Site Feasibility: Can This Site Deliver the Endpoint Assessments?

When a clinical development team prepares to launch a confirmatory clinical trial, what does a site feasibility questionnaire actually need to determine? In standard industry practice, candidate investigative sites routinely receive voluminous surveys centered on patient demographics, historical recruitment metrics, and high-level staffing headcounts. If an academic medical center or private research network answers that it can enroll 24 eligible patients within six months and checks the box indicating an available full-time study coordinator, sponsors frequently grant study startup approval. Yet across clinical trial operations, sites selected on recruitment projections alone routinely encounter severe delivery bottlenecks as soon as protocol procedures commence.

The fundamental operational reality is straightforward: recruitment capability and endpoint delivery capability are distinct operational domains governed by entirely different constraints. A clinical site may possess a vast registry of patients diagnosed with moderate-to-severe atopic dermatitis, rheumatoid arthritis, or early Alzheimer's disease. However, if that site lacks a clinician certified to administer the protocol's primary severity scale on patient visit days, operates on a hospital IT network that blocks the transmission of electronic patient-reported outcome (ePRO) diaries, or cannot process venous blood samples in a refrigerated centrifuge prior to a Thursday courier deadline, the site cannot deliver the trial. The consequence of that failure is not merely operational delay; undelivered assessments create missing data and can introduce intercurrent events that the estimand has to address.

The regulatory anchor for rigorous site evaluation is established law and international guidance, rather than discretionary administrative preference. Under 21 CFR 312.50, the sponsor is legally responsible for selecting qualified investigators, providing them with the necessary information to conduct the investigation properly, and ensuring proper monitoring. Under 21 CFR 312.53, the sponsor must obtain from each investigator a signed Form FDA 1572—which explicitly identifies the clinical sites, facilities, and clinical laboratories to be utilized—alongside a curriculum vitae or statement of qualifications prior to permitting participation. The regulation further mandates that the sponsor select monitors qualified by training and experience to oversee trial progress.

The ICH E6(R3) Good Clinical Practice guideline (consolidated Step 4, June 2026) uses should language and is not a U.S. statute; applicable local regulation takes precedence. Section 3.7.1 (Investigator Selection) states that the sponsor is responsible for selecting the investigator(s)/institution(s), and that each investigator should be qualified by education, training, and experience and should demonstrate adequate resources and facilities to conduct the trial properly. In Sections 2.2.1 and 2.2.2, ICH E6(R3) states that the investigator should be able to demonstrate a potential for recruiting the proposed number of eligible participants and should have sufficient time, an adequate number of available and qualified staff, and adequate facilities for the foreseen duration of the trial. Principle 7 states that trial processes should be operationally feasible and avoid unnecessary complexity, procedures, and data collection. A questionnaire cannot rescue an assessment schedule that no realistic site can execute; that failure belongs to protocol design. When a protocol is sound, an endpoint-derived questionnaire is operational practice for documenting those selection duties for a specific protocol—not a form that 21 CFR 312.50 or 312.53 prescribes by format.

Furthermore, ICH E8(R1) General Considerations for Clinical Studies treats feasibility considerations—including qualified investigators and site personnel, equipment and facilities required to conduct the study, the targeted patient population, and the ability to enrol a sufficient number of participants—as critical-to-quality factors whose integrity is fundamental to protecting participants and to the reliability of study results. Protocol adherence is a separate operational quality topic; it does not replace those feasibility checks. A questionnaire that never asks how each endpoint will be delivered leaves the sponsor without documented evidence that a candidate site has the resources and facilities the measurements require, which is the practical basis for selecting, conditionally selecting with named remediation, or rejecting the site.

This article is the pre-selection decision: whether a candidate site can deliver the protocol's assessments. It does not rewrite executable endpoint definitions in ICH M11 protocol language, lock ICH E8(R1) critical-to-quality factors before protocol freeze, control ClinRO rater drift after a site is already trained, or document COA administration evidence. Rater coverage and device logistics are selection questions here; they become monitoring and training questions after SIV.

