Network meta-analysis (NMA) is a new statistical approach which comes from head to head meta-analysis. Hence, NMA inherits all methodology challenges of head to head meta-analysis and with increased complexity results due to more intervention treatments involved. The issue of sample size and statistical power in individual trial and head to head meta-analysis is widely emphasized currently; however, they are not been paid due attention in NMA. This article aims to introduce the theory, computational principles and software implementation using examples with step by step approach.
The robustness of results of statistical analysis would be altered on the condition of repeated update of traditional meta-analysis and cumulative meta-analysis. In addition, the cumulative meta-analysis lacks estimation of the sample size. While trail sequential analysis (TSA), which introduces group sequential analysis in meta-analysis, can adjust the random error and ultimately estimate the required sample size of the systematic review or meta-analysis. TSA is performed in TSA software. In the present study, we aimed to introduce how to use the TSA software for performing meta-analysis.
This article introduces the methods about how to use the summary statistics such as the sample median, minimum and maximum to estimate the sample mean and standard deviation for continuous outcomes. For the purpose of illustration, we also apply the existing methods to a real data example.
With increasing amount of attention being paid to single case randomized controlled trial (N-of-1 trials), sample size estimation has become an important issue for clinical researchers. This paper mainly introduces the model and hypothesis of N-of-1 trials. Based on the hypothetical model, sample size estimation methods of fixed model and random model are proposed. The premises of the model application, formulas and examples are then given. It is expected in case of conduction N-of-1 trials, the correct methods are used to estimate sample size and improve the research quality of N-of-1 trials.
Sample size calculation is an important factor to evaluate the reliability of the diagnostic test. In this paper, a case study of the clinical diagnostic test of artificial intelligence for identification of liver contrast-enhanced ultrasound was performed to conduct two-category and multi-categories studies. Based on sensitivity and specificity, the sample size was then estimated in combination with the statistical characteristics of disease incidence, test level and one/two-sided test. Eventually, the sample size was corrected by integrating the factors of the proportion of training/test dataset and the dropout rate of cases in the medical image recognition system. Moreover, the application of Sample Size Calculator, MedCalc, PASS, and other software can accelerate sample size calculation and reduce the amount of labor.
Stepped wedge cluster randomized trials (SW-CRT) is a kind of cluster randomized controlled trial mainly applied in the field of public health policy that has emerged in recent years, which has gradually attracted the attention of workers in the field of health and wellness. At present, this trial method is not widely used at home and abroad, and there are various ways of sample size calculation and statistical analysis. This paper describes the principles, categories, and differences between SW-CRT and traditional randomized controlled trials, and outlines sample size calculation and statistical analysis methods. In general, SW-CRT is characterized by clustering, cross-design, and measurement of results at multiple time points. In terms of sample size calculation, it is necessary to distinguish between clusters with the same and different sizes, and commonly used sample size calculation procedures can be implemented in Stata, R, and SAS software, as well as in fixed online websites, including the "Steppedwedge" program, the "swCRTdesign" program, the "Swdpwr" program, the "CRTpowerdist" program, and the "Shiny CRT Calculator" tool and so on. Based on the design characteristics of SW-CRT, the researcher should also consider the confounding factors of time effects and repeated measurements of result. Therefore, the statistical analysis methods are often based on generalized linear mixed model (GLMM) and generalized estimating equations (GEE). However, most of the above models have been proposed based on cross-sectional studies, there is a lack of statistical methods for queue design and SW-CRT with transitional period now, and more comprehensive methodological exploration is still needed in the future.
ObjectiveTo explore the parameter selection of different sample size estimation methods and the differences in estimation results in single-group target value clinical trials with rate as the outcome evaluation index. MethodsWe conducted a literature review to assess the method of target value selection for single-group target value clinical trials. Then, different values of target value (P0), clinical expected value (P1), and class II error level (β) were set through numerical simulation. Sample size results estimated using different sample size estimation methods were obtained using PASS software. The coefficient of variation, range/mean, analysis of variance and other methods were used to compare the differences between different methods. ResultsAnalysis of the data simulation results showed: when the expected value P1 was fixed, the sample size first decreased rapidly and then decreased slowly along with the increase or decrease of the targeted value P0 on both sides of the sample size limit value. When the difference between P0 and P1 was within 0.15, the ratio before and after correction could be controlled within 0.9. When the difference between P0 and P1 was more than 0.6, the ratio before and after correction approached 0.5. When P0+P1≈1, the ratio of different standard error choices (Sp0 or Sp1) to the estimated sample size was close to 1. When 0.65<P0+P1<1.35, the ratio of different standard error choices (Sp0 or Sp1) to the estimated sample size was about 3:1. When the confidence was 0.8, P0 and P1 were between 0.25 and 0.75 and between 0.20 and 0.80, respectively. We found little difference among the sample sizes estimated using these five methods (CV<0.10, range/mean<0.2). ConclusionThere are some differences among different sample size estimation methods, however, when P0 and P1 values are around 0.5, the differences between different methods are small, suggesting that appropriate methods should be selected for sample size estimation.