COMMENT|ONLINE FIRST
Global Burden of Disease Study 2021 estimates: implications for health policy and research
Zachary J Ward, Sue J Goldie
Lancet Published: April 17, 2024
DOI: https://doi.org/10.1016/S0140-6736(24)00812-2
Over the past three decades, the Global Burden of Diseases, Injuries, and Risk Factors Study (GBD) has produced several iterations of global estimates for various disease metrics.1 The latest iteration, GBD 2021, published in The Lancet as a series of Articles, includes estimates of the global disease burden including incidence, prevalence, and disability-adjusted life-years (DALYs) for 371 diseases and injuries, at both the country level and subnational locations between 1990 and 2021.2
The GBD 2021 Diseases and Injuries Collaborators2 found that, although global DALYs increased from 2·63 billion (95% uncertainty interval 2·44–2·85) in 2010 to 2·88 billion (2·64–3·15) in 2021, the increase was largely due to population growth and ageing. Indeed, age-standardised DALY rates were estimated to have decreased between 2010 and 2021, with a decrease for communicable, maternal, neonatal, and nutritional diseases of 12·9% (5·1–19·2), with the biggest decreases seen for HIV/AIDS (decrease of 47·8% [43·3–51·7]) and diarrhoeal diseases (decrease of 47·0% [39·9–52·9]), but much smaller decreases were seen for non-communicable diseases (6·4% [3·5–9·5]).
Although health gains were estimated for other leading communicable diseases, these GBD estimates are the first since the COVID-19 pandemic, making this iteration an important opportunity to assess both the acute and longer-term health impacts of the pandemic. In addition to the enormous loss of life experienced globally, the myriad of effects of post-COVID-19 condition (long COVID; which can affect nearly every organ system) continue to affect the health of many people worldwide.3 Understanding and estimating the disease burden of these various sequelae will be an important challenge in the coming years to more fully characterise the cumulative global impact of the COVID-19 pandemic, with future work able to build on the initial estimates provided in this study.
In addition to quantifying the burden of emerging diseases, GBD provides estimates of the disease burden from 371 diseases and injuries, highlighting how the disease landscape varies by location and has changed over time. Macro trends in causes such as obesity, cardiovascular disease, and opioids started before the COVID-19 pandemic, and will continue to affect global health in years to come. GBD 2021's comprehensive assessment of the relative burden attributable to each disease can help identify the largest contributors to morbidity and mortality and how they vary by context (eg, location, sex, and age) and over time, which can inform long-term research priorities and investments for both public funding agencies and private philanthropy efforts.
GBD estimates can also be used to track progress towards existing targets, at global, regional, and country levels. For example, GBD 2021 can be used to assess progress towards the UN Sustainable Development Goals. However, the iterative nature of GBD can complicate tracking progress with respect to a baseline year. Because each new iteration of GBD re-estimates values for all previous years, the baseline values (eg, for 1990) can change from iteration to iteration. Although the updated historical estimates from more recent iterations of GBD might be more accurate than earlier estimates, updating the baseline values can complicate tracking progress over time. Therefore, policy makers should be aware of this iterative feature of the GBD estimates when setting targets and evaluating progress.
GBD has spent three decades estimating disease burden, but actionable guidance on how to reduce this burden is also needed to make progress in global health. Although structural models (eg, microsimulation models, agent-based models, and discrete event simulation models) that focus on specific diseases do not have the comprehensive scope of the GBD analytical approach, the causal associations specified in these models mean that they are better suited to the types of counterfactual policy analyses that are often of interest to decision makers and that can be used to supplement the epidemiological estimates available from GBD. In addition to performing counterfactual analyses, structural models can offer more robust predictions for complex systems than aggregate statistical models,4 and also have an advantage with regard to methodological transparency, with more straightforward interpretations of model assumptions, structures, and parameters than non-parametric approaches, such as Gaussian process regression.5
The estimates of disease burden by country and subnational locations provided as part of GBD 2021 are also important for characterising health disparities both across and within countries. However, the current methodology has some limitations in assessing the disease burden of multimorbidity, which might be underestimated because the microsimulation modelling approach assumes independence of comorbidities.2 These estimates can help to shed light on the relative burden of each disease in a population (ie, marginal prevalence), but might not necessarily capture the total burden of disease experienced by individuals or subgroups in the population (ie, correlation of multiple diseases or conditions). The clustering of disease burdens (especially non-communicable diseases) among individuals or within subgroups has implications for the design and implementation of interventions to improve population health and health equity, and could be an important area of methodological innovation for future iterations of GBD, as has been acknowledged by the authors.2
The sustained high burden of non-communicable diseases suggests that prevention and management together with health system strengthening will continue to be crucial policy priorities in the years to come. Although the effects of the COVID-19 pandemic highlight that large, unforeseen shocks can occur, many important risk factors and disease outcomes follow more stable trajectories than infectious disease epidemics, such as demographic trends (eg, fertility patterns and ageing populations), increasing BMI estimates, and non-communicable diseases such as cardiovascular disease and cancer.
To better inform policy efforts to address these issues, different types of modelling (ie, estimation and counterfactual policy analysis) and types of evidence (ie, intervention effectiveness, costs, and feasibility) based on robust empirical data should be synthesised. Although estimates of disease burden are necessary, alone they are not sufficient for public health priority setting and planning. The effectiveness and costs of potential interventions to ameliorate various disease burdens are also crucial for policy considerations. For example, although a particular disease might have a large burden, that does not necessarily imply that it should receive the largest amount of resources, especially if there are few (or even no) cost-effective interventions available. Although GBD 2021 estimates of disease burden can be an important input for policy making, resource allocations that optimise global health outcomes also need to consider the cost-effectiveness and feasibility of specific interventions, and how they might differ by setting.
