Clinical Trials in Practice: Evidence on Patient Outcomes and Site Integration

Clinical Trials in Practice: What the Evidence Says About Site Integration and Patient Outcomes

Randomized clinical trials are medicine's most reliable instrument for determining what actually works. That much is settled. What is far less settled, and what a landmark 2024 JAMA analysis called attention to, is the degree to which clinical trials remain structurally separated from everyday healthcare delivery. As Angus et al. observed in JAMA (July 2024) [1], "major issues with both the clinical trials enterprise and the lack of integration of clinical trials with healthcare delivery compromise medicine's ability to best serve society." The authors described this separation as a "house divided," where the mechanisms that generate clinical knowledge operate largely apart from the systems meant to apply it.

For practice leaders, site staff, and investigators weighing whether to participate in trials, that framing matters. The question is not whether clinical research produces useful knowledge. It does. The question is whether participating in that research is associated with measurable benefits for the patients in your practice, and what it takes to do it well.

This article draws on peer-reviewed evidence, regulatory guidance, and published industry data to address both questions.

Why the Gap Between Research and Clinical Care Has Consequences

The disconnect Angus and colleagues identified in JAMA [1] is not new, but its scale is worth naming precisely. Despite decades of randomized trial data accumulating in journals, a significant proportion of eligible patients never access trials at all. A widely cited participation figure examined and reanalyzed by Ken Getz of Tufts CSDD in 2023 indicates that recruitment of eligible patients into clinical trials has remained a persistent challenge for more than three decades [2].

That pressure runs in both directions. Practices that do not participate in trials forgo both the sponsor-budget reimbursement that funds trial activities and the operational expertise that comes from running GCP-regulated research. Sponsor budgets typically cover startup, per-visit and per-procedure fees, pharmacy handling, and closeout costs; whether that produces net revenue for a site depends on staffing, overhead, and negotiated rates. The GCP training, structured monitoring, delegation oversight, and quality-documentation disciplines that come from active trial conduct are difficult to build without running trials.

Drug development costs add further urgency to the efficiency question. The Tufts Center for the Study of Drug Development estimated in a 2014 analysis, based on data from 1995 to 2007, that the capitalized cost of developing a new approved drug stands at approximately $2.6 billion in 2013 dollars [3]. That figure, while widely cited, reflects a specific methodology and time period, and has been both challenged and updated by subsequent research; it should be understood as an order-of-magnitude benchmark rather than a precise current estimate. What is less contested is that delays at any trial phase compound costs significantly. A 2024 Tufts CSDD study drawing on 409 clinical trial budgets found that the direct daily cost of conducting a trial has changed substantially since the most commonly cited 1990s-era estimates [4].

For sites, the message is not that every practice should run trials to save pharmaceutical companies money. It is that the financial and operational infrastructure required to run trials has real value, and practices that build it early gain advantages that compound over time.

The Evidence on Research-Active Practices and Patient Outcomes

The question of whether patients in research-active institutions fare better has attracted considerable scientific scrutiny, and the findings are more instructive than a simple yes or no.

A population-based study published in Gut examined outcomes across more than 200,000 patients with colorectal cancer treated in NHS hospitals, comparing those at institutions with high rates of clinical trial participation against those with low or no participation. Patients treated in hospitals where at least 16% of patients were enrolled in interventional trials showed lower postoperative mortality and improved five-year survival [5]. Specifically, the study found that hospitals sustaining high trial participation for four or more consecutive years showed a 1.5-percentage-point lower postoperative mortality rate (from 6.5% to 5.0%) and a 3.8-percentage-point improvement in five-year survival rates (from 41.0% to 44.8%), compared with hospitals with zero participation [5]. The authors note that these associations were observed across NHS hospitals of varying sizes, not only academic centers, and that improvement in outcomes preceded and increased with the level of sustained participation. This is an observational association; the study did not establish causation.

