Covariate adjustment method for hierarchical outcomes improves statistical power in win ratio analyses
Photo by Clayton Robbins / Unsplash
Key Takeaway
Consider covariate adjustment for prognostic variables in win ratio analyses of hierarchical outcomes to improve statistical efficiency.
This methodological research paper evaluated a new ordinal regression-based covariate adjustment method for the win ratio in randomized clinical trials with hierarchical composite outcomes. The approach was compared against three alternative methods (probability index models, inverse probability weighting, and a randomization-based estimator) using simulation studies and application to EMPEROR-Preserved trial data. The study did not involve a new clinical trial population or sample size.
Covariate adjustment consistently increased statistical power when adjusting for prognostic baseline variables, with gains comparable to or greater than those in conventional Cox models. The analysis found no power loss when adjusting for non-prognostic covariates. The ordinal approach performed similarly to existing methods while providing interpretable covariate effect estimates. Adjusting for baseline values of quantitative components yielded power gains according to the baseline-to-follow-up correlation.
No safety or tolerability data were reported as this was methodological research. The findings are based on simulation studies and application to one trial dataset, not results from a new clinical trial of a medical intervention. The power gains represent methodological findings from simulations, not clinical efficacy results.
For practice, the paper recommends broader adoption of covariate adjustment and the ordinal method in randomized trials using hierarchical outcomes. However, clinicians should interpret these findings as methodological improvements that require validation when applied to specific trial designs and populations. The research addresses statistical analysis methods rather than causal effects of medical interventions.
More on Heart Failure
Beta-blockers show no mortality benefit in MI patients with preserved ejection fraction
Beta-blockers do not lower death risk for heart attack patients with strong heart pumps
This meta-analysis of 23,524 adults with myocardial infarction and left ventricular ejection fraction ≥ 40% found that beta-blockers did not…
A large review found beta-blockers did not reduce death or heart failure in adults with heart attacks and normal heart function.
Deep learning model screens for transthyretin amyloid cardiomyopathy using ECG images
AI spots hidden heart disease from routine ECGs
This observational study developed a deep learning model to identify transthyretin amyloid cardiomyopathy from ECG images. The model showed …
A new AI tool finds signs of a dangerous heart disease in routine ECGs years before symptoms worsen, offering hope for earlier treatment in …
NICE-Support programme improves frailty and quality of life in stable heart failure patients
Nutrition and exercise program improves frailty in heart failure
This randomized controlled trial in 240 adults with stable heart failure tested a 12-week multimodal support programme versus standard care.…
A 12-week program combining nutrition, exercise, and support significantly reduced frailty and improved quality of life in adults with heart…
Systematic review and meta-analysis shows remote monitoring reduces mortality and hospitalizations in heart failure
Remote monitoring reduces death risk for heart failure patients
This systematic review and meta-analysis evaluated remote patient monitoring in heart failure using data from approximately 23,000 participa…
Remote patient monitoring may lower death and hospitalization risks for people living with heart failure in recent clinical trials.
View Original Abstract ↓
BackgroundHierarchical composite outcomes, analyzed using the win ratio, are increasingly used in randomized clinical trials (RCTs). However, methods for covariate adjustment in this context are underdeveloped, despite evidence that adjusting for prognostic variables can increase statistical power.
ObjectivesIntroducing a new covariate adjustment method for hierarchical outcomes using ordinal logistic regression, comparing it with existing approaches, and assessing whether adjustment improves power in randomized trials with hierarchical outcomes.
MethodsWe developed an ordinal regression-based method for covariate adjustment of the win ratio and compared it with three alternatives: probability index models, inverse probability weighting, and a randomization-based estimator. Methods were applied to the EMPEROR-Preserved rial and tested through extensive simulations involving two common hierarchical outcome structures: time-to-event composites, and composites combining time-to-event with quantitative measures. Simulations assessed impacts on estimates, standard errors, and power across prognostic and non-prognostic settings.
