Global, regional, and national age-specific progress towards the 2020 milestones of the WHO End TB Strategy: a systematic analysis for the Global Burden of Disease Study 2021

1Despite being a preventable and largely curable disease, tuberculosis remains a major contributor to the global burden of disease. Tuberculosis causes more than 1 million deaths annually and was the leading cause of death due to a single infectious agent in 2019. Since the early 1990s, global initiatives to address tuberculosis have grown more prominent, resulting in declines in the global burden of this disease. Recent progress has been unimpressive, however, as global annual rates of decline in tuberculosis incidence over the past 10 years have ranged between 1% and 2%., The WHO End TB Strategy represents a renewed global resolve to accelerate progress by aiming to reduce deaths due to tuberculosis by 95% and cut tuberculosis incidence by 90% between 2015 and 2035, while also seeking to have zero tuberculosis-affected households experiencing catastrophic financial costs due to the disease by 2035. Rigorous evaluations of the trends in the global burden of tuberculosis are therefore crucial for assessing progress towards achieving the End TB Strategy targets and identifying specific drivers of marked national progress.

2Evidence before this study In the context of the WHO End TB Strategy, several groups have generated estimates of tuberculosis incidence and mortality, including the WHO Global Tuberculosis Programme and the Global Burden of Diseases, Injuries, and Risk Factors Study (GBD), to monitor global and national-level progress towards eliminating tuberculosis. 2020 marked a crucial year for evaluating progress as the first quantitative milestones of the End TB Strategy passed. We searched PubMed with the search string “(“tuberculosis” OR “TB”) AND (“burden” OR “estimates”) AND (“End TB”) AND (“target*” OR “milestone*”)”, with no language restrictions, for publications from Jan 1, 2015, to Dec 1, 2023. Our search identified seven studies presenting population-based tuberculosis burden estimates to assess progress towards the End TB targets for subsets of countries; of these, only one study examined differences in progress by age. The single study analysing progress by age used GBD 2019 tuberculosis estimates to evaluate temporal trends in Cambodia. Additionally, the 2022 Global Tuberculosis Report from WHO illustrated that tuberculosis deaths decreased by only 5·9%, while the tuberculosis incidence rate dropped by 10%, between 2015 and 2021, falling well short of the targeted milestones. However, as yet, no global, systematic, age-specific study has been done to investigate progress towards the 2020 End TB Strategy mortality and incidence milestones, which outlined a 35% reduction in tuberculosis deaths and 20% reduction in the tuberculosis incidence rate between 2015 and 2020, with additional assessments of the role of key risk factors in achieving these milestones. Added value of this study We comprehensively examined the burden of tuberculosis for five separate age groups in 204 countries and territories from 1990 to 2021. We focused on examining temporal trends in tuberculosis incidence and mortality from 2015 to 2020 to assess global progress towards the End TB Strategy milestones, with an emphasis on exploring progress by age for the first time. We also present age-specific, risk-deleted mortality estimates that represent deaths due to tuberculosis that would have been observed if the combined effects of all evaluated risk factors were removed, to provide insights into the role of addressing risk factors in achieving the End TB targets. Last, we analysed vital registration data only (2015–21 inclusive) to assess the potential impact of the COVID-19 pandemic on age-specific tuberculosis mortality in countries with available cause-specific data. Implications of all the available evidence Global tuberculosis control programmes could consider closely examining the 15 countries that achieved the 2020 incidence milestone and the 17 countries that achieved the 2020 mortality milestone, many of which were in western sub-Saharan Africa and eastern Europe, to better understand drivers of their marked progress. Many of these countries have implemented innovative approaches to active case finding while implementing social protection interventions such as advanced tuberculosis surveillance in high-risk areas and economic incentives for patients. We observed unequal progress by age groups and found particularly slow progress in reducing the tuberculosis burden in older adults (aged >50 years), indicating that tuberculosis in this age group should be more widely recognised and monitored to achieve WHO's 2035 End TB Strategy targets. One approach for reducing the tuberculosis burden in older adults and achieving the 2035 End TB targets is addressing risk factors for tuberculosis. Our assessment of risk-deleted mortality suggests that the world would have achieved the 2020 End TB mortality milestone, with a 36% reduction in global tuberculosis deaths and the largest benefits in older adults, if the risk factors of smoking, alcohol use, and diabetes were removed. Although complete elimination of such risk factors is highly unlikely, this assessment illustrates the magnitude of potential improvement. Finally, our secondary analysis of countries with complete vital registration data showed variable impacts of the COVID-19 pandemic on tuberculosis mortality in 2020 and 2021; some countries and age groups reported fewer than expected deaths due to tuberculosis, whereas others reported more deaths than expected. Additional data from countries with a high tuberculosis burden are urgently needed to understand the generalisability of these findings.

3Evidence before this study

4In the context of the WHO End TB Strategy, several groups have generated estimates of tuberculosis incidence and mortality, including the WHO Global Tuberculosis Programme and the Global Burden of Diseases, Injuries, and Risk Factors Study (GBD), to monitor global and national-level progress towards eliminating tuberculosis. 2020 marked a crucial year for evaluating progress as the first quantitative milestones of the End TB Strategy passed. We searched PubMed with the search string “(“tuberculosis” OR “TB”) AND (“burden” OR “estimates”) AND (“End TB”) AND (“target*” OR “milestone*”)”, with no language restrictions, for publications from Jan 1, 2015, to Dec 1, 2023. Our search identified seven studies presenting population-based tuberculosis burden estimates to assess progress towards the End TB targets for subsets of countries; of these, only one study examined differences in progress by age. The single study analysing progress by age used GBD 2019 tuberculosis estimates to evaluate temporal trends in Cambodia. Additionally, the 2022 Global Tuberculosis Report from WHO illustrated that tuberculosis deaths decreased by only 5·9%, while the tuberculosis incidence rate dropped by 10%, between 2015 and 2021, falling well short of the targeted milestones. However, as yet, no global, systematic, age-specific study has been done to investigate progress towards the 2020 End TB Strategy mortality and incidence milestones, which outlined a 35% reduction in tuberculosis deaths and 20% reduction in the tuberculosis incidence rate between 2015 and 2020, with additional assessments of the role of key risk factors in achieving these milestones.

5Added value of this study

6We comprehensively examined the burden of tuberculosis for five separate age groups in 204 countries and territories from 1990 to 2021. We focused on examining temporal trends in tuberculosis incidence and mortality from 2015 to 2020 to assess global progress towards the End TB Strategy milestones, with an emphasis on exploring progress by age for the first time. We also present age-specific, risk-deleted mortality estimates that represent deaths due to tuberculosis that would have been observed if the combined effects of all evaluated risk factors were removed, to provide insights into the role of addressing risk factors in achieving the End TB targets. Last, we analysed vital registration data only (2015–21 inclusive) to assess the potential impact of the COVID-19 pandemic on age-specific tuberculosis mortality in countries with available cause-specific data.

7Implications of all the available evidence

8Global tuberculosis control programmes could consider closely examining the 15 countries that achieved the 2020 incidence milestone and the 17 countries that achieved the 2020 mortality milestone, many of which were in western sub-Saharan Africa and eastern Europe, to better understand drivers of their marked progress. Many of these countries have implemented innovative approaches to active case finding while implementing social protection interventions such as advanced tuberculosis surveillance in high-risk areas and economic incentives for patients. We observed unequal progress by age groups and found particularly slow progress in reducing the tuberculosis burden in older adults (aged >50 years), indicating that tuberculosis in this age group should be more widely recognised and monitored to achieve WHO's 2035 End TB Strategy targets. One approach for reducing the tuberculosis burden in older adults and achieving the 2035 End TB targets is addressing risk factors for tuberculosis. Our assessment of risk-deleted mortality suggests that the world would have achieved the 2020 End TB mortality milestone, with a 36% reduction in global tuberculosis deaths and the largest benefits in older adults, if the risk factors of smoking, alcohol use, and diabetes were removed. Although complete elimination of such risk factors is highly unlikely, this assessment illustrates the magnitude of potential improvement. Finally, our secondary analysis of countries with complete vital registration data showed variable impacts of the COVID-19 pandemic on tuberculosis mortality in 2020 and 2021; some countries and age groups reported fewer than expected deaths due to tuberculosis, whereas others reported more deaths than expected. Additional data from countries with a high tuberculosis burden are urgently needed to understand the generalisability of these findings.

9The year 2020 marked a crucial timepoint for evaluating progress towards eliminating tuberculosis. The first quantitative milestones in the End TB Strategy called for a 35% reduction in deaths due to tuberculosis and a 20% reduction in the tuberculosis incidence rate by 2020. The 2022 Global Tuberculosis Report from WHO illustrated that tuberculosis deaths decreased by only 5·9%, while the tuberculosis incidence rate decreased by 10%, between 2015 and 2021, falling well short of the 2020 milestones. Moreover, studies examining global trends towards WHO targets have yet to fully consider age-specific differences in progress. The current evidence suggests that although the tuberculosis burden remains substantially high in children,, older age groups generally have the highest rates of tuberculosis mortality and incidence., A report of the Global Burden of Diseases, Injuries, and Risk Factors Study 2019 (GBD 2019) based on data from Cambodia also indicated that temporal trends of reductions in the tuberculosis burden become slower with increasing age. Several factors could explain these marked age-related differences, including age-specific diagnostic challenges,, age-specific mixing patterns, immune senescence and other age-related immune dysfunctions, and comorbidities common in older patients masking tuberculosis symptoms. Although largely unexplored at the global level, age-specific differences in risk factor prevalence might further augment these age-related discrepancies. Identification of age-specific progress towards the End TB Strategy milestones, combined with assessments of the role of key risk factors, could therefore help inform targeted interventions to accelerate progress towards eliminating tuberculosis.

10Leveraging the GBD 2021 framework, we aimed to examine the levels and trends in the global burden of tuberculosis to investigate age-specific attainment of the 2020 WHO End TB Strategy mortality and incidence milestones for 204 countries and territories. Considering that modifiable tuberculosis risk factors, such as alcohol consumption, smoking, and diabetes, can augment preventive interventions, we supplemented our analysis by examining age-specific risk-deleted tuberculosis mortality (ie, the tuberculosis mortality rate that would have been observed if risk factors for tuberculosis mortality were removed) to highlight the need to address reductions in risk factor exposure as part of any holistic response to achieving the End TB Strategy targets. Last, we examined the potential impact of the COVID-19 pandemic on tuberculosis mortality by drawing on vital registration data from those countries with available data. This manuscript was produced as part of the GBD Collaborator Network and in accordance with the GBD Protocol.

11We have previously published detailed methods of the GBD analytical framework, and tuberculosis burden estimation in GBD.,, Here, we summarise the methodology for estimating tuberculosis mortality and morbidity, as well as key risk factors, with more detailed descriptions of the modelling strategy provided in appendix 1 (pp 3–26). In compliance with the Guidelines for Accurate and Transparent Health Estimates Reporting, input data sources and codes for each step of the estimation process are available on the Global Health Data Exchange.

12The GBD 2021 Cause of Death database contained all available vital registration, surveillance system, and verbal autopsy data from 1980 to 2020. Leveraging the database, we included 22 603 site-years of vital registration data, 825 site-years of sample-based vital registration data, 680 site-years of mortality surveillance data, and 9 site-years of minimally invasive tissue sample (MITS) diagnoses data for modelling tuberculosis mortality among individuals without HIV co-infection. Country-specific data sources and citations are included in the Global Health Data Exchange. We processed raw cause of death data to account for completeness and differences in coding schemes, to redistribute deaths from unspecified codes to more specific underlying causes of death, and to reassign misclassified HIV deaths.

13We used the Cause of Death Ensemble modelling (CODEm) strategy to generate tuberculosis mortality estimates in individuals without HIV co-infection by location, year, age, and sex. CODEm is a hierarchical modelling platform that uses an ensemble of different modelling methods for rates or cause fractions with varying choices of covariates (eg, smoking prevalence, alcohol consumption, and Healthcare Access and Quality [HAQ] Index) that perform best with out-of-sample predictive validity testing. appendix 1 (p 13) provides example countries illustrating that final tuberculosis mortality estimates presented throughout this analysis largely follow trends in input data. Finally, we established age-specific tuberculosis deaths among individuals with HIV infection using a population attributable fraction (PAF) approach (appendix 1 pp 14–15).

14We simultaneously modelled age-sex-specific tuberculosis incidence, prevalence, and cause-specific mortality using the DisMod-MR 2.1 (disease-model-Bayesian meta-regression) modelling tool. DisMod-MR 2.1 is a Bayesian disease modelling tool that leverages all available morbidity and mortality data, the epidemiological relationships between disease parameters, and spatial relationships to output internally consistent disease burden estimates. We provide details of case definitions, input data sources, and data processing strategies in appendix 1 (pp 2–3, 16–26).

15Briefly, we identified all population-based tuberculosis prevalence surveys via comprehensive reviews of the literature. Similarly to previous GBD iterations, we used a Bayesian meta-regression tool to adjust prevalence surveys that used smear-positive tuberculosis as the case definition rather than bacteriologically confirmed tuberculosis. We further recalibrated surveys that used symptoms only as the screening method over both symptoms and chest X-ray (appendix 1 pp 22–23). Next, we maximised data informing DisMod-MR by predicting age-sex-specific incidence for countries with low-quality data ratings for cause of death data. We estimated incidence for these countries through a mortality-to-incidence ratio approach using countries with high-quality data ratings on cause of death data as inputs into a Bayesian meta-regression analysis where the HAQ Index was the primary covariate. We then linked location-specific predicted mortality-to-incidence ratios with tuberculosis death estimates to obtain estimated age-sex-specific incidence (appendix 1 p 20).

