Menu
Menu
Menu
Menu
Background: The UN’s Sustainable Development Goals (SDGs) are grounded in the global ambition of “leaving no one behind”. Understanding today’s gains and gaps for the health-related SDGs is essential for decision makers as they aim to improve the health of populations. As part of the Global Burden of Diseases, Injuries, and Risk Factors Study 2016 (GBD 2016), we measured 37 of the 50 health-related SDG indicators over the period 1990-2016 for 188 countries, and then on the basis of these past trends, we projected indicators to 2030. Methods: We used standardised GBD 2016 methods to measure 37 health-related indicators from 1990 to 2016, an increase of four indicators since GBD 2015. We substantially revised the universal health coverage (UHC) measure, which focuses on coverage of essential health services, to also represent personal health-care access and quality for several non-communicable diseases. We transformed each indicator on a scale of 0-100, with 0 as the 25th percentile estimated between 1990 and 2030, and 100 as the 975th percentile during that time. An index representing all 37 health-related SDG indicators was constructed by taking the geometric mean of scaled indicators by target. On the basis of past trends, we produced projections of indicator values, using a weighted average of the indicator and country-specific annualised rates of change from 1990 to 2016 with weights for each annual rate of change based on out-of-sample validity. 24 of the currently measured health-related SDG indicators have defined SDG targets, against which we assessed attainment. Findings Globally, the median health-related SDG index was 567 (IQR 319-668) in 2016 and country-level performance markedly varied, with Singapore (868, 95% uncertainty interval 846-889), Iceland (860, 841-876), and Sweden (856, 818-878) having the highest levels in 2016 and Afghanistan (109, 96-119), the Central African Republic (110, 88-138), and Somalia (113, 95-131) recording the lowest. Between 2000 and 2016, notable improvements in the UHC index were achieved by several countries, including Cambodia, Rwanda, Equatorial Guinea, Laos, Turkey, and China; however, a number of countries, such as Lesotho and the Central African Republic, but also high-income countries, such as the USA, showed minimal gains. Based on projections of past trends, the median number of SDG targets attained in 2030 was five (IQR 2-8) of the 24 defined targets currently measured. Globally, projected target attainment considerably varied by SDG indicator, ranging from more than 60% of countries projected to reach targets for under-5 mortality, neonatal mortality, maternal mortality ratio, and malaria, to less than 5% of countries projected to achieve targets linked to 11 indicator targets, including those for childhood overweight, tuberculosis, and road injury mortality. For several of the health-related SDGs, meeting defined targets hinges upon substantially faster progress than what most countries have achieved in the past. Interpretation GBD 2016 provides an updated and expanded evidence base on where the world currently stands in terms of the health-related SDGs. Our improved measure of UHC offers a basis to monitor the expansion of health services necessary to meet the SDGs. Based on past rates of progress, many places are facing challenges in meeting defined health-related SDG targets, particularly among countries that are the worst off. In view of the early stages of SDG implementation, however, opportunity remains to take actions to accelerate progress, as shown by the catalytic effects of adopting the Millennium Development Goals after 2000. With the SDGs’ broader, bolder development agenda, multisectoral commitments and investments are vital to make the health-related SDGs within reach of all populations.
This analysis of the health-related SDGs is based on the GBD study, which measures the health of populations on an annual basis. GBD produces age-specific, sex-specific, and country-specific estimates (including selected subnational units) of cause-specific mortality and morbidity, risk factor exposure, mortality and morbidity attributable to these risks, and a range of health system characteristics, from 1990 to the most recent year. Various summary measures are computed, including disability-adjusted life-years (DALYs) and healthy life expectancy. GBD uses highly standardised and validated approaches applied to all available data sources adjusted for major sources of bias. Further details on GBD 2016, which covers 1990–2016, are available elsewhere.28, 29, 30, 31, 32 As with all revisions of the GBD study, GBD 2016 provides an update of the full time series from 1990–2016 based on methodological improvements and newly identified data sources; subsequently, the full time series on the health-related SDGs published here as part of GBD 2016 supersedes previous GBD studies. The GBD 2016 study and this analysis comply with the Guidelines for Accurate and Transparent Health Estimates Reporting (GATHER).33 Further detail on the estimation and data sources used for all indicators are available in appendix 1. In this updated analysis we cover 37 of 50 health-related SDG indicators (table). Additional details on data and methods for estimating each indicator are in appendix 1. Appendix 2 outlines the 13 indicators not presently measured (pp 10–12). The addition of new causes, risks, and health indicators are considered by the GBD Scientific Council for each annual cycle of the GBD. For GBD 2016, four health-related SDG indicators were added: vaccine coverage (SDG indicator 3.b.1); two violence indicators (prevalence of physical or sexual violence [SDG indicator 16.1.3] and childhood sexual abuse [SDG indicator 16.2.3]); and well-certified death registration (SDG indicator 17.19.2c). Health-related goals, targets, and health-related SDG indicators used in the present analysis and further details regarding any indicator