Clinical trial site selection determines whether a study starts with an enrollment advantage or spends months recovering from optimistic assumptions. The best sites are not always the most famous investigators or the hospitals with the longest sponsor relationship. They are the sites that can find eligible patients, enroll them ethically, retain them, produce clean data, and handle the operational load of the protocol.
Traditional site selection still relies heavily on investigator databases, past relationships, and self-reported feasibility surveys. Those inputs matter, but they leave gaps. Data-driven site identification clinical trials teams now combine real-world evidence, claims, EHRs, registry data, competing-study intelligence, and site-performance history before activation.
Why clinical trial site selection fails
Site selection fails when sponsors confuse interest with capacity. A principal investigator may be enthusiastic, experienced, and respected, but still lack enough eligible patients, coordinators, pharmacy support, lab capacity, or competing-study bandwidth.
Enrollment risk is often visible before first patient in. Protocols may require biomarkers that are rarely tested in routine care. Inclusion criteria may exclude common comorbidities. Visit schedules may be too burdensome. A site may treat many patients with the disease but few who match the trial's line-of-therapy, lab-value, or prior-treatment requirements.
The cost of getting this wrong is high. Tufts CSDD research on 151 global Phase II/III trials reported that 11 percent of activated, ready-to-recruit investigative sites enrolled no patients, while 41 percent missed target enrollment. Even when exact rates vary by therapeutic area and sponsor, the operational pattern is familiar: too many sites are opened to compensate for weak feasibility, creating more monitoring burden and higher fixed cost.
The limits of feasibility questionnaires
Feasibility questionnaires collect structured site input, but they are vulnerable to optimism. Investigators may estimate eligible patient counts from memory. Coordinators may not have time to run detailed chart reviews. Sites may overstate capacity because they want access to the study.
Questionnaires also miss catchment dynamics. A hospital may see a large number of patients, but many may live too far away for repeated visits. A site may have a strong investigator but weak referral relationships. A specialist clinic may see the disease but not the right severity stage.
A better feasibility process asks sites for evidence. How many patients with the target diagnosis were seen in the past year? How many had the required biomarker? How many met prior-treatment criteria? How many are still in follow-up? How quickly can the site run pre-screening? What competing trials are active?
Using real-world evidence for site feasibility
Real-world evidence changes site feasibility because it allows sponsors to estimate eligible populations before activation. EHRs, claims, registries, laboratory databases, and pharmacy records can show where patients are, how they are treated, and how many likely match the protocol.
For example, an oncology study may need patients with a specific cancer type, biomarker status, prior therapy exposure, performance status, and lab thresholds. A diabetes trial may need patients on particular medication combinations with defined HbA1c ranges and renal function. A cardiology trial may need recent hospitalization, imaging, medication use, and follow-up availability.
Real-world data will not perfectly reproduce trial eligibility. Some fields are missing. Some criteria require clinician judgment. But even imperfect structured data is better than guessing from annual disease counts.
Kapsule uses de-identified patient and market data for exactly this kind of feasibility reasoning in African markets: identifying where relevant patient populations are concentrated, which facilities generate usable records, and which countries deserve deeper operational diligence.
Site identification in emerging markets
Emerging markets may improve access to relevant patient populations and support more representative enrollment, but only when patient opportunity is matched with ethics, regulatory, staffing, laboratory, pharmacy, data, and monitoring readiness. Sponsors should avoid assuming that high disease burden automatically equals trial readiness.
Africa illustrates the distinction: high disease burden in many therapeutic areas does not automatically mean a site is ready for a sponsor-regulated interventional trial. Some facilities have strong investigators, ethics committees, laboratory capacity, and digital records. Others may have patient volume but limited research infrastructure.
A data-driven approach should score both patient opportunity and execution readiness.
- Patient opportunity: disease prevalence, eligible patient counts, treatment patterns, referral catchment, and diversity relevance.
- Infrastructure: ethics timelines, pharmacy, lab, imaging, cold chain, internet, source documentation, and monitoring readiness.
- Workforce: PI experience, coordinators, nurses, data managers, and backup staffing.
- Data maturity: EHR coverage, registry participation, lab-system integration, and source-data quality.
- Access: patient travel burden, reimbursement, translation needs, and community trust.
This framework prevents sponsors from overlooking capable sites outside traditional geographies while also avoiding underprepared activation.
Competing-study pressure
A site can be excellent and still be the wrong choice if it is saturated. Competing-study pressure matters in oncology, rare disease, immunology, and cardiometabolic indications where multiple sponsors chase similar patient pools.
