How to Measure Recruitment Performance by Research Site

A multi-Site clinical trial may have ten Research Sites recruiting under the same protocol. At first glance, comparing performance appears simple. Site A enrolled 18 participants. Site B enrolled 11. Site C enrolled 7. So Site A is performing best.

Not necessarily. Those numbers tell only part of the story. The Sites may have very different population sizes, recruitment budgets, geographic markets, referral volumes, staff capacity, activation dates, competing Studies, advertising costs, screening availability, and participant travel burdens.

That means Site performance should not be reduced to one number. A useful Site comparison needs to answer: how efficiently is each Site converting the recruitment opportunity available to it? That requires a deeper scorecard.

Enrollment Volume Matters — But It Is Not Enough

Enrollment is ultimately important. A Site that consistently enrolls participants is creating real Study value. But total enrollment alone does not explain how much recruitment support the Site received, how many referrals it handled, how quickly it responded, how many screenings occurred, how much money was spent, or whether it could handle additional volume.

A Site with 20 enrollments from 500 referrals is behaving differently from a Site with 15 enrollments from 100 referrals. Both may be valuable. But for different reasons.

Start With a Standard Site Funnel

Multi-Site comparisons become meaningful only when every Site uses the same recruitment stages. A practical structure might be:

Referral Received → Site Contact Attempted → Participant Contacted → Screening Scheduled → Screening Completed → Enrolled

If each Site defines these stages differently, Study-level reporting becomes unreliable. For example, Site A may count “screening scheduled” as “screening,” while Site B may count only completed screening visits. Those numbers cannot be compared directly. Standardization comes first.

Metric 1: Referrals Received

The first Site-level metric should be: how many referrals did the Site receive? Without this number, enrollment comparisons lack context. If Site A gets 300 referrals and Site B gets 90, and Site A enrolls 20 while Site B enrolls 12, Site A has more total enrollments — but Site B may be converting its smaller opportunity pool much more efficiently. Referral volume is therefore the denominator behind many later KPIs.

Metric 2: Referral-to-Contact Rate

Once a referral arrives, how many participants does the Site successfully reach? Referral-to-Contact Rate = Participants Contacted ÷ Referrals Received. This metric helps reveal Site responsiveness, contact quality, participant availability, lead aging, and follow-up effectiveness. A low contact rate may not automatically mean poor Site performance — it may also reflect invalid contact information, poor referral quality, delayed centralized routing, or a difficult participant population. But it identifies where further investigation is needed.

Metric 3: Time to First Contact

Two Sites may have identical contact rates but very different response times — Site A with a median of 3 hours, Site B with a median of 2.5 days. That difference matters. Time to first contact can influence participant recall, trust, response, appointment scheduling, and overall funnel momentum. This makes response time one of the most important operational Site KPIs.

Metric 4: Screening Appointment Rate

How many referrals result in a scheduled screening? Screening Appointment Rate = Screenings Scheduled ÷ Referrals Received. This metric begins to show whether the Site is converting recruitment opportunities into actual Study activity. A low rate may reflect referral quality, participant hesitation, scheduling limitations, travel barriers, slow follow-up, or Study complexity.

Metric 5: Screening Attendance Rate

Scheduling a screening is not the same as completing one. Track Screening Visits Completed ÷ Screening Visits Scheduled. This helps identify no-show problems, participant burden, reminder effectiveness, scheduling delays, and transportation issues. A Site may appear strong based on scheduled visits but weak on actual attendance.

Metric 6: Referral-to-Screening Rate

This is one of the clearest measures of Site recruitment efficiency. Referral-to-Screening Rate = Completed Screenings ÷ Referrals Received. It tells you how effectively the Site converts incoming recruitment opportunity into protocol-based evaluation. Suppose Site A has 200 referrals and 40 screenings (a 20% rate), while Site B has 100 referrals and 35 screenings (a 35% rate). Site A produced more screenings. Site B converted referrals more efficiently. Both insights matter.

Metric 7: Screen Failure Rate

Not every screened participant will enroll. That is normal. Screen Failure Rate = Screen Failures ÷ Completed Screenings. A high screen failure rate can reflect difficult protocol criteria, weak preliminary pre-screening, broad recruitment messaging, missing clinical information, Site-specific interpretation, or population mismatch. This metric should never be interpreted in isolation — a high screen failure rate may be a recruitment issue, or it may simply reflect a difficult protocol.

Metric 8: Screening-to-Enrollment Rate

This measures how many formally screened participants ultimately enroll. Screening-to-Enrollment Rate = Enrollments ÷ Completed Screenings. This KPI can help distinguish top-of-funnel problems from later Study-level problems. If the Site receives strong referrals and completes many screenings but very few participants enroll, the issue may be deeper than advertising.

