Related Experiment Video
Updated: May 30, 2025
Inverse Probability of Treatment Weighting Propensity Score using the Military Health System Data Repository and National Death Index
Published on: January 8, 2020
Prediction-powered Inference for Clinical Trials: application to linear covariate adjustment
Pierre-Emmanuel Poulet1,2,3, Maylis Tran1,2,3, Sophie Tezenas du Montcel1,2,3
1ARAMIS, Sorbonne Université, Paris Brain Institute (ICM Institut du Cerveau), INRIA, INSERM, AP-HP, Groupe Hospitalier Sorbonne Université, Paris, France.
Prediction-powered inference (PPI) enhances clinical trials by creating digital twins for patients, reducing sample size and control group needs. This method leverages machine learning for statistically valid treatment effect estimation.
Area of Science:
- Biostatistics
- Clinical Trials
- Machine Learning
Background:
- Traditional statistical estimation methods in clinical trials can be inefficient.
- Leveraging machine learning for prediction-powered inference (PPI) offers a novel approach.
- Existing PPI methods can be enhanced for clinical trial applications.
Purpose of the Study:
- To introduce and evaluate Prediction-Powered Inference Plus Plus (PPI++) in the context of clinical trials.
- To demonstrate how PPI++ can provide statistically valid treatment effect estimates.
- To explore the implications of PPI++ for optimizing clinical trial design and resource allocation.
Main Methods:
- Utilizing disease progression models to generate prognostic scores for participants based on baseline covariates.
- Applying the PPI paradigm to create 'digital twins' of treated patients for comparison with untreated controls.
- Conducting theoretical analysis of the estimator's properties, including asymptotic unbiasedness and variance derivation.
- Performing simulations to validate the theoretical findings.
Main Results:
- The proposed PPI++ estimator is asymptotically unbiased for the Average Treatment Effect.
- An explicit formula for the variance of the PPI++ estimator has been derived.
- Simulations confirm the theoretical properties and practical utility of the method.
- Application to an Alzheimer's disease clinical trial indicates significant sample size reduction potential.
Conclusions:
- PPI++ offers a statistically valid method to incorporate machine learning predictions into clinical trial analysis.
- This approach can lead to more efficient clinical trials, requiring smaller sample sizes and fewer controls.
- PPI++ facilitates the direct application of large-scale disease prediction models to clinical trial settings.
- The method holds promise for improving the efficiency and feasibility of clinical research, particularly in complex diseases like Alzheimer's.
More Related Videos
07:31Implementation of a Real-Time Psychosis Risk Detection and Alerting System Based on Electronic Health Records using CogStack
Published on: May 15, 2020
04:09Predicting Treatment Response to Image-Guided Therapies Using Machine Learning: An Example for Trans-Arterial Treatment of Hepatocellular Carcinoma
Published on: October 10, 2018
Related Concept Videos
01:16Clinical Trials
There are four phases in a clinical trial. A phase one...
01:23Types of Biopharmaceutical Studies: Controlled and Non-Controlled Approaches
Non-controlled studies, commonly employed for initial exploration, lack a control group, rendering them susceptible to biases and external influences. In contrast,...
01:12Hazard Ratio
For example, in a clinical trial...
01:24Testing a Claim about Population Proportion
There are two methods of testing a claim about a population proportion: (1) Using the sample proportion from the data where a binomial distribution is approximated to the normal distribution and (2) Using the binomial probabilities calculated from the data.
The first method uses normal distribution as an approximation to the binomial distribution. The requirements are as follows: sample size is large...
01:03Prediction Intervals
However, the point estimate is most likely not the exact value of the population parameter, but close to it. After calculating point estimates, we construct interval estimates, called confidence intervals or prediction intervals. This prediction interval comprises a range of values unlike the point estimate and is a better predictor of the observed sample value, y.
01:13Sensitivity, Specificity, and Predicted Value
Sensitivity is the...