Statistical Power and Performance of Strategies to Analyze Composites of Survival and Duration of Ventilation in Clinical Trials
Bibliographic record
Abstract
BACKGROUND: Patients with acute hypoxemic respiratory failure are at high risk of death and prolonged time on the ventilator. Interventions often aim to reduce both mortality and time on the ventilator. Many methods have been proposed for analyzing these endpoints as a single composite outcome (days alive and free of ventilation), but it is unclear which analytical method provides the best performance. Thus, we aimed to determine the analysis method with the highest statistical power for use in clinical trials. METHODS: Using statistical simulation, we compared multiple methods for analyzing days alive and free of ventilation: the t, Wilcoxon rank-sum, and Kryger Jensen and Lange tests, as well as the proportional odds, hurdle-Poisson, and competing risk models. We compared 14 scenarios relating to: 1) varying baseline distributions of mortality and duration of ventilation, which were based on data from a registry of patients with acute hypoxemic respiratory failure and 2) the varying effects of treatment on mortality and duration of ventilation. RESULTS AND CONCLUSIONS: All methods have good control of type 1 error rates (i.e., avoid false positive findings). When data are simulated using a proportional odds model, the t test and ordinal models have the highest relative power (92% and 90%, respectively), followed by competing risk models. When the data are simulated using survival models, the competing risk models have the highest power (100% and 92%), followed by the t test and a ten-category ordinal model. All models struggled to detect the effect of the intervention when the treatment only affected one of mortality and duration of ventilation. Overall, the best performing analytical strategy depends on the respective effects of treatment on survival and duration of ventilation and the underlying distribution of the outcomes. The evaluated models each provide a different interpretation for the treatment effect, which must be considered alongside the statistical power when selecting analysis models.
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 | 0.000 | 0.001 |
| 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 | 0.000 |
| 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 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".