Derivation and validation of pragmatic clinical models to predict hospital length of stay after cardiac surgery in Ontario, Canada: a population-based cohort study
Bibliographic record
Abstract
BACKGROUND: Cardiac surgery is resource intensive and often requires multidisciplinary involvement to facilitate discharge. To facilitate evidence-based resource planning, we derived and validated clinical models to predict postoperative hospital length of stay (LOS). METHODS: We used linked, population-level databases with information on all Ontario residents and included patients aged 18 years or older who underwent coronary artery bypass grafting, valvular or thoracic aorta surgeries between October 2008 and September 2019. The primary outcome was hospital LOS. The models were derived by using patients who had surgery before Sept. 30, 2016, and validated after that date. To address the rightward skew in LOS data and to identify top-tier resource users, we used logistic regression to derive a model to predict the likelihood of LOS being more than the 98th percentile (> 30 d), and γ regression in the remainder to predict continuous LOS in days. We used backward stepwise variable selection for both models. RESULTS: Among 105 193 patients, 2422 (2.3%) had an LOS of more than 30 days. Factors predicting prolonged LOS included age, female sex, procedure type and urgency, comorbidities including frailty, high-risk acute coronary syndrome, heart failure, reduced left ventricular ejection fraction and psychiatric and pulmonary circulatory disease. The C statistic was 0.92 for the prolonged LOS model and the mean absolute error was 2.4 days for the continuous LOS model. INTERPRETATION: We derived and validated clinical models to identify top-tier resource users and predict continuous LOS with excellent accuracy. Our models could be used to benchmark clinical performance based on expected LOS, rationally allocate resources and support patient-centred operative decision-making.
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 | 0.015 | 0.041 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.002 |
| Bibliometrics | 0.001 | 0.002 |
| Science and technology studies | 0.002 | 0.001 |
| Scholarly communication | 0.002 | 0.000 |
| Open science | 0.002 | 0.001 |
| 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".