148 research outputs found
Power formulas for mixed effects models with random slope and intercept comparing rate of change across groups
No full textWe have previously derived power calculation formulas for cohort studies and clinical trials using the longitudinal mixed effects model with random slopes and intercepts to compare rate of change across groups [Ard & Edland, Power calculations for clinical trials in Alzheimer’s disease. J Alzheim Dis 2011;21:369–77]. We here generalize these power formulas to accommodate 1) missing data due to study subject attrition common to longitudinal studies, 2) unequal sample size across groups, and 3) unequal variance parameters across groups. We demonstrate how these formulas can be used to power a future study even when the design of available pilot study data (i.e., number and interval between longitudinal observations) does not match the design of the planned future study. We demonstrate how differences in variance parameters across groups, typically overlooked in power calculations, can have a dramatic effect on statistical power. This is especially relevant to clinical trials, where changes over time in the treatment arm reflect background variability in progression observed in the placebo control arm plus variability in response to treatment, meaning that power calculations based only on the placebo arm covariance structure may be anticonservative. These more general power formulas are a useful resource for understanding the relative influence of these multiple factors on the efficiency of cohort studies and clinical trials, and for designing future trials under the random slopes and intercepts model
Recommended from our members
Power and Probability Calculations in Longitudinal Outcome Measures and Image Analysis, with Applications to Biomedical Data
No full textIn this dissertation, we aim to utilize modern statistical frameworks to perform power and probability calculations in longitudinal outcome measures and image analysis. Chapter 1 and Chapter 2 primarily focus on longitudinal outcome measures data. In practice, two different analysis plans are commonly applied to these data: linear mixed effects model and repeated measures analysis. In Chapter 1, we are interested in generalizing current power formulas for linear mixed effects models to accommodate missing data due to study subject attrition, and unequal sample size and variance parameters across groups.For repeated measures analysis, a covariance structure needs to be specified when modeling the data and computing type I error and power. In Chapter 2, we describe a parsimonious covariance structure for repeated measures analysis that is useful for modeling longitudinal repeated measures of chronic progressive conditions and derive the power calculation formulas. In image analysis, finding the peak height distribution and power for peak detection are known to be challenging due to the spatial aspect of the data. In Chapter 3, we propose a novel way to approximate the power for peak detection using Gaussian random field theory (RFT) and demonstrate scenarios where the approximation works well. We also apply our formulas to 2D and 3D simulated datasets, and the 3D data is induced by real fMRI data to measure performance in a realistic setting.The main limitation of RFT-based image analysis is the model assumptions, and these assumptions are known to be difficult to check and even not appropriate in many application settings. In Chapter 4, we seek to relax the stationarity assumption and study the peak height distribution of non-stationary Gaussian random fields. The explicit formula for the peak height distribution is derived for 1D smooth Gaussian random fields and efficient numerical algorithms are proposed as a general solution for computing the peak height distribution in applications
Recommended from our members
From Sleep to Aging: Statistical Tools for Understanding Time-Dependent Health Patterns in Clinical Trials
No full textThe objective of this dissertation is to develop and extend statistical methods for analyzing repeated measures data in cognitive aging and sleep health. Challenges arise when these repeated measures show short-term temporal dependence, latent heterogeneity, and complex correlation structures, where the use of traditional methods may reduce the efficiency and accuracy of estimation. This work consists of four main statistical applications. The first chapter examines sex differences in cognitive decline using Cox models applied to the National Alzheimer’s Coordinating Center (NACC) cohort. Our results suggest that there may exist potential diagnostic bias, such that the observed onset of cognitive decline in women is delayed due to their verbal memory advantage. The second chapter focuses on the use of composite measures in clinical trial planning. We assess the power and efficiency of using Preclinical Alzheimer’s Cognitive Composite (PACC) and its variants as clinical trial endpoints for patients with amnestic mild cognitive impairment (aMCI). Simulations informed by NACC data show that the use of reciprocal standard deviation weighting can result in inefficient cognitive composites in this population. Therefore, we advocate a careful examination of the statistical properties of the components before they are used in a standard composite. The third chapter investigates efficiency gains in Generalized Estimating Equations (GEE) models using a chronic progressive working correlation structure. For data demonstrating a chronic progressive pattern, variance tends to increase as time increases. Simulations show improved power and reduced standard errors when the correlation structure reflects the chronic progressive pattern. The fourth chapter models daily sleep duration in shift workers using a two-stage autoregressive Gamma Hidden Markov Model (HMM). Latent sleep sufficiency states are estimated via HMM, followed by autoregressive mixed-effects modeling conditioned on decoded state history. Simulation and empirical analysis support an accurate recovery of the autoregressive structure
F4–02–01: Evidence of learning or placebo effects relevant to clinical trials of mild cognitive impairment
No full textRecommended from our members
α-Synuclein Seed Amplification Assay Amplification Parameters and the Risk of Progression in Prodromal Parkinson Disease.