16We subsequently modelled these data, together with age-sex-specific case notifications for locations with high-quality data ratings on causes of death data, population-based tuberculin surveys, and estimates of tuberculosis excess mortality rate, remission, and cause-specific mortality rate, in DisMod-MR 2.1 to generate all-form tuberculosis morbidity estimates that were internally consistent. To further improve internal consistency of modelling in DisMod, we computed all-age and both-sexes priors of tuberculosis duration using a combination of tuberculosis duration data from a systematic review of studies during the pre-chemotherapy era and the HAQ Index. We then used DisMod's statistical triangulation approach, using the estimated tuberculosis duration priors and age-sex-specific tuberculosis mortality, prevalence, and incidence data, to derive final age-sex-specific tuberculosis duration estimates (appendix 1 pp 27–28). Finally, we disaggregated tuberculosis incidence by HIV status by applying the fraction of new all-form tuberculosis cases that were HIV and tuberculosis co-infections to our all-form tuberculosis cases estimated from DisMod-MR.

17GBD has previously published detailed methodology for risk factor estimation. In summary, age-sex-specific PAFs in adults were computed within the comparative risk assessments framework of GBD with the following inputs: prevalence estimates for exposure to risk factors; the relative risk of tuberculosis mortality as an outcome of exposure to each risk factor; and the theoretical minimum risk exposure level (TMREL), defined as the level of exposure that would minimise risk for each individual in a population. We computed tuberculosis mortality attributable in adults (aged ≥15 years) to each risk factor by multiplying the PAF by the number of tuberculosis deaths for each risk–outcome pair. We subsequently derived risk-deleted mortality rates to quantify tuberculosis deaths that would have been observed had the risk factors been set to their corresponding TMRELs. The risk-deleted tuberculosis mortality rates are therefore computed by multiplying the observed tuberculosis mortality by one minus the PAF for a risk factor or combination of risk factors. The objective of the risk-deleted analysis is to illustrate the potential magnitude of addressing tuberculosis risk factors across countries and age groups. For this analysis, we quantified tuberculosis PAFs for alcohol consumption, smoking, and diabetes among adults without HIV co-infection. We provide complete details of the methods for estimation of the tuberculosis risk factors in appendix 1 (pp 39–67).

18We evaluated the potential age-specific impact of the COVID-19 pandemic on tuberculosis mortality without HIV co-infection using vital registration data for all countries with available cause-specific mortality data in 2020 and 2021. This secondary analysis included data from 41 countries that reported at least ten tuberculosis deaths in 2019 for each age group of interest, representing four of the seven GBD super-regions (appendix 2 pp 300–04), with 15 countries from the high-income super-region, 14 in central Europe, eastern Europe, and central Asia, ten in Latin America and the Caribbean, and two in southeast Asia, east Asia, and Oceania that had sufficient cause-specific mortality data. Using these data as inputs, we fitted quasi-Poission regression models to data from 2015 to 2019 with population size as an offset to estimate the expected number of tuberculosis deaths in 2020–21 for each country. The quasi-likelihood was selected to account for overdispersion in count data. This analysis was stratified by adults and children combined (aged <65 years), elderly individuals (≥65 years), and all ages combined. We selected these age groups on the basis of evidence illustrating that global tuberculosis diagnoses during the pandemic declined relatively more in individuals aged 65 years and older than in those younger than 65 years. We subsequently compared age-specific observed tuberculosis deaths to predicted tuberculosis deaths in 2020 and 2021 to identify potential excess tuberculosis deaths. Comparisons were made by computing the difference in observed to expected tuberculosis deaths (excess tuberculosis deaths) and the respective ratio.

19We aggregated tuberculosis incidence and mortality estimates by HIV status to present all-form tuberculosis burden estimates for five age groups (<5 years, 5–14 years, 15–49 years, 50–69 years, and ≥70 years) throughout this analysis. Next, we used the GBD world population age standard to derive age-standardised rates for all-form tuberculosis incidence and mortality. For changes over time, we provide annualised rates of change (ARC) as the difference in the natural log of the incidence and mortality rates at the start and end of the time interval divided by the number of years in the interval. We present tuberculosis incidence-specific and mortality-specific ARCs for 1990–2010 and 2010–21. To evaluate attainment of the 2020 WHO End TB interim milestones, we derived percentage changes in all-form tuberculosis rates and counts of both incidence and mortality from 2015 to 2020. At each modelling step described, parameter uncertainty was incorporated by randomly drawing 500 samples from each age-sex-location-year-specific parameter distribution and propagating this uncertainty forward across each step of the analysis. Consistent with the GBD framework, we computed 95% uncertainty intervals (UIs) for all estimates based on the 2·5th and 97·5th percentiles of the final 500 draws. More detailed tuberculosis burden results by HIV status, age, and sex across locations and years are available in the GBD Results Tool. Counts are rounded to three significant figures while rates and percentages are rounded to one decimal place, except where greater precision was needed to differentiate between values.

20The funder of the study had no role in study design, data collection, data analysis, data interpretation, writing of the report, or the decision to submit the manuscript for publication.

21In 2021, we estimated that there were 9·40 million (95% UI 8·36–10·5) tuberculosis incident cases and 1·35 million (1·23–1·52) deaths due to tuberculosis globally (table 1). HIV and tuberculosis co-infection constituted 1·00 million (0·90–1·13; 10·7% [10·3–11·0]) of the 9·40 million all-form tuberculosis incident cases and 205 000 (158 000–248 000; 15·1% [11·9–17·7]) of the 1·35 million all-form global tuberculosis deaths. The age-standardised incidence rate for all-form tuberculosis was 115 (102–128) per 100 000 population and the age-standardised mortality rate was 16·2 (14·8–18·2) per 100 000 population in 2021 (table 1). From 1990 to 2021, the global age-standardised tuberculosis incidence rate decreased by 37·0% (34·5–39·1), while the global age-standardised tuberculosis mortality rate decreased by 61·1% (52·6–66·3).

22The global age distribution for the number of all-form tuberculosis incident cases in 2021 demonstrated a drop-off in cases after age 5 years, but a rapid spike in cases at age 15–24 years, followed by steady decreases until age 70–74 years, after which there was a rapid drop in cases (appendix 2 p 4). The corresponding age distribution of tuberculosis incidence rates showed a similar pattern until age 20–24 years, when the rate levelled off until a rapid increase from age 50–69 years, followed by a small drop (appendix 2 p 5). In 2021, 3·8% (95% UI 3·6–4·2) of incident cases were in those younger than 5 years, 4·7% (3·8–5·9) were in those aged 5–14 years, 54·9% (52·4–58·9) were in those aged 15–49 years, 26·5% (23·4–29·5) were in those aged 50–69 years, and 10·1% (9·2–11·0) were in those aged 70 years and older (table 1).

