Enhancement and external validation of algorithms using diagnosis codes to identify invasive <i>Escherichia coli</i> disease
Bibliographic record
Abstract
Objective To assess the predictive accuracy of code-based algorithms for identifying invasive Escherichia coli (E. coli) disease (IED) among inpatient encounters in US hospitals.Methods The PINC AI Healthcare Database (10/01/2015–03/31/2020) was used to assess the performance of six published code-based algorithms to identify IED cases among inpatient encounters. Case-confirmed IEDs were identified based on microbiological confirmation of E. coli in a normally sterile body site (Group 1) or in urine with signs of sepsis (Group 2). Code-based algorithm performance was assessed overall, and separately for Group 1 and Group 2 based on sensitivity, specificity, positive and negative predictive value (PPV and NPV) and F1 score. The improvement in performance of refinements to the best-performing algorithm was also assessed.Results Among 2,595,983 encounters, 97,453 (3.8%) were case-confirmed IED (Group 1: 60.9%; Group 2: 39.1%). Across algorithms, specificity and NPV were excellent (>97%) for all but one algorithm, but there was a trade-off between sensitivity and PPV. The algorithm with the most balanced performance characteristics included diagnosis codes for: (1) infectious disease due to E. coli OR (2) sepsis/bacteremia/organ dysfunction combined with unspecified E. coli infection and no other concomitant non-E. coli invasive disease (sensitivity: 56.9%; PPV: 56.4%). Across subgroups, the algorithms achieved lower algorithm performance for Group 2 (sensitivity: 9.9%–61.1%; PPV: 3.8%–16.0%).Conclusions This study assessed code-based algorithms to identify IED during inpatient encounters in a large US hospital database. Such algorithms could be useful to identify IED in healthcare databases that lack information on microbiology data.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.004 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".