Optimal and maximin sample sizes for multicentre cost-effectiveness trials

Abstract

This paper deals with the optimal sample sizes for a multicentre trial in which the cost-effectiveness of two treatments in terms of net monetary benefit is studied. A bivariate random-effects model, with the treatment-by-centre interaction effect being random and the main effect of centres fixed or random, is assumed to describe both costs and effects. The optimal sample sizes concern the number of centres and the number of individuals per centre in each of the treatment conditions. These numbers maximize the efficiency or power for given research costs or minimize the research costs at a desired level of efficiency or power. Information on model parameters and sampling costs are required to calculate these optimal sample sizes. In case of limited information on relevant model parameters, sample size formulas are derived for so-called maximin sample sizes which guarantee a power level at the lowest study costs. Four different maximin sample sizes are derived based on the signs of the lower bounds of two model parameters, with one case being worst compared to others. We numerically evaluate the efficiency of the worst case instead of using others. Finally, an expression is derived for calculating optimal and maximin sample sizes that yield sufficient power to test the cost-effectiveness of two treatments.

Association of HLA‐DQA1 gene polymorphisms with the risk of children primary nephrotic syndrome in Chinese population

“…HLA genes area unit extremely polymorphic and play important roles in various diseases . Studies of smaller sample size from totally different populations have antecedently reportable association between SNPs in HLA‐DQA1 and HLA‐DQB1 and childhood‐onset steroid‐sensitive nephrotic syndrome, which was limited by sample sizes and was restricted to different population . This study aims to investigate whether SNPs of HLA‐DQA1 and HLA‐DQB1 are associated with risk of PNS in 501 PNS patients and 532 healthy controls of Chinese Han population.…”

Bayesian sample size determination for cost-effectiveness studies with censored data

“…An alternate parameterization utilizes the bivariate normal distribution for cost-effectiveness, similar to that published in sample size determination methods for cluster-randomized and multicenter cost-effectiveness trials [ 5 , 6 ]. Although costs and effectiveness are often non-normal, the convenience of modeling and interpretability has made the bivariate normal a common and robust tool for assessing cost-effectiveness [ 21 ], particularly for large trials where asymptotic properties can be invoked [ 22 ].…”

“…Comparing treatments on the basis of cost-effectiveness is increasingly important for clinical trials, and recent publications have considered approaches to assess power for cost-effectiveness studies [ 1 – 3 ]. Often such power calculations assume normally distributed costs and effectiveness, which is convenient as it allows for closed form formulae for sample size determination [ 4 – 6 ]. The bivariate normal approach has been utilized in recent methodological work allow for novel study designs, such as the case of large multicenter trials with or without cluster randomization, where approaches such as maximin have been utilized to obtain optimal sample sizes for cost-effectiveness studies [ 5 , 6 ].…”

“…Often such power calculations assume normally distributed costs and effectiveness, which is convenient as it allows for closed form formulae for sample size determination [ 4 – 6 ]. The bivariate normal approach has been utilized in recent methodological work allow for novel study designs, such as the case of large multicenter trials with or without cluster randomization, where approaches such as maximin have been utilized to obtain optimal sample sizes for cost-effectiveness studies [ 5 , 6 ]. However, for a variety of reasons, observations are often censored, most commonly as a result of the data originating from a survival study.…”

Power Analysis for Two-Level Multisite Randomized Cost-Effectiveness Trials

“…For example, often the analysis units for education studies are students, who are nested within classrooms that are nested within schools, and the intervention could be at any of these levels (Hedges & Rhoads, 2010; What Works Clearinghouse, 2014). Recent work in health science (Manju et al, 2014, 2015; Manju et al, 2019) provides formulas and an interactive computer program to calculate power for two-level RCETs, where the treatment is either at the first level (e.g., patients) or the second level (e.g., health care providers). These methods consider unconditional models that do not include any covariates in the analyses and assume individual-level cost data are always available.…”