A separate umbrella review examined outcomes for patients who personally enrolled in randomized controlled trials, drawing on six systematic reviews covering 380 comparisons across the included studies, of which 368 were analyzed by outcome category. Of those 368 analyzed comparisons, 69 (18.7%) showed statistically significant benefit for participants over non-participants, 264 (71.7%) showed no statistically significant difference, and 35 (9.5%) favored non-participation [6]. Critically, none of the six included systematic reviews concluded that participation was overall harmful, even though individual comparisons within those reviews showed outcomes favoring non-participants. The distinction matters: a review may contain specific disease-treatment combinations where non-participants fared better on one endpoint, without the broader review concluding a net harmful effect across its full scope. For patients with metastatic breast cancer, a retrospective cohort analysis of 806 patients at Samsung Medical Center found that clinical trial enrollment was an independent predictor of overall survival, with a hazard ratio of 0.75 (95% CI, 0.59-0.95) for death among enrolled patients compared with non-enrolled patients [7]. The authors noted that this association was not maintained in subgroup analysis of triple-negative breast cancer patients, and that selection effects from eligibility criteria may contribute to the observed survival difference.

The evidence is genuinely mixed, and careful readers will note that much of it reflects patient selection: trial participants tend to have better functional status and fewer comorbidities than the general population. A 2024 JAMA meta-analysis by Iskander and colleagues, examining whether trial participation itself generates survival benefit beyond the treatment received, found the participation effect remains unproven in the cancer context [8]. What is better established is the institutional association from the Gut study: hospitals with high sustained trial participation show measurably better outcomes for all their patients with colorectal cancer, not only those enrolled [5].

A related body of literature suggests that institutions actively engaged in research show greater adherence to clinical guidelines and more consistent application of evidence in routine care, though the mechanistic pathway for this remains debated [9].

Operational Implications for Sites and Practices

Integrating clinical trial activity into a medical practice is not simply a matter of signing on as a site. It requires changes to how a practice manages patient data, how it trains and retains staff, and how it tracks documentation.

Patient Identification and Eligibility Screening

For most practices, the largest barrier to trial participation is not willingness to conduct research, it is the systematic identification of eligible patients. Traditional approaches rely on clinicians recognizing a match during a consultation. That works inconsistently. A more structured approach uses electronic health record data to map patient characteristics against protocol eligibility criteria, surfacing candidates before a trial visit rather than during one.

FDA guidance on real-world evidence, developed in response to the 21st Century Cures Act (2016) [10], addresses the use of electronic health records and administrative data as potential sources of real-world data to support trial designs. FDA's framework for this purpose identifies EHR data as a category of real-world data that may be used in certain contexts to support regulatory decision-making [10]. Sites interested in using EHR-linked approaches for patient screening should assess their data infrastructure against applicable FDA guidance and the specific protocol's requirements before assuming compatibility.

FHIR-based standards for exchanging structured patient data have been incorporated into federal interoperability requirements through the ONC Cures Act Final Rule (2020) [19], which requires certified health IT systems to support HL7 FHIR Release 4 APIs. This creates a technical pathway for sites using certified EHR systems to query structured patient data against defined criteria. Implementation depth and data completeness vary across EHR platforms, and sites should validate any screening workflow against their specific system's FHIR endpoint capabilities and the protocol's eligibility requirements before deployment.

Staff Training and GCP Compliance

ICH E6(R3), issued as final FDA guidance in September 2025 [11], represents the most substantive revision to Good Clinical Practice standards since E6(R2) in 2016. Published by the International Council for Harmonisation, it replaces the process-based structure of E6(R2) with a principles-driven framework that explicitly accommodates decentralized trial elements, digital health technologies, and remote data collection [11]. For investigators, the guidance clarifies that delegation of trial-specific activities to other persons or parties is permissible, but oversight responsibility remains with the principal investigator [11]. As an ACRP commentary on E6(R3) noted in 2025, the guidelines reflect a shift toward more participant-centered approaches, including provisions for wearables, electronic consent, and remote monitoring [12].

Sites that intend to conduct trials under E6(R3) need training structures that reflect these expanded oversight responsibilities, not just historical GCP curricula built around E6(R2).

Protocol Amendments and Document Management

Approximately 60% of all trial protocols require amendments during a study, and about one-third of those amendments are described as avoidable, according to Tufts CSDD findings summarized in a HHS review of clinical trial cost barriers [13]. The primary source data for this estimate was collected from industry-sponsored trials across multiple therapeutic areas and is widely referenced in the clinical trial operations literature. Protocol amendments create documentation burden for sites: updated procedures must be implemented, staff retrained, and changes logged in the investigator site file (ISF), the site-maintained portion of essential trial documents. Sites that lack organized regulatory document workflows often struggle to manage amendment cycles without disruption to patient care.