Sponsors should review public trial registries, investigator participation, expected enrollment windows, standard-of-care changes, and referral-network overlap. If several studies are open for the same line of therapy, adding another site may not improve enrollment.
Competition also affects staff. Coordinators managing multiple complex trials may not have time for another demanding protocol. Pharmacy and lab teams can become bottlenecks even when the investigator is committed.
Measure competition at the patient-pathway level. Two trials at different hospitals can still compete if they draw referrals from the same oncology network. A private clinic and a public teaching hospital may compete for the same specialist's time. A registry study may seem low burden but still consume coordinator capacity during screening and follow-up.
Sponsors should also consider standard-of-care changes. A new reimbursed therapy, diagnostic guideline, or national treatment programme can change the eligible population during a trial. Good site selection includes a view of the policy and market environment, alongside current patient counts.
Protocol fit matters more than site reputation
A prestigious site may not fit a protocol. The best site for a first-line community-acquired infection study may be a high-volume district hospital, not an academic centre. The best site for a pharmacogenomics study may be a facility with strong laboratory linkage and longitudinal records. The best site for a decentralized follow-up model may be a clinic network with reliable patient contact information.
Sponsors should build protocol-specific scoring. The scorecard for an oncology biomarker study should differ from the scorecard for a vaccine trial, an observational RWE study, or a rare disease registry.
The protocol also determines the patient-retention risk. A study requiring frequent visits, imaging, or long follow-up needs sites that can support retention. That connects site selection directly to patient recruitment clinical trials: recruitment success includes completed participation, not only first consent.
Protocol fit should be tested through patient journeys. How does a patient reach diagnosis? Where are labs performed? Who pays for imaging? How often do patients change providers? Which visits can be aligned with routine care? A site that looks strong on paper may fail if the protocol conflicts with the way patients actually move through the health system.
This matters in multi-country studies. The same diagnosis can have different treatment pathways in Kenya, Nigeria, South Africa, Brazil, and the United States. Site selection should account for these differences before enrollment targets are assigned.
A site-selection scorecard
A clinical trial site selection scorecard should combine quantitative and qualitative evidence.
Core domains include:
- Eligible patient estimate from real-world data or chart review
- Historical enrollment in similar studies
- Screen-failure risk against protocol criteria
- Competing-study burden
- Ethics and contracting timelines
- Investigator and coordinator capacity
- Laboratory, imaging, pharmacy, and cold-chain readiness
- EHR/source-data quality
- Patient travel burden and retention support
- Diversity contribution relative to the study plan
A scorecard is a way to put the argument on paper. If a team chooses a lower-scoring site because of strategic value, the reason belongs in the record.
The same scorecard can separate sites that are ready now from sites worth building for later. Some sites may not be ready for a pivotal study today but could become valuable with training, equipment, contracting support, or a smaller observational study first. Sponsors that invest in site development can create future capacity in geographies with high unmet need.
For African sites, this pathway matters. A hospital may have strong patient volume and motivated clinicians but limited research administration. A phased approach can begin with retrospective data review, then registry participation, then interventional readiness. That builds evidence and trust without overloading the site too early.
Monitoring site performance after activation
Site selection does not end at activation. Early performance monitoring compares expected and actual screening, screen-failure reasons, consent rates, query rates, protocol deviations, and retention. If a site misses early indicators, sponsors need to intervene quickly.
Underperformance is not always the site's fault. The protocol may be too narrow, referral materials may be unclear, lab turnaround may be slow, or patient reimbursement may be inadequate. A good sponsor distinguishes fixable barriers from poor fit.
The monitoring plan also needs exit criteria. Keeping a non-enrolling site open for political reasons drains budget and distracts study teams. Closing or pausing a site can be the right decision when evidence shows the assumptions were wrong.
Practical takeaways
Clinical trial site selection should begin with patient evidence and end with operational proof. Ask where eligible patients actually receive care. Test protocol criteria against real records. Review competing trials. Validate site capacity. Budget for patient access and retention. Monitor performance quickly after activation.
Do not open sites for the sake of a long site list. Open the right sites early enough that the trial does not spend its first year learning what feasibility could have shown.
Kapsule provides access to structured, de-identified health records covering over 75 million patients across 14 African countries. Contact our team to discuss how African real-world data can support trial feasibility, country selection, and site identification.
This article is intended for informational purposes only and does not constitute legal, medical, or regulatory advice. Readers should obtain independent professional counsel for their specific circumstances.