Using GBD 2023 estimates for 204 countries and territories, this analysis quantified changes in the morbidity gap (life expectancy minus healthy life expectancy, HALE) from 1990 to 2023 and decomposed those changes by disease, injury, and risk factor. Globally, the morbidity gap increased from 8.8 years (95% UI 6.7–11.2) in 1990 to 10.7 years (8.2–13.7) in 2023, an absolute rise of 1.9 years and a 21.9% relative increase. In 2023, an average of 14.5% of life was spent in poor health versus 13.6% in 1990. Point estimates suggested widening morbidity gaps in 203 of 204 countries and territories.
The study used GBD 2023 annual estimates of mortality, morbidity, and population for 204 locations. Data inputs included vital registration, censuses, household surveys, sample registration systems, and disease registries. Where data were incomplete, standard GBD modelling frameworks were applied, including CODEm for cause-specific mortality, ST-GPR for spatiotemporal estimation, and DisMod-MR for non-fatal outcomes. Population denominators were derived from GBD demographic models to ensure consistency. Analyses were conducted in Python, Stata, and R; results represent means and 95% uncertainty intervals from 250 or more draws from posterior distributions.
Life expectancy is defined as average years expected from birth under period mortality rates. HALE denotes expected years of life in full health, estimated by adjusting life expectancy for years lived with disability (YLDs) using disability weights from population surveys. The morbidity gap equals life expectancy minus HALE; the proportional morbidity gap is that difference divided by life expectancy. To decompose the morbidity gap, the authors assigned to each cause group its proportional share of YLDs, adjusted for co-occurrence of non-fatal causes to avoid double counting. The same approach was applied at GBD Level 3 causes and to risk-factor decompositions.
Between 1990 and 2023 global life expectancy increased from 64.6 years (95% UI 64.5–64.8) to 73.8 years (73.6–74.1). HALE rose from 55.9 years (53.5–57.9) to 63.1 years (60.2–65.6). Despite increases in HALE, gains in life expectancy outpaced gains in HALE in nearly all locations, producing an expansion of the morbidity gap from 8.8 years to 10.7 years over the study period. The COVID-19 pandemic (defined here as 2020–22) temporarily disrupted trends: life expectancy declined to 71.9 years in 2021 and HALE declined to 61.7 years, with little net change in the global morbidity gap during that interval.
Morbidity gaps widened across almost all geographic aggregations and countries. In 2023, morbidity gaps were largest in the high Socio-demographic Index (SDI) quintile and smallest in the low SDI quintile. Countries with longer life expectancy tended to have larger morbidity gaps, and both life expectancy and the morbidity gap showed a positive relationship with SDI. However, the relationship between life expectancy and the proportional morbidity gap (the share of life spent in poor health) was less consistent.
Across locations, age-specific morbidity gaps widened across the adult life course rather than concentrating solely in the final years of life. The analysis reports trends by sex and found expansion of morbidity across sexes and geographies. Detailed age-sex decomposition results and country-level time series are included in GBD outputs and appendices.
A small set of largely non-fatal, chronic conditions accounted for a majority of unhealthy years. In 2023, musculoskeletal disorders (especially low back pain), mental disorders (depressive and anxiety disorders), sense organ diseases (age-related hearing loss), unintentional injuries (falls), and other non-communicable diseases collectively contributed 57.4% of years lived in poor health globally. The decomposition used YLD rates multiplied by person-years lived across ages to allocate the number of years of the morbidity gap attributable to each cause, with adjustment for comorbidity.
The leading risk factors contributing to the global morbidity gap in 2023 included high fasting plasma glucose, high body-mass index, and child and maternal malnutrition. Additional modifiable risks—such as tobacco use and other behavioural, metabolic, and nutritional exposures—also contributed, with the mix of leading risks varying by geography and SDI level. Risk-specific decompositions were produced using GBD attributable burden methods and the same person-years framework used for cause decompositions.
The expansion of the morbidity gap across nearly all locations indicates that survival improvements have generally outpaced reductions in non-fatal health loss. Most populations now spend a decade or more of life in poor health across the lifespan. The authors argue that progress towards healthy ageing should be measured not only by mortality reductions but also by declines in years lived in poor health. Policy implications include prioritising targeted reduction of modifiable risk factors and strengthening delivery of rehabilitation, mental health, and long-term chronic care services to align healthspan more closely with lifespan.
The study uses GBD 2023 modelled estimates and inherits the assumptions and limitations of those methods, including reliance on modelling where primary data are incomplete. The COVID-19 pandemic (2020–22) introduced substantial disruptions to life expectancy and HALE in many settings. The funder (the Gates Foundation) had no role in study design, analysis, interpretation, or writing. The authors note that full methodological detail and appendices are available in GBD capstone publications and that analytic code can be requested from the GBD team.