Regulatory and Documentation Requirements for Investigator Sites

Principal investigators carry significant regulatory obligations under FDA regulations. For drug trials, 21 CFR Part 312 [14] establishes requirements for investigational new drug applications and the responsibilities of investigators participating in them. For device trials, 21 CFR Part 812 [15] governs investigational device exemptions and site conduct. These are distinct regulatory regimes with different requirements, and sites should not assume that experience in one automatically prepares them for the other.

Core investigator obligations include:

• Maintaining adequate facilities and qualified staff to conduct the trial
• Ensuring that informed consent is obtained and documented per 21 CFR Part 50 [16]
• Retaining trial records for the periods required under the applicable regulation (under 21 CFR 312.62, drug study records must be retained for two years following the date on which a marketing application is approved; if no application is filed or if an application is not approved, records must be retained for two years after the investigation is discontinued and FDA has been notified; device study record retention under 21 CFR 812.140 follows a different schedule tied to specific study events)
• Reporting adverse events according to the protocol and applicable regulatory requirements
• Permitting FDA inspection of trial records upon request

Under ICH E6(R3) [11], sites are expected to implement risk-proportionate oversight and controls proportionate to trial complexity. The guidance establishes a risk-based quality management framework, but it is important to note that the primary quality management system responsibility under E6(R3) rests with the sponsor. Investigators are responsible for site-level oversight of delegated activities, and for ensuring that any site-level service providers they engage operate within GCP expectations [11]. Conflating sponsor-level quality system obligations with investigator-level oversight responsibilities leads to misallocation of effort and potential compliance gaps.

The Declaration of Helsinki [17], adopted by the World Medical Association in 1964 and last revised in 2024, provides the ethical framework underlying most global clinical research regulation. It is an ethical standard, not a directly binding legal instrument in most jurisdictions. Helsinki and ICH GCP articulate participant-protection principles also reflected in FDA regulations. FDA has issued E6(R3) as guidance [11], while binding US obligations arise from regulations including 21 CFR Part 50 [16] and 21 CFR Part 56. The EMA similarly notes that protection of clinical trial subjects in the European Union is consistent with Declaration of Helsinki principles, implemented through Regulation (EU) 536/2014 [22]. Sites conducting FDA-regulated research should be aware that compliance with FDA's domestic regulations is the operative legal obligation; familiarity with Helsinki itself, while professionally valuable, does not substitute for regulatory compliance.

Technology and AI in Site Operations

Artificial intelligence in clinical trial site operations is most credibly applied to problems with specific, measurable throughput. Patient screening is one such problem: structured EHR queries against protocol eligibility criteria can identify potential candidates systematically, reducing dependence on individual clinician recognition during routine consultations.

McKinsey's 2025 analysis of clinical trial delivery trajectories identified predictive analytics for site selection and performance management as among the higher-impact operational applications available to sponsors [18]. Practices that generate and maintain structured, queryable data about their patient population are better positioned to respond efficiently to sponsor feasibility questionnaires. The ability to produce rapid patient-count estimates against eligibility criteria is a practical advantage in feasibility discussions, independent of whatever technology supports it.

Regulatory document generation is another area where AI-assisted tools have seen adoption. Protocol drafting, informed consent form preparation, and clinical study report narratives all involve substantial repetitive structure where templated AI generation is proposed to reduce initial drafting time, though workflow-specific validation data for these applications in GCP contexts remains limited in the published literature. Sites evaluating AI tools for trial operations should treat them as decision-support aids rather than autonomous decision-makers, and should require documented validation evidence appropriate to the workflow before deployment. For high-stakes functions such as protocol-deviation adjudication, safety signal assessment, and eligibility determination, the investigator's personal responsibility for the conduct and oversight of the trial under ICH E6(R3) [11] means that delegating the final judgment to an unvalidated system would constitute a direct compliance risk, not merely an operational preference.

Kitsa's KScreener product, described by the company as a FHIR-based patient matching and pre-screening tool, is designed to address the patient identification workflow gap at sites and practices participating in clinical trials (kitsa.ai). KScribe, Kitsa's AI-powered regulatory document generation platform, supports protocol, ICF, and related document drafting workflows (kitsa.ai/regulatory-document-generation).

