunei Darussalam 64.4 63.4 63.0 - 64.5 12.4 12.0 - 13.6 65.4 64.7 - 66.9 12.6 11.9 - 14.4 74.3 79.5 10.9 14.3 14.6 18.0 60 55 - 63 Bulgaria 64.4 61.2 60.8 - 61.6 12.2 12.0 - 12.7 67.7 67.4 - 68.3 15.1 14.8 - 15.7 67.4 74.8 6.3 7.1 9.3 9.4 61 50 - 64 Bahrain 64.4 63.9 63.0 - 64.8 11.6 11.0 - 12.3 64.9 63.8 - 66.0 12.6 11.8 - 13.3 70.7 73.6 6.8 8.7 9.7 11.8 62 57 - 66 Hungary 64.1 60.4 59.6 - 61.2 11.7 11.4 - 12.3 67.9 67.5 - 68.5 15.5 15.3 - 16.3 66.3 75.1 5.9 7.2 9.0 9.6 63 56 - 67 Lithuania 64.1 60.6 59.7 - 61.6 13.4 13.1 - 14.2 67.5 66.9 - 68.3 16.2 15.9 - 17.0 67.0 77.9 6.4 10.4 9.5 13.3 64 60 - 69 TFYR Macedoniaa 63.7 61.8 61.2 - 62.6 11.7 11.4 - 12.5 65.6 65.1 - 66.5 13.5 13.2 - 14.3 69.8 74.1 8.0 8.5 11.4 11.5 65 60 - 72 Azerbaijan 63.7 60.6 59.9 - 61.4 12.7 12.4 - 13.5 66.7 66.1 - 67.7 15.7 15.4 - 16.7 67.8 75.3 7.2 8.6 10.6 11.4 66 55 - 80 Qatar 63.5 64.2 63.4 - 65.1 10.8 10.1 - 11.4 62.8 61.2 - 64.3 10.2 8.9 - 11.3 71.5 74.6 7.4 11.8 10.3 15.8 67 61 - 78 Cook Islands 63.4 62.2 61.0 - 63.4 12.2 11.5 - 12.8 64.5 63.1 - 65.9 13.7 13.0 - 14.5 69.3 73.3 7.0 8.8 10.1 12.0 68 61 - 82 Kuwait 63.2 63.0 61.6 - 64.3 11.1 10.2 - 12.0 63.4 61.6 - 65.1 11.8 10.6 - 12.9 71.9 75.2 8.9 11.9 12.4 15.8 69 65 - 77 Estonia 63.1 58.1 57.3 - 59.0 11.2 10.9 - 11.9 68.1 67.4 - 69.0 15.8 15.5 - 16.6 64.4 75.3 6.3 7.2 9.8 9.5 70 68 - 76 Ukraine 63.0 58.5 58.1 - 59.0 11.5 11.4 - 12.0 67.5 67.1 - 68.0 15.5 15.3 - 16.1 64.4 74.2 5.8 6.9 9.1 9.3 71 64 - 82 Paraguay 63.0 60.7 59.2 - 62.0 14.2 13.3 - 15.1 65.3 63.9 - 66.7 16.0 15.1 - 17.0 69.6 74.1 8.9 8.8 12.9 11.8 72 62 - 86 Oman 63.0 61.8 60.8 - 62.8 10.6 9.9 - 11.3 64.1 62.8 - 65.3 12.1 11.2 - 12.8 70.4 73.8 8.6 9.7 12.2 13.1 73 67 - 82 Turkey 62.9 64.0 62.9 - 65.0 16.2 15.5 - 16.8 61.8 60.6 - 63.0 15.2 14.6 - 15.9 69.7 69.9 5.7 8.1 8.2 11.6 74 64 - 82 Colombia 62.9 60.3 59.2 - 61.5 13.5 12.7 - 14.2 65.5 64.2 - 66.7 15.4 14.6 - 16.2 68.1 74.1 7.8 8.6 11.5 11.6 75 65 - 83 Tonga 62.9 61.4 60.2 - 62.6 11.5 10.9 - 12.2 64.3 62.9 - 65.6 13.3 12.6 - 14.1 68.3 72.9 6.8 8.6 10.0 11.8 76 68 - 83 Sri Lanka 62.8 59.3 58.3 - 60.3 12.7 12.1 - 13.4 66.3 65.2 - 67.4 16.0 15.4 - 16.7 65.8 73.4 6.5 7.1 9.9 9.7 77 67 - 85 Suriname 62.7 60.2 58.9 - 61.4 14.4 13.8 - 15.2 65.2 64.0 - 66.5 15.5 14.6 - 16.3 68.2 73.6 7.9 8.3 11.6 11.3 78 