Definition
Blinding in clinical trials (also called masking): keeping participants, investigators, outcome assessors and sometimes data analysts unaware of which treatment each participant receives, so that this knowledge cannot influence their care, behavior, outcome assessment or analysis. Blinding protects a randomized trial against bias after the treatment has been assigned.
- Single blind, double blind and triple blind describe how many groups are kept unaware, but the labels are used inconsistently, so trials should report exactly who was blinded.
- Blinding is not allocation concealment: concealment protects the assignment before randomization, blinding protects it afterwards.
- When blinding is impossible, blinded outcome assessors and objective endpoints limit bias.
Updated November 10, 2026
Where you learn this: randomization, blinding and un-blinding procedures, and how to monitor them on site, are covered in the VIARES Clinical Research Associate course.
Contents
What is blinding in clinical trials?
Blinding hides the group assignment from the people involved in a clinical trial. After randomization, each participant is in the experimental group or the control group. In blinded trials, those who could be influenced by that knowledge do not learn it until the data are final.
Schulz and Grimes define it as keeping participants, investigators (usually the health care providers) or assessors, the people collecting outcome data, unaware of the assigned intervention. ICH E9 uses “blinding” and “masking” as synonyms. Outside medicine, the same idea is called a blind experiment or blinded experiment.
Blind studies are standard in drug trials. Our page on the double-blind study covers the most common design.
Why blinding matters: the importance of blinding in clinical trials
According to ICH E9, blinding limits conscious and unconscious bias when knowledge of treatment could influence recruitment, care, participants’ attitudes, endpoint assessment, withdrawals or the exclusion of data.
- ✓ Performance bias: an investigator who knows the active treatment may give one group extra care or co-interventions.
- ✓ Detection (ascertainment) bias: an assessor who knows the arm may score a symptom, an X-ray or a borderline event differently.
- ✓ Reporting bias: study participants who know they got the active drug may report more improvement, part of the placebo effect. Patient blinding also helps compliance and retention.
- ✓ Analysis bias: data analysts who know the arms could favor one when choosing analyses. A pre-specified statistical analysis plan helps.
How much does lack of blinding change results? Wood and colleagues, including Altman DG and Schulz KF, combined three meta-epidemiological studies covering 146 meta-analyses and 1,346 trials. In trials with a subjective outcome, lack of blinding was linked to exaggerated treatment effect estimates (ratio of odds ratios 0.75). For objective outcomes, such as all-cause mortality, they found little evidence of bias.
That is why a systematic review rates the risk of bias of each included trial, and why randomised controlled trials with subjective endpoints need blinding most.
Types of blinding: single blind or double blind, triple blind and open-label
ICH E9 defines three levels. In an open-label study, everyone knows the treatment. In a single-blind trial, the investigator knows but the participant does not, or vice versa. In a double-blind trial, neither the participant nor any investigator or sponsor staff involved in treatment or clinical evaluation knows. ICH E9 calls double blind the optimal approach. Triple blind usually adds the data analysts.
The labels are less clear than they look. Devereaux PJ and colleagues asked 91 internal medicine physicians which groups are blinded in single, double and triple blind trials. They gave 10, 17 and 15 different interpretations, and 25 textbooks gave 5, 9 and 7 different definitions.
CONSORT 2025, whose authors include Moher D and Schulz KF, therefore asks reports to state who was blinded after assignment, such as participants, care providers, outcome assessors and data analysts (item 20a), and how (item 20b). In short: report who was blinded, not just the label.
| Design | Who is unaware of the treatment | Typical use |
|---|---|---|
| Open-label | Nobody (outcome assessors can still be blinded) | Surgery vs drug, objective endpoints |
| Single blind | Usually participants | Investigators must know, e.g. to deliver a procedure |
| Double blind | Participants, investigators, site staff, usually outcome assessors | Placebo- or active-controlled drug trials |
| Triple blind | As double blind, plus data analysts or a review committee | Subjective outcomes, sensitive analysis choices |
| Double-dummy | Participants and investigators, although the two products look different | Two active treatments with different forms or schedules |
Blinding vs allocation concealment vs randomization
Schulz and Grimes note that many researchers confuse blinding with allocation concealment. Each protects a different moment:
- ✓ Randomization creates the sequence by chance, so a confounding variable is unlikely to differ systematically between arms.
