Mining medical narratives on geriatric falls to predict post-fall hospitalization via survival model and language models
Bibliographic record
Abstract
A bstract Timely admission to the emergency department is a crucial determinant of patient outcomes. Conversely, unnecessary hospital admissions can overburden health systems and induce anxiety and stress among patients, their families, and care-givers. This study examines these implications in the geriatric population by investigating the hypothesis that delay time, i.e. the interval between injury and hospital admission, is associated with patient outcomes post admission. As delay times are not typically captured in electronic medical records, we leverage a large database from an open challenge where short narratives describing the patient injuries and treatment were made publicly available. Accordingly, we developed prognostic survival models based on large-language models that predict time to an adverse outcome using features extracted from the textual narratives, as well as additional data provided by the challenge, e.g. data about patients’ baseline characteristics, conditions of their injuries. To this end, we found that models incorporating textual embeddings achieved a dynamic area under the curve (D-AUC) of 0.713-0.715, compared to 0.637–0.649 for models without textual features, when evaluated on an external cohort. This study provides preliminary evidence that the textual data collected at the time of patients triage can be useful for patient prognosis. Future studies will examine how the same data, collected right after fall events could be used to project patient’s recovery progress.
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 | 0.003 | 0.002 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.003 | 0.004 |
| Research integrity | 0.001 | 0.002 |
| 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".