18 Site Selection and Feasibility
A clinical trial exists only on paper until it enrolls patients, and patients are enrolled at clinical sites: hospitals, academic medical centers, private practices, or specialty clinics where investigators conduct the research. Industry surveys consistently show that patient recruitment is among the most common reasons for trial delays, and the wrong sites are where slow enrollment and poor data quality can begin. Historically, a Tufts CSDD benchmark of more than 150 trials run between 2008 and 2010 found that roughly 11% of investigative sites fail to enroll a single patient, with an additional 37% under-enrolling relative to their targets (Tufts Center for the Study of Drug Development 2013; Getz 2012). For a significant portion of activated sites, the substantial investment in startup and training yields little to no scientific data. More recent Tufts CSDD benchmarking is somewhat more encouraging: comparing site-activation and enrollment metrics across 2012, 2019, and 2023, Lamberti and colleagues found that actual enrollment exceeded planned enrollment for a majority of studies and that timelines were often shorter than expected, though achievement varied by region and site type (Lamberti et al. 2024).
For a Phase III trial requiring 1,000 patients, a sponsor might need 100 or more sites across multiple countries, each capable of identifying eligible patients, conducting the study procedures competently, maintaining regulatory compliance, and producing high-quality data. Finding sites that meet all these criteria, and that are genuinely committed to the trial, is harder than it might appear.
18.1 Site Identification
The search for sites begins with investigators who have access to the target patient population. A rheumatoid arthritis trial needs rheumatologists. An advanced lung cancer trial needs oncologists at centers that see late-stage disease. A rare disease trial may need the handful of specialists worldwide who see enough patients to contribute meaningfully.
Access to patients matters most: a distinguished investigator who rarely sees the target population will not enroll patients. Geographic distribution matters for global trials, where regulatory authorities may require data from relevant populations and where sponsors want to understand drug effects across regions, and prior experience with the sponsor, the therapeutic area, or clinical research generally can indicate capability.
Databases of investigators, clinical trial registries, publication records, and prior trial performance all inform site identification, as do sponsor relationships with key opinion leaders and clinical networks. What those sources turn up ranges from large academic medical centers (AMCs) and integrated health systems to small, independent physician practices and site networks (see Figure 18.1).
The pool of available investigators is shrinking. A 2004 survey found that about 13% of physicians actively served as clinical investigators (Harris Interactive 2004); the 2024 WCG Clinical Research Site Challenges Report documents the number of unique U.S. investigators falling roughly 78%, from a pandemic-inflated peak of over 28,000 in 2020 to approximately 6,200 in 2024, alongside a 57% decrease in U.S. trial starts over the same period (WCG Clinical 2024). That headline overstates the structural trend: measured from the pre-pandemic level of about 23,000 investigators, the decline is closer to 73%, still a steep erosion of the base. The 2024 investigator and trial-start counts are also preliminary, drawn from a partial-year ClinSphere snapshot that exaggerates the single-year drop from 2023; WCG’s more recent site-challenges reporting has continued to show the same downward direction of travel.
The two trends in Figure 18.2 are distinct. The investigator pool surged with the pandemic in 2020 and then crashed, while new trial starts held relatively stable, peaking near 2,000 in 2021, before falling to fewer than 800 in 2024. Several forces drive that decline. COVID-19-related clinical activity, which peaked in 2021, has dropped substantially. Biopharma funding slowed sharply in 2022 and only partially rebounded in 2023, remaining well below its 2020-21 pandemic peak and constraining emerging-biopharma companies (IQVIA Institute for Human Data Science 2024). With the cost of bringing a single novel asset to launch, failures included, now estimated at $2.6-2.7 billion (Deloitte Centre for Health Solutions 2026; DiMasi, Grabowski, and Hansen 2016), sponsors have grown risk-averse, concentrating capital on existing late-stage pipelines instead of new early-phase starts. Mounting regulatory hurdles and administrative burdens, meanwhile, have pushed study start-up past 60 days at most sites, discouraging new starts during periods of economic uncertainty (WCG Clinical 2024).
Other pressures are structural. Increasing clinical demands leave physicians with limited bandwidth for research, while research team staffing shortages further strain site capacity. This has exacerbated the “one and done” phenomenon, where approximately 66% of physicians who participate in a clinical trial do so only once before exiting the research ecosystem (WCG Clinical 2024). The consequences of this investigator attrition for recruitment and retention are discussed further in Chapter 19.
