Why Site Selection Decides the Trial

Part of Clinical Research Associate (CRA) Training: Site Identification and Feasibility.

About 5 minutes. Free, in your language, no account needed.

Why this matters

Almost every problem you will spend the next two years fixing was decided before the first participant was enrolled: which sites were chosen.

A weak site does not get better under monitoring. It absorbs monitoring. Selection is the one point at which the problem can still be avoided cheaply.

What good selection prevents

Choosing the right sites reduces:

• Slow recruitment, the leading cause of trial delay.
• Protocol deviations, which cluster heavily at a minority of sites.
• High query volumes and late data entry.
• Regulatory findings, which attach to the sponsor as well as the site.
• Poor data integrity, which is the one problem that cannot be fixed later.

A trial with 40 sites usually has 5 that produce most of its problems. Selection is where you try to avoid recruiting those 5.

The stages

Site selection normally runs through seven stages. The CRA supports all of them except the final decision, which is the sponsor's.

1. Site identification — building a list of candidates.
2. Initial feasibility — sending and collecting questionnaires.
3. Feasibility review and validation — checking what the sites claimed.
4. Pre-study visit, sometimes called a site qualification visit — the in-depth evaluation.
5. Risk assessment and decision — scoring and recommending.
6. Qualification and activation preparation — getting a selected site ready.
7. Documentation, reporting and handover.

Modules 3 and 4 walk through these. Module 5 covers contracts and activation.

Where candidate sites come from

Sponsors and CROs build the initial list from:

• Preferred site lists — sites with a track record on previous studies.
• Known high-performing investigators in the therapeutic area.
• Publication records, which indicate genuine expertise.
• Recommendations from medical liaisons or clinical experts.
• Trial registry data, showing which sites run studies in this indication.
• Competitive intelligence — which sites are already active in similar studies. This cuts both ways: it shows capability, and it warns of competition for the same patients.
• Deliberate expansion into regions or populations the trial needs but has not previously used.

The last one deserves attention. Trials cluster in a small number of countries, which means much of the world's patient population is under-represented in the evidence that eventually treats them. A site without a track record is not the same as a weak site. It is a site whose capability has to be assessed rather than assumed.

The CRA's role, honestly stated

You do not approve sites. The sponsor selects. Your job is to produce an objective, evidence-based evaluation good enough for the sponsor to decide well.

That means:

• Understanding the protocol's real demands before you assess anyone against them.
• Gathering evidence rather than impressions.
• Being specific about what a site can and cannot do.
• Recording risks, including risks at sites you are recommending.
• Resisting pressure — including your own enthusiasm — to make a promising site look better than the evidence supports.

A recommendation that reads "excellent site, strongly recommend" and turns out to be wrong damages your credibility for years. A recommendation that says "recommend, conditional on a second coordinator being funded" survives contact with reality.

The two questions behind everything

Every criterion in the next two lessons is a way of asking one of two questions.

Can this site do the protocol? Do they have the patients, the people, the equipment, the systems and the time?

Will this site do it right? Do they have the processes, the oversight, the documentation habits and the compliance history?

A site can pass one and fail the other. The site with a huge patient population and a disengaged investigator is a common and expensive failure mode. So is the meticulous site with no patients.

In practice

Leah is asked to evaluate four candidate sites for VT-301.

Site A is a large teaching hospital with a famous investigator, 400 eligible patients on its database, and three competing diabetes trials already running.

Site B is a small clinic with 90 eligible patients, no trial in the indication before, an investigator who has been a sub-investigator on four studies, and a full-time coordinator with nothing else booked.

The instinct is to prefer Site A. Leah writes down what the evidence supports. Site A's 400 patients are shared across four studies and its coordinator team is already stretched. Site B's 90 patients are uncontested and its staff have capacity.

She recommends both, with different conditions: Site A conditional on a named coordinator being allocated to VT-301 specifically, and Site B conditional on a training plan and a first monitoring visit within six weeks of first enrollment.

Eighteen months later Site B has enrolled 14 and Site A has enrolled 11.

Reputation is not capacity. Ask what is actually free.

Key takeaways

• Most problems a monitor spends years managing were decided at site selection, and a weak site absorbs monitoring rather than improving under it.
• Selection runs through seven stages from identification to handover, and the CRA supports all of them except the sponsor's final decision.
• A site with no track record is not the same as a weak site; its capability must be assessed rather than assumed, and expanding into under-represented regions requires exactly that assessment.
• The CRA's product is evidence, not endorsement, and a conditional recommendation that names its risks is more useful than an enthusiastic one.
• Every criterion answers one of two questions: can this site do the protocol, and will it do it right — and sites commonly pass one and fail the other.