Metric 9: Referral-to-Enrollment Rate

This is one of the strongest Site comparison metrics. Referral-to-Enrollment Rate = Enrollments ÷ Referrals Received. It compresses multiple stages of the recruitment funnel into one downstream conversion measure. Suppose Site A has 300 referrals and 18 enrollments (6%), while Site B has 120 referrals and 14 enrollments (11.7%). Site A enrolled more people. Site B converted referrals almost twice as efficiently. This is exactly why absolute enrollment volume should not be the only comparison.

Metric 10: Enrollment Velocity

Recruitment is also about time. A Site enrolling 10 participants in two months may be operationally stronger than a Site enrolling 12 participants in eight months. Enrollment velocity can be measured as enrollments per week or per month after Site activation. This helps normalize performance across Sites that began recruiting at different times.

Metric 11: Time From Referral to Screening

A Site may contact participants quickly but schedule screenings slowly. That creates another bottleneck. Track the time from referral date to screening date. Long delays can increase participant loss, no-shows, competition with other Studies, and frustration. This metric gives a better view of Site capacity.

Metric 12: Cost Per Screening by Site

When recruitment spend can be attributed geographically or by Site, Cost Per Screening = Recruitment Spend Allocated to Site ÷ Completed Screenings. This provides a deeper economic metric than CPL. Two Sites may have the same CPL but very different Cost Per Screening.

Metric 13: Cost Per Enrollment by Site

Where attribution allows, Cost Per Enrollment = Recruitment Spend Allocated to Site ÷ Attributed Enrollments. This can be especially useful in centralized recruitment programs. But it needs context — a Site may appear expensive because its geography has higher media costs, the Study population is scarce locally, travel burden is high, or fewer recruitment channels are available. Cost should inform decisions. It should not replace judgment.

Metric 14: Site Capacity

Performance is not just about what has happened. It is also about what the Site can handle next. Track available screening appointments, coordinator capacity, physician availability, recruitment backlog, and current enrollment pace. A Site can be highly efficient but temporarily unable to absorb additional referrals. Sending more volume may actually reduce performance.

Metric 15: Referral Aging

How long do referrals sit before being acted upon? Useful categories might include less than 24 hours, 1–2 days, 3–5 days, and 5+ days. Referral aging helps detect operational backlog. This is especially useful in centralized recruitment programs.

Metric 16: Unable-to-Contact Rate

Unable-to-Contact Rate = Unreachable Referrals ÷ Total Referrals. Compare this metric between Sites. If one Site has 12% unreachable and another has 45%, the difference deserves investigation — possible causes include follow-up speed, number of attempts, communication channels, participant language, referral age, and staff workflow.

Metric 17: Participant Decline Rate

Track how often participants actively decline. This can reveal travel burden, Study visit requirements, compensation expectations, Site communication, Study complexity, or a recruitment-message mismatch. The reason for decline is often more useful than the rate itself.

Use Disposition Codes

Every non-progressing referral should ideally receive a structured disposition — unable to contact, declined, outside geography, screening not scheduled, screening no-show, screen failure, enrolled, duplicate, Site capacity, or not appropriate after review. This turns lost participants into operational data.

Compare Sites Using Conversion, Not Just Volume

A useful Site dashboard should present both. For example:

Site	Referrals	Screenings	Enrollments	Referral→Screen	Referral→Enroll
A	300	60	20	20%	6.7%
B	140	48	18	34.3%	12.9%
C	90	26	11	28.9%	12.2%

Now the picture changes. Site A has the highest enrollment volume. Site B is converting referrals most efficiently. Site C has lower scale but similar downstream efficiency. That is much more useful than simply ranking 20/18/11 enrollments.

Normalize for Time

Sites rarely activate simultaneously. Comparing cumulative enrollment without accounting for activation date can be misleading. A Site active for eight months should not automatically be compared directly with one active for six weeks. Track weeks active, enrollments per active month, screening per active month, and referrals per active month. This creates fairer comparisons.

Normalize for Recruitment Opportunity

Sites also receive different opportunity levels. A large metropolitan Site may receive ten times more digital exposure than a rural Site. Performance should therefore be viewed through both volume and conversion. Do not punish a Site for having a smaller market if it converts extremely well.

Geography Matters

Site performance can be influenced heavily by geography — population density, average drive time, competing Studies, language, advertising cost, socioeconomic factors, and transportation. A Site with a difficult geography may perform strongly operationally even if total enrollment is lower. This is why geographic targeting and Site performance should be analyzed together.

Site Performance and Marketing Performance Are Different

This distinction is critical. Marketing performance includes CTR, CPL, lead quality, pre-screen completion, and Cost Per Referral. Site performance begins after the referral arrives — it includes response time, contact, scheduling, screening, and enrollment. Do not blame the Site for poor advertising. Do not blame advertising for poor Site follow-up. A strong dashboard separates both.

A Simple Ownership Model

Marketing / central recruitment owns acquisition, pre-screening, referral routing, and referral quality. The Research Site owns referral response, participant contact, screening scheduling, screening execution, and enrollment. Both contribute to participant experience, funnel performance, feedback, and optimization. Clear ownership reduces unproductive disputes.