In that same 2004 survey, among physicians not involved in clinical trials, 38% cited lack of opportunity as the main reason for not serving as a Principal or Sub-Investigator, while 32% indicated that the time commitment is prohibitive (Harris Interactive 2004). Those constraints, compounded by the current exit of experienced investigators, make site identification a search for a rare combination: expertise, patient access, and operational bandwidth.
Historically, clinical trials have often failed to represent the diverse populations that will ultimately use the investigated therapies, which compromises the external validity of the results and can obscure important differences in safety or efficacy across demographic groups. The Food and Drug Omnibus Reform Act of 2022 (FDORA) requires sponsors to submit Diversity Action Plans for Phase III trials (and certain other pivotal studies) to the FDA, but only for studies that begin enrolling more than 180 days after FDA publishes final guidance on the plans (U.S. Congress 2022). That statutory requirement stands, but its implementing guidance has been in flux: FDA’s June 2024 draft Diversity Action Plans guidance was withdrawn following a January 2025 executive order on diversity, equity, and inclusion, then restored under court order weeks later, and its final version was not issued by the statutory deadline of June 26, 2025 (Grossi 2025; U.S. Food and Drug Administration 2024a). A separate December 2025 final guidance, Enhancing Participation in Clinical Trials, kept the emphasis on representative enrollment while dropping the “diversity” framing (U.S. Food and Drug Administration 2025). Until the final Diversity Action Plans guidance appears, the submission requirement has not taken effect, so sponsors should confirm the current posture rather than treat the mandate as settled.
Site selection is the primary operational lever for achieving diversity goals, which means engaging sites embedded in underrepresented communities: community hospitals, federally qualified health centers (FQHCs), or practices in geographic areas with high demographic density of the target populations. Those sites may lack the dedicated research infrastructure (study coordinators, -80°C freezers, secure storage) found in major academic centers, so the assessment must weigh a site’s potential capability with appropriate support alongside its current capability. The goal is a site network that geographically and demographically mirrors the epidemiology of the disease.
18.2 Feasibility Assessment
Once potential sites are identified, a feasibility assessment evaluates whether each site can realistically conduct the study. It typically involves questionnaires completed by site staff and may be supplemented by in-person or virtual visits. Table 18.1 summarizes the questions it asks and the answers that count as red flags.
| Category | Key Questions | Red Flags |
|---|---|---|
| Patient Access | How many target patients per year? Competing trials? | <10 eligible patients/year; Multiple competing studies |
| Investigator | Prior trial experience? Enrollment track record? | No research experience; History of poor enrollment |
| Infrastructure | Dedicated research staff? Equipment? Data systems? | No study coordinator; Inadequate facilities |
| Regulatory | IRB timeline? Institutional requirements? | IRB backlog >3 months; Complex approval layers |
| Commitment | PI availability? Staff turnover? | Part-time PI; High coordinator turnover |
Each row calls for judgment. A site’s count of eligible patients has to be checked against the trials it is already running; investigator review covers licensure as well as enrollment track record; infrastructure means space for study procedures and laboratory capabilities along with dedicated research staff; and regulatory readiness turns on the efficiency of the local IRB and on any institutional requirements that could delay startup. Commitment is gauged from PI availability and coordinator turnover history, since a part-time PI or high turnover can undermine execution even at an otherwise well-resourced site.
18.3 The Enrollment Prediction Problem
Sites are notoriously over-optimistic about how many patients they will enroll. An investigator who claims to see 50 eligible patients per month and predicts enrolling 30 may ultimately enroll 3.
The gap has several sources. Investigators may not accurately know how many of their patients meet eligibility criteria: the criteria are often more restrictive than general clinical practice. Patients who appear eligible on paper may decline participation or prove ineligible upon screening, competing trials may draw away others, and staff turnover or other operational issues may limit enrollment capacity.
Experienced sponsors treat site predictions with skepticism and routinely discount them, often steeply (Getz 2008). Historical performance in prior trials is a more reliable guide than investigator optimism.