23The age distribution for tuberculosis deaths in 2021 showed a similar drop-off in deaths after age 5 years, with numbers rapidly increasing after age 19 years before levelling off at age 55–69 years and then sharply decreasing (appendix 2 p 4). For the tuberculosis mortality rate distribution, the rate dropped after age 5 years before steadily increasing for the remaining age groups until age 80–84 years when it levelled off (appendix 2 p 5). In 2021, 4·5% (95% UI 3·5–5·0) of global tuberculosis deaths were in children younger than 5 years, 1·5% (1·5–1·6) were in those aged 5–14 years, 36·4% (35·7–35·9) were in those aged 15–49 years, 33·2% (32·8–33·6) were in those aged 50–69 years, and 24·5% (24·6–24·9) were in those aged 70 years and older (table 1).

24In 2021, tuberculosis incidence rates were greater than 100 per 100 000 population in children younger than 5 years in 34 countries, and in children aged 5–14 years in 14 countries (appendix 2 pp 6, 8, 20–85). Tuberculosis incidence rates exceeded 500 per 100 000 population in those aged 15–49 years in nine countries, in those aged 50–69 years in 33 countries, and in those aged 70 years and older in 50 countries (appendix 2 pp 10, 12, 14, 20–85).

25In 2021, tuberculosis mortality rates exceeded 25 per 100 000 population in children younger than 5 years in 20 countries, in those aged 5–14 years in one country, and in those aged 15–49 years in 37 countries (appendix 2 pp 7, 9, 11, 20–85). Among older adults, tuberculosis mortality rates exceeded 300 per 100 000 in those aged 50–69 years in eight countries and in those aged 70 years and older in 44 countries (appendix 2 pp 13, 15, 20–85).

26Our assessment of the global ARC in tuberculosis incidence showed that children younger than 5 years and those aged 5–14 years had the largest ARCs, decreasing by 4·1% (95% UI 3·8–4·3) and 3·7% (3·4–4·0), respectively, between 2010 and 2021, while ARCs in the adult age groups decreased by around 2% in the same period (table 1). Similarly, children younger than 5 years and those aged 5–14 years had the largest ARCs for tuberculosis mortality rates, decreasing by 7·9% (5·8–9·6) and 6·6% (5·6–7·7), respectively, between 2010 and 2021, while ARCs in the adult age groups decreased by 4·0–4·7% in that period (table 1).

27At the super-region level, central Europe, eastern Europe, and central Asia (incidence ARC –4·3% [95% UI –4·8 to –3·8]; mortality ARC –8·7% [–9·3 to –8·1]) and sub-Saharan Africa (incidence ARC –4·1% [–4·2 to –3·8]; mortality ARC –6·1% [–7·2 to –5·1]) had the largest age-standardised ARCs in 2010–21 (table 1). We found that 23 countries had ARCs for age-standardised tuberculosis incidence rates that exceeded 4·5% in 2010–21, 11 of which were in sub-Saharan Africa and nine were in central Europe, eastern Europe, and central Asia (appendix 2 pp 20–85). Similarly, 23 countries had ARCs for age-standardised tuberculosis mortality rates that exceeded 7·0%; of these, 13 were in central Europe, eastern Europe, and central Asia and six were in sub-Saharan Africa.

28At the global level, the all-age tuberculosis incidence rate declined by 6·26% (95% UI 5·27 to 7·25) between 2015 and 2020 (table 2; figure 1). When stratified by sex, the tuberculosis incidence rate decreased by 7·9% (6·5 to 9·2) in females and by 4·9% (4·1 to 5·9) in males (appendix 2 p 18). Among both sexes, children younger than 5 years (16·5% [14·8 to 18·4]) and those aged 5–14 years (16·2% [14·2 to 17·9]) had the largest declines in incidence rate, while the percentage decline ranged between 5·72% and 8·48% for the other age groups (table 2).

29The number of global tuberculosis deaths decreased by 11·9% (95% UI 5·77 to 17·0) from 2015 to 2020 (table 2). The percentage change in the number of global tuberculosis deaths was greater in females (13·8% [8·4 to 18·0]) than in males (10·5% [0·1 to 18·1]; appendix 2 p 18). Among both sexes, children younger than 5 years (35·3% [26·7 to 41·7]) and those aged 5–14 years (29·5% [25·5 to 34·1]) had the largest declines in the number of deaths, while individuals aged 70 years and older (3·29% [–5·56 to 9·07]) had the lowest decline (figure 1).

30Among all countries, 15 achieved a 20% decrease in the all-age tuberculosis incidence rate between 2015 and 2020: eight of these were in western sub-Saharan Africa, three were in eastern sub-Saharan Africa, one was in central Europe, two were in central Asia, and one was in Andean Latin America (figure 2A; appendix 2 pp 86–149). When stratified by age groups, 66 countries achieved a 20% reduction in the all-age tuberculosis incidence rate in children younger than 5 years, as did 50 countries in those aged 5–14 years, 22 countries in those aged 15–49 years, 12 countries in those aged 50–69 years, and ten countries in those aged 70 years and older (appendix 2 pp 86–149). The only region that achieved a 20% decrease in the incidence rate was western sub-Saharan Africa (21·6% [95% UI 19·7 to 23·3]; appendix 2 pp 86–149). By 2021, 38 countries had achieved a 20% reduction in the all-age tuberculosis incidence rate (21 in sub-Saharan Africa; 11 in central Europe, eastern Europe, and central Asia; three in high-income countries; and three in the remaining GBD super-regions), with the sub-Saharan Africa super-region becoming the only super-region to achieve a 20% reduction (appendix 2 pp 150–213). Among the 20 high tuberculosis burden countries in GBD 2021, three countries (Tanzania, Kenya, and Nigeria) achieved a 20% decrease in the all-age incidence rate in 2020, with an additional three high tuberculosis burden countries (Myanmar, Uganda, and Zimbabwe) meeting the target in 2021 (appendix 2 p 16).

31Eastern Europe (43·8% [95% UI 40·7 to 45·9]) was the only region that achieved a 35% decline in the number of tuberculosis deaths between 2015 and 2020 (appendix 2 pp 86–149). 19 countries and territories achieved this milestone: six in eastern Europe, four in central Europe, four in central Asia, two in the Caribbean, one in Andean Latin America, one in Australasia, and one in western sub-Saharan Africa (figure 2B; appendix 2 pp 86–149). Central Europe, eastern Europe, and central Asia was the only GBD super-region to achieve a 35% reduction in the number of tuberculosis deaths by 2020 (38·8% [36·5–40·6]; table 2). Regarding age groups, 95 countries achieved a 35% decrease in the number of tuberculosis deaths in children younger than 5 years, as did 55 countries in those aged 5–14 years, 26 countries in those aged 15–49 years, 20 countries in those aged 50–69 years, and six countries in those aged 70 years and older. In 2021, two additional countries (ie, 19 in total) achieved a 35% decrease in tuberculosis deaths (appendix 2 pp 150–213). None of the countries with a high tuberculosis burden achieved a 35% decrease in tuberculosis deaths in 2020 or 2021, but the Democratic Republic of the Congo (23·7% [11·9–34·9]), Myanmar (22·9% [6·94–33·0]), and Tanzania (22·5% [10·7–34·0]) had the largest reductions between 2015 and 2021 (appendix 2 p 17).

32After removing the combined effects of three tuberculosis risk factors (smoking, alcohol use, and diabetes), the number of global all-form tuberculosis deaths would decrease from 1·39 million (95% UI 1·28–1·54) to 1·00 million (0·703–1·23) in 2020 (table 3). Comparing the observed number of tuberculosis deaths in 2015 to the risk-deleted number of tuberculosis deaths in 2020 represents a 36·5% (21·5–54·8) decrease. The corresponding percentage declines would be 31·0% (15·9–48·8) in those aged 15–49 years, 40·5% (22·3–62·6) in those aged 50–69 years, and 41·0% (25·4–62·0) in those aged 70 years and older. If the effects of risk factors were removed, 13 additional regions would have joined eastern Europe in achieving a 35% decline in the number of tuberculosis deaths. Furthermore, the cumulative number of countries achieving a 35% decrease would have increased from 15 to 132 in 2020 and 143 in 2021 (appendix 2 pp 214–99). A 35% reduction would have been achieved in 104 countries for individuals aged 15–49 years, in 144 countries for those aged 50–69 years, and in 169 countries for those aged 70 years and older in 2020.

33For the 41 countries included in our analysis of COVID-19 impact, 50 900 (95% CI 49 700 to 52 400) tuberculosis deaths without HIV co-infection were expected across all ages in 2020, compared to an observed 45 500 tuberculosis deaths (figure 3A; appendix 2 pp 300–04). This corresponded to 5340 (4070 to 6920) fewer tuberculosis deaths than expected (negative excess tuberculosis deaths) and an observed to expected (OE) ratio of 0·90 (0·87 to 0·92). Individuals aged 65 years and older contributed more to the negative excess deaths (–4330 [–5310 to –3370], OE ratio 0·80 [0·77 to 0·84]) than did those younger than 65 years (–1000 [–2150 to –127], OE ratio 0·97 [0·93 to 1·00]). We estimated that 28 of 41 countries had negative excess tuberculosis deaths, with a combined contribution of 6000 (4770 to 7400) fewer tuberculosis deaths and an OE ratio of 0·86 (0·83 to 0·89). Those aged 65 years and older also contributed more negative excess deaths (–4340 [–5310 to –3370], OE ratio 0·78 [0·75 to 0·82]) than did those younger than 65 years (–1660 [–2580 to –817], OE ratio 0·93 [0·89 to 0·96]) in these 28 countries. 13 countries had excess tuberculosis deaths, with a combined contribution of 669 (–63 to 1030) excess tuberculosis deaths and an OE ratio of 1·08 (0·99 to 1·13). However, in these countries individuals younger than 65 years contributed almost all the excess tuberculosis deaths (660 [–60 to 1010], OE ratio 1·11 [0·99 to 1·17] vs eight [–160 to 126], OE ratio 1·00 [0·92 to 1·07], in those aged ≥65 years).

34There were 20 countries reporting data through the year 2021 (figure 3B; appendix 2 pp 300–04). In these countries, 39 600 (95% CI 38 300 to 41 100) tuberculosis deaths without HIV co-infection were expected compared to an observed 39 000 tuberculosis deaths (appendix 2 pp 300–04). This corresponded to 657 (–713 to 2180) fewer tuberculosis deaths than expected, with an OE ratio of 0·98 (0·95–1·02). There were marked differences by age as people younger than 65 years contributed excess tuberculosis deaths (2290 [1100 to 3370], OE ratio 1·10 [1·04 to 1·15]), whereas those aged 65 years and older had negative excess tuberculosis deaths (–2950 [–3920 to –2100], OE ratio 0·81 [0·76 to 0·86]). Only five countries contributed negative excess deaths, with a combined contribution of 2310 (1100 to 3680) fewer tuberculosis deaths and an OE ratio of 0·90 (0·85 to 0·95). In those countries, people aged 65 years and older contributed negative excess deaths (–2880 [–3860 to –2090], OE ratio 0·72 [0·66 to 0·78]), whereas those younger than 65 years did not (574 [–467 to 1602], OE ratio 1·04 [0·97 to 1·13]). The remaining 15 countries had a combined contribution of 1650 (737 to 2100) excess tuberculosis deaths and an OE ratio of 1·10 (1·04 to 1·13). Among these 15 countries, people younger than 65 years contributed almost all excess tuberculosis deaths (1720 [824 to 2097], OE ratio 1·16 [1·07 to 1·20]), whereas those aged 65 years and older contributed 71 (–213 to 395) fewer deaths than expected (OE ratio 0·99 [0·93 to 1·04]).

35Despite accelerated global progress in reducing the burden of tuberculosis in the past decade, we observed that the first WHO End TB interim milestones in 2020 were not achieved at the global level: global tuberculosis incidence decreased by only 6% and global tuberculosis deaths decreased by 12% between 2015 and 2020. The substantial gap in reaching the global milestones is in keeping with observations from other studies.,,, We further observed that progress was differential with respect to age such that individuals younger than 15 years had the sharpest declines while the older adult age groups had only minimal reductions. Our analysis of risk-deleted mortality suggests that reducing tuberculosis risk factors will be important in any holistic strategy for reducing tuberculosis mortality among adults and achieving the optimistic End TB milestones.

36The finding that only 15 countries achieved the 2020 tuberculosis incidence milestone in 2020 underscores that substantial challenges remain in identifying all individuals with tuberculosis in communities and initiating treatment. When individuals with tuberculosis remain undiagnosed, they might unknowingly propagate infections to healthy contacts. Tuberculosis control programmes could consider enhancing interventions that have previously been shown to improve detection and mitigate delays in treatment initiation, including active case finding, tuberculosis outreach screening with communication campaigns, and the use of new rapid tuberculosis diagnostic tools. Tuberculosis control programmes could also consider exploring the programmes running in sub-Saharan Africa, since 11 of the 15 countries that achieved the incidence milestone in 2020 were in this region. For example, Nigeria achieved a 22% decrease (appendix 2 pp 86–149), indicating that its novel active case-finding interventions with advanced surveillance in tuberculosis hotspots, increase in diagnostic equipment and partnering with the private sector, and economic incentives for patients were effective in helping to achieve the milestone. In Tanzania, which achieved a 26% reduction (appendix 2 pp 86–149), the national tuberculosis programme scaled up community-based interventions aiming to link individuals to tuberculosis services with proven effectiveness while introducing an effective nationwide quality-improvement initiative in tuberculosis case finding. The marked progress in Cameroon (30% reduction in incidence) and Kenya (23% reduction in incidence; appendix 2 pp 86–149) might also point to the success of intensified tuberculosis case-finding programmes that were reported to increase case detection in these respective countries. In addition to these innovative case-finding programmes, many countries in sub-Saharan Africa saw marked increases in antiretroviral therapy coverage for people with HIV between 2015 and 2020 that might have further contributed to the substantial reductions in tuberculosis incidence in the region.

