Choosing Sites on Evidence, Not Enthusiasm

Part of Clinical Project Management (PM) Training: Site Selection and Activation.

About 2 minutes. Free, in your language, no account needed.

Feasibility asks the wrong question well

A feasibility questionnaire asks how many eligible patients a site sees. Sites answer honestly and are usually wrong, because the question invites a count of the condition rather than a count of people who would actually randomize.

The better questions are about evidence of past behavior. How many participants did you randomize in your last trial in this indication, against how many you predicted? How long did your last contract take to execute? Who does your consent discussions, and are they there next year?

What actually predicts a good site

A coordinator with capacity. One experienced coordinator running four trials is a constraint no amount of patient volume overcomes.

A PI who is present. Oversight that exists in fact, not only on the delegation log.

Prior performance in the same indication. The nearest thing to a reliable predictor.

Access to the population. Referral pathways, not just a catchment area.

The cost of a site that never enrolls

It is not zero and it is not small. Contract negotiation, ethics submission, site initiation, drug shipped and stored, monitoring visits, and close-out — all incurred for no data. Several of those costs land whether or not a single participant is randomized.

Choosing ten sites where four would have done is a common and expensive habit, and it also dilutes the monitoring attention each site receives.

Say no early

A site that looks marginal at feasibility is almost never rescued by activation. Declining a site is a decision you can take cheaply in month one and cannot take cheaply in month nine, once the contract is signed and drug is on the shelf.