No full textOBJECTIVES: Tools are needed to evaluate the risk of developing Parkinson disease (PD) in at-risk populations. In this study, we examine differences in alpha-synuclein seed amplification assay (αSyn-SAA) qualitative results and amplification parameters between nonmanifesting carriers (NMCs) of PD-related pathogenic variants, prodromal PD, and PD and the risk of developing a synucleinopathy in participants with prodromal PD. METHODS: Cross-sectional and longitudinal CSF αSyn-SAA results from participants in the Parkinsons Progression Markers Initiative were analyzed. αSyn-SAA positivity and amplification parameters (maximum fluorescence [Fmax], time-to-threshold [TTT], time-to-50% Fmax [T50], and area under the curve [AUC]) were compared between NMCs, participants with prodromal PD, and participants with PD, and their relationship with the likelihood of phenoconversion in participants with prodromal PD was investigated. RESULTS: Samples from 1,027 participants were analyzed (159 healthy controls [HCs], 247 NMCs, 96 participants with prodromal PD, and 525 participants with PD). TTT and T50 were faster, and AUC was higher in αSyn-SAA+ participants with prodromal PD and PD than αSyn-SAA+ NMCs and HC participants (Kruskal-Wallis χ2 = 4.15-13.96, p < 0.0002-0.04). Participants with prodromal PD with positive αSyn-SAA tests and faster TTT had higher rates of phenoconversion (log-rank p = 0.001 and log-rank test-for-trend p < 0.0001). There were no changes in 48 participants with prodromal PD with longitudinal assays. DISCUSSION: αSyn-SAA positivity and faster seed amplification are associated with a greater risk of developing PD in at-risk individuals and may aid in predicting phenoconversion
P1‐321: Biological standards to control variability in cerebrospinal fluid beta‐amyloid readings
No full textThe Chronic Progressive Repeated Measures (CPRM) Model for Clinical Trials Comparing Change Over Time in Quantitative Trait Outcomes
No full textRepeated measures analysis is a common analysis plan for clinical trials comparing change over time in quantitative trait outcomes in treatment versus control. Mixed model for repeated measures (MMRM) assuming an unstructured covariance of repeated measures is the default statistical analysis plan, with alternative covariance structures specified in the event that the MMRM model with unstructured covariance does not converge. We here describe a parsimonious covariance structure for repeated measures analysis that is specifically appropriate for longitudinal repeated measures of chronic progressive conditions. This model has the parsimonious features of the mixed effects model with random slopes and intercepts, but without restricting the repeated measure means to be linear with time. We demonstrate with data from completed trials that this pattern of longitudinal trajectories spreading apart over time is typical of Alzheimer’s disease. We further demonstrate that alternative covariance structures typically specified in statistical analysis plans using MMRM perform poorly for chronic progressive conditions, with the compound symmetry model being anticonservative, and the autoregressive model being poorly powered. Finally, we derive power calculation formulas for the chronic progressive repeated measures model that have the advantage of being independent of the design of the pilot studies informing the power calculations. When data follow the pattern of a chronic progressive condition. These power formulas are also appropriate for sizing clinical trials using MMRM analysis with unstructured covariance of repeated measures
Power formulas for mixed effects models with random slope and intercept comparing rate of change across groups
No full textPlatinum Group Element Mineralization in the Reinfjord Ultramafic Complex - A geochemical and petrological study of the PGE-enriched parts of the RF-4 drill core
No full textThe Reinfjord Ultramafic Complex (RUC), a part of the Seiland Igneous Province (SIP), is located in Nord-Troms, Norway. The RUC constitutes parts of deep seated magmatic conduit system that comprises a low tenor PGE-reef. In the RF-4 drill core, the reef is located at a depth of approximately 62 m, where the Pd+Pt concentration reaches around 0.8 ppm. This thesis places its emphasis on investigating the PGE mineralization of the RF-4 drill core, in order to obtain knowledge regarding the formation mechanisms of the reef. The thesis also compares the results with the mineralization found in the RF-1 drill core. Chemical data was provided by ICP spectrometry, SEM and EPMA, while preliminary investigations was done by optical microscopy.