66 - 88 Mauritius 62.7 59.0 57.9 - 60.2 10.2 9.8 - 11.2 66.3 65.7 - 67.3 13.5 13.2 - 14.4 66.7 74.1 7.7 7.7 11.6 10.4 79 68 - 86 Dominican Republic 62.5 62.1 60.7 - 63.6 17.1 16.1 - 18.2 62.9 61.8 - 63.7 16.1 15.2 - 16.8 71.4 72.8 9.2 9.9 12.9 13.6 80 74 - 84 Romania 62.3 58.8 58.4 - 59.2 12.0 11.8 - 12.6 65.8 65.4 - 66.5 14.6 14.3 - 15.3 65.1 73.5 6.4 7.6 9.8 10.4

76

Annex Table A (continued). Disability-adjusted life expectancy (DALE) and life expectancy (LE) at birth and age 60, by sex, WHO Member States, 1999

Expected years Per cent of total Disability-adjusted life expectancy (DALE) in years Life expectancy lost to disability life expectancy at birth (years) at birth (DLE) lost to disability Total Males Females Uncertainty population At Uncertainty At age Uncertainty Uncertainty At age Uncertainty Rank range Member State at birth birth range 60 range At birth range 60 range Males Females Males Females Males Females 81 72 - 83 China 62.3 61.2 60.7 - 62.0 11.6 11.3 - 12.3 63.3 62.8 - 64.2 13.5 13.2 - 14.4 68.1 71.3 6.9 8.0 10.2 11.2 82 72 - 88 Latvia 62.2 57.1 55.9 - 58.2 11.4 11.0 - 12.2 67.2 66.4 - 68.2 15.9 15.5 - 16.7 63.6 74.6 6.5 7.4 10.2 9.9 83 79 - 90 Belarus 61.7 56.2 55.4 - 57.1 10.1 9.8 - 10.8 67.2 66.7 - 68.0 15.1 14.8 - 15.9 62.4 74.6 6.2 7.3 9.9 9.9 84 79 - 95 Algeria 61.6 62.5 61.4 - 63.5 12.9 12.3 - 13.6 60.7 59.4 - 62.0 12.0 11.3 - 12.6 68.2 68.8 5.7 8.1 8.4 11.7 85 70 - 101 Niue 61.6 61.0 59.2 - 62.6 12.2 11.7 - 13.3 62.2 60.4 - 63.8 13.2 12.7 - 15.2 68.3 70.9 7.3 8.7 10.7 12.2 86 74 - 100 Saint Kitts and Nevis 61.6 58.7 57.4 - 59.9 12.8 12.2 - 13.3 64.4 63.2 - 65.6 14.3 13.7 - 15.0 65.0 71.2 6.3 6.8 9.7 9.5 87 79 - 97 El Salvador 61.5 58.6 57.4 - 59.7 13.9 13.1 - 14.6 64.5 63.2 - 65.7 15.8 14.9 - 16.6 66.9 73.0 8.3 8.5 12.4 11.6 88 82 - 91 Republic of Moldova 61.5 58.5 58.0 - 59.0 10.7 10.6 - 11.2 64.5 64.0 - 65.2 13.0 12.8 - 13.7 64.8 71.9 6.3 7.4 9.7 10.3 89 81 - 96 Malaysia 61.4 61.3 60.2 - 62.1 9.7 9.2 - 10.2 61.6 60.5 - 62.7 9.7 9.1 - 10.3 67.6 69.9 6.3 8.3 9.4 11.9 90 82 - 96 Tunisia 61.4 62.0 61.2 - 62.9 11.2 10.8 - 11.7 60.7 59.7 - 61.8 10.3 9.8 - 10.8 67.0 67.9 5.0 7.2 7.4 10.6 91 84 - 95 Russian Federation 61.3 56.1 55.4 - 56.9 10.5 10.3 - 11.2 66.4 65.8 - 67.2 14.9 14.6 - 15.7 62.7 74.0 6.6 7.6 10.5 10.3 92 83 - 100 Honduras 61.1 60.0 58.8 - 61.2 15.0 14.2 - 15.7 62.3 61.1 - 63.5 14.4 13.5 - 15.1 68.2 70.8 8.2 8.5 12.0 12.0 93 85 - 100 Ecuador 61.0 59.9 58.9 - 60.9 12.6 11.9 - 13.2 62.1 61.1 - 63.3 12.9 12.3 - 13.7 67.4 70.3 7.5 8.2 11.1 11.6 94 82 - 104 Belize 60.9 58.5 56.9 - 60.1 13.6 12.7 - 14.4 63.3 61.5 - 65.0 15.2 14.3 - 16.2 69.6 75.0 11.1 11.6 15.9 15.5 95 90 - 103 Lebanon 60.6 61.2 60.2 - 62.2 10.1 9.6 - 10.6 60.1 58.8 - 61.2 9.2 8.7 - 9.7 66.2 67.3 5.1 7.2 7.7 10.7 96 89 - 104 Iran, Islamic Republic of 60.5 61.3 60.2 - 62.3 11.9 11.3 - 12.5 59.8 58.6 - 61.1 10.9 10.2 - 11.6 66.8 67.9 5.5 8.1 8.2 11.9 97 87 - 103 Samoa 60.5 58.7 57.5 - 59.8 9.5 9.0 - 10.1 62.3 60.9 - 63.6 12.3 11.7 - 13.0 65.4 70.7 6.7 8.4 10.2 11.9 98 90 - 109 Guyana 60.2 57.1 55.8 - 58.6 15.4 14.5 - 16.4 63.3 61.9 - 64.7 16.8 15.8 - 17.8 65.6 72.2 8.4 8.9 12.9 12.3 99 92 - 107 Thailand 60.2 58.4 57.1 - 59.6 13.7 12.9 - 14.5 62.1 60.9 - 63.3 13.9 13.1 - 14.7 66.0 70.4 7.6 8.3 11.6 11.8 100 94 - 104 Uzbekistan 60.2 58.0 57.4 - 58.8 11.5 11.3 - 12.2 62.3 61.6 - 63.1 13.4 13.1 - 14.3 65.8 71.2 7.7 8.9 11.7 12.6 101 95 - 108 Jordan 60.0 60.7 59.8 - 61.5 9.5 9.1 - 10.0 59.3 58.2 - 60.4 8.9 8.5 - 9.4 66.3 67.5 5.6 8.2 8.4 12.1 102 94 - 107 Albania 60.0 56.5 55.8 - 57.4 10.1 9.9 - 10.9 63.4 62.7 - 64.4 13.9 13.6 - 14.7 65.1 72.7 8.6 9.3 13.3 12.8 103 96 - 112 Indonesia 59.7 58.8 57.5 - 60.1 16.3 15.3 - 17.2 60.6 59.3 - 61.8 15.8 15.0 - 16.6 66.6 69.0 7.8 8.4 11.7 12.2 104 95 - 114 Micronesia, Federated States of 59.6 58.7 57.2 - 60.0 11.1 10.4 - 11.8 60.6 59.0 - 62.0 11.5 10.7 - 12.4 66.4 70.1 7.8 9.5 11.7 13.5 105 101 - 114 Peru 59.4 58.0 56.9 - 59.0 12.3 11.7 - 13.0 60.8 59.6 - 62.0 13.1 12.3 - 13.9 65.6 69.1 7.6 8.2 11.6 11.9 106 95 - 116 Fiji 59.4 57.7 56.1 - 59.1 8.3 8.0 - 9.1 61.1 59.8 - 62.3 9.8 9.5 - 10.8 64.0 69.2 6.3 8.1 9.8 11.7 107 100 - 114 Libyan Arab Jamahiriya 59.3 59.7 58.7 - 60.7 9.7 9.2 - 10.2 58.9 57.6 - 60.2 9.3 8.7 - 10.0 65.0 67.0 5.3 8.1 8.2 12.1 108 102 - 112 Seychelles 59.3 56.4 55.7 - 57.3 8.6 8.3 - 9.4 62.1 61.5 - 63.0 11.7 11.4 - 12.4 64.9 70.6 8.4 8.4 13.0 11.9 109 99 - 117 Bahamas 59.1 56.7 55.1 - 58.1 11.3 10.5 - 12.1 61.6 59.9 - 63.4 13.0 12.0 - 14.0 67.0 73.6 10.3 12.0 15.4 16.3 110 101 - 114 Morocco 59.1 58.7 57.9 - 59.6 11.5 11.0 - 12.0 59.4 58.4 - 60.4 11.4 10.8 - 12.0 65.2 66.8 6.4 7.4 9.8 11.0 111 102 - 115 Brazil 59.1 55.2 54.4 - 56.1 11.8 11.5 - 12.7 62.9 62.2 - 63.9 14.8 14.4 - 15.8 63.7 71.7 8.5 8.8 13.3 12.3 112 103 - 116 Palau 59.0 57.4 56.1 - 58.5 8.0 7.5 - 8.5 60.7 59.2 - 61.9 9.7 9.1 - 10.4 64.5 69.7 7.1 9.0 11.0 13.0 113 104 - 116 Philippines 58.9 57.1 56.0 - 58.1 10.3 9.7 - 10.9 60.7 59.4 - 61.9 12.4 11.6 - 13.1 64.1 69.3 7.1 8.7 11.0 12.5 114 105 - 117 Syrian Arab Republic 58.8 58.8 57.7 - 59.9 9.7 9.2 - 10.2 58.9 57.6 - 60.2 10.0 9.4 - 10.6 64.6 67.1 5.8 8.2 8.9 12.3 115 107 - 118 Egypt 58.5 58.6 57.7 - 59.5 11.8 11.2 - 12.2 58.3 57.1 - 59.6 11.7 11.1 - 12.4 64.2 65.8 5.6 7.5 8.8 11.4 116 110 - 119 Viet Nam 58.2 56.7 55.6 - 57.9 9.7 9.1 - 10.4 59.6 58.4 - 60.9 10.8 10.1 - 11.5 64.7 68.8 8.0 9.2 12.3 13.3 117 110 - 119 Nicaragua 58.1 56.4 55.3 - 57.4 11.1 10.4 - 11.8 59.9 58.7 - 61.1 12.5 11.7 - 13.2 64.8 68.8 8.4 8.9 13.0 13.0 118 110 - 121 Cape Verde 57.6 54.6 53.0 - 56.2 11.4 10.6 - 12.3 60.6 58.8 - 62.4 15.3 14.2 - 16.4 64.2 71.8 9.6 11.2 15.0 15.5 119 115 - 122 Tuvalu 57.4 57.1 55.7 - 58.3 10.3 9.7 - 10.9 57.6 56.2 - 58.8 9.4 8.8 - 10.0 63.9 65.5 6.8 7.9 10.6 12.1 120 116 - 123 Tajikistan 57.3 55.1 53.5 - 56.5 12.3 11.4 - 13.2 59.4 57.9 - 60.9 15.6 14.7 - 16.4 65.1 70.1 10.1 10.6 15.5 15.2

77

Annex Table A (continued). Disability-adjusted life expectancy (DALE) and life expectancy (LE) at birth and age 60, by sex, WHO Member States, 1999

Expected years Per cent of total Disability-adjusted life expectancy (DALE) in years Life expectancy lost to disability life expectancy at birth (years) at birth (DLE) lost to disability Total Males Females Uncertainty population At Uncertainty At age Uncertainty Uncertainty At age Uncertainty Rank range Member State at birth birth range 60 range At birth range 60 range Males Females Males Females Males Females 121 117 - 125 Marshall Islands 56.8 56.0 54.4 - 57.4 10.7 10.0 - 11.4 57.6 55.9 - 59.0 11.1 10.3 - 12.0 64.0 67.1 7.9 9.5 12.4 14.2 122 120 - 124 Kazakhstan 56.4 51.5 50.9 - 52.2 8.8 8.7 - 9.5 61.2 60.8 - 62.0 13.1 12.8 - 13.9 58.8 69.9 7.2 8.7 12.3 12.4 123 119 - 125 Kyrgyzstan 56.3 53.4 52.6 - 54.2 9.6 9.4 - 10.4 59.1 58.3 - 60.1 12.4 12.1 - 13.3 61.6 69.0 8.2 9.9 13.3 14.3 124 122 - 127 Pakistan 55.9 55.0 53.8 - 56.3 