To build a questionnaire that tests endpoint-delivery capability, trial teams should abandon generic template banks and derive questions directly from the protocol's schedule of activities and endpoint definitions. Every primary, secondary, and safety outcome detailed in ICH M11 Section 3 relies on a specific sequence of human actions, technical tools, environmental conditions, and temporal windows. Translating each endpoint into concrete site capability probes requires evaluating four operational dimensions:

• Execution personnel: Exactly who administers, performs, or oversees the measurement? Is specialized board certification, licensed clinician status, or formal instrument qualification required, and who provides backup coverage when the primary assessor is absent?

• Instrumentation and modality: What equipment, device, or electronic platform captures the primary observation? Is the hardware sponsor-provisioned, commercially available, or research-dedicated, and are calibration and maintenance certificates current?

• Operational and environmental constraints: What physical space and schedule boundaries govern data capture? Does the protocol require a quiet, distraction-free room for cognitive batteries, a measured corridor for motor assessments, or strict fasting visit windows?

• Transfer and chain of custody: How does raw data move from the clinic floor to the central repository? Are biological specimens subject to centrifugation speed, temperature, and courier stability cutoffs, and can digital records navigate site firewall restrictions?

The following matrix maps primary endpoint modalities to specific questionnaire fields, required verification artifacts, and operational red flags that must trigger conditional remediation or disqualification:

This endpoint-derived architecture is not theoretical; it represents established practice among rigorous research networks. In the publicly documented feasibility questionnaire developed by the North American Prodromal Synucleinopathy (NAPS) Consortium (2020), questionnaire fields are constructed directly around the study's clinical assessments. Rather than asking only about generic laboratory access, the NAPS instrument requires candidate sites to specify currently certified staff to administer the Clinical Dementia Rating (CDR) and the Movement Disorder Society Unified Parkinson's Disease Rating Scale (MDS-UPDRS), research-dedicated MRI and PET/SPECT scanners by manufacturer and model, a 4°C refrigerated centrifuge, a -80°C freezer, a 10-foot hallway or room that can be marked for a timed walk test, a quiet room capable of supporting 3 to 3.5 hours of test administration, IATA-certified personnel able to ship samples on dry ice, and experience exporting polysomnograms and uploading them to a central reader.

By demanding objective evidence for each measurement requirement, the NAPS questionnaire operationalizes the core directive of ICH E8(R1): identifying critical-to-quality factors before trial initiation and confirming that participating institutions possess the operational capability to preserve measurement integrity.

Consider a hypothetical Phase 3 trial, Study TE-301, evaluating a novel targeted biologic for moderate-to-severe atopic dermatitis across a 24-week double-blind treatment period. The protocol specifies three key endpoints that determine study success:

• Primary efficacy endpoint: The proportion of participants achieving an Investigator's Global Assessment (vIGA-AD) score of clear (0) or almost clear (1) with at least a 2-grade reduction from baseline at Week 16. Protocol rules require this ClinRO score to be assigned by a qualified investigator certified on the validated scale.

• Key secondary endpoint: The proportion of participants achieving at least a 4-point improvement from baseline in the weekly mean Peak Pruritus Numerical Rating Scale (PP-NRS) at Week 16, recorded nightly by participants on sponsor-provisioned handheld eCOA smartphones.

• Exploratory biomarker endpoint: Serum cytokine kinetics measured from venous blood draws at Weeks 0, 4, 12, 16, and 24. Blood samples must be centrifuged at 4°C within 30 minutes of collection, transferred to cryovials, frozen immediately at -80°C, and shipped on dry ice to the central bioanalytical laboratory. Due to weekend central lab closure and courier customs processing, samples cannot be delivered on Saturdays or Sundays, and the central lab accepts no shipments dispatched on Thursdays or Fridays.

Candidate Site 104, a renowned urban academic dermatology center, completes a traditional feasibility questionnaire. The responses appear exceptional on paper: the site projects 24 evaluable patients within six months, cites a dedicated research coordinator with eight years of experience, and confirms full GCP certification under the Principal Investigator (PI). In standard practice, Site 104 would be fast-tracked for selection. However, when evaluated through an endpoint-to-capability questionnaire module, three critical operational conflicts emerge:

1. ClinRO rater availability: The questionnaire asks the site to identify primary and secondary certified vIGA-AD raters and their clinic availability. Site 104 discloses that only the PI holds vIGA-AD certification. The PI holds clinical research hours exclusively on Tuesdays and Thursdays, spending Mondays and Wednesdays in general dermatology clinics and Fridays performing dermatologic surgery. No sub-investigator has been trained or certified on the scale.

