Burden of disease scenarios for 204 countries and territories, 2022–2050: a forecasting analysis for the Global Burden of Disease Study 2021

2021 GBD Forecasting Collaborators, Jing Sun

Research output: Contribution to journalArticlepeer-review

Abstract

Background Future trends in disease burden and drivers of health are of great interest to policy makers and the public
at large. This information can be used for policy and long-term health investment, planning, and prioritisation. We
have expanded and improved upon previous forecasts produced as part of the Global Burden of Diseases, Injuries,
and Risk Factors Study (GBD) and provide a reference forecast (the most likely future), and alternative scenarios
assessing disease burden trajectories if selected sets of risk factors were eliminated from current levels by 2050.
Methods Using forecasts of major drivers of health such as the Socio-demographic Index (SDI; a composite measure
of lag-distributed income per capita, mean years of education, and total fertility under 25 years of age) and the full set
of risk factor exposures captured by GBD, we provide cause-specific forecasts of mortality, years of life lost (YLLs),
years lived with disability (YLDs), and disability-adjusted life-years (DALYs) by age and sex from 2022 to 2050 for
204 countries and territories, 21 GBD regions, seven super-regions, and the world. All analyses were done at the
cause-specific level so that only risk factors deemed causal by the GBD comparative risk assessment influenced future
trajectories of mortality for each disease. Cause-specific mortality was modelled using mixed-effects models with SDI
and time as the main covariates, and the combined impact of causal risk factors as an offset in the model. At the allcause mortality level, we captured unexplained variation by modelling residuals with an autoregressive integrated
moving average model with drift attenuation. These all-cause forecasts constrained the cause-specific forecasts at
successively deeper levels of the GBD cause hierarchy using cascading mortality models, thus ensuring a robust
estimate of cause-specific mortality. For non-fatal measures (eg, low back pain), incidence and prevalence were
forecasted from mixed-effects models with SDI as the main covariate, and YLDs were computed from the resulting
prevalence forecasts and average disability weights from GBD. Alternative future scenarios were constructed by
replacing appropriate reference trajectories for risk factors with hypothetical trajectories of gradual elimination of risk
factor exposure from current levels to 2050. The scenarios were constructed from various sets of risk factors:
environmental risks (Safer Environment scenario), risks associated with communicable, maternal, neonatal, and
nutritional diseases (CMNNs; Improved Childhood Nutrition and Vaccination scenario), risks associated with major
non-communicable diseases (NCDs; Improved Behavioural and Metabolic Risks scenario), and the combined effects
of these three scenarios. Using the Shared Socioeconomic Pathways climate scenarios SSP2-4.5 as reference and
SSP1-1.9 as an optimistic alternative in the Safer Environment scenario, we accounted for climate change impact on
health by using the most recent Intergovernmental Panel on Climate Change temperature forecasts and published
trajectories of ambient air pollution for the same two scenarios. Life expectancy and healthy life expectancy were
computed using standard methods. The forecasting framework includes computing the age-sex-specific future
population for each location and separately for each scenario. 95% uncertainty intervals (UIs) for each individual
future estimate were derived from the 2·5th and 97·5th percentiles of distributions generated from propagating
500 draws through the multistage computational pipeline.
Findings In the reference scenario forecast, global and super-regional life expectancy increased from 2022 to 2050, but
improvement was at a slower pace than in the three decades preceding the COVID-19 pandemic (beginning in 2020).
Gains in future life expectancy were forecasted to be greatest in super-regions with comparatively low life expectancies
(such as sub-Saharan Africa) compared with super-regions with higher life expectancies (such as the high-income
super-region), leading to a trend towards convergence in life expectancy across locations between now and 2050. At
the super-region level, forecasted healthy life expectancy patterns were similar to those of life expectancies. Forecasts
for the reference scenario found that health will improve in the coming decades, with all-cause age-standardised
DALY rates decreasing in every GBD super-region. The total DALY burden measured in counts, however, will increase
in every super-region, largely a function of population ageing and growth. We also forecasted that both DALY counts
and age-standardised DALY rates will continue to shift from CMNNs to NCDs, with the most pronounced shifts
occurring in sub-Saharan Africa (60·1% [95% UI 56·8–63·1] of DALYs were from CMNNs in 2022 compared with
35·8% [31·0–45·0] in 2050) and south Asia (31·7% [29·2–34·1] to 15·5% [13·7–17·5]). This shift is reflected in the
leading global causes of DALYs, with the top four causes in 2050 being ischaemic heart disease, stroke, diabetes, and
chronic obstructive pulmonary disease, compared with 2022, with ischaemic heart disease, neonatal disorders, stroke,and lower respiratory infections at the top. The global proportion of DALYs due to YLDs likewise increased from
33·8% (27·4–40·3) to 41·1% (33·9–48·1) from 2022 to 2050, demonstrating an important shift in overall disease
burden towards morbidity and away from premature death. The largest shift of this kind was forecasted for subSaharan Africa, from 20·1% (15·6–25·3) of DALYs due to YLDs in 2022 to 35·6% (26·5–43·0) in 2050. In the
assessment of alternative future scenarios, the combined effects of the scenarios (Safer Environment, Improved
Childhood Nutrition and Vaccination, and Improved Behavioural and Metabolic Risks scenarios) demonstrated an
important decrease in the global burden of DALYs in 2050 of 15·4% (13·5–17·5) compared with the reference
scenario, with decreases across super-regions ranging from 10·4% (9·7–11·3) in the high-income super-region to
23·9% (20·7–27·3) in north Africa and the Middle East. The Safer Environment scenario had its largest decrease in
sub-Saharan Africa (5·2% [3·5–6·8]), the Improved Behavioural and Metabolic Risks scenario in north Africa and the
Middle East (23·2% [20·2–26·5]), and the Improved Nutrition and Vaccination scenario in sub-Saharan Africa
(2·0% [–0·6 to 3·6]).
Interpretation Globally, life expectancy and age-standardised disease burden were forecasted to improve between
2022 and 2050, with the majority of the burden continuing to shift from CMNNs to NCDs. That said, continued
progress on reducing the CMNN disease burden will be dependent on maintaining investment in and policy emphasis
on CMNN disease prevention and treatment. Mostly due to growth and ageing of populations, the number of deaths
and DALYs due to all causes combined will generally increase. By constructing alternative future scenarios wherein
certain risk exposures are eliminated by 2050, we have shown that opportunities exist to substantially improve health
outcomes in the future through concerted efforts to prevent exposure to well established risk factors and to expand
access to key health interventions.
Funding Bill & Melinda Gates Foundation.
Original languageEnglish
Article numberhttps://doi.org/10.1016/S0140-6736(24)00685-8
Pages (from-to)2204-2256
Number of pages54
JournalThe Lancet
Volume403
Issue number2204-2256
Publication statusPublished - 16 May 2024

Fingerprint

Dive into the research topics of 'Burden of disease scenarios for 204 countries and territories, 2022–2050: a forecasting analysis for the Global Burden of Disease Study 2021'. Together they form a unique fingerprint.

Cite this