Modeling Probability-Based Injury Severity Scores in Logistic Regression Models: The Logit Transformation Should Be Used
Bibliographic record
Abstract
BACKGROUND: The International Classification of Disease Injury Severity Score (ICISS) and the Trauma Registry Abbreviated Injury Scale Score (TRAIS) are trauma injury severity scores based on probabilities of survival. They are widely used in logistic regression models as raw probability scores to predict the logit of mortality. The aim of this study was to evaluate whether these severity indicators would offer a more accurate prediction of mortality if they were used with a logit transformation. METHODS: Analyses were based on 25,111 patients from the trauma registries of the four Level I trauma centers in the province of Quebec, Canada, abstracted between 1998 and 2005. The ICISS and TRAIS were calculated using survival proportions from the National Trauma Data Bank. The performance of the ICISS and TRAIS in their widely used form, proportions varying from 0 to 1, was compared with a logit transformation of the scores in logistic regression models predicting in-hospital mortality. Calibration was assessed with the Hosmer-Lemeshow statistic. RESULTS: Neither the ICISS nor the TRAIS had a linear relation with the logit of mortality. A logit transformation of these scores led to a near-linear association and consequently improved model calibration. The Hosmer-Lemeshow statistic was 68 (35-192) and 69 (41-120) with the logit transformation compared with 272 (227-339) and 204 (166-266) with no transformation, for the ICISS and TRAIS, respectively. CONCLUSIONS: In logistic regression models predicting mortality, the ICISS and TRAIS should be used with a logit transformation. This study has direct implications for improving the validity of analyses requiring control for injury severity case mix.
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 | 0.002 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.001 |
| 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".