Abstract TP189: Identifying Stroke Patients At Risk For Cognitive Impairment And Dementia Using Electronic Health Record Data And Machine Learning
Bibliographic record
Abstract
Background: Stroke patients are at high risk of developing cognitive impairment and dementia. Failure to identify cognitive impairment in time could hasten the progression of dementia and impact the rehabilitation plan. Therefore, a reliable method is needed to determine a stroke survivor’s susceptibility to post-stroke cognitive impairment and dementia (PSCID). Method: We conducted a retrospective cohort study of cryptogenic stroke (CS) patients from January 1, 2017, to February 28, 2022, using EHR data to query for PSCID onset. A machine learning (ML) model was created to forecast the occurrence of PSCID during follow-up. The features used in the model included elements from EHR that were found to be associated with PSCID in other studies, including sociodemographic, medical, and mental morbidities. Logistic Regression, Random Forest (RF) Classifier, and Gradient Boosting are examples of ML algorithms that were used. The final model used RF Classifier due to its superior performance. Results: Of 390 CS patients (62±16 years, 56.4% male) included in the analysis, 110 (28.2%) had documented PSCID in EHR following the initial stroke. We evaluated our model in a repeated (n=100) 10-fold cross validation scheme and used Synthetic Minority Oversampling Technique (SMOTE) to compensate for class imbalance. We identified the most informative ten features using an Extra Tree classifier, which are age, transient ischemic attack, sex, smoking, diabetes, hypertension, atherosclerosis, anemia, and atrial fibrillation. Using these features, our RF classifier reached a performance of 71.2±6.5% accuracy, 70±4.9% precision, 75±12% recall, and 0.777±0.082 AUC. Conclusion: Our model could predict the CS patients at risk for PSCID with reasonable accuracy using only ten features. Future work should involve a larger cohort along with more advanced machine learning algorithms to enhance the prediction performance.
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 | 0.005 | 0.011 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 0.001 |
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".