modifications, and inclusion in the health-related MDG index or health-related non-MDG index Detailed descriptions of the data and methods used to estimate each health-related SDG indicator are in appendix 1. DALY=disability-adjusted life-year. GBD=Global Burden of Disease. HAQ Index=Healthcare Access and Quality Index. IOTF=International Obesity Task Force. JMP=Joint Monitoring Programme. MDG=Millennium Development Goal. NCDs=non-communicable diseases. SDG=Sustainable Development Goal. SEV=summary exposure value. WaSH=water, sanitation, and hygiene. PM2·5=fine particulate matter smaller than 2.5 μm. Vaccine coverage (SDG indicator 3.b.1), defined as “proportion of the target population covered by all vaccines included in their national programme”, became a separate indicator as part of the March, 2017, revision to the SDG framework.5 We report on this indicator by using the geometric mean of the coverage of three-dose diphtheria, pertussis, and tetanus (DPT3); three-dose polio; first-dose measles vaccine; and for countries where the vaccine(s) are included in the national schedule: BCG vaccine, three-dose pneumococcal conjugate vaccine (PCV3), three-dose Haemophilus influenzae type b vaccine (Hib3), three-dose hepatitis B vaccine (delivered as part of pentavalent vaccines), and two-dose or three-dose rotavirus vaccine. To account for the scale-up period for newly introduced vaccines, we include new vaccines in the geometric mean only 3 years after the introduction year in each country. We also added two violence indicators in GBD 2016: age-standardised prevalence of physical or sexual violence experienced by populations in the last 12 months (SDG indicator 16.1.3) and age-standardised prevalence of women and men aged 18–29 years who experienced sexual violence by age 18 years (SDG indicator 16.2.3). The UN definition for SDG indicator 16.1.3 includes psychological violence, but due to limited data availability and highly variable definitions of self-reported psychological violence, we restricted this measurement to physical and sexual violence. As part of GBD 2016, we developed a data quality measure to reflect the proportion of well-certified deaths by a vital registration (VR) system among a country’s total population, which corresponds with the third component of 17.19.2 (referred to as SDG indicator 17.19.2c). Well-certified deaths were determined by three measures: (1) completeness of death registration; (2) fraction of deaths not assigned to major garbage codes (ie, causes that cannot or should not be underlying causes of death); and (3) fraction of deaths assigned to detailed GBD causes. More detail on this measure can be found elsewhere 29 and in appendix 1. We also refined the measurement of several previously included health-related indicators. First, SDG indicator 16.1.2 (conflict mortality) now exclusively focuses on deaths due to conflict and terrorism. Second, we revised the exposure period from lifetime to 12 months for SDG indicator 5.2.1 (intimate partner violence) to match the UN SDG definition. Third, we limited our measurement of SDG indicator 6.2.1b (hygiene) to access to a handwashing facility, which also aligns more directly with the UN SDG target. Fourth, we extended the measurement of SDG indicator 3.8.1 (coverage of essential health services, or UHC tracer interventions) to include the individual components of the HAQ Index,20 which is based on risk-standardised death rates from 32 causes amenable to personal health care.34, 35 This revised approach expands the range of potential health services, particularly those for NCDs, captured by this summary measure. The previous UHC tracer indicator included only maternal and child health and selected infectious disease interventions (malaria, HIV, and tuberculosis).6 Last, a subset of indicators have undergone substantial revision due to data improvements, methodological improvements, or both, implemented in GBD 2016, including alcohol consumption and child growth failure (ie, under-5 stunting and wasting). Further detail on these updates can be found in appendix 1, as well as accompanying GBD 2016 papers.28, 29, 30, 31, 32 We projected the health-related SDG indicators on the basis of past trends. We first calculated for each location the annual rate of change between 1990 and 2016 for each individual year in natural-log space or, for indicators bounded between 0 and 1 (eg, intervention coverage, percentage of population) in logit-space. We then calculated the weighted median annualised rate of change for each country using the following weighting function: The value of ω denotes how much weight is given to recent years compared with past years when calculating the median annualised rate of change. To determine the appropriate value of ω for each SDG indicator, we did an out-of-sample predictive validity test in which we held out data for all countries from 2008 to 2016 and predicted values for this time period using the data from 1990 to 2007. We tested values of ω ranging from 0 to 2 in increments of 0·2 and chose the indicator-specific value of ω that minimised the root mean squared error (RMSE) in the held out data (2008–16). This was used to project each indicator to 2030. Appendix 1 provides the indicator specific values of ω used and further details on methods. For HIV, we used an alternative approach. In many countries, antiretroviral therapy (ART) coverage, through large internal investments, substantial development assistance via programmes such as the President’s Emergency Plan for AIDS Relief (PEPFAR),36 and reductions in drug prices, has been scaled up considerably. If past trends are used to project future coverage, many countries would be projected to achieve 100% coverage by 2030. This ignores health system constraints in scaling up ART. For ART coverage, our projections were a function of projected ART price based on data from the Global Price Reporting Mechanism (GPRM),37 projected government