2. eCOA device provisioning and connectivity: The questionnaire asks the site to describe its device check-in/out workflow and clinic wireless infrastructure. Site 104 reports that while the main hospital campus has robust Wi-Fi, institutional cybersecurity policies enforce a mandatory 24-hour captive portal re-login that provisioned Android study devices cannot navigate. Furthermore, the site coordinator has no designated secure storage for backup handsets or replacement SIM cards.

3. Central laboratory shipping constraints: The questionnaire requires the site to cross-reference patient visit scheduling against the central lab's shipping cutoff. To meet its enrollment targets, Site 104 plans to conduct 50% of patient study visits at its suburban satellite clinic on Thursday afternoons. When queried regarding satellite facility capabilities, the site reveals that the satellite clinic possesses only an ambient benchtop centrifuge and a -20°C freezer. It lacks a 4°C refrigerated centrifuge, lacks a -80°C freezer, and has no contracted dry-ice courier service.

If the sponsor selects Site 104 without resolving these gaps, the study can reproduce the post-selection failure patterns documented by Bruneau et al. (2024). At the Site Initiation Visit (SIV), the team can discover that Thursday patient visits at the satellite clinic cannot process biomarker samples in accordance with protocol cold-chain specifications. The site then either reschedules visits—creating visit-window protocol-deviation risk—or stores samples at -20°C over the weekend, outside the protocol's -80°C processing specification.

Simultaneously, whenever the PI travels for academic conferences or handles surgical emergencies, Week 16 primary endpoint visits may be scored by an uncertified sub-investigator—breaking the protocol's certified-rater rule—or rescheduled outside the hypothetical ±3-day visit window. When participants struggle with eCOA device connectivity at the clinic, coordinators may skip on-site synchronization, delaying compliance tracking and losing diary entries.

The statistical consequences under ICH E9(R1) (Estimands and Sensitivity Analysis) are distinct from a mere operational delay. Intercurrent events are events after treatment initiation that affect the interpretation or existence of the measurements associated with the clinical question of interest, and they have to be addressed when the estimand is defined. Data that would be meaningful for that estimand but were not collected present a missing-data problem for the statistical analysis and should be distinguished from data that do not exist after an intercurrent event. If Week 16 vIGA-AD scores are never collected, the analysis has a missing-data problem; if visits cannot occur because a certified rater or shipping window was never available, that operational failure can also affect whether the measurement exists. Those consequences are why undelivered assessments are not a startup nuisance. They are not a prediction that any one imputation method will decide a marketing application. Sensitivity analyses that stay aligned to the same estimand are covered separately in Missing Data vs Intercurrent Events.

Instead of blind approval or outright rejection, the endpoint-derived questionnaire enables a precise, documented Conditional Selection with three non-negotiable pre-SIV remediation milestones:

• Milestone 1 (ClinRO): A designated sub-investigator must complete formal sponsor-approved vIGA-AD rater qualification training at least 21 days prior to SIV in this hypothetical, establishing verified primary and secondary investigator rater coverage across weekdays.

• Milestone 2 (eCOA): Hospital IT must approve and whitelist an isolated cellular IoT gateway or dedicated guest SSID in the clinic research suite to ensure frictionless device onboarding, verified by a successful test upload prior to SIV.

• Milestone 3 (Laboratory logistics): Site 104 must execute a binding protocol operational plan stipulating that all study visits requiring blood draws must be scheduled exclusively at the main urban facility on Monday through Wednesday mornings, ensuring sample processing in the verified 4°C centrifuge, storage at -80°C, and courier dry-ice dispatch prior to the Thursday morning cutoff.

If Site 104 fails to fulfill these verified artifacts 14 calendar days prior to the planned SIV date, conditional selection is revoked, and the study spot is reallocated to an alternate qualified institution. This structured triage protects the trial from costly downstream rescue operations.