11.3 10.5 - 12.1 56.8 54.6 - 57.9 12.6 11.9 - 13.2 62.6 64.9 7.6 8.2 12.1 12.6 125 123 - 129 Kiribati 55.3 53.9 52.4 - 55.3 9.4 8.7 - 10.1 56.6 55.0 - 58.0 11.0 10.3 - 11.7 61.4 65.5 7.4 8.9 12.1 13.6 126 123 - 129 Iraq 55.3 55.4 54.4 - 56.4 9.2 8.7 - 9.8 55.1 53.9 - 56.2 8.2 7.6 - 8.8 62.0 63.4 6.2 7.7 10.0 12.2 127 123 - 133 Solomon Islands 54.9 54.5 53.0 - 55.8 8.8 8.2 - 9.5 55.3 53.7 - 56.7 9.2 8.6 - 9.9 62.0 64.0 7.5 8.7 12.2 13.7 128 125 - 133 Turkmenistan 54.3 51.9 50.6 - 53.3 9.0 8.7 - 10.0 56.7 55.3 - 58.0 10.9 10.6 - 11.8 61.0 65.3 9.1 8.6 14.9 13.2 129 126 - 132 Guatemala 54.3 52.1 51.1 - 53.1 9.1 8.6 - 9.8 56.4 55.4 - 57.5 10.1 9.5 - 10.7 60.2 64.7 8.1 8.3 13.4 12.8 130 127 - 134 Maldives 53.9 54.4 53.0 - 55.9 12.1 11.3 - 13.0 53.3 51.8 - 54.7 11.5 10.8 - 12.2 63.3 62.6 8.9 9.3 14.0 14.9 131 127 - 135 Mongolia 53.8 51.3 49.7 - 52.7 11.8 11.0 - 14.4 56.3 54.7 - 57.7 14.3 13.4 - 15.1 58.9 64.8 7.7 8.5 13.0 13.1 132 129 - 135 Sao Tome and Principe 53.5 52.1 51.1 - 53.3 11.4 11.1 - 12.5 54.8 53.8 - 55.8 11.7 11.4 - 12.6 62.1 64.9 10.0 10.1 16.1 15.5 133 129 - 136 Bolivia 53.3 52.5 51.3 - 53.7 11.6 10.9 - 12.3 54.1 52.8 - 55.2 11.2 10.6 - 11.9 60.7 62.2 8.3 8.1 13.6 13.1 134 130 - 136 India 53.2 52.8 52.1 - 53.5 10.6 10.4 - 11.3 53.5 52.8 - 54.3 12.1 11.8 - 12.8 59.6 61.2 6.8 7.7 11.3 12.5 135 130 - 138 Vanuatu 52.8 51.3 49.8 - 52.7 8.0 7.4 - 8.6 54.4 52.8 - 55.8 9.2 8.5 - 9.8 58.7 63.0 7.4 8.6 12.7 13.7 136 133 - 138 Nauru 52.5 49.8 48.7 - 50.9 3.6 3.1 - 4.0 55.1 53.8 - 56.2 5.9 5.4 - 6.4 56.4 63.3 6.6 8.1 11.6 12.9 137 134 - 139 Democratic People's Republic of Korea 52.3 51.4 49.8 - 53.1 9.6 8.7 - 10.6 53.1 51.3 - 55.0 11.6 10.7 - 12.6 58.0 60.7 6.6 7.6 11.3 12.5 138 135 - 139 Bhutan 51.8 51.4 50.0 - 52.7 11.4 10.7 - 12.0 52.2 50.7 - 53.6 12.6 12.0 - 13.3 59.6 60.8 8.2 8.7 13.8 14.2 139 136 - 139 Myanmar 51.6 51.4 50.0 - 52.6 12.5 11.7 - 13.3 51.9 50.5 - 53.2 12.3 11.6 - 12.9 58.4 59.2 7.1 7.4 12.1 12.4 140 140 - 142 Bangladesh 49.9 50.1 48.7 - 51.3 9.9 9.2 - 10.5 49.8 48.3 - 51.2 10.5 9.8 - 11.1 57.5 58.1 7.4 8.3 12.9 14.3 141 140 - 142 Yemen 49.7 49.7 48.1 - 51.1 8.5 7.9 - 9.2 49.7 48.2 - 51.1 8.2 7.5 - 8.8 57.3 58.0 7.6 8.3 13.2 14.3 142 140 - 142 Nepal 49.5 49.4 48.1 - 50.7 10.3 9.6 - 10.9 49.5 48.2 - 50.9 10.0 9.4 - 10.7 57.3 57.8 