37We observed that only 17 countries achieved the End TB mortality milestone in 2020, indicating continued suboptimal treatment outcomes globally, potentially owing to poor early linkage to treatment, inadequate support for people with tuberculosis to ensure treatment adherence, unidentified drug resistance, and inadequate prevention of advanced HIV disease. Previous research has indicated that treatment adherence interventions (eg, patient counselling and education, reminders, and digital health technologies), and social protection interventions (eg, cash transfers and economic incentives), can increase treatment adherence and prevent mortality. A new tuberculosis treatment regimen for drug-resistant tuberculosis, endorsed by WHO, that is efficacious, safe, and shorter than previous regimens could have a role in reducing tuberculosis deaths. Additional research should examine the marked progress in reducing tuberculosis deaths in the 17 countries that achieved the End TB mortality milestone in 2020. For instance, national social protection interventions in Moldova (which achieved a 48% reduction) and Ecuador (which achieved a 41% reduction; appendix 2 pp 86–149) markedly increased treatment adherence, potentially prevented drug resistance, and might have played a role in the countries' progress.

38Most importantly, our results showed differential progress with respect to age in reducing the burden of tuberculosis. We found that children (aged <15 years) had the largest reductions in tuberculosis burden, with a 16% reduction in the incidence rate and 34% decrease in deaths between 2015 and 2020. This finding is in keeping with a recent study done in Cambodia, which used GBD 2019 tuberculosis estimates to show greater reductions in the tuberculosis burden in children than in older adults. The substantial progress in reducing the tuberculosis burden in children might reflect recent trends showing marked decreases in child mortality, global improvements in child nutrition outcomes,, increasing tuberculosis case finding and services for children,,, and continued reductions in vertical HIV transmission., Despite these improvements, the burden of tuberculosis in children remains high, with an estimated 89 800 deaths in 2020. With recent evidence suggesting that most tuberculosis transmission among children occurs outside the household, integrating effective household contact-tracing programmes, with community-based strategies (eg, routine or mass screening, targeted preventive therapy, and environmental interventions) will be particularly important for continued progress.

39While individuals aged 50 years and older constituted 37% of all tuberculosis incident cases and 58% of all tuberculosis deaths in 2020, these age groups have shown slow progress in reducing the tuberculosis burden since 2015. These results suggest that tuberculosis in older adults poses a major challenge to tuberculosis control, and achieving the End TB targets will therefore require targeting these age groups. Previous studies have shown several challenges in controlling tuberculosis in older individuals. For example, older individuals are more likely to progress to active tuberculosis, they do not have classical clinical features, which has been shown to delay diagnosis and treatment, and they are more likely to have adverse reactions to treatment. Studies show that screening and treating older adults for latent tuberculosis is essential for meeting the 2025 interim End TB milestones. However, the high probability of adverse reactions to treatment, potentially contributing to higher likelihoods of interruptions to tuberculosis preventive therapy in this age group, underscores the importance of developing shorter, less toxic treatments, along with early diagnosis, for achieving the End TB targets.

40Our study highlights that one opportunity for reducing the tuberculosis burden among older adults and achieving the End TB targets is by better addressing tuberculosis risk factors. We found that eliminating tuberculosis risk factors would reduce global tuberculosis deaths by 37% in 2020 compared with 2015, with the largest reductions expected in individuals aged 50 years and older. Implementation of WHO's recent framework for collaborative action on tuberculosis and comorbidities across health systems will be important for addressing tuberculosis risk factors. The framework provides an evidence-based and patient-centred approach for coordinating with risk factor control initiatives throughout the continuum of tuberculosis care, including prevention, diagnosis and treatment, and care after the completion of tuberculosis treatment. A successful case study of the programme is Pakistan, which integrated smoking cessation support with routine tuberculosis care in 2017 and found that a quarter of participants stopped smoking. Mexico's innovative tuberculosis and diabetes partnership is another example that provides bidirectional services across the continuum of care and has been shown to improve tuberculosis treatment outcomes while screening 85% of all diagnosed tuberculosis cases for diabetes.,

41However, eliminating risk factors would still leave 1·00 million global tuberculosis deaths in 2020. This estimate underscores that other approaches, combined with addressing risk factors, will be important for reaching the upcoming End TB milestones. These approaches include considering the social determinants of tuberculosis, such as poverty, poor housing, overcrowding, undernutrition, and poor access to health care, which create environments conducive to widespread transmission. Tuberculosis subsequently augments existing social and economic vulnerabilities as it creates substantial losses in productivity and catastrophic financial costs for people living in poverty. The universal health coverage agenda will therefore be important to ending tuberculosis as it will ensure uninterrupted availability of and access to tuberculosis diagnostic tests and treatment for people at highest risk of tuberculosis.

42Finally, our secondary analysis of the impact of the COVID-19 pandemic on tuberculosis burden underscores the need for continued research and data. Although many models have predicted hundreds of thousands of additional tuberculosis deaths due to the COVID-19 pandemic,, our analyses of the available vital registration data have so far indicated variable impact in 2020 and 2021. Studies have shown substantial disruptions to tuberculosis health services, but a review of the impact of the COVID-19 pandemic on tuberculosis indicated that there are limited empirical data on the effects of the pandemic on tuberculosis outcomes. In the context of our analysis of 41 countries, we found many countries reporting lower than expected deaths, while a few countries had more deaths than expected in the absence of the COVID-19 pandemic, with Russia reporting an 8% increase in deaths in 2020 and a 15% increase in 2021, and Brazil reporting a 10% increase in 2021 (appendix 2 300–04). This analysis provides some indication that previously described disruptions in tuberculosis services and treatment in both Russia, and Brazil might have led to increases in tuberculosis mortality. However, this hypothesis does not explain why almost all the observed excess deaths occurred in individuals younger than 65 years. For example, this age group contributed all excess tuberculosis deaths in Russia and the USA in 2020 and 462 of the 480 excess deaths in Brazil in 2021 (appendix 2 300–04). If disruptions in tuberculosis services led to increases in tuberculosis mortality, then we would have seen increases in tuberculosis deaths across all age groups, including older age groups (≥65 years). Since this was not observed in our study, one potential explanation for the excess deaths in individuals younger than 65 years might be misclassification of COVID-19 deaths as tuberculosis deaths. A recent study in Brazil found marked levels of misclassified COVID-19 deaths in garbage codes and hypothesised that there are additional misclassified COVID-19 deaths in other respiratory diseases. Several other studies have also shown potential misclassified COVID-19 deaths in non-COVID-19 respiratory conditions,, which might be amplified in locations with scarce or inadequate SARS-CoV-2 diagnostic capabilities. Further investigations and scrutiny of cause of death data are urgently needed to understand the degree to which COVID-19 deaths were misassigned to tuberculosis and other respiratory conditions. Additionally, although several studies have shown disruptions to tuberculosis services, many found that tuberculosis treatment success rates did not change during the pandemic., With additional evidence showing that the average duration until cure or death is 3 years for untreated patients with tuberculosis, this timeframe might have also provided opportunities for tuberculosis service recovery programmes,,, to mobilise and avert mortality.