The PGE host rock in RF-4 is a dunite, containing mostly olivine, with minor clinopyroxene and orthopyroxene. Interstitial carbonates and amphibole is also present. The RF-4 drill core revealed a wide variety of PGM's, where bismuthotellurides, arsenides and sulfurarsenides are the most common phases. Other phases comprise antimonides and PGE alloys. The PGM's are often BMS associated, where both pentlandite and pyrrhotite are common hosts. The presence of the PGE-arsenides-and sulfurarsenides, set the mineralization in RF-4 apart from that of RF-1. Possible scenarios explaining the differences in PGM assemblage are localized gain of As from surrounding rocks, removal by hydrothermal fluids or vertical displacement of As-rich zones due to tectonics.
The author suggest that the PGM's formed from sulfide fractionation, and the concentration process is thus magmatic. Further, the author propose the possibility of PGM's exsolving from an iss, due the PGM's often showing an affinity to Cu-rich pyrrhotite. PGM's can also be found associated with serpentine, sometimes showing remainders of BMS, indicating high degrees of serpentinization. Serpentinization is also responsible for the breakdown of chalcopyrite, leading to the formation of native Cu. Stability field calculations with constant bulk composition a.k.a a pesudosection was done for a mineral assemblage comprising amphibole, orthopyroxene, olivine and magnesite, which represents a secondary volatile phase. The model propose that re-mobilization of PGE and Au could happen at high P ( bar) and mid-to-high T ( K) conditions. Pd would be more soluble in the fluid compared to Pt, and the Pd-cluster in RF-1 could thus be the result of such processes
The MAX Statistic is Less Powerful for Genome Wide Association Studies Under Most Alternative Hypotheses
No full textGenotypic association studies are prone to inflated type I error rates if multiple hypothesis testing is performed, e.g., sequentially testing for recessive, multiplicative, and dominant risk. Alternatives to multiple hypothesis testing include the model independent genotypic c2 test, the efficiency robust MAX statistic, which corrects for multiple comparisons but with some loss of power, or a single Armitage test for multiplicative trend, which has optimal power when the multiplicative model holds but with some loss of power when dominant or recessive models underlie the genetic association. We used Monte Carlo simulations to describe the relative performance of these three approaches under a range of scenarios. All three approaches maintained their nominal type I error rates. The genotypic c2 and MAX statistics were more powerful when testing a strictly recessive genetic effect or when testing a dominant effect when the allele frequency was high. The Armitage test for multiplicative trend was most powerful for the broad range of scenarios where heterozygote risk is intermediate between recessive and dominant risk. Moreover, all tests had limited power to detect recessive genetic risk unless the sample size was large, and conversely all tests were relatively well powered to detect dominant risk. Taken together, these results suggest the general utility of the multiplicative trend test when the underlying genetic model is unknown
- …
