Development and Validation of a Rule-Based Natural Language Processing Algorithm to Identify Falls in Inpatient Records of Older Adults: Retrospective Analysis
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Background: In order to address fall underestimation by the International Classification of Diseases (ICD) in clinical settings, information from clinical notes could be incorporated via natural language processing (NLP) as a possible solution. However, its application to inpatient notes has not been fully investigated. Objective: This study aims to develop and validate a rule-based NLP algorithm to identify falls based on inpatient admission notes from older patients. Methods: This retrospective study used 12-year electronic inpatient records of patients aged ≥65 years from public hospitals in Hong Kong. A random sample of 1000 patients was drawn to develop the NLP algorithm. Manual review was the gold standard for assessing the algorithm's performance, with sensitivity, specificity, precision, and F1-score calculated at the record, episode, and patient levels. In addition, the study compared the number of falls identified by ICD codes and clinical notes independently and combined. Results: Our rule-based NLP algorithm showed excellent performance, with a sensitivity, specificity, precision, and F1-score of 93.3%, 99.0%, 87.5%, and 0.903 at the record and episode levels, and 92.9%, 98.3%, 89.7%, and 0.912 at the patient level. The combined identification strategy using ICD codes and the NLP method provided the most comprehensive capture of fall-related episodes and fallers. Conclusions: The NLP method proved efficient and accurate in detecting falls from clinical notes in inpatient episodes. For comprehensive capture of fall episodes and fallers, we recommend the combined use of the NLP algorithm and ICD codes, which should be applied in future fall epidemiology studies and clinical practice for identifying high-risk groups of fall interventions.
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.
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 | 0.001 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| 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 it