Analytical flowchart for the development of the GBD 2019 cause of death database
- 01. Standardize Input Data
- 01.1 Disaggregation
- 01.2 State Splitting
- 01.4 Calculate non-maternal deaths
- 02. Map to GBD Cause list
- 02.1 India Urban/Rural splitting
- 02.2 State splitting
- 03. Age-sex splitting
- 03.1 Generate global age-sex weights by cause
- 04. Correction of miscoded causes
- 05. Redistribution
- 05.1 Redistribute HIV-related garbage
- 05.2 Regress garbage codes vs. non-garbage
- 05.3 VA anemia adjustment
- 06. HIV misclassification correction
- 07. Scale strata to province
- 08. Restrictions post-redistribution
- 09. Drop VR country-years or mark as non-representative
- 10. Cause aggregation
- 11. Remove shocks and HIV\AIDS maternal adjustments
- 11.1 Remove HIV\AIDS, shocks from denominator HIV\AIDS in cause list
- 11.2 Remove HIV\AIDS deaths from maternal mortality sources
- 11.3 HIV\AIDS correction of sibling history, census, and survey data
- 11.4 HIV\AIDS correction of other maternal mortality data
- 12. Noise reduction
- 12.1 VR and CR noise reduction
- 12.2 VA noise reduction
- ICD 10 subtotal recalculation
- Map incidence data using nonfatal cause list
- Map mortality data to GBD cause list
- Age-sex splitting
- Cause disaggregation
- Redistribute cancer data
- Combine matching incidence and mortality data
- Retain best matched MI data
- Linear step models with validation
- Retain best incidence data
- Combine MI estimates with best incidence data
- Modeling ST-GPR
Anorexia nervosa
Aortic Aneurysm
- Age sex splitting
- CodCorrect
- CODEm
- Garbage code redistribution
- ICD mapping
- Noise reduction
- Standardize input data
Appendicitis
- Standardize input data
- ICD mapping
- Age sex splitting
- Garbage code redistribution
- Noise reduction
- CODEm
- CoDCorrect
Ascariasis
Asthma
- Age sex splitting
- CodCorrect
- CODEm
- Garbage code redistribution
- ICD mapping
- Noise reduction
- Standardize input data
Atrial fibrillation & flutter
- Cause of death estimation
- DisMod-MR 2.1
- Claims data code adjustment
- Dismod-MR 2.1 (model 1) Step 2
- Dismod-MR 2.1 (model 1) Step 4
- Dismod-MR 2.1 (model 1) Step 6
- Interpolate to all years
- Nonfatal age sex splitting
- Excess mortality rate calculation
- Final burden estimation
Bulimia nervosa
Cancers
- Cause of death estimation
- 1. Format incidence data
- 1. Format mortality data
- 2. ICD 10 subtotal recalculation
- 3. Map incidence data to GBD cause using nonfatal map
- 3. Map mortality data to GBD cause using mortality map
- 4. Age\sex split incidence data using proportions from CI5, NORDCAN, and SEER
- 4. Age\Sex split mortality data using proportions from COD
- 5. Cause disaggregation
- 6. Redistribute cancer data
- 7. Select mortality-incidence data sources
- 7. Select incidence data sources
- 8. Combine matching incidence and mortality data
- 8. Linear model for MI ratio
- 8. ST-GPR
- 9. Combine MI estimates with best incidence data
- 11. Prep mortality estimates for CoD upload
- 12. CODEm
- 14. CodCorrect
- Liver Cancer subtypes