What Sponsors and CROs Look for When Selecting Sites

Site selection decisions by sponsors and CROs are consequential for practices that want to grow their trial portfolio. Published research and industry analysis point consistently to a set of site characteristics that inform enrollment performance assessments, though these factors are typically evaluated together rather than ranked against each other.

Prior enrollment history in a given therapeutic area is consistently referenced in feasibility literature as a relevant predictor of future performance in that area. A 2024 Tufts CSDD study benchmarking site activation and patient enrollment across phase II and III studies found that enrollment achievement varied substantially across geographic regions and site types [20]. These findings support using site-specific historical performance data alongside regional benchmarks, rather than relying on regional averages alone, when evaluating site suitability. Patient population size and diagnostic mix matter independently: a cardiovascular specialty practice with a large panel of documented heart failure patients is inherently better positioned for heart failure trials than a general practice of similar overall size. A 2017 BMJ Open study evaluating a standardized, multi-step site selection method demonstrated a structured alternative to selection based primarily on reputation and informal judgment, using quantitative performance indicators and structured feasibility data to categorize sites before activation [21].

Site staff stability and regulatory compliance history also factor in sponsor feasibility reviews, particularly as risk-based monitoring under ICH E6(R3) [11] calibrates monitoring intensity based on site performance and quality indicators. These factors are interconnected. A site with strong enrollment history in oncology, stable coordinator staff, and clean inspection records across three consecutive trials is not merely the sum of those attributes: it is a site that has demonstrated organizational capacity, and that distinction carries weight with sponsors evaluating first-time site relationships.

Kitsa's KScout product, described by the company as a site selection and research site intelligence tool, is built to support sponsor-side feasibility analysis by providing structured views of site characteristics relevant to enrollment potential and operational readiness (kitsa.ai).

Key Takeaways

• Research-active hospitals show associations with better patient outcomes across entire patient populations, not only enrolled trial participants. An NHS-based population study found that hospitals sustaining high trial participation for four or more years were associated with a 3.8-percentage-point improvement in five-year colorectal cancer survival rates, compared with hospitals with no participation [5]. This is an observational finding, not proven causation.
• An umbrella review covering 368 analyzed comparisons across six systematic reviews found that 18.7% showed benefit associated with RCT participation, 71.7% showed no significant difference, and 9.5% favored non-participation. None of the included reviews identified a harmful effect [6].
• ICH E6(R3) [11] has updated investigator obligations and flexibility, including provisions for decentralized elements, remote monitoring, and digital health technology. Sponsor-level quality management system responsibilities are distinct from investigator-level site oversight obligations.
• Protocol amendments affect approximately 60% of trials [13], creating a persistent documentation burden that sites without organized regulatory workflows often struggle to absorb.
• FHIR-based API standards, required for certified EHR systems under the ONC Cures Act Final Rule [19], create a technical pathway for structured patient screening against eligibility criteria, though implementation depth varies by platform and requires validation against specific protocol requirements.
• Drug and device investigator record-retention obligations follow different regulatory schedules under 21 CFR 312.62 [14] and 21 CFR 812.140 [15] respectively; sites conducting both types of studies should track these separately.
• Sites with documented enrollment history, stable staff, and strong GCP compliance records consistently appear in sponsor feasibility criteria as indicators of organizational readiness for trial participation.

