| English | 中文 |
Reviewed: 2026-10-26. See the GBD method for data meaning, uncertainty and normalization, and the first-download guide for operations. Official product availability is distinct from adapter support and verified local acquisition.
GBD 2023 disease, injury and risk estimates cover 1990–2023 and 204 countries and territories, with selected subnational estimates. The study reports 375 diseases/injuries and 88 risks; the cause-of-death study covers 292 causes. Valid combinations have constraints, so multiplying these totals does not give an export row count. Sources: disease and risk paper, cause-of-death paper.
The table also uses the IHME tools guide, GBD 2021 edition to identify product types and channels. That older guide does not establish current IDs, years or account export permissions.
| Product | Outputs and interpretation | Official route or evidence | Project boundary |
|---|---|---|---|
| Cause-specific deaths | Counts, rates and applicable proportions | Results → Cause; 2023 cause-of-death paper, 1990–2023 | One global all-cause count website export verified; other combinations and run-tool remain unverified |
| YLLs | Years of life lost, counts/rates | Results → Cause; 2023 burden study | Measure parameter available; current selections need checking, not individually tested |
| YLDs | Years lived with disability, counts/rates | Results → Cause; 2023 burden study | Same boundary; not individually verified |
| DALYs | Disability-adjusted life years, counts/rates | Results → Cause; 2023 burden study | Do not add DALYs to their YLL/YLD components |
| Incidence and prevalence | Counts and applicable rates/proportions | Results → Cause; tools guide and disease methods | Check cause/age validity; export not verified |
| Risk-attributable burden | Attributable deaths, YLLs, YLDs, DALYs and applicable PAFs | Results → Risk; 2023 risk study | No independent risk ID argument; changing context is insufficient; use the website or adapt the interface |
| Risk exposure | SEV; pollutant concentrations require the corresponding exposure product | SEV entry; risk study and tools guide | No verified dedicated interface; SEV is not PM2.5 concentration |
| HALE | Healthy life expectancy in years | Corresponding Results entry; 2023 burden study | Verify dedicated dimensions, not automatically supported by the cause adapter |
| All-cause mortality and life expectancy | Demographic mortality, life expectancy and related estimates | Round-specific GHDx product; demographic paper | Verify tables, years and channel separately; life expectancy is not a rate per 100,000 |
| Population, life tables and fertility | Age/sex population and demographic indicators | Tools guide; Results navigation or GHDx record | Check release and period per product; population export unverified; population_group is not a count switch |
| Other topics | Etiology, injury nature and impairment results where provided | Tools guide and current topic product | Dedicated arguments are missing; manual selection/adaptation required |
Results commonly export CSV, while some products use separate records or asynchronous download tasks. Preserve estimates and available uncertainty bounds. Rounded website cells are not full-precision exports. The source catalog indexes inputs; it does not make every input microdataset publicly downloadable.
| Dimension | Suggested scope | Checks |
|---|---|---|
| Release round | Prefer GBD 2023 for this task | Use one round over the study period; verify the version ID rather than copying historical defaults |
| Time | Usually 1990–2023 for disease/risk studies; 2000–2023 may suit a research task | Demographic, life-table and topic periods may differ |
| Location | Global, region, country; add provided subnational locations when needed | Avoid overlapping totals; do not promise arbitrary county, city or grid data |
| Sex | Male, female or both | Do not combine both-sex totals with components; some indicators restrict sex |
| Age | Age-specific, all-age or age-standardized, requested separately | The study commonly reports 25 fine age groups, but product selections can include aggregates or restrictions |
| Cause/risk | A small set of target leaf causes or risks | Avoid overlapping hierarchies |
| Measure | Applicable deaths, YLLs, YLDs, DALYs, incidence or prevalence | Defines the estimand, not merely the unit |
| Metric | Applicable number, rate or percent | Check units/denominators; age-standardized rates cannot reconstruct counts using actual population |
A has passed. B and C are candidate plans, not completed downloads. Define the research task before expanding the scope.
| Stage | Scope | Expected combinations | Purpose |
|---|---|---|---|
| A: verified connectivity check | GBD 2023; 2023; Global; All causes; All ages; Both; Deaths; Number | 1 | Fresh result, ZIP acquisition and CSV content verified |
| B: small China table | 2019/2023; China; ischemic heart disease; All ages; Both; Deaths; Number and Rate | Up to 4 | Check country, years and count/rate distinctions after verifying options |
| C: candidate health-model baseline | 2000–2023; China; six candidate causes; male/female; 15 groups from 25–29 to 95+; Deaths; age-specific Rate | Up to 4,320 | Age-specific baseline mortality; confirm risk-function, cause and age applicability |
C’s candidate causes are ischemic heart disease, stroke, COPD, tracheal/bronchus/lung cancer, lower respiratory infections and type 2 diabetes. This is not a universal PM2.5-model cause set. Adapt the plan if the selected risk function requires children, other ages or different cause definitions.
Count combinations as locations × years × causes × sexes × ages × measures × metrics. Inapplicable combinations may reduce actual rows. The estimate and bounds are fields in a record, not three extra records. Do not begin with every cause, age and location.
On 2026-10-26, scope A passed the official website route: fresh result checked at 21:00, export submitted at 21:02, 717-byte ZIP acquisition confirmed at 21:05, and one-record/18-column CSV validation completed at 21:05:42. Times are Beijing time (UTC+8). Follow the verified guide. No task CDP session was confirmed for a real ihme_gbd_fetch_results call.
The user handled access verification. No authentication state was copied or internal endpoint used to replace verification. Only the new export was inspected, not existing research data.