If i am doing a prospective randomised study on fracture pt and i have selected 90 patients and i have to divide them into 3 groups. But the patients will be recruited as they will come to the clinci over a period of time. How will i ensure randomisation and equal number of 30 in 3 groups?
fracture pt will be divided into 3 groups
There are 90 items in your list (30 ones, 30 twos and 30 threes). Here are the first 10 in random order. (You can do more at https://www.random.org/lists/). Give the first patient the assignment in the first row. Then keep going down the list as they come in.
- For true randomisation equal numbers in all three groups are not possible. Any way of predetermining group assingment will create a bias. In the proposed randomisation by Elizabeth E LeClair time and date of addmission are biases since they can affect the age of the patients or their activity profile (ie: working age people have a greater chance of sustaining a fracture during the weekend when they have more free time to venture in sports and other activities whereas retired people have equal chances every day). Depending on the type of fracture you are looking at, however, this might not have a significant impact.
I would recommend “block randomization“ as the easiest way with blocks of 6 or 9 people if you don’t have to consider confounders. If you want one or more variables to be equally distributed in three group, ” stratified randomization“ is a better option. (baseline characteristics of all subjects must be available before allocation.) Please see the reference linked below. It contains links to useful online software for randomination.
Article An overview of randomization techniques: An unbiased assessm...
- Badges
- Science topic
- Similar topics
- Correction