7.9 8.3 13.7 14.3 143 142 - 144 Gambia 48.3 47.2 46.3 - 48.2 9.9 9.3 - 10.6 49.4 48.4 - 50.4 11.7 11.2 - 12.4 56.0 58.9 8.8 9.5 15.7 16.1 144 143 - 145 Gabon 47.8 46.6 45.4 - 47.6 10.3 9.8 - 10.8 49.0 47.8 - 50.1 12.3 11.8 - 12.8 54.6 57.5 8.0 8.5 14.6 14.8 145 144 - 147 Papua New Guinea 47.0 45.5 44.3 - 46.8 8.2 7.5 - 8.9 48.5 47.1 - 49.8 8.7 8.0 - 9.4 53.4 56.6 7.8 8.1 14.7 14.3 146 145 - 147 Comoros 46.8 46.1 45.1 - 47.1 8.9 8.3 - 9.6 47.5 46.5 - 48.5 9.8 9.4 - 10.4 56.0 58.1 9.9 10.6 17.7 18.3 147 145 - 150 Lao People's Democratic Republic 46.1 45.0 43.5 - 46.5 8.9 8.0 - 9.7 47.1 45.5 - 48.6 8.8 7.9 - 9.6 54.0 56.6 9.0 9.5 16.6 16.7 148 146 - 150 Cambodia 45.7 43.9 42.6 - 45.1 7.4 6.6 - 8.2 47.5 46.1 - 48.9 9.3 8.7 - 10.0 52.2 55.4 8.3 7.9 15.8 14.2 149 147 - 151 Ghana 45.5 45.0 43.8 - 46.2 9.9 9.3 - 10.6 46.0 44.8 - 47.2 10.2 9.6 - 10.8 54.2 55.6 9.2 9.6 16.9 17.2 150 148 - 152 Congo 45.1 44.3 43.1 - 45.5 10.7 10.0 - 11.3 45.9 44.6 - 47.1 12.8 12.2 - 13.4 53.6 55.2 9.3 9.3 17.4 16.9 151 149 - 153 Senegal 44.6 43.5 42.5 - 44.5 8.8 8.1 - 9.5 45.6 44.6 - 46.7 11.3 10.6 - 11.9 53.5 56.2 10.0 10.6 18.7 18.8 152 150 - 153 Equatorial Guinea 44.1 42.8 41.7 - 43.9 9.4 8.9 - 10 45.4 44.4 - 46.6 11.0 10.5 - 11.7 51.4 55.4 8.6 9.9 16.7 17.9 153 149 - 156 Haiti 43.8 42.4 41.0 - 43.6 7.4 6.8 - 8.0 45.2 43.7 - 46.7 8.0 7.3 - 8.7 50.5 55.0 8.2 9.8 16.2 17.8 154 153 - 157 Sudan 43.0 42.6 41.2 - 43.7 5.6 5.1 - 6.0 43.5 42.1 - 44.6 6.0 5.6 - 6.5 53.1 54.7 10.5 11.2 19.8 20.5 155 153 - 156 Côte d'Ivoire 42.8 42.2 41.2 - 43.3 11.9 11.5 - 12.5 43.3 42.3 - 44.4 12.7 12.2 - 13.2 47.3 48.3 5.1 5.0 10.8 10.3

78 Annex Table A (continued). Disability-adjusted life expectancy (DALE) and life expectancy (LE) at birth and age 60, by sex, WHO Member States, 1999

Expected years Per cent of total Disability-adjusted life expectancy (DALE) in years lost to disability life expectancy Life expectancy at birth (DLE) lost to disability at birth (years) Total Males Females Uncertainty population At Uncertainty At age Uncertainty Uncertainty At age Uncertainty Rank range Member State at birth birth range 60 range At birth range 60 range Males Females Males Females Males Females 156 154 - 158 Cameroon 42.2 41.5 40.4 - 42.5 9.6 9.0 - 10.2 43.0 41.8 - 44.2 11.9 11.3 - 12.5 49.9 52.0 