18.4 Financial and Contractual Negotiations
Between a qualified site and an activated one lies the negotiation of the Clinical Trial Agreement (CTA) and the site budget, and it is a consistent bottleneck: industry benchmarks show that the cycle time from site selection to site initiation is driven more often by contract execution than by regulatory approval (WCG Clinical 2024).
The CTA is a legally binding contract governing the relationship between sponsor and institution, addressing liability, indemnification, intellectual property rights, data ownership, and publication policy. Negotiations can be protracted, particularly with large academic medical centers that adhere to rigid institutional policies regarding indemnification and subject injury. For global trials, these agreements must also align with local laws regarding data privacy and anti-bribery regulations.
The site budget requires a granular accounting of both per-patient costs and overhead fees. A well-constructed one covers per-subject costs for clinical procedures, investigator time, and pharmacy services; startup fees for the administrative effort of regulatory submission and staff training; pass-through costs for variable expenses such as IRB fees and participant travel; and an institutional overhead surcharge applied to direct costs.
A related exercise, especially for U.S. sites, is the coverage analysis. Before a trial opens, the site reviews each protocol-required procedure to determine which costs may be billed to a participant’s insurance as routine clinical care and which must be paid by the sponsor as research costs. In the United States this determination rests on the Medicare Clinical Trial Policy (National Coverage Determination 310.1, “Routine Costs in Clinical Trials”) and, for commercial plans, the Affordable Care Act’s requirement that the routine patient costs of qualifying clinical trials be covered. Getting it wrong exposes the institution to double-billing and false-claims liability, so it demands specialized revenue-cycle expertise. Coverage analysis is one component of the study start-up work (alongside budgets and contracts) that sites most often identify as their leading source of start-up delay (WCG Clinical 2024).
Operational Bottlenecks by Site Type
The administrative burden of study start-up falls unevenly across the research ecosystem. According to the 2024 WCG Clinical Research Site Challenges Report, “Study Start-up” remains the single largest operational obstacle for larger institutions (academic medical centers, health systems, and community hospitals), with 43% of these sites identifying it as a primary challenge (WCG Clinical 2024). In contrast, smaller independent sites, physician practices, and site networks report start-up as a top challenge only 24% of the time, focusing instead on recruitment and retention as their primary burden (39%).
The timelines diverge accordingly. Approximately 60% of independent sites and physician practices report being able to initiate studies in under 60 days; 77% of academic medical centers, health systems, and community hospitals report start-up timelines exceeding 60 days, often because of the complexity of managing multiple internal stakeholders, the lack of centralized research administration, and bureaucratic regulatory layers (WCG Clinical 2024). Across all site types in the United States, “Budgets and Contracts” remain the single largest bottleneck, cited by 77% of sites as the primary contributor to start-up delays (see Figure 18.3).
Sponsors design their selection and activation strategy around these profiles, often favoring independent networks for rapid enrollment and academic centers for complex therapeutic areas or highly specialized patient populations.
Budget negotiation requires balancing fiscal discipline with the reality that sites must cover their costs to remain viable. Underfunding a site invariably shows up later as slow data entry and as deprioritization against better-funded trials.
18.5 Site Initiation
Once a site is selected and regulatory approvals are obtained, it must be initiated: prepared to enroll patients and conduct the study. Figure 18.4 shows the full site lifecycle from identification through closeout.
%%{init: {'flowchart': {'nodeSpacing': 25, 'rankSpacing': 35, 'padding': 8, 'diagramPadding': 30, 'subGraphTitleMargin': {'top': 6, 'bottom': 14}}}}%%
flowchart TB
subgraph row1[" "]
direction LR
A["Site<br/>Identification"] --> B["Feasibility<br/>Assessment"] --> C["Site<br/>Selection"] --> D["Contracts<br/>& Budget"]
end
subgraph row2[" "]
direction LR
E["IRB/EC<br/>Approval"] --> F["Site Initiation<br/>Visit"] --> F2["Screening &<br/>Enrollment"] --> G["Treatment &<br/>Follow-up"]
end
subgraph row3[" "]
direction LR
I["Last Patient<br/>Visit"] --> J["Site Close-Out<br/>Visit"] --> K["Archiving"]
end
M["Monitoring Visits<br/>(recurring throughout<br/>conduct)"]
row1 --> row2 --> row3
row2 -. ongoing .- M
The Site Initiation Visit (SIV) formalizes training. During the SIV, study personnel review the protocol in detail, walk through study procedures, demonstrate equipment, set up data systems, and ensure that essential supplies are in place. The investigator and key site staff must participate; their engagement (or lack thereof) often signals future performance.