Why do traditional site feasibility questionnaires fail so consistently? Over the past two decades, academic and industry investigators have extensively evaluated feasibility methodologies. The consensus findings highlight significant structural flaws in conventional self-reported surveys. In a seminal research statement published by the American Society of Clinical Oncology (ASCO), Kurbegov et al. (JCO Oncology Practice, 2021) characterized current site feasibility assessments as costly, inefficient, unnecessarily burdensome, and resource-intensive. The statement's conclusion also calls the process inconsistent, labor intensive, and of uncertain effectiveness, and it recommends minimizing and standardizing questions rather than expanding the form.

Similarly, in the 2024 task-force report by the Site Enablement League, Bruneau et al. demonstrated that clinical research sites frequently complete feasibility questionnaires based on incomplete or preliminary protocol synopses. Under commercial pressure to secure study allocations, sites routinely provide optimistic enrollment projections derived from raw electronic health record (EHR) diagnostic codes (e.g. billing ICD-10 counts). These high-level estimates fail to account for restrictive exclusion criteria, required biomarker washouts, or intensive patient visit schedules. When asked whether they possess adequate staff and equipment, sites instinctively check affirmative boxes, assuming operational details can be sorted out during startup.

Meanwhile, the commercial search landscape rewards vendor glossaries and contract research organization (CRO) promotional guides—such as materials from Advarra and Credevo—that define feasibility questionnaires around generic checklists of facilities, standard equipment, and patient availability. While useful as introductory overviews, these commercial baselines provide no objective method to distinguish high-predictive evidence from unverified assertions.

To establish an effective feasibility evaluation framework, trialists must systematically separate low-predictive conventional claims from high-predictive, artifact-backed signals across key operational domains:

By requiring verifiable audit artifacts for high-predictive signals, the study startup team shifts the evaluation from subjective optimism to objective operational capability. A site that cannot produce calibration records, rater schedules, or pre-screened cohort filters before selection is signaling that its startup execution will be fraught with delays.

A rigorous questionnaire must culminate in a transparent, standardized decision framework. Evaluating endpoint capabilities is not an exercise in collecting trivia; it is an operational triage that categorizes candidate sites into three distinct pathways:

• Green — Unconditional Selection: The site provides complete, verifiable audit artifacts for every endpoint-delivery requirement. Primary and backup raters are certified and scheduled; physical testing facilities meet protocol space and quietness standards; equipment is calibrated within specification; and laboratory courier logistics perfectly align with visit schedules and stability windows. The site proceeds immediately to contract execution and SIV scheduling.

• Yellow — Conditional Selection: The site exhibits remediable capability gaps that do not represent permanent structural defects. Examples include scheduling secondary rater certification, procuring an approved IATA dry-ice courier contract, configuring an isolated guest Wi-Fi network for eCOA provisioning, or establishing a protected clinic block for specialized visits. Conditional selection requires a documented Remediation Plan with a named owner, concrete deliverables, and a binding deadline at least 14 days prior to SIV.

• Red — Disqualification / Rejection: The site exhibits non-remediable structural deficits that make protocol compliance impossible. Examples include the complete absence of mandatory specialized hardware (e.g. lacking a 3.0T MRI scanner or certified lumbar puncture suite), geographic isolation that prevents courier delivery within sample stability windows (e.g. PBMC viability requires 4-hour isolation, but transit takes 12 hours), institutional policies prohibiting mandatory procedures, or irreconcilable investigator schedule conflicts. The site is disqualified, protecting both the trial and the site from inevitable failure.

