Randomization method of participants is stratified randomization to reduce the possibility of heterogeneity due to pre-pregnancy body mass index in intervention and control groups as much as possible. Also, considering the sample size of less than 100 people in each class and also that pregnant mothers appear in the study for 6-10 weeks continuously and are not known before, it seems a suitable method. Therefore, for a sample size of 320 people; there are four categories that are determined based on pre-pregnancy body mass index (underweight, normal, overweight and obese) and include at least 80 people (40 intervention and 40 control). Then, in each category with block randomization, 80 relevant people are placed in the intervention or control group. In this way, we have 20 quadruple blocks, which is the most common block size, and the order in which pregnant mothers are placed in them is 6, and the number related to the sequence of each block is from 1 to 6, and only one of research members is aware of the size, sequence, or order of blocks. In the next step, the responsible person who does not participate in any of the research steps, will select a sequence of random numbers using R software. The first number is assigned to the block with the same sequence number, and according to the special order of the block, pregnant mothers in each category are assigned to the intervention or control group; prior to the allocation, the assigned group for each participant in the study should not be specified in order to allocation concealment. If the second number in the obtained sequence of random numbers is not related to the sequence of blocks of each class (1 to 6), is discarded until we reach a number that introduces the sequence number of the next block and again in the special order of that block, pregnant mothers will be placed in the intervention or control group. In this way, the adequate sample size will be determined. Numbers related to the sequence of blocks may be repeated several times in a row.
R software:
https://cran.r-project.org/web/packages/randomizeR/randomizeR.pdf
Uschner D, Schindler D, Hilgers RD, Heussen N. randomizeR: an R package for the assessment and implementation of randomization in clinical trials. J Stat Softw. 2018 Jun 19;85(8):1-22
Random sequencing is performed to allocate samples by random permutation blocks using R software and blockrand package. The program codes are as follows:
## stratified by obesity status, 80 in stratum, 2 treatments
library(blockrand)
obese <- blockrand(n=80, id.prefix='obese', block.prefix='obese',stratum='obese')
overweight <- blockrand(n=80, id.prefix='overw', block.prefix='overw',stratum='overweight')
normal <-blockrand(n=80, id.prefix='norm', block.prefix='norm',stratum='normal')
underweight <-blockrand(n=80, id.prefix='undrw', block.prefix='underw',stratum='underweight')
my.study <- rbind(obese,overweight,normal,underweight)
Snow G, Snow MG. Package ‘blockrand’. The Comprehensive R Archive Network. 2013.