- ✓ Concealment hides the next group assignment from whoever enrolls participants, so nobody can steer a patient into an arm. It is possible in every trial.
- ✓ Blinding hides the assignment afterwards, during treatment, data collection and analysis. It is not always possible.
Blinding methods: how blinding is achieved in practice
ICH E9 says a double blind clinical trial requires treatments that cannot be told apart by appearance, taste or other features. Common blinding methods:
- ✓ Matching placebo: an inactive product that looks, tastes and is packed like the active treatment.
- ✓ Double-dummy: if two treatments cannot be made identical, each gets a matching placebo, and participants take active A plus placebo B, or placebo A plus active B (ICH E9 glossary).
- ✓ Over-encapsulation: an active comparator is hidden in an opaque capsule. ICH E9 warns this may change pharmacokinetics, so bioequivalence may need to be shown.
- ✓ Sham procedures: a simulated procedure in the control arm. ICH E9 notes that dummy operations can raise ethical difficulties.
- ✓ Blinded outcome assessment: an independent assessor, central reader or endpoint committee. ICH E6(R3) says adjudication committees should typically be blinded, even in trials that are not.
- ✓ Coded supplies and IRT: the investigational medicinal product carries kit numbers, and an interactive response technology (IRT or IWRS) system assigns kits without showing the treatment.
ICH E6(R3) asks sponsors of blinded trials for a process to blind individuals and to prevent and detect inappropriate unblinding. Procedures must say who was unblinded, when and why, who stays blinded, and which safeguards protect the blind. Many sponsors collect these rules in a blinding plan.
Some roles are unblinded by design, such as the unblinded monitors that ICH E6(R3) mentions. They must not share what they know with blinded study staff.
When blinding is not possible
ICH E9 accepts that the double-blind ideal is sometimes out of reach, for example when surgery is compared with drug therapy. Behavioral and many device trials face the same problem. In single blind or open-label trials, ICH E9 recommends:
- ✓ Blind the outcome assessor: clinical assessments by staff not involved in treatment who stay blind.
- ✓ Objective endpoints: a primary outcome that is as objective as possible.
- ✓ Central randomization: the decision to enroll comes before the investigator knows the next treatment.
Open treatment with blinded endpoint assessment is sometimes called a PROBE design (prospective randomized open, blinded endpoint). In the US, 21 CFR 314.126 requires well-controlled studies to minimize bias by subjects, observers and analysts, for example by blinding.
Unblinding: emergency code breaks and accidental unblinding
ICH E9 says the blind for one subject should be broken only when the physician considers the treatment essential to know for that subject’s care. Under ICH E6(R3) section 2.11, the investigator must be able to unblind without undue delay in an emergency, and must promptly document and explain any premature unblinding to the sponsor, whether accidental, for an emergency or due to a serious adverse event.
- ✓ Protect the others: the code break, usually in IRT, reveals one participant’s treatment and keeps all other assignments hidden.
- ✓ Watch for side effects: a typical adverse effect or lab change can reveal the treatment. ICH E9 suggests blinding investigators to selected lab results.
- ✓ Interim looks: comparing arms mid-trial breaks the blind. ICH E9 notes that an independent data monitoring committee may review or conduct the interim analysis.
ICH E6(R3) lists verifying that blinding is maintained among the tasks of site monitoring. Maintaining blinding is part of Good Clinical Practice.
Example from trial practice: blinded vs unblinded assessors
The multiple sclerosis clinical trial (Noseworthy et al., 1994)
The Canadian cooperative trial compared cyclophosphamide and plasma exchange regimens with placebo, using sham plasma exchange and placebo medication. At each assessment, every patient was examined by two neurologists: one blinded, one unblinded.
Using the blinded neurologists’ assessments, the primary analysis, neither active regimen beat placebo. The unblinded neurologists’ scores showed an apparent treatment benefit at 6, 12 and 24 months for the plasma exchange group. Physician blinding prevented a false-positive conclusion about efficacy.
Hypothetical CRA monitoring visit
Hypothetical example. In a double-blind multicenter trial, the IRT report shows one emergency code break at a site since the last visit. The CRA checks:
- ✓ the medical reason in the source notes and the case report form, against the protocol criteria;
- ✓ that the investigator documented the break and informed the sponsor promptly;
- ✓ who at the site learned the treatment, and that the outcome assessor stayed blinded;
- ✓ that no unblinded records, such as pharmacy logs, sit in files the blinded staff use.
Any gap goes into the visit report and is followed up as a deviation.
Where you learn this at VIARES
Section 6 of the VIARES Clinical Research Associate course, “Clinical Trial Design & Clinical Research Roles and Responsibilities”, covers trial design; its assignment asks learners to explain designs such as the double-dummy design and their bias-reducing effect.
Section 10, “Investigational Medicinal Product Management”, explains the randomization process and its documentation requirements, including un-blinding procedures. It also shows how to monitor IP-related aspects during monitoring visits, such as drug accountability and randomisation of patients.
In the Clinical Study Coordinator I course (80–100 hours), Section 6 “Clinical Research Process and Study Design” covers study designs and randomization.
FAQ
What does blinding mean in a clinical trial?
Blinding means that participants, investigators, outcome assessors or data analysts do not know which treatment each participant receives. This stops that knowledge from influencing care, symptom reports and outcome assessment.
What is the difference between single blind and double blind?
In a single blind trial, usually only the participants do not know their treatment. In a double blind trial, the participants and the investigators and staff who treat and assess them are all unaware. Because the labels vary, reports should state exactly who was blinded.
What is a triple blind study?
A triple blind study usually keeps participants, investigators and the data analysts or a review committee unaware of the group assignment. Definitions differ between textbooks, so the report should name each blinded group.
Can a study be double blind without a placebo?
Yes. Two active treatments can be compared double blind if they are made to look identical, for example by over-encapsulation, or with a double-dummy design in which each treatment has its own matching placebo.
Who can break the blind in a clinical trial?
The investigator can break the blind for one participant when knowing the treatment is essential for that participant’s care, usually through the IRT system. The break must be documented and explained to the sponsor promptly, and other assignments stay hidden.
Sources
- ICH E9: Statistical Principles for Clinical Trials, sections 2.3.1 Blinding, 3.4 and Glossary (double-dummy)
- ICH E6(R3): Guideline for Good Clinical Practice (2025), sections 2.11, 3.9.8, 3.11.4.5.2, 3.15.2, 3.16.1 and 4.1
- 21 CFR 314.126: Adequate and well-controlled studies (eCFR)
- Hopewell S, Chan AW, Collins GS, et al. CONSORT 2025 statement: updated guideline for reporting randomised trials. BMJ 2025;389:e081123 (items 20a and 20b)
- Schulz KF, Grimes DA. Blinding in randomised trials: hiding who got what. Lancet 2002;359:696–700
- Devereaux PJ, Manns BJ, Ghali WA, et al. Physician interpretations and textbook definitions of blinding terminology in randomized controlled trials. JAMA 2001;285:2000–2003
- Noseworthy JH, Ebers GC, Vandervoort MK, et al. The impact of blinding on the results of a randomized, placebo-controlled multiple sclerosis clinical trial. Neurology 1994;44:16–20
- Wood L, Egger M, Gluud LL, Schulz KF, Jüni P, Altman DG, et al. Empirical evidence of bias in treatment effect estimates in controlled trials with different interventions and outcomes: meta-epidemiological study. BMJ 2008;336:601–605