8.4 9.0 16.9 17.3 157 154 - 158 Benin 42.2 41.9 40.9 - 42.9 9.6 9.0 - 10.3 42.6 41.5 - 43.6 10.6 9.9 - 11.2 51.3 53.3 9.4 10.7 18.4 20.1 158 157 - 159 Mauritania 41.4 40.2 39.2 - 41.2 9.2 8.6 - 9.9 42.5 41.5 - 43.5 11.0 10.3 - 11.7 49.5 53.0 9.3 10.5 18.8 19.7 159 158 - 161 Togo 40.7 40.0 38.8 - 41.3 9.5 8.8 - 10.1 41.4 40.1 - 42.6 11.0 10.5 - 11.6 48.9 50.8 8.9 9.4 18.2 18.6 160 158 - 161 South Africa 39.8 38.6 37.7 - 39.5 6.8 6.4 - 7.3 41.0 39.9 - 42.1 9.3 8.9 - 9.8 47.3 49.7 8.7 8.8 18.4 17.6 161 160 - 163 Chad 39.4 38.6 37.2 - 39.8 9.2 8.6 - 9.9 40.2 38.8 - 41.5 10.6 10.0 - 11.2 47.3 50.1 8.7 9.9 18.4 19.8 162 160 - 164 Kenya 39.3 39.0 37.9 - 40.2 9.2 8.6 - 13.4 39.6 38.4 - 41.0 12.0 11.5 - 12.5 47.3 48.1 8.4 8.5 17.7 17.6 163 162 - 169 Nigeria 38.3 38.1 36.9 - 39.2 8.7 8.1 - 9.4 38.4 37.1 - 39.6 10.1 9.5 - 10.7 46.8 48.2 8.7 9.7 18.5 20.2 164 162 - 170 Swaziland 38.1 37.8 36.5 - 39.0 8.1 7.7 - 8.6 38.4 36.9 - 39.9 9.5 9.0 - 10.0 45.8 46.8 8.0 8.4 17.5 17.9 165 163 - 170 Angola 38.0 37.0 35.7 - 38.1 8.9 8.3 - 9.6 38.9 37.7 - 40.0 10.8 10.1 - 11.4 46.3 49.1 9.3 10.2 20.0 20.7 166 163 - 170 Djibouti 37.9 37.7 36.2 - 38.8 6.9 6.4 - 7.4 38.1 36.6 - 39.3 7.9 7.6 - 8.2 45.0 45.0 7.3 7.0 16.2 15.5 167 164 - 170 Guinea 37.8 37.0 36.1 - 38.0 8.5 7.9 - 9.2 38.5 37.5 - 39.5 9.6 9.0 - 10.3 46.2 48.9 9.2 10.4 19.9 21.2 168 163 - 173 Afghanistan 37.7 36.7 34.9 - 38.5 7.9 7.0 - 8.8 38.7 36.9 - 40.5 7.9 7.2 - 8.7 45.3 47.2 8.5 8.4 18.9 17.9 169 164 - 171 Eritrea 37.7 38.5 37.6 - 39.5 8.2 7.6 - 8.7 36.9 35.9 - 37.9 7.9 7.4 - 8.4 46.6 46.5 8.2 9.6 17.5 20.6 170 166 - 174 Guinea-Bissau 37.2 36.8 35.6 - 37.9 9.1 8.5 - 13.4 37.5 36.4 - 38.6 10.0 9.4 - 10.6 45.0 47.0 8.2 9.5 18.1 20.2 171 167 - 175 Lesotho 36.9 36.6 35.3 - 38.0 9.9 9.3 - 10.4 37.2 35.7 - 38.7 11.3 10.8 - 11.9 44.1 45.1 7.5 8.0 17.0 17.7 172 169 - 176 Madagascar 36.6 36.5 35.5 - 37.4 6.7 6.1 - 7.2 36.8 35.7 - 37.7 6.6 6.0 - 7.2 45.0 47.7 8.5 10.9 19.0 22.9 173 169 - 178 Somalia 36.4 35.9 34.4 - 37.2 6.1 5.6 - 6.5 36.9 35.3 - 38.1 7.5 7.2 - 7.9 44.0 44.7 8.2 7.8 18.6 17.4 174 171 - 177 Democratic Republic of the Congo 36.3 36.4 35.5 - 37.3 7.3 6.8 - 7.9 36.2 35.4 - 37.3 7.8 7.3 - 8.4 45.1 46.5 8.7 10.3 19.3 22.1 175 172 - 178 Central African Republic 36.0 35.6 34.6 - 36.7 8.8 8.3 - 9.3 36.5 35.3 - 