Accuracy of the modified Global Burden of Disease International Classification of Diseases coding methods for identifying sepsis: a prospective multicentre cohort study
Authors' conclusion
The modified GBD method demonstrated low accuracy in identifying sepsis; with the implicit plus method being the most accurate.
Study context
- Population
- ED patients aged ≥ 16 years with high sepsis risk, nine hospitals in NSW, Australia6,869 screened
- Participants
- 450 patients
- Comparison
- Modified Global Burden of Disease (GBD) International Classification of Diseases (ICD) coding methods and Angus method vs clinical sepsis
- Clinical sepsisn = 215reference
- Outcome
- Accuracy of ICD coding methods in identifying sepsis vs clinical sepsis
Methodological detailsprospective, multicentre, observational
- Prospective
- Multicentre
- Observational
Performance
As reported in the abstract
Sepsis identification
| Test or model | Value | Sensitivity | Specificity | False positives | False negatives |
|---|---|---|---|---|---|
| Explicit | 24.0% | 41.4% | 91.9% | 19 | 126 |
| Implicit | 38.9% | 58.1% | 78.7% | 50 | 90 |
| Implicit plus | 49.3% | 67.4% | 67.2% | 77 | 70 |
| Angus | 37.8% | 55.8% | 79.1% | ||
| Clinical sepsis | 47.8% |
Value
Explicit24.0%
Implicit38.9%
Implicit plus49.3%
Angus37.8%
Clinical sepsis47.8%
Sensitivity
Explicit41.4%
Implicit58.1%
Implicit plus67.4%
Angus55.8%
Clinical sepsis— not stated in the abstract
Specificity
Explicit91.9%
Implicit78.7%
Implicit plus67.2%
Angus79.1%
Clinical sepsis— not stated in the abstract
False positives
Explicit19
Implicit50
Implicit plus77
Angus— not stated in the abstract
Clinical sepsis— not stated in the abstract
False negatives
Explicit126
Implicit90
Implicit plus70
Angus— not stated in the abstract
Clinical sepsis— not stated in the abstract
- Agreement between clinical sepsis and all ICD coding methods low (Cronbach α)
- α = 0.52-0.56
Further findings
Sources of error
- False negatives without documented pathogen
- 44.3%-55.6%
- False positives: unspecified urinary tract infection, hypotension and acute kidney failure commonly assigned codes
Original publication
Accuracy of the modified Global Burden of Disease International Classification of Diseases coding methods for identifying sepsis: a prospective multicentre cohort study. Crit Care. Published 2 June 2025. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →