Game 7 · Clinical trials
How do you keep a trial fair?
Randomisation, allocation concealment, blinding, intention-to-treat and per-protocol analysis.
Read with: Clinical trials78 DNB questions
Play
Keep the trial fair from the first patient to the last analysis
Each round tests one safeguard: predicting allocations from a block list, the group sizes simple randomisation gives, who can be blinded, and how intention to treat and per protocol differ. Every third round is a short question. You have three lives.
Loading the first round…
Randomisation
Simple, block and stratified randomisation
Random allocation means a chance process that nobody can predict decides who gets which treatment. A computer-generated list or a random-number table does this. Alternate patients, date of birth, hospital number and day of admission are predictable, so they are not randomisation.
Simple randomisation is a coin toss for each patient. It cannot be predicted, but in a small trial the groups can be unequal. The number of patients given A follows a binomial distribution.
P(k patients get A) = C(n, k) × 0.5ⁿ
- n
- number of patients randomised
- k
- number allocated to treatment A
- C(n, k)
- number of ways of choosing k patients out of n
With 20 patients, one group has 7 or fewer in 26% of trials, and an exact 10 : 10 split happens only 17.6% of the time. Block randomisation fixes this: a block of 4 always holds two A and two B, in one of 6 orders (AABB, ABAB, ABBA, BAAB, BABA, BBAA). The cost is predictability. Once two of one drug have been given in a block, the rest of the block is known. ICH E9 therefore advises that investigators should not know the block length, and two or more block lengths chosen at random serve the same purpose.
Stratified randomisation first divides patients by an important prognostic factor such as severity or centre, then randomises in blocks within each stratum, so that factor is balanced. Minimisation balances several factors at once by assigning each new patient to the group that reduces imbalance. Factors nobody has measured are left to randomisation itself.
Two kinds of hiding
Allocation concealment is not blinding
| Feature | Allocation concealment | Blinding |
|---|---|---|
| What is hidden | The next allocation, before a patient is assigned | The treatment received, after assignment |
| Prevents | Selection bias at enrolment | Performance bias and detection bias |
| Always possible? | Yes, in every randomised trial | No. A surgeon cannot be unaware of the operation |
| Methods | Central telephone or web system. Sequentially numbered, opaque, sealed containers | Identical placebo, double dummy, sham procedure, blinded outcome assessor |
| CONSORT 2025 | Items 18 and 19 | Items 20a and 20b |
ICH E9 defines a double-blind trial as one in which neither the subject nor any of the investigator or sponsor staff involved in the treatment or clinical evaluation of the subjects are aware of the treatment received. In a single-blind trial one side knows. In an open-label trial everyone knows. Triple-blind has no agreed definition and usually means that outcome assessors or data analysts were also unaware. Physicians read these labels in many different ways (Devereaux et al., 2001), so CONSORT 2025 item 20a asks authors to say who was blinded: participants, care providers, outcome assessors or data analysts.
When the two treatments look different, as a tablet and an injection do, the double-dummy technique gives every patient one active drug and one matching placebo. When the treatment itself cannot be hidden, as in surgery or physiotherapy, the outcome assessor can still be kept unaware.
Analysis
Intention to treat and per protocol
Risk difference = events in group A ÷ patients in group A − events in group B ÷ patients in group B
- Intention to treat
- every randomised patient, counted in the group they were randomised to, whatever they took
- Per protocol
- only patients who followed the protocol: those who stopped or crossed over are left out
ITT keeps the groups as randomisation made them. ICH E9 calls the set as close as possible to this ideal the full analysis set. Per protocol removes patients whose adherence may depend on both the treatment and their prognosis, so ICH E9 warns its bias may be severe. In a superiority trial, ITT is the primary analysis and usually underestimates the effect, which is a conservative error. In a non-inferiority trial, non-adherence and crossover make the groups look alike, which favours the claim, so ITT is not conservative. Both analyses are reported, and both should support the conclusion.
For worked examples and the CONSORT flow diagram, read the clinical trials chapter.
Writing the answer
How to write this in a 10-mark DNB answer
- Define randomisation in one line and say why it is done: it balances known and unknown prognostic factors between groups.
- List the methods with one line each: simple, block (draw a block of 4 and its 6 orders), stratified, minimisation, cluster.
- Draw a two-column table of allocation concealment against blinding: what is hidden, when, what bias it prevents, methods.
- Name who can be blinded (participants, care providers, outcome assessors, analysts) and say that the trial should state who was blinded (CONSORT 2025 item 20a).
- Define ITT and per protocol, work a small made-up example with both, and say which is primary for superiority and why both are needed for non-inferiority.
- End with ICH E9 and CONSORT 2025 as the references.
The one thing to remember
Randomisation makes the groups comparable. Concealment stops anyone choosing who goes where, blinding stops anyone treating or judging the groups differently, and intention to treat keeps the groups as they were randomised.
Past papers
DNB questions this game helps you answer
Questions from NBE theory papers for DNB, DrNB and Diploma courses, most often asked first. The tag after each question is the specialty and the year it was set.
- ×3Methods of randomization.
- ×2Block randomization.
- ×2a) Non-parametric tests. b) Principles of randomization. c) Names of randomized multi institutional trials with respect to coronary artery bypass surgery and their salient findings.
- ×2Intention to Treat Analysis.
- ×1Allocation concealment and blinding. [4]
- ×1Randomization and blinding in clinical trials.
- ×1Randomization techniques in clinical trials.
- ×1Randomization.
- ×1What is randomization in a clinical trial? Mention the types of randomization. Discuss the merits and demerits of randomization.
- ×1What is randomized controlled study? How would you do randomization?
- ×1Enumerate and define different types of blinding in a clinical trial.
- ×1How will you ensure blinding in such group of patients?
- ×1What are various confounding factors in clinical trials? Briefly discuss the double-blind design.
- ×1Describe different randomization techniques.
- ×1Randomisation techniques.
- ×1Randomization in experimental study.
- ×1What is randomization and its different types? What are the advantages of randomization?
- ×1In clinical trials: Describe various blinding strategies.
- ×1Intention to treat versus per protocol analysis.
- ×1What is Intention-to-treat Analysis?
- ×1What is intention to treat analysis and per protocol analysis? Mention the strength and weakness of both the analysis types.
- ×1What is randomization? What are its benefits? Write briefly about the different types of randomization.
- ×1Describe role of concealment (blinding) in randomised control studies.
- ×1What is Randomised Controlled Double Blind trial? Give one example.
- ×1Enlist steps in conducting a randomized controlled trial. Describe intention to treat analysis and per protocol analysis.
- ×1Blinding & Randomization.
- ×1Describe randomization.
- ×1Discuss the role of randomization in clinical trials. What are the different types of randomization?
- ×1What are the advantages and disadvantages of permuted block randomisation over simple randomisation?
- ×1Bias and Blinding.
- ×1Compare and contrast blinding and concealment in randomised controlled trials.
- ×1What is the importance of blinding in trials? How would you achieve blinding in study of a new malaria vaccine?
- ×1Randomization and concealment of randomization.
- ×1Randomization in clinical trials.
- ×1What is randomization?
- ×1What is the purpose of randomization in a randomized controlled trial?
- ×1Why and how is allocation concealment done in a randomized controlled trial?
- ×1Add a note on blinding in research.
- ×1Blinding.
- ×1Double blind randomised controlled trial.
- ×1How is blinding done and what is the difference between single and double blind studies?
- ×1Intention to treat analysis.
- ×1Per protocol analysis.
- ×1What is block randomization? What does it achieve?
- ×1What is the need for randomization in a clinical trial?
- ×1Blinding in randomized controlled trial.
- ×1Outline the differences between Intent to Treat and Per Protocol analyses.
- ×1Allocation concealment.
- ×1Enumerate the methods used for randomization.
- ×1Methods of Randomization and points of consideration when selecting end points.
- ×1Randomization in Randomized Control Trials.
- ×1Randomization methods in a clinical research.
- ×1Randomization methods.
- ×1What is a double blind study?
- ×1What is double blind drug trial? What is double blind double dummy drug trial? How are they different from other clinical drug trails?
- ×1Intention to Treat Analysis and Co-efficient of Correlation.
- ×1Write a commentary on per-protocol vis-a-vis intention to treat analysis.
- ×1Block randomization
- ×1How is blinding different from allocation concealment?
- ×1Randomization and blinding in a clinical trial.
- ×1What is randomization and why is it required?
- ×1Types and importance of blinding in randomized trials.
- ×1Double Blind Randomization.
- ×1What are allocation concealment and blinding in randomized trials? How can these be achieved?
- ×1Blinding in research studies.
- ×1a) How do you calculate sample size for a research study? b) What is the difference between consent form and assent form? c) What is the difference between 'review article' and 'metanalysis'? d) Double-blinding in research studies.
- ×1Explain the concepts of blinding, allocation concealment and randomization in clinical trials.
- ×1Methods of randomization
- ×1Various methods of randomization.
- ×1What is randomization
- ×1a) Confidence interval. b) Methods of randomization.
- ×1What is blinding?
- ×1What are the various methods of randomization in an interventional study?
- ×1Double blind randomized trial.
- ×1Concepts underlying “Odd’s Ratio”, “Multiple Logistic Regression Analysis” and “Intention to treat analysis”.
- ×1What do you understand by double blind placebo controlled trial and its significance?
- ×1What is randomization? What are various methods/tools of randomization in clinical medicine? What are the benefits of randomization?
- ×1What are the features of the different phases of clinical trials? What is the importance of randomization and stratification?
Read next