Related Experiment Video
Updated: Jul 22, 2026
Observational Study Protocol for Repeated Clinical Examination and Critical Care Ultrasonography Within the Simple Intensive Care Studies
Published on: January 16, 2019
Validation of an ICD-Based Algorithm to Identify Sepsis: A Retrospective Study
Shi-Tong Diao1, Run Dong1, Jin-Min Peng1
1Medical Intensive Care Unit, State Key Laboratory of Complex Severe and Rare Diseases, Peking Union Medical College Hospital, Peking Union Medical College & Chinese Academy of Medical Sciences, Beijing, People's Republic of China.
A modified International Classification of Diseases (ICD)-10 algorithm accurately identifies hospitalized sepsis patients. This sepsis surveillance tool shows higher validity than previous criteria, aiding resource allocation and health policy.
Area of Science:
- Medical informatics
- Public health surveillance
- Clinical epidemiology
Background:
- Sepsis surveillance is crucial for effective resource allocation, prevention strategies, and health policy development.
- Accurate identification of sepsis cases is essential for understanding population health burden.
Purpose of the Study:
- To validate a modified International Classification of Diseases (ICD)-10 based algorithm for identifying hospitalized patients with sepsis.
- To compare the performance of the modified ICD-10 algorithm against established sepsis criteria.
Main Methods:
- Retrospective analysis of a prospective, single-center cohort of adult patients admitted with suspected or confirmed infections.
- Validation of a modified ICD-10 algorithm against the Sepsis-3 definition using Sequential Organ Failure Assessment (SOFA) scores.
- Calculation of sensitivity, specificity, positive predictive value, negative predictive value, and AUROCs for various criteria.
Main Results:
- The modified ICD-10 criteria identified 50% of patients, while Sepsis-3 criteria identified 61%.
- In ICU settings, modified ICD-10 criteria demonstrated high sensitivity (84.47%) and specificity (88.57%).
- In non-ICU settings, modified ICD-10 criteria showed sensitivity of 64.19% and specificity of 80.00%.
- The modified ICD-10 algorithm achieved an AUROC of 0.76, comparable to eSOFA (0.75) and superior to Angus's and Martin's criteria (0.62).
Conclusions:
- The modified ICD-10 algorithm exhibits higher validity compared to Angus's and Martin's criteria for sepsis identification.
- The validity of the modified ICD-10 criteria is comparable to the eSOFA criteria.
- This modified ICD-10 algorithm offers an accurate method for estimating population-based sepsis burden in China.
