Can machine learning improve randomized clinical trial analysis?. (October 2021)

• Record Type: Journal Article
• Authors: Romero, Juan; Chiang, Sharon; Goldenholz, Daniel M.
• Is Part Of: Seizure. Volume 91(2021)
• Page Start: 499
• Page End: 502
• Publication Date: 2021-10
• Subjects: Machine learning -- Randomized controlled trial -- Medication -- Modeling; Epilepsy -- Periodicals; Epilepsy -- Periodicals; Seizures -- Periodicals; Épilepsie -- Périodiques; Electronic journals; Electronic journals; 616.853
• DOI: 10.1016/j.seizure.2021.07.033
• Languages: English
• ISSNs: 1059-1311

Abstract

Highlights: Improvements in trial analysis methods are needed to accelerate drug development. Machine learning may be capable of improving upon current trial analysis methods. A logistic regression model was able to outperform the most efficient trial method. The new model can be used at low computational cost. Abstract: Purpose: Recently a realistic simulator of patient seizure diaries was developed that can reproduce effects seen in randomized clinical trials (RCTs). RCTs suffer from high costs and statistical inefficiencies. Using realistic simulation and machine learning this study aimed to identify a more statistically efficient outcome metric. Methods: Five candidate deep learning architectures with 54 permutations of hyperparameters were compared to the traditional standard, median percent change (MPC). Each were also tested for type 1 error. All models had similar outcomes, with appropriate low levels of type 1 error. Results: The simplest model was equivalent to a logistic regression of a histogram of individual percentage changes in seizure rate, requiring 21-22% less patients to discriminate drug from placebo at 90% power. This model was referred to as LPC. Conclusion: Future studies to validate LPC may enable faster, cheaper and more efficient clinical trials.