Following the SIV, the site is authorized to begin screening and enrolling patients. The time from site selection to the point at which a site is cleared to enroll (the site activation time), and from there to the first patient enrolled, can range from weeks to months depending on regulatory complexity, contract negotiations, and site responsiveness.
A global network needs training delivered consistently across languages and time zones, usually through a combination of e-learning platforms and regional investigator meetings. Communication has to strike a balance: enough information to keep sites aligned, without overwhelming staff with unnecessary updates. Performance management systems track real-time metrics (enrollment rates, data quality, query resolution times) to identify the sites that need additional support or, in extreme cases, closure. Sponsors treat high-performing sites as long-term partners, and investment in those relationships through transparent communication and professional support improves enrollment, data quality, and site retention across the drug development program.
Global trials add complications of their own. Regulatory requirements differ by country, and language barriers affect training and documentation. Cultural factors influence patient willingness to participate in research, expectations about investigator relationships, and approaches to adverse event reporting. Standard of care may differ across regions, affecting choice of comparator and interpretation of results, while patient populations may differ in disease characteristics, comorbidities, and concomitant medications. Some regions may have limited research infrastructure requiring more intensive support, while others may have mature research environments but intense competition for investigators and patients. All of it feeds back into trial design and analysis.
18.6 Sites New to Research
Table 18.1 lists “no research experience” as a red flag, and for a sponsor that needs fast enrollment it is one. For the trial population it is often the opposite. Patients who live far from academic centers can join a trial only if a site near them takes part, and many such sites have never run one. This section covers what it takes to bring a hospital or clinic with no trial experience into a study: what it lacks, who supplies it, and how a sponsor oversees it.
The literature has no standard term or definition. Industry writing uses “research-naive” or “trial-naive” loosely for community hospitals and clinics with little or no trial history. “Rural” is no better defined: a scoping review of pediatric trials in rural and community settings found no common definition of rurality, and about a third of the articles described their setting too briefly to tell (Watson et al. 2022).
Where the Gap Is
In cancer, the size of the gap is known. Using Commission on Cancer accreditation data for 2013 to 2017, Unger and colleagues estimated that 7.1% of adult cancer patients took part in a treatment trial and at least 21.9% in some kind of research study. Treatment trial participation was 21.6% at NCI-designated comprehensive cancer centers, 5.4% at academic programs without NCI designation, 5.7% at integrated network programs and 4.1% at community programs. The programs outside NCI designation made up 96.4% of the accredited sites, and the authors point to the infrastructure and staff support that NCI centers receive and other programs lack (Unger et al. 2024).
The facilities to close the gap already exist. Chapter 19 cites the estimate that nearly 17% of U.S. residents aged 35 and older live more than 100 miles from an NCI-funded trial site. The same study found that counting the Commission on Cancer accredited programs as well cuts that share to 1.6%; for people below the poverty line, the figures are 20.2% and 1.9%. Over 97% of NCI cancer center and NCORP facilities lie in urban core areas, against 85% of accredited programs, and only three NCI-funded sites lie outside urban cores, against 179 accredited programs (Shriver et al. 2025). Most of the hospitals that could bring trials to rural patients treat cancer already. What they lack is trial capacity.
For children, the gap shows up as distance to a subspecialist. In 2019, depending on the subspecialty, between 2% and 53% of US children lived 80 or more miles in a straight line from a pediatric subspecialist, measured from the subspecialist’s address to the child’s zip code (Turner, Ricketts, and Leslie 2020). A 2025 erratum corrected the paper’s sampling and coding errors, and the corrected range is the one quoted here.