43Our results showed that many countries had negative excess (ie, lower than expected) deaths due to tuberculosis throughout the pandemic. For example, we found negative excess tuberculosis deaths in both 2020 and 2021 when aggregating across 41 of 204 countries, with notable reductions in the Philippines (a 14% reduction in 2020 and a 10% reduction in 2021), Peru (a 37% decrease in 2020), South Korea (a 10% decrease in 2020), and many other countries contributing negative excess deaths in both years (appendix 2 300–04). Other studies with empirical data have also reported similar reductions during the pandemic, with one study in South Africa reporting a decrease in undiagnosed tuberculosis deaths; another in Madurai, India, reporting a 50% decrease in deaths due to infectious diseases (which included deaths from tuberculosis); a study in Taiwan (province of China) reporting continued gradual decreases in tuberculosis mortality during the pandemic years; and a recent study in Indonesia showing a pre-pandemic to pandemic all-cause mortality rate ratio of 0·97 (95% CI 0·91–1·04) among people diagnosed with tuberculosis. These results suggest that public health measures for COVID-19 and mask wearing might have helped to prevent tuberculosis transmission, as observed with influenza and respiratory syncytial virus. Interestingly, our results showed that in most countries with reductions in tuberculosis mortality, the decreases were driven by the group aged 65 years and older. This observation provides some support for the hypothesis that individuals with tuberculosis and SARS-CoV-2 co-infection are more prone to dying from COVID-19 than individuals with SARS-CoV-2 infection only., Considering that the cumulative percentage of the global population with SARS-CoV-2 infection at least once during the first 2 years of the pandemic was higher than 40%, that mortality due to COVID-19 is highest among older people, and the high fatality rates in individuals with tuberculosis and SARS-CoV-2 co-infection,, older individuals with co-infection, particularly those who were undiagnosed with tuberculosis at the beginning of the pandemic, might have died from COVID-19. Emerging evidence from India's national tuberculosis prevalence survey between 2019 and 2021 provides some support for potential reductions in transmission, as the study showed a marked reduction in tuberculosis prevalence coinciding with the peak in COVID-19 deaths during the second wave in the country. Although our analysis of vital registration data from 41 countries overall showed a variable impact of the COVID-19 pandemic on tuberculosis mortality, continued research is needed to accurately quantify this association, with a focus on potential misclassification of COVID-19 deaths as tuberculosis deaths and on the potential reduction in the numbers of patients with tuberculosis owing to patients with tuberculosis dying from COVID-19.

44Insufficient empirical data on the impact of the COVID-19 pandemic on the tuberculosis burden is the primary reason we observed discrepancies in estimates between GBD 2021 and WHO's 2022 Global Tuberculosis Report for the years 2020 and 2021. Although both groups had similar estimates in 2019 (eg, both estimated approximately 1·4 million deaths), the tuberculosis results diverge in 2020–21, with GBD 2021 estimating 1·39 million (95% UI 1·28–1·54) deaths in 2020 and 1·35 million (1·23–1·52) deaths in 2021, whereas WHO estimated 1·49 million deaths in 2020 and 1·59 million deaths in 2021. Tuberculosis incidence estimates for 2021 were slightly lower in GBD 2021 (9·40 million [95% UI 8·36–10·5] cases) than in WHO's 2022 report (10·6 million [95% CI 9·9–11] cases) but with overlapping uncertainty intervals. WHO relied on mathematical models to generate tuberculosis incidence and death estimates for 26 countries in 2020–21. Due to the scarcity of empirical data, WHO made several assumptions, including assuming that any decreases in notification were solely attributed to delays in tuberculosis diagnosis and treatment. For GBD 2021, we did not make any changes to our modelling approach for estimation in 2020–21 as we could not validate potential assumptions with empirical data. For the years 2020–21, we used the same estimation framework as described in the methods to generate estimates for other years. The differing estimation approaches in 2020–21 subsequently also caused differences between GBD 2021 and WHO's 2022 report in estimates of the percentage changes towards the End TB milestones in 2021 (deaths: 5·9% for WHO vs 14·0% for GBD 2021; incidence: 10·0% for WHO vs 8·0% for GBD 2021; appendix 2 pp 150–213). As more empirical data on tuberculosis deaths and incidence during the pandemic become available, GBD and WHO estimates are expected to align more.

45This study has multiple limitations. The primary limitation is data availability as there were gaps in data across countries, age groups, and years. In the absence of data, estimates were dependent on covariates with evidence of a biological relationship or strong relationship with tuberculosis, spatial relationships, and out-of-sample predictive validity assessments. The absence of data is reflected in larger uncertainty intervals. These data limitations also create challenges in our statistical triangulation approach, particularly in sub-Saharan Africa where data are more scarce, where our model attempts to triangulate between tuberculosis mortality and prevalence. In this study, we were unable to quantify tuberculosis mortality attributable to all important risk factors, with indoor air pollution and malnutrition not included. We plan to include malnutrition as a risk factor in future iterations of GBD as novel robust data establishing a causal link become available. We were also unable to quantify the burden attributable to the social determinants of tuberculosis (eg, poverty and educational attainment). Additionally, PAF estimations assume that there is a causal relationship between an exposure and an outcome without confounding, that exposure removal does not impact the distribution of unrelated risk factors, and that a feasible intervention to eliminate exposure exists. However, these assumptions might not always hold true in practice. Studies have also highlighted the potential for bias in Levin's formula for PAF calculation that we used in this analysis,, but the bias when using adjusted relative risks, as we have done here, is often small. Moreover, although we accounted for potential biases when quantifying the relationship between risk factors and tuberculosis through the inclusion of bias covariates, these covariates might not entirely identify and rectify bias if all or most input studies are inherently biased. Finally, we could not quantify the impact of the COVID-19 pandemic on the tuberculosis burden owing to insufficient empirical evidence of impact. Although the Institute for Health Metrics and Evaluation has estimated that there were 18 million excess all-cause deaths for the years 2020–21, there is a scarcity of data on what fraction of these, if any, are due to tuberculosis. Our assessment of the available vital registration data indicated that the COVID-19 pandemic did not significantly change tuberculosis mortality in 2020–21 in most countries. However, this analysis was restricted to 41 countries that had available data, which subsequently restricts the generalisability of our findings. The release of more data from other countries, particularly in settings with a high tuberculosis burden, and more years of data are urgently needed to understand the generalisability of our findings. In the next iteration of GBD, we plan to include new vital registration data, emerging prevalence surveys, and other empirical data sources to systematically quantify the impact of COVID-19 on tuberculosis incidence and mortality.

