Blinded and unblinded sample size reestimation in crossover trials balanced for period

Michael J. Grayling; Adrian P. Mander; James M. S. Wason

doi:10.1002/bimj.201700092

Power and sample size requirements for GEE analyses of cluster randomized crossover trials

Statistics in Medicine ◽

10.1002/sim.7995 ◽

2018 ◽

Vol 38 (4) ◽

pp. 636-649 ◽

Cited By ~ 8

Author(s):

Fan Li ◽

Andrew B. Forbes ◽

Elizabeth L. Turner ◽

John S. Preisser

Keyword(s):

Sample Size ◽

Crossover Trials ◽

Cluster Randomized

Download Full-text

Sample size re-estimation in crossover trials: application to the AIM HY-INFORM study

Trials ◽

10.1186/s13063-019-3724-6 ◽

2019 ◽

Vol 20 (1) ◽

Cited By ~ 1

Author(s):

Julie Wych ◽

Michael J. Grayling ◽

Adrian P. Mander

Keyword(s):

Sample Size ◽

Repeated Measures ◽

Interim Analysis ◽

Blood Pressure Response ◽

Sample Size Calculation ◽

Primary Treatment ◽

Design Stage ◽

Crossover Designs ◽

Crossover Trials ◽

Randomised Controlled

Abstract Background Crossover designs are commonly utilised in randomised controlled trials investigating treatments for long-term chronic illnesses. One problem with this design is its inherent repeated measures necessitate the availability of an estimate of the within-person standard deviation (SD) to perform a sample size calculation, which may be rarely available at the design stage of a trial. Interim sample size re-estimation designs can be used to help alleviate this issue by adapting the sample size mid-way through the trial, using accrued information in a statistically robust way. Methods The AIM HY-INFORM study is part of the Informative Markers in Hypertension (AIM HY) Programme and comprises two crossover trials, each with a planned recruitment of 600 participants. The objective of the study is to test whether blood pressure response to first line antihypertensive treatment depends on ethnicity. An interim analysis is planned to reassess the assumptions of the planned sample size for the study. The aims of this paper are: (1) to provide a formula for sample size re-estimation in both crossover trials; and (2) to present a simulation study of the planned interim analysis to investigate alternative within-person SDs to that assumed. Results The AIM HY-INFORM protocol sample size calculation fixes the within-person SD to be 8 mmHg, giving > 90% power for a primary treatment effect of 4 mmHg. Using the method developed here and simulating the interim sample size reassessment, if we were to see a larger within-person SD of 9 mmHg at interim, 640 participants for 90% power 90% of the time in the three-period three-treatment design would be required. Similarly, in the four-period four-treatment crossover design, 602 participants would be required. Conclusions The formulas presented here provide a method for re-estimating the sample size in crossover trials. In the context of the AIM HY-INFORM study, simulating the interim analysis allows us to explore the results of a possible increase in the within-person SD from that assumed. Simulations show that without increasing the planned sample size of 600 participants, we can reasonably still expect to achieve 80% power with a small increase in the within-person SD from that assumed. Trial registration ClinicalTrials.gov, NCT02847338. Registered on 28 July 2016.

Download Full-text

SO063POOR REPORTING AND CONDUCT OF CLUSTER AND CROSSOVER RANDOMISED TRIALS IN THE HAEMODIALYSIS POPULATION

Nephrology Dialysis Transplantation ◽

10.1093/ndt/gfaa139.so063 ◽

2020 ◽

Vol 35 (Supplement_3) ◽

Author(s):

Prachi Kaushal ◽

Daniel March ◽

Helen Eborall ◽

Laura Gray ◽

James Burton

Keyword(s):

Sample Size ◽

Cluster Size ◽

Primary Outcome ◽

Statistical Significance ◽

Attrition Rate ◽

Controlled Trials ◽

Cochrane Central Register ◽

Randomised Trials ◽

Stepped Wedge ◽

Crossover Trials

Abstract Background and Aims There are fewer trials conducted in nephrology than any other speciality. Many trials fail to recruit to target and are therefore underpowered, resulting in unclear evidence and unlikely to be implemented into clinical practice. Carefully designed trials could improve the treatment options for haemodialysis patients. These could include designs other than individually randomised parallel group designs. We have reviewed cluster, stepped wedge and crossover randomised trials carried out in prevalent haemodialysis patients. Method A search for randomised controlled trials (RCTs) in haemodialysis patients was conducted across five databases (MEDLINE, EMBASE, CINAHL, Cochrane Central Register of Controlled Trials and ClinicalTrials.gov) in October 2019. The first publication of a stepped wedge trial in haemodialysis was in 2013, so we searched from 2013-2019. Inclusion criteria were RCTs including adults receiving at least 3 months of haemodialysis. Once eligible studies were identified, we compared the sample size sections to the CONSORT requirements and extracted data on these parameters and assessed recruitment targets, attrition rates and primary outcome result. Results We identified 911 RCTs of which 13 (1.4%) cluster randomised trials, including one stepped wedge, and 145 (15.9%) crossover trials were identified. Data were extracted for 10 randomly selected crossover trials. The total sample size required was not reported in 46% (n=6) cluster trials and 60% (n=6) crossover trials, with only 37.5% (n=3) of cluster trials and 50% (n=4) of crossover trials providing a justification for their sample size. In cluster trials, 13% (n=1) stated number of clusters required and 25% (n=2) gave the average cluster size, none of these stated the anticipated variance in cluster size and only 12.5% (n=1) stated the intracluster correlation (ICC) used to calculate the sample size, with none taking into account uncertainty in ICC. In 31% (n=4) of the cluster trials, the dialysis centre was the cluster and in the other 69% (n=9), the dialysis shift was the cluster. For crossover trials, within participant variability was only accounted for in 12.5% (n=1) of trials. A CONSORT flow diagram was reported in 70% of the trials assessed, with 39% recruiting to target. There were also some differences seen between crossover and cluster designs. The average number of participants at the start of a cluster trial was 470, compared to 26 in crossover trials. Where able to assess, cluster trials had an average attrition rate of 25% this was lower in crossover trials, where the average attrition rate was 14%. Eighty five percent (n=11) of cluster trials had a primary outcome reaching statistical significance, however only 20% (n=2) of crossover trials reached statistical significance. Conclusion Cluster and crossover randomised trials are poorly reported. There was insufficient information provided for sample size calculations to be replicated in majority of trials. The average attrition rate was low in crossover trials, however a large proportion of these trials did not have a primary outcome reaching statistical significance, suggesting these trials could have been underpowered. Improvements in the design, conduct and reporting of cluster and crossover trials in the haemodialysis population is urgently required.

Download Full-text