Site Rankings Can Be Dangerous

Creating a leaderboard from best Site to worst Site can oversimplify performance. It may encourage Sites to optimize for the wrong behaviors. A stronger model categorizes Sites — high volume/high conversion is a strong candidate for additional budget; high volume/low conversion warrants investigating Site operations or referral quality; low volume/high conversion may be a potentially underfunded market; and low volume/low conversion needs deeper evaluation. This framework creates strategic decisions rather than competition.

The Four-Quadrant Site Model

Imagine a simple chart with referral volume on the horizontal axis and referral-to-enrollment conversion on the vertical axis. Sites then fall into four groups.

Quadrant 1 — Scale (high volume, high conversion): protect and expand.

Quadrant 2 — Opportunity (low volume, high conversion): consider more recruitment investment.

Quadrant 3 — Inefficiency (high volume, low conversion): investigate Site workflow and referral quality.

Quadrant 4 — Challenge (low volume, low conversion): reassess geography, Site capability, or strategy.

This is much more useful than a simple enrollment ranking.

Response Time Should Be Visible at Study Level

A CRO or SMO should know that Site A has a 3-hour median response, Site B has 18 hours, and Site C has 2.4 days. This information can explain significant downstream differences. If response time is invisible, the Sponsor may incorrectly conclude that geography or marketing is responsible.

Capacity Should Affect Budget Allocation

Suppose Site A performs extremely well but is already at maximum screening capacity. Sending another 100 referrals may reduce performance. Site B has slightly lower conversion but substantial unused capacity. The optimal Study-level decision may be to shift incremental budget toward Site B. The goal is not maximizing one Site. It is maximizing Study enrollment.

Performance Should Drive Dynamic Allocation

Traditional multi-Site recruitment often begins with equal budget per Site. A stronger model begins with a rational baseline and then adjusts according to market size, referral quality, Site conversion, capacity, recruitment velocity, and cost. Budget becomes dynamic.

Use Rolling Performance Windows

Cumulative metrics can hide recent changes. A Site may have performed poorly early and strongly recently, or the opposite. Useful views include the last 7 days, the last 30 days, and Study-to-date. This helps identify trends.

Small Samples Need Caution

Do not overreact to tiny numbers. A Site with 2 referrals and 1 enrollment shows a 50% conversion rate — that does not automatically mean the Site is exceptional. Sample size matters. A good dashboard should show both rate and volume.

A Practical Site Performance Scorecard

For each Site, track volume (referrals, screening appointments, completed screenings, enrollments), conversion (contact rate, referral-to-screening rate, screening-to-enrollment rate, referral-to-enrollment rate), speed (time to first contact, time to screening, enrollment velocity), cost (Cost Per Referral, Cost Per Screening, Cost Per Enrollment), operations (referral aging, Site capacity, unable-to-contact rate, no-show rate), and context (geography, active time, recruitment channel, language). That creates a much more complete picture.

Build Dashboards for Action

A dashboard should not merely report. It should indicate what to do. For example: a Site with high conversion and high capacity should increase volume. A Site with good leads but slow response should improve follow-up. A Site with strong contact but high screen failure should review pre-screen criteria. A Site with high cost and low downstream conversion should reassess geography or recruitment investment. The dashboard should support decisions.

Feedback Should Flow Back to Recruitment

Site performance data should not remain inside the Site. It should return upstream. For example, if a Site reports high screen failure from one campaign, that information can influence pre-screening, messaging, audience, geography, and budget allocation. The recruitment funnel becomes a loop, not a one-way pipeline.

Multi-Site Recruitment Is a System

One Site does not operate independently from the Study. A Site may have excellent conversion but poor volume. Another may have high volume but weaker conversion. Another may have strong enrollment but no remaining capacity. The central recruitment strategy needs to coordinate all three. That is why Site-level measurement matters.

The Goal Is Not to Find the “Best Site”

The real goal is to understand what each Site needs to contribute most effectively to the Study. That may mean more budget, less budget, more referrals, operational support, wider geography, tighter targeting, additional staff, or faster follow-up. Performance measurement should guide improvement, not simply ranking.

Frequently Asked Questions

What is the most important Research Site recruitment metric?

There is no single metric. Referral volume, response time, referral-to-screening conversion, enrollment, velocity, and Site capacity all provide different information.

Should CROs compare Sites by total enrollment?

Yes, but not alone. Enrollment should be evaluated alongside referral volume, activation time, conversion, geography, cost, and capacity.

What is referral-to-enrollment rate?

It is the percentage of Site referrals that ultimately become enrolled participants.

Why is response time a Site performance metric?

Because delayed Site contact can reduce participant engagement and lower downstream screening and enrollment.

Should recruitment budgets be equal across Sites?

Not necessarily. Budget may be more effective when adjusted according to geography, Site capacity, referral conversion, and Study-level performance.