46In conclusion, despite substantial progress in reducing the global burden of tuberculosis, our results show that the world did not achieve the 2020 interim End TB incidence and mortality milestones. Tuberculosis control programmes could consider closely evaluating those countries that achieved the 2020 milestones to better understand the drivers of their marked progress. We also observed that the pace of decline was unequal across age groups. Targeting the burden of disease in older adults (aged >50 years) will be crucial for achieving the upcoming End TB targets as this population represents a large share of the tuberculosis burden and has experienced minimal progress. Our analyses indicate that addressing risk factors, by collaborating with risk factor control initiatives, will also be important to reducing tuberculosis mortality among older adults and could be crucial for achieving the next End TB milestones. Finally, the impact of the COVID-19 pandemic on the global tuberculosis burden remains heterogenous and uncertain. Empirical data on the effects of the COVID-19 pandemic on tuberculosis are urgently needed.

47To download the data used in these analyses, please visit the Global Health Data Exchange website at https://ghdx.healthdata.org/record/ihme-data/gbd-2021-tuberculosis-incidence-mortality-1990-2021.

48J R Andrews reports grants or contracts from the US National Institutes of Health. Z Basharat reports other financial interests as an employee of Alpha Genomics, outside the submitted work. P J G Bettencourt reports other financial or non-financial interests as a project reviewer at Botnar Foundation, outside the submitted work. L Cegolon reports grants or contracts from ORCHESTRA project WP5 (grant agreement number 101016167); payment or honoraria for lectures, presentations, speakers bureaus, manuscript writing, or educational events from the University of Trieste (Italy) and University Pavia (Italy) academic lectures; and support for attending meetings or travel from the National Congress of the Italian Society of Occupational Medicine (Genoa: September, 2022; Turin: September, 2023) and the National Congress of the Italian Society of Public Health Medicine (Rome: May, 2023), all outside the submitted work. I Ilic reports support for this study from the Ministry of Education, Science and Technological development, Serbia (project number 175042, 2011–23). M Ilic reports support for this study from the Ministry of Education, Science and Technological development, Serbia (project number 451-03-47/2023-01/200111). N E Ismail reports unpaid leadership or fiduciary roles in board, society, committee, or advocacy groups, with the Malaysian Academy of Pharmacy as the Bursary and Council Member, outside the submitted work. K Krishan reports non-financial support from the UGC Centre of Advanced Study, CAS II, awarded to the Department of Anthropology, Panjab University (Chandigarh, India), outside the submitted work. M Lee reports support for this study from the Ministry of Education of South Korea and the National Research Foundation of Korea (NRF-2021R1I1A4A01057428) and Bio-convergence Technology Education Program through the Korea Institute for Advancement Technology (KIAT) funded by the Ministry of Trade, Industry and Energy (number P0017805). R J Maude reports support for this study from the Wellcome Trust, (grant number 220211) as it provides core funding for Mahidol Oxford Tropical Medicine Research Unit and contributes to their salary. L Monasta reports support for this study from the Italian Ministry of Health (Ricerca Corrente 34/2017), and payments made to the Institute for Maternal and Child Health IRCCS Burlo Garofolo. S Nomura reports support for this study from the Ministry of Education, Culture, Sports, Science and Technology of Japan (grant 21H03203) and from Precursory Research for Embryonic Science and Technology from the Japan Science and Technology Agency (grant JPMJPR22R8). V C F Pepito reports grants or contracts from the International Initiative for Impact Evaluation (3ie) and Sanofi Consumer Healthcare. L F Reyes reports grants or contracts, consulting fees, and payments or honoraria from MSD, GSK, and Pfizer, and support for attending meetings from GSK, outside the submitted work. O Rezahosseini reports grants or contracts from Rigshospitalet and A.P.Møllers fonden. J Ross reports grants or contracts from the US National Institutes of Health, the US Department of Veterans Affairs, and Merck. J P Silva reports salary support for this study from the Portuguese Foundation for Science and Technology. J A Singh reports consulting fees from AstraZeneca, Crealta/Horizon, Medisys, Fidia, PK Med, Two labs, Adept Field Solutions, Clinical Care Options, ClearView Healthcare Partners, Putnam Associates, Focus Forward, Guidehouse consulting (formerly Navigant), Spherix, MedIQ, Jupiter Life Science, UBM, Trio Health, Medscape, WebMD, Practice Point Communications, and the US National Institutes of Health and the American College of Rheumatology; payment of honoraria for lectures, presentations, speakers bureaus, manuscript writing, or education events as a member of the speaker's bureau of Simply Speaking; support for attending meetings as a past steering committee member of OMERACT (an international organisation that develops measures for clinical trials and receives funding from 12 pharmaceutical companies); participation on a data safety monitoring board or advisory board with the US Food and Drug Administration Arthritis Advisory Committee; leadership or fiduciary role in other board, society, committee, or advocacy groups, paid as a past steering committee member of OMERACT, unpaid as a Co-Chair of the Veterans Affairs Rheumatology Field Advisory Committee, and unpaid as an editor and Director of the UAB Cochrane Musculoskeletal Group Satellite Center on Network Meta-analysis; stock or stock options in atai Life Sciences, Kintara Therapeutics, Intelligent Biosolutions, Acumen Pharmaceuticals, TPT Global Tech, Vaxart Pharmaceuticals, Aytu BioPharma, Adaptimmune Therapeutics, GeoVax Labs, Pieris Pharmaceuticals, Enzolytics, Seres Therapeutics, Tonix Pharmaceuticals Holding, and Charlotte's Web Holdings; and previous stock options in Amarin, Viking, and Moderna Pharmaceuticals, all outside the submitted work. L K Stafford reports support for this study from the Bill & Melinda Gates Foundation and the Institute for Health Metrics and Evaluation (IHME). Z Wang reports grants or contracts from the Fred Hollows Foundation, Fonds de recherche du Québec - Santé, China Scholarship Council, and McGill University Global Health Scholars Program; consulting fees from the Fred Hollows Foundation; support for attending meetings or travel from McGill University (Graduate Mobility Award) and the Fred Hollows Foundation; and a leadership or fiduciary role in other board, society, committee, or advocacy groups, paid or unpaid as a member of the trainee advisory committee with the Consortium of Universities for Global Health, outside the submitted work. H Zhang reports grants or contracts from WHO research grants. M Zielińska reports other financial interests as an AstraZeneca employee, outside the submitted work.