Trick or Treat …. In April!
Take one small envelope but…
Don’t open it! Teaching and Leading EBM 2019
Randomization: simple, stratified and block
Juan M. Lozano, MD, MSc .Describe three methods of randomization: o Simple. o Stratified. o Block. .Describe the method of randomization used in a paper. .Demonstrate interactive teaching strategies and tools. Randomized controlled trial
“A clinical trial that involves at least one test treatment and one control treatment, concurrent enrollment and follow-up of the test- and control-treated groups, and in which the treatments to be administered are selected by a random process.”
National Library of Medicine Random allocation
Allocates subjects to the interventions (experimental or control) through chance. Each subject has a known probability (usually equal) of receiving any of the interventions. Random allocation random sampling. Why?
Creates study groups with similar prognosis = it controls bias by making the groups homogeneous at baseline. Equal distribution of the benefits and risks of the interventions (which are usually unknown). Baseline for using statistical methods. “The cookie group evaluated teachers significantly better than the control group. Course material was considered better and summation scores evaluating the course overall were significantly higher in the cookie group.” Medical Education 2018 doi: 10.1111/medu.13627 My PICOTT question
P In participants at Duke Teaching and Leading EBM 2019, I does receiving complimentary dark chocolate bites, C as compared with milk chocolate bites, O increase their satisfaction with the workshop? T Type of question: intervention T Type of study: RCT Simple randomization: Non-random allocation: . Coin toss, throwing a dice . Order of entry into the . Random number table: clinic/hospital o Textbook . Day of the week, month o Computer generated . Last digit of document (ID, . Random allocation software clinical record, etc.) https://www.openepi.com/ 1 = Dark; 20 participants 2 = Milk; 30 participants Probability of 20 heads in 50 flips of a coin: 6% Trick or Treat …. In April!
Take your small envelope and…
open it! Table 1. Baseline characteristics of participants – N (%) Characteristic Dark (N = 20) Milk (N = 30) Female Living in NC Pediatrician How do the “pros” do it? Other methods of randomization:
. Stratified . Block Tab 5, first document
Researchers randomized patients stratifying according to trial site (10 EDs in the U.S.) and duration of symptoms (<48 h vs ≥48 h); which of the following statements is correct about such stratification?
A. Stratification made equal the number of patients randomized to probiotic vs. placebo. B. It was necessary to know each patient’s duration of symptoms before randomizing her / him to the study interventions. C. This stratification intended to ensure good baseline balance between the treatment groups regarding gender. D. The trial had 12 strata (10 sites plus 2 duration of symptoms). Stratified randomization Stratified randomization
. Keeps balance between groups in terms of selected participants’ baseline characteristics . Participants are randomized within each stratum . The total number of strata is the product of all categories within each stratum: . Center: 4 (A, B, C, D) . Gender: 2 (M, F) 16 strata . Obesity: 2 (Y, N) The trial also used “permuted blocks with random block sizes”; which one of the following statements is correct about this block randomization?
A. Each participant had a 1:1 probability of being allocated to probiotic or to placebo. B. Participants were randomly allocated in blocks of varying size (2, 4, 6, etc.) at a time, with all patients in each block randomized together to probiotic or to placebo. C. Block randomization contributed to achieve good baseline balance regarding gender. D. Using blocks of different size threatens the concealment of allocation. Block randomization
. Keeps the number of participants in each treatment group within a desired ratio (usually 1:1 or “equal”) . Participants are randomized within each block . Varying block sizes preserve the concealment of allocation (especially important in “open” trials). C. Random selection of blocks (i.e., 1, 3, A. Block size of 4 4, 4, 3, 6, 1, 2, 5, 2 – 40 participants)
B. Possible permutations with repetition (i.e., 2 to control group, 2 to treatment group) Take home messages
Randomization: Each participant has a known probability (usually equal) of receiving any of the interventions
Type Balance regarding Stratified Selected baseline characteristics Block Number (ratio) of participants per group Thank you!!