What a First-Time Site Lacks
Direct evidence on first-time sites is thin. The most detailed assessment comes from a network that was building capacity on purpose. When the ECHO IDeA States Pediatric Clinical Trials Network assessed its 17 sites and 7 affiliates while the network was being built, every principal investigator reported prior trial experience, but only 38% to 42% of study coordinators reported experience in the relevant domains. Most sites had an on-site pharmacy (93%) and a research pharmacy for investigational agents (79%), and 92% to 96% could store and ship biosamples, yet about 17% did not use an electronic medical record. The barriers the sites named were knowledge gaps among the regional affiliates about local study conduct and follow-up, little protected research time, grant budgeting and contracting, and the need for trusted community champions (Annett et al. 2021). These were sites of an organized network with academic leads, so a hospital starting alone is likely to lack more.
The pediatric scoping review found the same leading barrier from the other side. Across 27 articles covering 31 U.S. studies, insufficient staff and resources was the barrier cited most often, followed by family distrust, lack of family time, burdensome or frightening procedures, and transportation. Only six of the studies were predominantly or exclusively rural, only two examined rural retention, and none compared strategies head to head or measured their cost-effectiveness (Watson et al. 2022).
One study measured what first-time sites had to build. In 2017 the same network joined a pharmacokinetic study run by the Pediatric Trials Network. Of its 18 sites, 4 had been active before the network existed and 14 were activated between April 2018 and May 2019, and together they recruited 382 participants by the study’s close in September 2019. Of the 20 principal investigators, 7 (35%) were early-stage investigators and 7 (35%) were new to pharmacokinetic research, as were 8 of the 18 coordinators. To take part, 10 sites (56%) added study staff, 7 (39%) added equipment or space, 12 (67%) built new specimen-processing workflows, 5 (28%) wrote new SOPs, and 16 (89%) built new EHR queries to find eligible patients. Sites rated the effect on their research team’s confidence at 8.1 on a 0 to 10 scale (Knight et al. 2022). The paper reports activation dates and enrollment, not comparisons with experienced sites, and does not report how many participants were rural.
From these sources and from the start-up steps described earlier in this chapter, a site’s first trial typically lacks five things: a coordinator who knows GCP and the sponsor’s systems; a regulatory file and the habit of keeping it; a contract and budget office that has negotiated a clinical trial agreement before; a billing process that can separate research costs from routine care under a coverage analysis; and a pharmacy that can receive, store and account for investigational product. Each of these can be supplied from outside the site, but someone has to be named to supply it.
Who Does What
Table 18.2 sets out a division of labor that puts the specialized work on the party that already has the expertise, and leaves the site the tasks only it can do. It is a planning template, not a description of measured practice.
| Task | Sponsor, CRO or coordinating center | Site investigator and coordinator | Host institution |
|---|---|---|---|
| Readiness assessment | Visits or calls the site; checks staff, space, pharmacy, records, internet access | Answers honestly, including what is missing | Confirms support from administration |
| Regulatory file | Supplies a pre-built file structure and templates (delegation log, training log, Form 1572) | Keeps the file current from the first day | Provides secure storage |
| Training | GCP, protocol, EDC and safety reporting training before activation; repeated after amendments | Completes training and documents it | Protects research time for staff |
| IRB | Arranges reliance on a single IRB where the trial allows it | Completes local paperwork | Signs the reliance agreement |
| Contract and budget | Offers a standard template and a budget that covers start-up and coordinator time | Reviews feasibility of the budget | Signs; sets up the research account |
| Coverage analysis and billing | Provides or funds the coverage analysis | Flags research-only procedures at each visit | Bills routine care and research correctly |
| Investigational product | Ships, sets up accountability logs, trains on storage | Dispenses and records | Pharmacy receives and stores |
| First enrolled patient | Reviews the first consent and first visits promptly, remotely or on site | Calls the monitor before and after the first visit | |
| Ongoing oversight | Central review of data and deviations; visits targeted to the site’s risks | Answers queries quickly |
Oversight in Proportion to Risk
ICH E6(R3) asks sponsors to scale oversight to the risks that matter for participant safety and the reliability of results (International Council for Harmonisation 2025), and site inexperience is a risk factor in its own right. In practice that argues for more attention early (review of the first consents, the first eligibility decisions and the first safety reports) and less once the site has shown it can do the work. Central monitoring of the kind described in Chapter 22 can follow a small site between visits, because the data reach the sponsor whether or not a monitor travels.
