Question 1

Start with the experts and sites you already know, but don't stop there

Accepted Answer

A practical starting point is your internal pool of known experts and sites. The issue is not that teams start there. It’s that too many treat it as the strategy rather than the foundation for site selection.

Look first at the people and sites you already know, contracted with, and trust. That includes experts who helped shape the trial, served on advisory boards, or worked on the program earlier. This is an efficient way to start because you've already got relationships, contracts, and some working history.

But going back to the same sites again and again and expecting a better outcome is a bad habit, not a strategy. Begin by asking how big your internal pool really is and whether those sites are experts in the indication you plan to study. Then expand beyond that base so familiarity doesn't turn into false confidence.

Question 2

Build a big funnel first, then narrow it with evidence that matters

Accepted Answer

You can build a thorough site selection approach like a funnel. At the top, cast a wide net. Build a broad list of sites and investigators with relevant therapeutic area or indication experience. Then narrow that list by asking targeted questions for more details:

• Do they just publish papers, or do they actually run studies?
• Do they treat enough patients to be useful?
• Could they serve as a referral funnel even if they're not the main enrolling site?
• Are they experienced, emerging, or both?

This process matters because a shortlist can still mean hundreds or even thousands of names, which isn't scalable for most sponsors. The goal isn't to create a long list that looks impressive. The goal is to rank and prioritize the right candidates so qualification efforts are focused and useful.

Question 3

Use a Venn diagram of data—not one source—to get a complete picture

Accepted Answer

Confidence comes from layering data, not from trusting a single input. A complete view of a site should combine:

• Internal data, which tells you whether you've worked with the site before and what that relationship was actually like.
• Public data, such as ClinicalTrials.gov, registries, and PubMed, which shows who has studied what and who is publishing in the space.
• Purchased data, such as claims, prescription, electronic medical record (EMR), and electronic health record (EHR) data, which helps show where patients are seen and how treatment patterns actually work.

Some key site performance data isn’t public at all and may only be visible through a sponsor’s own systems or trusted partners with operational data. Each source answers a different question, but none is enough on its own.

When teams use a Venn diagram of data, they move from guesswork to triangulation, which creates a more confident shortlist and a stronger final decision.

Question 4

Don't take self-reported site data at face value. It's usually aspirational.

Accepted Answer

Qualification questionnaires have value, but they're best used as an important input rather than a final answer.

Self-reported site data is often aspirational. Sites are sharing what they believe they can deliver, but those claims still need to be validated against objective evidence.

That's understandable, but it's also consequential if taken at face value.
An investigator may say they can bring in a large number of patients, but that claim needs a gut check. Can it be backed up by claims data, EMR or EHR evidence, or prior recruitment performance?

A site might see the right patients and still struggle to recruit, screen, randomize, or retain them. Strong teams use questionnaires to uncover missing details, then bounce those answers back against objective data before they rely on the picture being presented.

Question 5

Validate where patients are treated, not just where they live or show up on a map

Accepted Answer

One of the biggest mistakes in site selection is confusing where patients live with where they actually get treatment, and how their treatment pathways really work.
Prevalence maps, census data, and demographic patterns can be helpful, but they can also mislead teams into selecting sites in the wrong places. A patient may live in one community but travel to a very different health system for specialized care. That gap matters.

If you only study disease prevalence by geography, you may miss the real treatment pathway. Better decisions come from combining census, claims, EMR, EHR, prescription, and social determinants data to understand not just who patients are, but where they enter care and where they stay in care.

Site selection improves when you follow the treatment journey, not just the ZIP code.

Question 6

Use historical performance to prioritize which sites deserve closer attention

Accepted Answer

If you're not layering in historical performance, you're missing context on how a site can rise to the occasion.

Experience and profile matter. Capacity and capability matter. But performance can help show whether a site is likely to turn potential into results. It helps answer the practical question at the center of every trial: Can this site recruit patients?

Past performance isn't a guarantee of future success, but it's still one of the strongest signals for deciding which sites deserve deeper scrutiny first. Yet powerful performance data is often underused by larger teams, and not available to smaller teams. When a site has never been graded in this class, you need more scrutiny.

And when it's consistently delivered before, that history deserves meaningful weight in your ranking, too. Trends in performance can be important indicators of not just probable performance, but probable growth as well.

Question 7

Focus on enrollment metrics that show whether a site can optimally recruit into this trial

Accepted Answer

Historical performance should be practical, specific, and tied to enrollment. The following measures reveal intelligence beyond what a site’s anecdotal or word-of-mouth reputation can:

• Patient recruitment timelines
• Patients per site per month
• Time to first patient in (FPI)
• Screening pace
• Randomization success
• Retention

These metrics show whether the site can effectively move people through the trial.

The first three to six months of enrollment are especially important because that period often sets the tone for the full recruitment curve. At the simplest level, ask: Can they get one patient, then two, then more?