The decision workflow governing this operational triage is illustrated in the following process diagram:

flowchart TD
  A["Protocol Endpoint Table & Schedule of Activities"] --> B["Generate Endpoint-Specific Feasibility Module"]
  B --> C["Site Submission of Questionnaire & Audit Artifacts"]
  C --> D{"Feasibility Triage Evaluation"}
  
  D -->|"All Evidence Verified & Fully Aligned"| E["GREEN: Unconditional Selection"]
  D -->|"Remediable Operational Gaps Identified"| F["YELLOW: Conditional Selection"]
  D -->|"Structural Deficits or Incompatible Logistics"| G["RED: Disqualification / Rejection"]
  
  F --> H["Issue Binding Remediation Plan with Named Owner"]
  H --> I{"Pre-SIV Milestone Audit (14 Days Prior)"}
  I -->|"All Milestones Met & Verified"| E
  I -->|"Milestones Unfulfilled"| G
  
  E --> J["Clear for Site Initiation Visit (SIV)"]
  J --> K["Seed Protocol Risk-Based Monitoring (RBM) Plan"]
  G --> L["Reallocate Spot to Alternate Candidate Site"]

In designing this scoring framework, sponsors should heed the warning issued in the ASCO research statement (Kurbegov et al., 2021): current feasibility assessments are unnecessarily burdensome, and questions should be minimized and standardized. The workable design is a lean standardized core plus a small protocol-specific endpoint module—not a longer form.

The solution is not to create an 80-page bespoke survey for every study. Rather, clinical operations teams should implement a lean, standardized core institutional profile (capturing general facility accreditation, standard IT infrastructure, and GCP credentials) and append only a concise, tailored Endpoint Delivery Module (1–2 pages) derived directly from the protocol's schedule of events. This modular approach minimizes administrative overhead while rigorously probing the exact operational dependencies that govern trial success.

The primary strategic value of an endpoint-derived feasibility questionnaire extends far beyond initial site selection. In modern clinical trial governance, questionnaire responses provide the foundational risk baseline that feeds directly into the trial's Risk-Based Monitoring (RBM) and Risk-Based Quality Management (RBQM) architecture.

The 2013 guidance Oversight of Clinical Investigations — A Risk-Based Approach to Monitoring (August 2013) contains nonbinding recommendations. FDA recommends that sponsors prospectively identify critical data and processes, perform a risk assessment of what could affect collection of those data or performance of those processes, and develop a monitoring plan focused on important and likely risks. Monitoring activities should focus on preventing or mitigating important sources of error in critical data and processes. The monitoring plan should describe the study, its objectives, and the critical data and procedures, with particular attention to data and procedures that are unusual relative to clinical routine and require site training. FDA also notes that sponsors may wish to track deviations related to collection or reporting of data necessary to interpret the primary endpoint, and that the relative experience of the clinical investigator is among the factors that can justify more intensive monitoring.

Similarly, the consolidated ICH E6(R3) guideline states that the sponsor should identify risks that may have a meaningful impact on critical-to-quality factors prior to trial initiation and throughout trial conduct (section 3.10.1.1). ICH E6(R3) also states that the monitoring strategy should ensure appropriate oversight and consider site capabilities and the potential burden, and that the plan should focus on aspects that are critical to quality. When feasibility evaluations capture site-level endpoint-delivery capabilities, they provide documented inputs for that risk assessment:

• Calibrating on-site Source Data Verification (SDV): If a site was conditionally approved under a yellow flag—for instance, where a secondary ClinRO rater was recently qualified—the monitoring plan can increase on-site review of source records and rater workbooks for that site's early randomized participants, whereas an established site with historical rater stability may receive more targeted sampling. Those intensity choices are sponsor risk-assessment outputs, not a universal source-data-verification quota.

• Configuring Centralized Statistical Monitoring (CSM): Feasibility findings seed specific Key Risk Indicators (KRIs) within centralized monitoring platforms. If a site reported tight courier schedules for central lab bioanalytical shipments, the centralized monitoring team establishes an automated KRI tracking specimen accessioning delays and hemolyzed sample rates, immediately alerting the CRA if transit times exceed protocol thresholds.

• Targeting eCOA and digital endpoint surveillance: Sites that reported potential wireless connectivity challenges or coordinator staffing constraints can be flagged for closer compliance surveillance. Centralized review can then watch diary completion and device synchronization against the protocol's own compliance rules, so a CRA can intervene before missing assessments accumulate. Numeric alert thresholds belong in the monitoring plan, not in a generic questionnaire template.

By bridging feasibility findings directly into the monitoring plan, the sponsor ensures that operational risks identified during study startup are not forgotten in an archived survey folder. Instead, they become active parameters in the trial's quality assurance framework, driving focused oversight where it matters most: safeguarding the data integrity of primary and key secondary endpoints.

