Sample size calculation based on precision for pilot sequential multiple assignment randomized trial (SMART)

Yan, Xiaoxi; Ghosh, Palash; Chakraborty, Bibhas

doi:10.1002/bimj.201900364

Cited by 13 publications

(10 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The power-based sample size calculation method should only be used if the goal of the study is to perform a hypothesis test 28 , as is typical in a confirmatory study. This method requires a working knowledge of the standardized proximal effect sizes of intervention categories (i.e.…”

Section: Power-based Sample Size Calculationmentioning

confidence: 99%

“…In order to operationalize the sample size calculation for pilot MRT, we suggest the sample size to be calculated in order to achieve certain level of accuracy or precision for the estimates of 's; see, e.g., Kelley et al 30 or Maxwell et al 31 , for discussions in the classical settings. The rationale and benefit of this approach in the relatively recent context of pilot SMARTs has been described in Yan et al 28 .…”

Section: Precision-based Sample Size Calculationmentioning

confidence: 99%

See 1 more Smart Citation

Micro-Randomized Trial with Flexible Design and Sample Size Considerations

Xu¹,

Yan²,

Figueroa³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

Technological advancements have made it possible to deliver mobile health interventions to individuals. A novel framework that has emerged from such advancements is the just-in-time adaptive intervention (JITAI), which aims to suggest the right support to the individuals when their needs arise. The micro-randomized trial (MRT) design has been proposed recently to test the proximal effects of these JITAIs. In an MRT, participants are repeatedly randomized to one of the intervention categories of various components, at a scale of hundreds or thousands of decision-points over the study. However, the extant MRT framework only considers the components with a fixed number of categories. We propose a novel extension of the MRT design, by allowing flexible addition of more categories to the components during the study, in a spirit analogous to adaptive platform trials. Accordingly, we develop a methodology to allow for simultaneous comparisons of varying numbers of categories over a given study period. The proposed methodology is motivated by collaboration on the DIAMANTE study, which learns to adaptively deliver multiple complex kinds of text messages to encourage physical activity among the patients with Diabetes and Depression. We apply a methodology similar to the generalized estimating equation approach on the longitudinal data arising from the proposed MRT, to develop novel test statistics for assessing the proximal effects and deriving the associated sample size calculators. We conduct simulation studies to evaluate the sample size calculators, based on both power and precision. We have developed an R shiny application of the sample size calculators.

show abstract

Section: Power-based Sample Size Calculationmentioning

confidence: 99%

Section: Precision-based Sample Size Calculationmentioning

confidence: 99%

Micro-Randomized Trial with Flexible Design and Sample Size Considerations

Xu¹,

Yan²,

Figueroa³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…The number of research participants was calculated through the statistical formula that stipulates the representative percentage of the total number of students registered in secondary education centers in Andalusia, Ceuta, and Melilla (N = 386.821). This formula is indicated with a confidence level of 95% and a maximum estimation error of 2.1%, which makes the selected sample statistically acceptable [50]. The age of the sample ranges from 12 to 16 years (M = 13.97 years; DT = 1.410), with 376 (49.8%) males and 379 (50.2%) females.…”

Section: Population and Samplementioning

confidence: 99%

Cyberbullying Analysis in Intercultural Educational Environments Using Binary Logistic Regressions

2021

View full text Add to dashboard Cite

The goal of this study is to analyze how religion, ethnic group, and race influence the likelihood of becoming either a cybervictim or cyberbully in intercultural educational environments. In the research, 755 students in secondary education were analyzed in the south of Spain through the Cyberbullying Scale for students with Cultural and Religious Diversity (CSCRD). The analyses were carried out using the Statistical Package SPSS and the STATA software. The results obtained from the Kruskal–Wallis H test showed significant differences according to these aspects, for both the cybervictim and cyberbully parameters. The results stemming from binary logistic regressions confirmed such differences and regarded those students who belong to the Muslim religion, the gypsy ethnic group and the Asian race as being more likely to become cybervictims. Furthermore, these analyses showed that Gypsy and Asian students were also more likely to be cyberbullies than other groups. The main conclusions state that minority groups are more likely to suffer cyberbullying in intercultural educational environments, and that students from these groups are also more likely to become cyberbullies.

show abstract

“…Furthermore, they assume asymptotic normality which may not hold due to the small sample size of embedded treatment sequences in a SMART. Yan et al (2020) developed a frequentist sample size calculation method for pilot SMARTs. In particular, they size a SMART so that the mean outcome of a DTR is within "a margin of error".…”

Section: Introductionmentioning

confidence: 99%

“…In summary, our method makes comparisons between > 2 embedded DTRs adjusting for multiplicity, rather than looking at them marginally. The authors of Yan et al (2020) emphasize if there is no hypothesis testing in the pilot SMART, then standard power analysis is not relevant. Therefore, their method fills an important gap for pilot SMARTs.…”

Section: Introductionmentioning

confidence: 99%

Bayesian Set of Best Dynamic Treatment Regimes and Sample Size Determination for SMARTs with Binary Outcomes

Artman,

Ertefaie,

Lynch

et al. 2020

Preprint

View full text Add to dashboard Cite

One of the main goals of sequential, multiple assignment, randomized trials (SMART) is to find the most efficacious design embedded dynamic treatment regimes. The analysis method known as multiple comparisons with the best (MCB) allows comparison between dynamic treatment regimes and identification of a set of optimal regimes in the frequentist setting for continuous outcomes, thereby, directly addressing the main goal of a SMART. In this paper, we develop a Bayesian generalization to MCB for SMARTs with binary outcomes. Furthermore, we show how to choose the sample size so that the inferior embedded DTRs are screened out with a specified power. We compare log-odds between different DTRs using their exact distribution without relying on asymptotic normality in either the analysis or the power calculation. We conduct extensive simulation studies under two SMART designs and illustrate our method's application to the Adaptive Treatment for Alcohol and Cocaine Dependence (ENGAGE)

show abstract

Sample size calculation based on precision for pilot sequential multiple assignment randomized trial (SMART)

Cited by 13 publications

References 29 publications

Micro-Randomized Trial with Flexible Design and Sample Size Considerations

Micro-Randomized Trial with Flexible Design and Sample Size Considerations

Cyberbullying Analysis in Intercultural Educational Environments Using Binary Logistic Regressions

Bayesian Set of Best Dynamic Treatment Regimes and Sample Size Determination for SMARTs with Binary Outcomes

Contact Info

Product

Resources

About