43 Clinical logic and reasoning algorithm (CLARA): an AI-enabled whatsapp chatbot for guiding antiarrhythmic drug therapy in atrial fibrillation
Bibliographic record
Abstract
Introduction Appropriate selection of antiarrhythmic drugs (AADs) in atrial fibrillation (AF) requires consideration of structural heart disease, renal and hepatic function, arrhythmia burden, and drug contraindications. CLARA is an artificial intelligence-powered decision support system that uses a WhatsApp chatbot interface integrated with an Excel-based clinical rules engine to guide AAD selection based on individualised patient profiles. Methods A synthetic dataset of 150,000 patients with atrial fibrillation was created with varied clinical variables including left ventricular ejection fraction (LVEF), left atrial (LA) size, renal and liver function, AF type, prior AAD exposure, and comorbidities. CLARA was programmed to deliver AAD recommendations. To assess performance, CLARA was tested against cardiologist consensus recommendations in a cohort of 100 complex AF cases, including patients with left ventricular dysfunction, significant structural heart disease, and multiple prior AAD failures (figure 1). Results CLARA achieved 91% concordance with expert cardiologist recommendations in selecting the appropriate class and dose of AAD. In 29 patients with LVEF <40%, CLARA correctly excluded Class IC drugs in all cases and recommended amiodarone or rate control. In 18 patients with significant left atrial enlargement, rhythm control was appropriately deprioritised. For patients with prior AAD intolerance (n=26), CLARA accurately identified contraindicated agents. Median response time was under 2 minutes per case. Conclusions/Implications CLARA demonstrates high clinical accuracy and responsiveness in recommending antiarrhythmic therapy in complex AF patients. Its integration of expert-derived algorithms with real-time chatbot interaction offers scalable, structured support for personalising rhythm control strategies. Future work will validate its use in prospective real-world AF populations and explore its integration with electronic health records.
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 | 0.003 | 0.011 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.002 | 0.001 |
| Open science | 0.001 | 0.002 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.014 | 0.003 |
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".