Development and validation of a rheumatoid arthritis case definition: a machine learning approach using data from primary care electronic medical records
Bibliographic record
Abstract
BACKGROUND: Rheumatoid Arthritis (RA) is a chronic inflammatory disease that is primarily diagnosed and managed by rheumatologists; however, it is often primary care providers who first encounter RA-related symptoms. This study developed and validated a case definition for RA using national surveillance data in primary care settings. METHODS: This cross-sectional validation study used structured electronic medical record (EMR) data from the Canadian Primary Care Sentinel Surveillance Network (CPCSSN). Based on the reference set generated by EMR reviews by five experts, three machine learning steps: 'bag-of-words' approach to feature generation, feature reduction using a feature importance measure coupled with recursive feature elimination and clustering, and classification using tree-based methods (Decision Tree, Random Forest, and Extreme Gradient Boosting). The three tree-based algorithms were compared to identify the procedure that generated the optimal evaluation metrics. Nested cross-validation was used to allow evaluation and comparison and tuning of models simultaneously. RESULTS: Of 1.3 million patients from seven Canadian provinces, 5,600 people aged 19 + were randomly selected. The optimal algorithm for selecting RA cases was generated by the XGBoost classification method. Based on feature importance scores for features in the XGBoost output, a human-readable case definition was created, where RA cases are identified when there are at least 2 occurrences of text "rheumatoid" in any billing, encounter diagnosis, or health condition table of the patient chart. The final case definition had sensitivity of 81.6% (95% CI, 75.6-86.4), specificity of 98.0% (95% CI, 97.4-98.5), positive predicted value of 76.3% (95% CI, 70.1-81.5), and negative predicted value of 98.6% (95% CI, 98.0-98.6). CONCLUSION: A case definition for RA in using primary care EMR data was developed based off the XGBoost algorithm. With high validity metrics, this case definition is expected to be a reliable tool for future epidemiological research and surveillance investigating the management of RA in CPCSSN dataset.
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 machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.045 | 0.110 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.005 | 0.003 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.003 | 0.001 |
| Open science | 0.003 | 0.002 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 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 source (direct Gemma or distilled Codex), 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".