Enrollment metrics only become useful when they're interpreted in the context of the current indication and protocol. A site may have enrolled well in one indication but have little relevance to the current study, so those metrics matter most when they're tied to the indication and protocol in front of you. Use performance as a key factor but always connect it back to protocol fit and patient access.

Question 8

Check capacity, staffing, and equipment before you assume a strong site is a good fit

Accepted Answer

A site can look impressive on paper and still be the wrong fit for your trial. That's why capacity and capability belong on every checklist to help confirm feasibility with data. You need to know whether the site:

• Has room to take on more work.
• Is already overloaded with active trials.
• Has the people, coordinators, equipment, and physical setup your protocol requires.

Some of those signals can be inferred early, but they often need to be validated directly through qualification: A site with the right patient population but the wrong staff structure may struggle badly, while a site with the right reputation but outdated equipment may not be trial-ready.

A good fit comes from the overlap of multiple factors, not a single attribute. The question is not just whether a site is strong in general. The question is whether it's built to execute this specific trial, with precisely this protocol, at this very moment.

Question 9

Look for startup speed and operational efficiency, not just headline recruitment numbers

Accepted Answer

Operational data separates strong sites from merely promising ones. Recruitment totals matter, but they're only part of the picture. Teams should also evaluate:

• Startup speed.
• Time to activation.
• Time to FPI.
• Recruitment to plan.
• Patient retention.
• Visit compliance.
• The site’s ability to execute consistently once the study begins.

In some cases, the clearest view of those signals comes from operational systems that capture how sites move through startup and execution.

These signals reveal how efficiently a site works day to day. They also help explain whether a performance pattern is driven by the site itself or by the protocol design. Strong teams use these objective performance signals differently than weaker teams do. They don't treat them as background detail. They use them to prioritize qualification and selection in a disciplined way.

That discipline matters because it keeps teams from confusing eventual enrollment with strong execution. A site that starts slowly or operates sloppily—even if it enrolls eventually—is still likely to create major downstream costs and timeline problems. By looking at the efficiency of speed with metrics rather than recruitment totals, you can select a site based on how they achieve their targets, not just what their targets were.

Question 10

Watch for the avalanche effect: One bad assumption can delay the whole trial

Accepted Answer

When teams over-index on a narrow set of inputs, the most common error is simple: They believe they'll get more patients faster than they actually will. That one bad assumption can trigger an avalanche:

• The wrong sites get pushed forward.
• Enrollment curves stretch out.
• Startup plans slip.
• More sites have to be added later.
• Trial windows move.
• Budgets swell.

Contract research organizations (CROs) and other partners must be re-engaged.
For smaller companies, that kind of miss can be existential. For larger sponsors, it creates major budget overruns and operational drag.

The lesson: False confidence usually starts small. It often begins with unvalidated claims, weak data triangulation, or too much trust in a narrow view. To avoid this assumption, site selection should have goals that are justified by clear evidence that supports the strategy. There will always be challenges meeting enrollment goals, but careful planning can prevent those challenges from becoming obstacles for sites in the future.

Question 11

Don't overlook quality and responsiveness: Reliability is part of performance

Accepted Answer

Some of the hardest things to validate through conventional means are compliance, follow-through, and day-to-day reliability. Yet these things can heavily influence whether a site's worth selecting.

A site that recruits well but creates a messy process (for example, one that causes handoffs and follow-ups to break down) may end up costing more than it's worth. Quality isn't just about major public compliance events. It's also about responsiveness, communication, and how smoothly the site works with oversight teams. Some of those quality and responsiveness signals only emerge through operational workflows, not public sources.

These are often hidden signals, but they matter. These incremental slowdowns in the process can drain time and resources even when a site looks strong on paper.

A responsive site is easier to activate, easier to manage, and often more dependable under pressure. Confidence grows when teams evaluate not just whether a site can enroll, but whether it can execute cleanly and consistently.

← Back to blog

A Practical Checklist for Building Confidence in Site Selection

How operational insight, performance data, and patient access signals can strengthen clinical trial site decisions

***Start* with the experts and sites you already know, but *don’t stop there***

Build a big funnel first, then narrow it with evidence that matters

Use a Venn diagram of data—not one source—to get a complete picture

Don’t take self-reported site data at face value. It’s usually aspirational.

Validate where patients are treated, not just where they live or show up on a map

Use historical performance to prioritize which sites deserve closer attention

Focus on enrollment metrics that show whether a site can optimally recruit into this trial

Check capacity, staffing, and equipment before you assume a strong site is a good fit

Look for startup speed and operational efficiency, not just headline recruitment numbers

Watch for the avalanche effect: One bad assumption can delay the whole trial

Don’t overlook quality and responsiveness: Reliability is part of performance

Better site decisions start with better confidence

Want to learn more? Check out this overview of Advarra’s site identification and feasibility services or contact us to ask an expert a specific question.

Ian Bailey

Advarra to your inbox