ResultsIn RCT data and simulations, covariate adjustment consistently increased power when adjusting for prognostic baseline variables. Gains were comparable to or greater than those in conventional Cox models, with no power loss for non-prognostic covariates. Our ordinal approach performed similarly to existing methods while providing interpretable covariate effect estimates. Adjusting for baseline values of quantitative components yielded power gains according to the baseline-to-follow-up correlation.
ConclusionsCovariate adjustment for prognostic variables meaningfully improves efficiency in win ratio analyses for hierarchical outcomes. Our ordinal method is easily implemented and facilitates covariate effect interpretation. We recommend the broader adoption of covariate adjustment and our ordinal method in randomized trials using hierarchical outcomes.
Evolocumab reduces cardiovascular events in diabetes without known atherosclerosis
Evolocumab cuts heart events in high-risk diabetes patients
Related in Cardiology
Cryoballoon ablation reduces PV automaticity retention compared to radiofrequency ablation in PAF
Ablation times differ but rhythm return looks the same for both methods
This RCT in 100 patients with paroxysmal atrial fibrillation undergoing initial catheter ablation found that cryoballoon ablation (CBA) resu…
Two common heart rhythm procedures show similar success rates for stopping fast beats, even though one takes longer and uses more X-ray time…
Structured Exercise Training Improves LVEF and LDL-C in HFmrEF Patients After PCI
Heart function improved more with exercise than medicine alone in stable heart disease patients
This single-center randomized trial of 120 stable CAD patients with HFmrEF (LVEF 40-49%) post-PCI found that adding a 52-week structured exe…
A 52-week program of structured exercise plus standard heart medicines helped heart function recover better than medicines alone in patients…
Vericiguat showed no significant effect on flow-mediated dilation in a small pilot trial of 26 heart failure patients.
Pilot trial shows vericiguat may lower heart failure blood pressure markers
This pilot, randomized, double-blind, placebo-controlled trial enrolled 26 patients with heart failure with reduced ejection fraction (HFrEF…
A small pilot trial found vericiguat reduced a heart failure blood pressure marker, though the study was too small to confirm this benefit.
Nurse-led care coordination fails to improve medication adherence in high-need patients
Nurse-led care coordination did not improve medication adherence for most common chronic conditions.
This pragmatic randomized controlled trial tested nurse-led care coordination versus usual care in high-need, high-cost commercially insured…
A large trial found that nurse-led care coordination did not improve how often people took their diabetes, heart, or thyroid medications com…
From Other Specialties
Phase 1 study traces ainuovirine metabolism and excretion in healthy men
A single dose of oral ainuovirine clears the body within two days in healthy men
This phase 1 open-label study in 6 healthy men examined the pharmacokinetics and mass balance of a single dose of radiolabelled ainuovirine.…
A single dose of oral ainuovirine clears the body within two days in healthy men, with most of the drug leaving unchanged through urine and …
Nirmatrelvir/Ritonavir in Breast Milk: Low Infant Exposure in Lactating Women
Paxlovid levels low in breast milk, early study finds
This Phase 1 open-label study in 8 healthy lactating women found that nirmatrelvir/ritonavir concentrations in breast milk were consistently…
A small study finds that nirmatrelvir/ritonavir (Paxlovid) passes into breast milk at low levels, suggesting minimal infant exposure.
Islatravir and lenacapavir fixed-dose combination shows similar bioavailability to single agents in healthy volunteers
New HIV pill works with or without food in early safety tests
This open-label phase 1 study enrolled 93 people without HIV to assess the oral relative bioavailability and safety of a fixed-dose combinat…
A new HIV pill containing islatravir and lenacapavir showed similar drug levels whether taken with food or on an empty stomach, supporting f…
Digital chest radiography with CAD shows variable sensitivity for asymptomatic tuberculosis in South African adults
Digital scans miss many silent tuberculosis cases in South Africa adults
This cohort study evaluated digital chest radiography (dCXR) read by blinded human readers and qXR CAD in 1353 adults in South Africa. The s…
New data from South Africa shows digital scans and computer tools miss many silent tuberculosis cases, raising questions about current scree…