health expenditure as source,38 and projected development assistance for health (DAH) for HIV or AIDS.38 We bounded ART projections with an ART coverage frontier produced on the basis of income per capita to reflect health system constraints. We then used projected ART coverage to project HIV incidence hazard and HIV incidence using Spectrum.39 Further detail on this method is in appendix 1. As in GBD 2015, we developed an overall health-related SDG index that is a function of the 37 health-related SDG indicators (referred to as the health-related SDG index), an index reflecting the 14 SDG health-related indicators previously included in the MDG monitoring framework (referred to as the MDG index), and one reflecting the 23 SDG health-related indicators not included in the MDGs (referred to as the non-MDG index). A variety of approaches exist to create indices from multidimensional data. As in GBD 2015,6 we adopted a preference-weighted approach that weights each indicator by expressed preferences for the relative importance of different indicators. We interpret the SDG targets to represent the expressed preferences of UN member states and thus assume that each SDG target should be treated equally. To combine indicators, we first transformed each indicator on a scale from 0 to 100. Scaling indicator values in this way allows comparisons to be made on the relative performance on very different SDG indicators and allows us to produce an overall health-related SDG index by calculating an arithmetic or geometric mean of the scaled values. For GBD 2016, we transformed each health-related SDG indicator on a scale from 0 to 100, in order from worst to best, with 0 being the 2·5th percentile value observed over the time period 1990–2030 (ie, including projected values) and 100 the 97·5th percentile value observed during this time. This was implemented in log-space for mortality and incidence rates. To estimate the health-related SDG index, we first computed the geometric mean of each scaled health-related SDG indicator for a given target, followed by the geometric mean of resulting values across all SDG targets (reflecting the preference-weighted approach described above). The geometric mean allows indicators with very high values to partly compensate for low values on other indicators (referred to as partial substitutability). To avoid problems with indicator values close to 0, when computing indices we applied a floor of 1 to all indicators. The same process was used to construct the MDG and non-MDG indices. Results of sensitivity analyses based on alternative approaches to create the SDG, MDG, and non-MDG indices are detailed in appendix 1. Of the 37 health-related indicators measured in GBD 2016, 24 had defined targets, with 21 having absolute targets to reach by 2030, and three featuring targets relative to 2015 levels (ie, SDG target 3.4, “By 2030, reduce by one third premature mortality from NCDs”). For these 24 indicators, we applied these thresholds to determine achievement by 2030 (or 2020, in the case of road injury mortality [SDG indicator 3.6.1]). 17 health-related indicators had targets citing “achieving elimination”, “ending epidemics”, or “reaching universal coverage or access”. For these indicators we set target thresholds as at least 99% for universal coverage or access and achieving a rate of 0·000005 or less for measures of morbidity (ie, ≤0·005 per 1000 or ≤0·5 per 100 000) and 0·5% for prevalence. The table details the target thresholds or relative reductions applied for each indicator with a defined target. Because some of these elimination targets have been operationalised in terms of reducing incidence or prevalence by 2030,40 we applied a more conservative 80% reduction threshold from 2015 to 2030 for indicators with established elimination SDG targets and compared these results. We also used a threshold of 90% or more for indicators with universal coverage or access in this conservative target attainment scenario.41 In addition to examining global patterns in SDG performance, we report on differences in the health-related SDG index and individual indicators across levels of development. To do this, we use the Socio-demographic Index (SDI), a summary measure of overall development that was originally introduced as part of GBD 2015.30 SDI is based on income per capita, mean years of education among populations 15 years and older, and total fertility rates, on a scale of 0 to 1. We use the SDI quintiles established in the GBD study to compare performance and progress on the health-related SDGs. More details on the estimation of SDI can be found in accompanying GBD 2016 publications.28, 29, 30, 31, 32 GBD produces 1000 draws for all indicator estimates by location, age, and sex (when relevant) and for all years from 1990 to 2016. These draws from the posterior distribution represent uncertainty in the underlying data sources as well as the various steps in the estimation process. Further details on this are provided in the accompanying GBD 2016 papers28, 29, 30, 31, 32 and in appendix 1 for each indicator. These 1000 draws are used to determine 95% uncertainty intervals (UIs) in each of the scaled SDG indicators, as well as the three indices, using simulation analysis. To estimate uncertainty in SDG indicators and indices for the projected values, we applied the median rate of change chosen from the out-of-sample validity test to each of the 1000 draws of the indicator to estimate 1000 draws of each indicator for the time period 2017–30. Additionally, for each of the 1000 draws we allow for year-to-year deviation from the median rate of change on the basis of the variance across all draws. The funder of the study had no role in the study design, data collection, data analysis, data interpretation, or writing of the report. The corresponding author had full access to all the data in the study and had final responsibility for the decision to submit for publication.
