When a Trial Is Open-Label but the Endpoint Is Blinded: Reading PROBE

PROBE stands for Prospective Randomized Open Blinded End-point: patients and clinicians know the assigned treatment, but an independent committee judges the outcomes without knowing it. This protects hard, objective endpoints such as death or an imaging-confirmed event when full blinding is impractical. It does not protect subjective, self-reported outcomes, because the unblinded patient and clinician have already shaped those before any committee sees them.

The problem PROBE tries to solve

Some treatments cannot be blinded. You cannot easily make a surgical procedure, an implanted device, or an intensive lifestyle program indistinguishable from its comparator. A fully open trial then risks detection bias, where knowing the assignment colors how the outcome is judged.

PROBE, which stands for Prospective Randomized Open Blinded End-point, is a design that accepts open treatment but tries to rescue the endpoint. It is a compromise built for situations where the ideal of full double-blinding is simply not available.

How the design works

In a PROBE trial, patients and their treating clinicians know which arm they are in, so the treatment is open-label. The outcomes, however, are sent to an independent committee that judges them without knowing the assignment. Randomization and allocation concealment still apply at entry.

The bet is that blinding the assessment of the outcome recovers much of the protection that blinding the treatment would have given, at lower cost and closer to everyday practice. Whether that bet pays off depends heavily on the kind of outcome being judged.

Where it works well

PROBE is most defensible for hard, objective endpoints that a blinded committee can adjudicate from records: death, a confirmed stroke, a laboratory-verified event, or an imaging read. For these, what matters is a consistent, assignment-blind judgment, and the committee supplies it.

Many large cardiovascular trials use this design for exactly this reason. When the endpoint is a documented event rather than an impression, a blinded adjudicator can neutralize most of the risk that open treatment would otherwise create.

Where it falls short

PROBE does not fix everything that blinding would. Because patients and clinicians know the assignment, behavior can still diverge: differential use of other treatments, differing intensity of follow-up, or differences in how subjective symptoms are reported.

A blinded committee cannot unbias a soft outcome that the unblinded patient has already shaped. For endpoints such as pain, quality of life, or anything self-reported, open treatment remains a real threat even when a blinded end-point committee signs off on the numbers.

What the evidence on assessors shows

The value of blinding the assessor is not a theoretical nicety. A systematic review of trials that used both blinded and non-blinded assessors for the same binary outcome found that non-blinded assessors exaggerated the odds ratio by roughly a third on average.

That exaggeration was driven by misclassifying only a small fraction of patients, which is a large distortion from a small nudge. It is precisely the kind of error a blinded end-point committee is meant to prevent, and it explains why the blinded adjudication in PROBE is not just paperwork.

How to read a PROBE trial

First, confirm it really is PROBE and not simply open-label with no blinded adjudication at all. Ask who adjudicated the outcomes and whether they were genuinely blind to the assignment.

Then ask whether the primary endpoint is the kind that blinded adjudication can protect: objective and record-based, or subjective and patient-shaped. A PROBE trial with a hard endpoint can be quite trustworthy. The same design reporting a self-rated outcome should be read as closer to open-label, with the caution that implies.