37.7 10.6 10.1 - 11.1 43.3 44.9 7.7 8.4 17.7 18.7 176 172 - 178 United Republic of Tanzania 36.0 35.9 35.1 - 36.8 7.8 7.2 - 8.4 36.1 35.2 - 37.1 9.2 8.7 - 9.8 44.4 45.6 8.5 9.5 19.1 20.8 177 172 - 180 Namibia 35.6 35.8 34.3 - 37.4 9.8 9.3 - 10.4 35.4 33.8 - 37.4 12.1 11.5 - 12.6 43.3 43.0 7.5 7.6 17.4 17.7 178 173 - 179 Burkina Faso 35.5 35.3 34.1 - 36.6 7.9 7.3 - 8.5 35.7 34.4 - 37.0 9.1 8.5 - 9.8 44.1 45.7 8.8 10.0 19.9 21.9 179 176 - 182 Burundi 34.6 34.6 33.0 - 36.2 7.6 6.9 - 8.3 34.6 32.8 - 36.3 9.4 8.8 - 10.0 43.2 43.8 8.6 9.2 19.9 21.1 180 178 - 183 Mozambique 34.4 33.7 32.3 - 35.3 8.3 7.7 - 8.9 35.1 33.5 - 36.9 10.7 10.1 - 11.4 41.8 44.0 8.1 8.9 19.3 20.3 181 179 - 183 Liberia 34.0 33.8 32.7 - 34.9 7.3 6.7 - 8.0 34.2 33.1 - 35.3 8.3 7.7 - 8.9 42.5 44.9 8.7 10.7 20.4 23.8 182 180 - 185 Ethiopia 33.5 33.5 32.5 - 34.5 7.5 7.0 - 8.1 33.5 32.3 - 34.7 8.6 8.0 - 9.2 41.4 43.1 7.9 9.5 19.1 22.1 183 181 - 186 Mali 33.1 32.6 31.6 - 33.7 7.7 7.1 - 8.3 33.5 32.5 - 34.5 9.0 8.4 - 9.7 41.3 44.0 8.7 10.5 21.0 23.8 184 181 - 187 Zimbabwe 32.9 33.4 32.3 - 34.5 8.8 8.3 - 9.4 32.4 31.3 - 33.7 10.1 9.6 - 10.6 40.9 40.0 7.5 7.6 18.4 18.9 185 181 - 187 Rwanda 32.8 32.9 31.6 - 34.3 6.9 6.2 - 7.6 32.7 31.3 - 34.3 7.4 6.9 - 8.1 41.2 42.3 8.4 9.6 20.3 22.6 186 183 - 187 Uganda 32.7 32.9 32.1 - 33.9 6.2 5.6 - 6.9 32.5 31.6 - 33.5 7.4 6.8 - 8.0 41.9 42.4 9.0 9.9 21.4 23.4 187 185 - 187 Botswana 32.3 32.3 31.7 - 32.9 6.1 5.6 - 6.6 32.2 31.6 - 33.0 9.7 9.3 - 10.0 39.5 39.3 7.2 7.1 18.2 18.0 188 188 - 189 Zambia 30.3 30.0 28.9 - 30.9 7.6 6.9 - 8.3 30.7 29.5 - 31.7 10.7 10.1 - 11.4 38.0 39.0 8.0 8.3 21.1 21.3 189 188 - 190 Malawi 29.4 29.3 28.3 - 30.2 6.8 6.2 - 7.5 29.4 28.4 - 30.4 8.3 7.7 - 8.9 37.3 38.4 8.0 9.0 21.3 23.3 190 189 - 190 Niger 29.1 28.1 27.1 - 29.0 6.6 5.8 - 7.4 30.1 29.0 - 31.1 9.6 8.8 - 10.6 37.2 40.6 9.0 10.5 24.3 25.8 191 191 Sierra Leone 25.9 25.8 24.5 - 26.8 6.0 5.4 - 6.7 26.0 24.8 - 27.1 6.0 5.3 - 6.7 33.2 35.4 7.4 9.5 22.4 26.7 (a) Uncertainty ranges show 10th and 90th of the uncertainty distribution of the indicator (rank or DALE). (b) DLE (Expected years lost to disability) is calculated as total life expectancy minus DALE. Per cent of life expectancy lost to disability is DLE/LE as a per cent. (c) The Former Yugoslav Republic of Macedonia

79