Global Burden of Disease Study 2019 (GBD 2019) - Nonfatal Health Outcomes 9 (Neonatal Encephalopathy - Onchocerciasis)

Neonatal Encephalopathy

• YLD estimation
  • Comorbidity correction (COMO)
  • Moderate to severe impairment splits
  • Mild impairment splits
• Neonatal prevalence estimation
  • Convert CFR to EMR
  • DisMod-MR 2.1 prevalence model
  • Age/sex split and identify outliers
  • Interpolate prevalence at 28 days
• Impairment estimation
  • DisMod-MR 2.1 impairment model
  • Mixed-effects regression of impairment proportion
  • Severity split (asymptomatic, mild, moderate)
  • Convert standardized mortality ratio data to excess mortality ratio estimates
  • Mixed-effects regression of CFR

---

Neonatal preterm birth

• Standardized mortality ratio (SMR) of cerebral palsy
• Convert SMR to EMR based on all-cause mortality estimates
• Fixed-effects regression using log neonatal mortality rate
• Subtract mild and moderate proportion from 1
• Severity split
• Dismod MR 2.1
• Comorbidity correction (COMO)

---

Neonatal sepsis

• Neonatal prevalence estimation
  • Adjustment of hospital data to account for multiple admissions, multiple diagnoses, and Healthcare Access and Quality
  • Convert CFR to EMR
  • Convert prevalence data to incidence
  • Crosswalk claims and literature data to hospital data using MR-BRT - derived adjustment factors
  • Age- and sex-split incidence data
  • MR-BRT model to identify outliers
  • DisMod-MR 2.1 prevalence model
  • Interpolate prevalence at 28 days
• Impairment estimation
  • Meta-analysis of mild, moderate, severe impairment proportions
  • Fixed-effects regression of CFR
  • Severity splits (asymptomatic, mild, moderate/severe)
  • Convert standardized mortality ratio data to excess mortality ratio estimates
  • DisMod-MR 2.1 impairment model
• YLD estimation
  • Mild impairment sequelae splits
  • Moderate to severe impairment sequelae splits
  • Comorbidity correction (COMO)

---

Non-alcoholic fatty liver disease\Non-alcoholic steatohepatitis (NFLP\NASH)

• Sex splitting
• Age splitting
• DisMod-MR 2.1
• Apply alcohol exclusions
• Comorbidity correction (COMO)

---

Non-melanoma skin cancer (basal cell and squamous cell carcinoma)

• Mapped duration of BCC\SCC between 1-5 years based on ATC
• Split C44 into basal cell and squamous cell carcinoma based on proportion from literature
• Dismod-MR 2.1
• Severity split into 3 levels using MEPs
• Comorbidity correction (COMO)

---

Non-rheumatic calcific aortic valve disease

• Age-sex splitting
• Dismod MR 2.1
• Asymptomatic proportion model
• Treatment proportion model
• Heart failure severity split model
• Severity split
• Comorbidity correction (COMO)
• Disability weights for each sequela

---

Non-rheumatic degenerative mitral valve disease

• Age-sex splitting
• Dismod MR 2.1
• Asymptomatic proportion model
• Treatment proportion model
• Heart failure severity split model
• Severity split
• Comorbidity correction (COMO)
• Disability weights for each sequela

---

Non-rheumatic valvular heart disease

---

Nonfatal Injuries

• LTD
  • Crosswalk EQ5D to SF-12
  • SF-12 or EQ5D scores translated into disability weights
  • Generate long-term disability weights
  • Mixed effects model long-term probabilities
  • Generate severity N-code ranks
• EN
  • Generate E\N matrices (Dirichlet regression)
• Pipeline
  • Incidence by Ecode
  • Apply incidence-to-mortality ratios to fatal discontinuities and state actor violence deaths
  • Apply incidence E\N matrices (if multiple N-codes, assign the most severe)
  • (% treated*treated duration) + (% untreated*untreated duration)
  • Calculate fraction of incidence that is long-term
  • Translate SMR to excess mortality
  • Calculate prevalence (incidence *duration)
  • Apply short-term disability weights
  • Upload results to epi database
• Crosswalking
  • Crosswalks with MR-BRT models
  • DisMod-MR incidence models
  • Derive outpatient incidence using outpatient coefficent

---

Obstetric fistula

• Age-sex splitting
• Comorbidity correction (COMO)
• DisMod-MR 2.1
• proportion vesicovaginal and rectovaginal
• Severity Splits
• Geographic exclusions

---

Onchocerciasis

• Extrapolation of all case prevalence and sequelae
• Combine acute and chronic skin disease level 2
• Multiply visual impairment and blindness by random value normally distributed with mean zero and standard deviation 0.1 to add uncertainty
• Multiply visual impairment estimates by a random value to generate moderate vision impairment, then subtract resulting estimates from visual impairment to obtain estimates of severe vision impairment
• Re-calculate prevalence
• Transform to logit space - logit(prevalence)
• Add uncertainty due to nod-mf conversion for OCP countries
• Normalize draws
• Add time trend uncertainty
• Re-expand normalized draws to adjusted scale
• Model population at risk in each focuse using a Poisson model. Covariates: year splines
• Model prevalence of onchocerciasis in the Americas among the population at risk in each focus using a glm model with binomial family, logit link, no intercept term and random effects on a combined foci cariable. Covariates: focie indicator variable, years since MDA start, age splines
• Sum cases across GBD location-years and divide by population
• Adjust prevalence in foci estimates by population at risk to get case total by foci
• Split sequelae using all-age all-sex global sequelae ratios in Africa estimates
• Estimate prevalence of asymptomatic onchocerciasis
• YLD calculation
• Comorbidity Correction

---