Autoinflammatory & rare systemic diseases
Development and validation of a text-mining-based algorithm for detection of rare hereditary periodic fever syndromes
Authors' conclusion
Text-mined EHR data contains valuable information to distinguish PFS from related diagnoses. Pending further validation in the pre-diagnostic setting, and a strategy to keep false positive rates manageable, machine learning models may contribute to identification of undiagnosed patients with PFS.
Study context
- Population
- Potential periodic fever syndrome (periodic fever syndromes (PFS)) cases identified from the electronic health record (EHR) of a tertiary care hospital and screened for Eurofever/PRINTO classification criteria, and controls with billing codes corresponding to overlapping diagnoses and/or diagnoses with symptoms also occurring in PFS; external validation in a secondary care hospital cohortCases and controls per cohort · Development cohort 194 cases (145 Familial Mediterranean fever (FMF)) and 14,435 controls · External validation cohort 120 cases (118 FMF) and 34,315 controls
- Index test or model
- Naïve Bayes model trained on the presence of 60 text-mined features to discriminate cases from controls
- Target condition
- Periodic fever syndromes (PFS)
Methodological detailsdevelopment and external validation of a diagnostic prediction model, text mining of the electronic health record (EHR), naïve Bayes model
- Development and external validation of a diagnostic prediction model
- Text mining of the electronic health record (EHR)
- Naïve Bayes model
Performance
As reported in the abstract
| Test or model | AUC ROC |
|---|---|
| Internal validation | 0.90 |
| External validation cohort | 0.88 [95%CI 0.84-0.91] |
AUC ROC
Internal validation0.90
External validation cohort0.88 [95%CI 0.84-0.91]
Further findings
Discrimination restricted to pre-diagnostic data in the external validation (50 PFS cases)
- AUC ROC (substantially decreased performance)
- 0.66 [95%CI 0.56-0.75]
Original publication
Development and validation of a text-mining-based algorithm for detection of rare hereditary periodic fever syndromes. Rheumatology (Oxford). Published 13 November 2026. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →