Critical Care Randomized Trials Demonstrate Power Failure: A Low Positive Predictive Value of Findings in the Critical Care Research Field
Bibliographic record
Abstract
BACKGROUND: We aimed to determine the post-hoc power of randomized controlled trials (RCTs) in critical care, and describe the implications for long-term positive (PPV) and negative predictive value (NPV) of statistically significant and non-significant findings respectively in the research field. METHODS: We reviewed three cohorts of RCTs. "Adult-RCTs" were 216 multicenter RCTs with a mortality outcome from a published systematic review. "Pediatric-RCTs" were 120 RCTs with a mortality outcome, obtained by search of picutrials.net. "Consecutive-RCTs" were 90 recent RCTs obtained by screening publications in 6 journals. Post-hoc power for each study was calculated at α 0.05 and 0.005, for measures of small, medium, and large effect-size, using G*Power software. Long-run expected PPV and NPV of critical care research field findings were then calculated. RESULTS: With α 0.05, post-hoc power for small effect-size was very low in all RCT-cohorts (eg, median 24% in Adult-RCTs). For medium effect-size, post-hoc power was low, except for Adult-RCTs (eg, median 9% in Pediatric-RCTs). For large effect-size, post-hoc power for non-human-animal Consecutive-RCTs was low (median 32%). With α 0.005, post-hoc power was even lower. The corollary was that both PPV and NPV were poor for small effect-size, unless α 0.005 was used. Even with α 0.005, with realistic (vs. optimistic) prior probability of the alternative hypothesis, the PPV was low (eg, in Adult-RCTs 57.1% vs. 92.3%). Adding mild bias (0.1) reduced the PPV even further. For medium effect-size both PPV and NPV were better; nevertheless, with α 0.05 and realistic prior probability of the alternative hypothesis the PPV was poor, and with α 0.005 and mild bias (0.1) the PPV was very low (eg, Adult-RCTs median 44.1%). CONCLUSIONS: To improve the predictive value of findings in the critical care research field, RCTs should be designed to have 80% power for realistic effect-size at α 0.005.
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 | 0.281 | 0.932 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.033 | 0.011 |
| Bibliometrics | 0.003 | 0.003 |
| Science and technology studies | 0.000 | 0.002 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.004 | 0.000 |
| Research integrity | 0.001 | 0.004 |
| Insufficient payload (model declined to judge) | 0.007 | 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; both teacher heads agree on what is shown here.
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".