Estimation of treatment effects following a sequential trial of multiple treatments.

When a clinical trial is subject to a series of interim analyses as a result of which the study may be terminated or modified, final frequentist analyses need to take account of the design used. Failure to do so may result in overstated levels of significance, biased effect estimates and confidence intervals with inadequate coverage probabilities. A wide variety of valid methods of frequentist analysis have been devised for sequential designs comparing a single experimental treatment with a single control treatment. It is less clear how to perform the final analysis of a sequential or adaptive design applied in a more complex setting, for example, to determine which treatment or set of treatments amongst several candidates should be recommended. This article has been motivated by consideration of a trial in which four treatments for sepsis are to be compared, with interim analyses allowing the dropping of treatments or termination of the trial to declare a single winner or to conclude that there is little difference between the treatments that remain. The approach taken is based on the method of Rao-Blackwellization which enhances the accuracy of unbiased estimates available from the first interim analysis by taking their conditional expectations given final sufficient statistics. Analytic approaches to determine such expectations are difficult and specific to the details of the design: instead "reverse simulations" are conducted to construct replicate realizations of the first interim analysis from the final test statistics. The method also provides approximate confidence intervals for the differences between treatments.

Applying methods for personalized medicine to the treatment of alcohol use disorder.

OBJECTIVE: Numerous behavioral treatments for alcohol use disorder (AUD) are effective, but there are substantial individual differences in treatment response. This study examines the potential use of new methods for personalized medicine to test for individual differences in the effects of cognitive behavioral therapy (CBT) versus motivational enhancement therapy (MET) and to provide predictions of which will work best for individuals with AUD. We highlight both the potential contribution and the limitations of these methods. METHOD: We performed secondary analyses of abstinence among 1,144 participants with AUD participating in either outpatient or aftercare treatment who were randomized to receive either CBT or MET in Project MATCH. We first obtained predicted individual treatment effects (PITEs), as a function of 19 baseline client characteristics identified a priori by MATCH investigators. Then, we tested for the significance of individual differences and examined the predicted individual differences in abstinence 1 year following treatment. Predictive intervals were estimated for each individual to determine if they were 80% more likely to achieve abstinence in one treatment versus the other. RESULTS: Results indicated that individual differences in the likelihood of abstinence at 1 year following treatment were significant for those in the outpatient sample, but not for those in the aftercare sample. Individual predictive intervals showed that 37% had a better chance of abstinence with CBT than MET, and 16% had a better chance of abstinence with MET. Obtaining predictions for a new individual is demonstrated. CONCLUSIONS: Personalized medicine methods, and PITE in particular, have the potential to identify individuals most likely to benefit from one versus another intervention. New personalized medicine methods play an important role in putting together differential effects due to previously identified variables into one prediction designed to be useful to clinicians and clients choosing between treatment options. (PsycInfo Database Record (c) 2021 APA, all rights reserved).

A permutation test for assessing the presence of individual differences in treatment effects.

An important goal of personalized medicine is to identify heterogeneity in treatment effects and then use that heterogeneity to target the intervention to those most likely to benefit. Heterogeneity is assessed using the predicted individual treatment effects framework, and a permutation test is proposed to establish if significant heterogeneity is present given the covariates and predictive model or algorithm used for predicted individual treatment effects. We first show evidence for heterogeneity in the effects of treatment across an illustrative example data set. We then use simulations with two different predictive methods (linear regression model and Random Forests) to show that the permutation test has adequate type-I error control. Next, we use an example dataset as the basis for simulations to demonstrate the ability of the permutation test to find heterogeneity in treatment effects for a predicted individual treatment effects estimate as a function of both effect size and sample size. We find that the proposed test has good power for detecting heterogeneity in treatment effects when the heterogeneity was due primarily to a single predictor, or when it was spread across the predictors. Power was found to be greater for predictions from a linear model than from random forests. This non-parametric permutation test can be used to test for significant differences across individuals in predicted individual treatment effects obtained with a given set of covariates using any predictive method with no additional assumptions.

Prior elicitation of the efficacy and tolerability of Methotrexate and Mycophenolate Mofetil in Juvenile Localised Scleroderma.

Background: Evidence is lacking for safe and effective treatments for juvenile localised scleroderma (JLS). Methotrexate (MTX) is commonly used first line and mycophenolate mofetil (MMF) second line, despite a limited evidence base. A head to head trial of these two medications would provide data on relative efficacy and tolerability. However, a frequentist approach is difficult to deliver in JLS, because of the numbers needed to sufficiently power a trial. A Bayesian approach could be considered. Methods: An international consensus meeting was convened including an elicitation exercise where opinion was sought on the relative efficacy and tolerability of MTX compared to MMF to produce prior distributions for a future Bayesian trial. Secondary aims were to achieve consensus agreement on critical aspects of a future trial. Results: An international group of 12 clinical experts participated. Opinion suggested superior efficacy and tolerability of MMF compared to MTX; where most likely value of efficacy of MMF was 0.70 (95% confidence interval (CI) 0.34-0.90) and of MTX was 0.68 (95% CI 0.41-0.8). The most likely value of tolerability of MMF was 0.77 (95% CI 0.3-0.94) and of MTX was 0.62 (95% CI 0.32-0.84). The wider CI for MMF highlights that experts were less sure about relative efficacy and tolerability of MMF compared to MTX. Despite using a Bayesian approach, power calculations still produced a total sample size of 240 participants, reflecting the uncertainty amongst experts about the performance of MMF. Conclusions: Key factors have been defined regarding the design of a future Bayesian approach clinical trial including elicitation of prior opinion of the efficacy and tolerability of MTX and MMF in JLS. Combining further efficacy data on MTX and MMF with prior opinion could potentially reduce the pre-trial uncertainty so that, when combined with smaller trial sample sizes a compelling evidence base is available.