The operational landscape of trial endpoints has expanded significantly with the advent of decentralized clinical trial (DCT) models and digital health technologies (DHTs). In September 2024, the FDA issued its final guidance, Conducting Clinical Trials With Decentralized Elements, establishing comprehensive regulatory expectations for trials utilizing remote visits, local healthcare providers (HCPs), and home-based endpoint collection.

A common misconception among study teams is that decentralized elements reduce the operational burden on clinical research sites. In reality, decentralized protocols change the site's task list. Under FDA's 2024 guidance, investigators are responsible for the conduct of a DCT and for protecting the rights, safety, and welfare of subjects under their care, including when trial-related activities occur away from the traditional site. Feasibility questionnaires for hybrid and decentralized studies therefore need to probe an additional set of site capabilities:

• Third-party contracted service coordination: DCTs frequently require investigative sites to coordinate with visiting home health nurses, local clinical laboratory facilities, mobile phlebotomy vendors, and direct-to-patient investigational product couriers. The questionnaire should verify that the coordinator team has the bandwidth and contractual mechanisms to manage those schedules.

• Oversight and reconciliation of remote clinical data: In the investigator-and-delegation discussion of that September 2024 guidance, investigators should review data from other trial personnel and local HCPs, as applicable, and follow up on any data that are missing, concerning, or appear to be in error. The feasibility questionnaire should ask whether the site has designated personnel and a workflow to review those external records.

• Digital health technology and device triage: When continuous physiological sensors or other remote devices are deployed, participants often encounter technical obstacles such as pairing failures, sensor detachment, or operating-system incompatibilities. The feasibility questionnaire should ask whether site staff can provide first-line device troubleshooting or whether technical escalations flow to a centralized vendor help desk.

FDA's DCT guidance also advises that specific issues related to the feasibility, design, implementation, or analysis of a DCT should be discussed early with the relevant FDA review divisions. Documented site-oversight and data-review workflows are operational evidence for that discussion; they are not a substitute for protocol design, and they do not constitute a regulatory determination that decentralization is adequate. Operating rules for collecting endpoints away from traditional sites are covered separately in Decentralized Endpoint Collection: FDA, EMA, and ICH E6(R3) Annex 2 Operating Rules.

While an endpoint-derived feasibility questionnaire is an indispensable tool for study startup, clinical trial leaders must maintain a clear understanding of its inherent operational boundaries:

• Self-reported evidence is screening intelligence, not empirical proof: Even when accompanied by CVs and calibration certificates, questionnaire responses represent claims made by the site. Physical inspection during Pre-Study Qualification Visits (PSVs) or detailed operational walkthroughs during SIV remain essential to corroborate that equipment is genuinely accessible, personnel are actively engaged, and workflows operate as described.

• A questionnaire cannot cure an infeasible protocol: If a trial's assessment schedule is fundamentally unworkable—for example, demanding that frail oncology or geriatric patients undergo five consecutive hours of invasive biopsies, cognitive tests, and pharmacokinetic draws within rigid 15-minute windows—no site questionnaire can prevent trial failure. ICH E6(R3) Principle 7 states that trial processes should be operationally feasible and avoid unnecessary complexity, procedures, and data collection, and Principle 8 adds that the protocol itself should be operationally feasible. The fix is protocol redesign, not a longer feasibility form.

• Feasibility mapping is a decision framework, not a static psychometric tool: Questionnaire fields inevitably vary across therapeutic indications, study phases, and regulatory jurisdictions. The framework outlined in this article provides a structured methodology for linking endpoint definitions to site capabilities, but it must be tailored by experienced clinicians, biostatisticians, and operational leads to fit the unique scientific questions of each protocol.

Ultimately, site feasibility is not an administrative hurdle to be cleared as quickly as possible. It is the first line of defense in protecting trial data integrity. By deriving questionnaire probes directly from the protocol's endpoint matrix, demanding verifiable audit artifacts, enforcing a disciplined three-tier triage, and feeding findings directly into the risk-based monitoring plan, sponsors transform site selection from an optimistic gamble into a rigorous, auditable science.