FAQ

References

1. Angus DC, Berry S, Lewis RJ, et al. "The Integration of Clinical Trials with the Practice of Medicine: Repairing a House Divided." JAMA. 2024;332(2):153-162. https://doi.org/10.1001/jama.2024.4088
2. Getz KA. "Rebooting the Statistic That 5% of Eligible Patients Participate in Clinical Trials." Applied Clinical Trials. Vol. 32, No. 3. March 2023. Applied Clinical Trials
3. DiMasi JA, Grabowski HG, Hansen RA. "Innovation in the Pharmaceutical Industry: New Estimates of R&D Costs." Journal of Health Economics. 2016;47:20-33. (2014 Tufts CSDD announcement: link)
4. Smith ZP, DiMasi JA, Getz KA. "New Estimates on the Cost of a Delay Day in Drug Development." Therapeutic Innovation & Regulatory Science. 2024;58(5):855-862. https://doi.org/10.1007/s43441-024-00667-w
5. Downing A, Morris EJ, Corrigan N, et al. "High Hospital Research Participation and Improved Colorectal Cancer Survival Outcomes: A Population-Based Study." Gut. 2017;66(1):89-96. PMC5256392. PMC5256392
6. Bouzalmate-Hajjaj A, Masso Guijarro P, Khan KS, Bueno-Cavanillas A, Cano-Ibanez N. "Benefits of Participation in Clinical Trials: An Umbrella Review." International Journal of Environmental Research and Public Health. 2022;19(22):15368. PMID: 36430100. https://doi.org/10.3390/ijerph192215368
7. Lee JY, Lim SH, Lee MY, Kim HS, Ahn JS, Im YH, Park YH. "The Impacts of Inclusion in Clinical Trials on Outcomes among Patients with Metastatic Breast Cancer (MBC)." PLOS ONE. 2016;11(2):e0149432. PMC4763476. https://doi.org/10.1371/journal.pone.0149432
8. Iskander R, Moyer H, Vigneault K, Mahmud SM, Kimmelman J. "Survival Benefit Associated With Participation in Clinical Trials of Anticancer Drugs: A Systematic Review and Meta-Analysis." JAMA. 2024;331(24):2105-2113. https://doi.org/10.1001/jama.2024.6281
9. Labad J, Monreal JA, Cabezas A, Safont G, Garcia-Portilla MP, Bobes J. "Quality of Care in Psychiatry Is Related to Research Activity." BMC Health Services Research. 2021;21(1):985. PMC8446072. PMC8446072
10. U.S. Congress. "21st Century Cures Act." Public Law 114-255. Signed December 13, 2016. FDA real-world evidence framework: FDA RWE
11. ICH E6(R3). "Good Clinical Practice (GCP) Guideline." International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use. Issued as final FDA guidance September 2025. FDA docket: FDA-2023-D-1955. FDA guidance
12. ACRP. "ICH E6(R3): Transforming the Future of Clinical Trials with Enhanced Efficiency, Safety, and Innovation: A Commentary." Association of Clinical Research Professionals. April 2025. ACRP commentary
13. U.S. Department of Health and Human Services, ASPE. "Examination of Clinical Trial Costs and Barriers for Drug Development." HHS/ASPE. ASPE report
14. U.S. Food and Drug Administration. "21 CFR Part 312: Investigational New Drug Application." Code of Federal Regulations, Title 21. eCFR
15. U.S. Food and Drug Administration. "21 CFR Part 812: Investigational Device Exemptions." Code of Federal Regulations, Title 21. eCFR
16. U.S. Food and Drug Administration. "21 CFR Part 50: Protection of Human Subjects." Code of Federal Regulations, Title 21. eCFR
17. World Medical Association. "Declaration of Helsinki: Ethical Principles for Medical Research Involving Human Subjects." Adopted June 1964; last revised October 2024. WMA
18. McKinsey & Company. "Achieving Best-in-Class Clinical Trial Delivery: The Road to 2035." McKinsey Life Sciences. October 2025. McKinsey
19. Office of the National Coordinator for Health Information Technology (ONC). "21st Century Cures Act: Interoperability, Information Blocking, and the ONC Health IT Certification Program Final Rule." Federal Register. May 1, 2020. ONC
20. Lamberti MJ, Dirks A, Kikuchi N, Patel Cervantes N, Getz KA. "Benchmarking Site Activation and Patient Enrollment." Therapeutic Innovation & Regulatory Science. 2024;58(4):696-703. https://doi.org/10.1007/s43441-024-00638-1
21. Hurtado-Chong A, Joeris A, Hess D, et al. "Improving Site Selection in Clinical Studies: A Standardised, Objective Multi-step Method and First Experience Results." BMJ Open. 2017;7(7):e014796. BMJ Open
22. European Medicines Agency. "Good Clinical Practice." EMA Human Regulatory Overview. EMA

Suggested Internal Links

• KScout: Site Selection and Research Site Intelligence
• KScreener: FHIR-Based Patient Pre-Screening
• KScribe: AI Regulatory Document Generation
• Blog: ICH E6(R3) GCP Compliance Checklist for Sites
• Blog: What Is a Clinical Trial Protocol?
• Blog: How AI Is Changing Clinical Trial Site Operations