Decentralized elements can let a small site take part without carrying every procedure. FDA’s 2024 guidance on decentralized elements describes trial visits conducted by telehealth, in participants’ homes, or with local health care providers (U.S. Food and Drug Administration 2024b), so a rural clinic can handle routine visits close to home while the specialized procedures stay at a larger center. Chapter 19 describes what those elements do for patients’ travel.
What Capacity Programs Have Measured
Several programs have tried to bring trials to community and rural settings. Table 18.3 lists those with published figures, and what each figure does and does not show.
| Program or study | What was measured | What it does not show |
|---|---|---|
| NCI Community Oncology Research Program (NCORP) | Launched 2014; in 2019, 7 research bases, 32 community sites, 14 minority/underserved sites (at least 30% minority or rural patients) and more than 1,000 affiliates (National Cancer Institute 2019). NCI’s NCORP director reported 71,184 accruals from August 2014 to April 2022 (McCaskill-Stevens 2022) | Accrual per newly added affiliate; rural share of accrual |
| Mayo Clinic, three campuses, 2016 to 2017 | 731 of 2,313 trial participants (32%) were rural, against 19% of the population of the nine states served; rural participants traveled 103 miles on average, urban 68 (Bharucha et al. 2021) | Whether regional sites, rather than referral to campuses, drove rural enrollment |
| Mayo Clinic decentralized trial program, 2024 to 2025 | Across 765 trials and 7,469 participants, the rural share rose from 16.6% to 24.9% between the first quarters of 2024 and 2025 (Haddad et al. 2026) | Cause: a descriptive study, and the authors do not attribute the change to the program |
| Academic molecular tumor board for rural community cancer centers | Of 151 patients reviewed, 27 were matched to a trial and 1 enrolled; distance to the trial center was the barrier cited most (70%) (Kumar et al. 2024) | Expertise without a local trial did not produce enrollment |
| ASCO and ACCC site self-assessment pilot | 62 sites; 82% identified opportunities for improvement (Guerra et al. 2023) | Any change in enrollment after the assessment |
| Iowa shared care of children with cancer, about 15 years | 137 community primary care physicians shared management of 226 children on oncology protocols; survival did not differ from 240 children treated by pediatric oncologists alone (Kisker et al. 1997) | Retrospective comparison, decades old |
| SWOG Pragmatica-Lung, 2024 update | About 20% of accrued patients lived in a rural location, with accrual spread across NCORP sites, academic sites and other rostered sites (National Cancer Advisory Board 2024) | Rural definitions varied; a single trial |
Two kinds of program have no outcome data at all. We found no peer-reviewed report of Project ECHO tele-mentoring producing trial activations or enrollment at community sites; the one curriculum aimed at trial access that we found reported preliminary survey results only (National Association of Chronic Disease Directors 2026). We also found no citable trial-access outcomes for networks serving American Indian and Alaska Native communities. Evidence can also be lost: a 2023 retrospective of an NCI program that placed radiation oncology trials in six institutions serving underserved patients could not collect patient data so long after the program ended, and relied on a survey of its principal investigators (Coleman et al. 2023).
What the Evidence Still Does Not Show
No study we found compares the activation time, accrual or protocol deviation rates of first-time sites with those of experienced sites, and none compares ways of supporting them. The ISPCTN pharmacokinetic study reports what new sites built and when they opened, not how they performed against others. Figures for NCORP’s share of national accrual and its minority accrual by site type appear in secondary sources that we could not check against the original evaluation. The advice above rests on the regulatory requirements, the barriers sites report, and the start-up steps every site must complete. A sponsor that brings new sites into a trial should record their start-up times, deviations and enrollment, because at present that is the only way anyone will know what works.
TipChecklist: Activating a Site with No Trial Experience
- Name one person at the sponsor, CRO or coordinating center as the site’s contact for the whole start-up.
- Assess readiness in person or by video, including pharmacy, records access and internet connection.
- Provide a pre-built regulatory file and walk the coordinator through it.
- Train on GCP, the protocol, EDC and safety reporting before the initiation visit, and record the training.
- Fund the coordinator’s time and the coverage analysis in the budget.
- Use a single IRB and a standard contract where the trial allows.
- Review the first consent, eligibility decision and safety report within days, not at the next routine visit.
- Track start-up time, deviations and enrollment for the site, and compare them with the trial’s experienced sites.