Power Estimation in Planning Randomized Two-Arm Pre-Post Intervention Trials with Repeated Longitudinal Outcomes

Journal of Biometrics & Biostatistics

article

Yirui Hu¹  Donald R Hoover² 

[1] Biomedical and Translational Informatics;Department of Statistics and Biostatistics and the Institute for Health, Health Care Policy and Aging Research, Rutgers University

关键词: Compound symmetry;    Power and sample size estimation;    Toeplitz correlation;    Optimal allocation;    Pre-post interventional study;    Generalized least squares;    Mixed model;   

DOI  :  10.4172/2155-6180.1000403

来源: Hilaris Publisher

【 摘 要 】

Background: Intervention effect on ongoing medical processes is estimated from clinical trials on units (i.e. personsor facilities) with fixed timing of repeated longitudinal measurements. All units start out untreated. A randomly chosensubset is switched to the intervention at the same time point. The pre-post switch change in the outcome between theseunits and unswitched controls is compared using Generalized Least Squares models. Power estimation for such studiesis hindered by lack of available GLS based approaches and normative data.Methods: We derive Generalized Least Squares variance of the intervention effect. For the commonly assumedcompound symmetry correlation structure, this leads to simple power formulas with important optimality properties. Tomaximize power given a constrained number of total time points, we investigate on the optimal pre-post allocation withthe local minimization of variance.Results: In four examples from nursing home and HIV patients, the Toepltiz within-unit correlation of repeatedmeasures differed from compound symmetry. We applied empirical Toeplitz based calculations for variance of theestimated intervention effect to these examples (each with up to seven longitudinal measures). Unlike what happenedunder compound symmetry, where power was often maximized with multiple observations being pre-intervention, forthese examples, having one pre-intervention measure tended to maximize power. Attempts to approximate the Toeplitzvariance structures with compound symmetry (to take advantage of the simpler formulas) resulted in overestimation ofpower for these examples.Conclusions: While compound symmetry correlation among repeated within-unit measures leads to simple powerestimation formulas, this structure often did not hold. There may be strong underestimation of variance of the interventioneffect estimate from incorporating short-term within-unit correlation estimates as a common compound symmetrycorrelation to approximate an unknown Toeplitz correlation without adequately accounting for the correlation betweenrepeated measures declining with time.

【 授权许可】

Unknown