Langford, Oliver

AbstractBackgroundThe residual approach has found wide application in researching cognitive resilience, a phenomenon conceptually understood as cognitive performance being better‐than‐typical for an individual, despite apparent AD pathology. The standard residual approach extracts information about an individual’s resilience from the residuals of a linear model predicting cognition. This approach is subject to several limiting assumptions. Here, we explore the implications of these assumptions and propose an alternative predictive modeling approach which circumvents them.MethodsOur alternative approach relies on training a machine learning model on a sample of participants who follow a “typical” clinical progression (Fig. 1A). Residuals are then extracted for a heterogeneous dataset that potentially includes resilient participants by comparing observed cognitive outcomes to those predicted by the “typical” model (Fig. 1B). Additionally, our estimate can be refined by effectively estimating and removing model error from the extracted residual by leveraging a second predictive model (Fig. 1C). Using simulated data, we explore the assumptions made by the standard approach and the consequences of their violation. We then contrast this against our approach.ResultsOur simulations demonstrated that the standard residual approach only extracted accurate resilience estimates if the predictors used in the model explained cognition sufficiently well, while being unrelated to resilience itself. That is, as the predictors explain more of the resilience, the residuals became increasingly determined by model error (Fig. 2). We found that residuals achieved with our approach are more precise and provide individualized estimates that are unaffected by the aforementioned trade‐off (Fig. 3).ConclusionsOur approach holds several advantages over the standard approach: The definition of resilience can be freely adjusted and enforced via different definitions of “typical”, it allows for incorporating complex, non‐linear models without removing relevant information from the residual, the error component of the residual can be estimated through cross‐validation, and all predictors that potentially explain cognition can be included into the model without losing resilience information from the residual. Our approach can be used to measure any phenomenon that can be operationalized as a deviation from a typical outcome and, thus, is also applicable for researching resistance to pathology.

AbstractBackgroundMixed model repeated measures (MMRM) is the most common analysis approach in clinical trials for Alzheimer’s disease. The model treats time as a categorical variable, which can be problematic when assessments occur off‐schedule, as including off‐schedule visits can induce bias, and excluding them ignores valuable information and violates the intention to treat principle. This problem has been exacerbated by clinical trial visits which have been delayed due to the COVID19 pandemic. Since MMRM entails a mean parameter for each group and visit, the model can be inefficient for longer trials of Preclinical Alzheimer’s. As an alternative to MMRM, we propose a constrained longitudinal data analysis with natural cubic splines that treats time as continuous and uses test version effects to model the mean over time.MethodWe use model selection techniques to compare the predictive value of spline models to slope models, MMRM, and models that assume a proportional treatment effect using data from several completed trials. These trials include the ADCS trial of donepezil in MCI, trials of intranasal insulin and a fyn kinase inhibitor in mild Alzheimer’s dementia, and the A4 trial of solanezumab in Preclinical Alzheimer’s. Simulation studies assess the power and Type I error of these models for trials in Preclinical Alzheimer’s, MCI, and mild dementia.ResultThe spline model with ADAS‐Cog version effect attained the best predictive value for the trials in mild dementia. The proportional model was preferred for the trial in MCI, but the proportional effect assumption was violated, and the treatment effect estimate was in the wrong direction. Considering that the spline with two degrees of freedom and version effect provide fits that are similar to MMRM in Figure 1, we concluded that two degrees of freedom appear to be sufficient to capture enough of the general trend to assess the treatment effect. AIC results from the A4 Study will be considered and included after the study is completed in Spring 2023.ConclusionThe spline model is shown to have better predictive value and precision in real datasets and has better power and Type I error control in a variety of simulation scenarios.

Abstract Mixed model repeated measures (MMRM) is the most common analysis approach used in clinical trials for Alzheimer's disease and other progressive diseases measured with continuous outcomes over time. The model treats time as a categorical variable, which allows an unconstrained estimate of the mean for each study visit in each randomized group. Categorizing time in this way can be problematic when assessments occur off‐schedule, as including off‐schedule visits can induce bias, and excluding them ignores valuable information and violates the intention to treat principle. This problem has been exacerbated by clinical trial visits which have been delayed due to the COVID19 pandemic. As an alternative to MMRM, we propose a constrained longitudinal data analysis with natural cubic splines that treats time as continuous and uses test version effects to model the mean over time. Compared to categorical‐time models like MMRM and models that assume a proportional treatment effect, the spline model is shown to be more parsimonious and precise in real clinical trial datasets, and has better power and Type I error in a variety of simulation scenarios.