Clinical trials in the mid-2020s are experiencing another stage of technological renewal and a deeper restructuring of the logic underlying drug development and trial conduct. Contemporary reviews increasingly emphasize that research success is determined not only by scientific novelty but also by the industry's capacity to responsibly integrate new approaches into routine operational activities. This means that the subject of discussion now encompasses both new tools and the maturity of all stakeholders involved in the organization of clinical trials [1, 3–8].

Such changes were already foreshadowed earlier. A 2023 publication already highlighted the growth in the number of trials, the increasing significance of real-world data, technological innovations, patient-centeredness, and rising study complexity. However, by 2026, these directions have come to be perceived not as general forecasts but as practical requirements for contemporary research [13].

According to the international service Statista, as of May 29, 2023, approximately 454,000 clinical trials across various phases were being conducted worldwide. By the end of 2025, this number exceeded 473,000, and according to WCG Clinical estimates, further growth is expected in 2026, particularly in oncology, neuroscience, and metabolic disorders [1, 7, 13]. For many decades, the largest number of studies have been aimed at finding treatments for cancer (112,899 studies in 2023), followed by neuropsychiatry (87,346), and cardiovascular diseases (63,757) [13]. Vaccines demonstrate the highest success rate (33.4%), whereas in oncology, only 3.4% of investigated molecules reach approval [13].

However, the growth in the number of studies is accompanied by unprecedented protocol complexity. According to data from the Tufts Center for the Study of Drug Development, since 2015, the number of mandatory procedures in Phase III studies has increased by 42%, the number of endpoints by 37%, and nearly one-third of collected data relates to procedures that do not support the primary endpoints [1]. Concurrently, 45% of research sites indicate that persistent operational problems affect their ability to participate in new studies [1].

It is precisely in this context that digital solutions, artificial intelligence (AI), hybrid models, and patient-centered approaches cease to be optional and become necessary elements of a sustainable research ecosystem. This review is devoted to the analysis of seven key trends shaping the development of clinical trials in 2026.

By 2026, artificial intelligence and machine learning have established themselves as actively implemented elements of the clinical research operational model. The industry has moved to a stage of full-scale integration of algorithms into planning, conduct, and analysis processes. The use of intelligent systems ensures effective problem-solving under conditions of increasing protocol complexity and growing data volumes, enabling rapid decision-making and accurate prediction of study progress [1, 2, 3, 5, 8].

The application of AI encompasses a wide range of practical tasks, from identifying new therapeutic targets to the strategic selection of countries and research sites. Algorithms conduct deep analysis of historical data, evaluate demographic factors, and assess participant dropout risks. The implementation of such solutions significantly minimizes manual data entry, reduces the likelihood of errors, and optimizes the workload on research site personnel [1, 4, 5, 8].

Predictive modeling and site selection. One of the most impressive applications of AI is predictive modeling for the selection of countries and research sites. As noted in the 2026 WCG report, the ability to synthesize large volumes of data, guided by specific intent, allows the use of years of experience and accumulated data for highly specific needs. By analyzing historical clinical trial data, evaluating similar protocol designs and their outcomes, and integrating demographic trends with disease prevalence, predictive algorithms can identify regions and site profiles most likely to attract the optimal participant population [1].

This approach not only improves recruitment accuracy but also enables the development of strategies that account for the expected burden on sites and participants, thereby contributing to reduced dropout rates. Similar conclusions are found in the Veeva Systems report (2026), which emphasizes that automated orchestration using AI allows data managers to transition from the role of "air traffic controllers" to a more strategic role, with 57% of data management professionals expecting that risk-based data management (RBDM) will influence the evolution of their role within the next two years [8].

Document workflow automation and ethical review. Russian legislation and regulatory frameworks in this area are also actively developing, establishing an ethical and legal foundation for innovation. A key reference point has been the "Code of Ethics for the Application of Artificial Intelligence in Healthcare" (current version 2.1, 2025), developed with the participation of the Russian Ministry of Health [15]. The document establishes strict requirements for transparency, explainability, and safety of algorithms at all stages of their lifecycle. The Code directly indicates the need for ethical review of AI systems and mandatory retention of control by the medical community, making it fundamental for conducting clinical trials in Russia [15].

A practical example of the implementation of these principles is the "LEC-AI" solution, developed by the N.N. Petrov National Medical Research Center of Oncology in collaboration with technology partners (Yandex Cloud). This system, based on a large language model (YandexGPT), automates the verification of document packages submitted to the local ethics committee (LEC). The tool conducts a preliminary audit of protocols and informed consent forms within minutes, identifying non-compliance with ethical standards. This not only accelerates the study start-up phase but also ensures compliance with the Code's requirements for the quality of ethical review and minimization of bias risks [1, 15].

Human-in-the-loop principle and regulatory standards. The fundamental element of AI architecture remains the human-in-the-loop principle. According to the Ministry of Health Code of Ethics and international standards, expert oversight is mandatory at critical decision-making stages [10, 15]. The specialist verifies system outputs, ensuring result interpretability, compliance with ethical norms, and retention of ultimate human responsibility. This interaction complements the physician's professional expertise with AI capabilities, without excluding the human factor from the data validation process.

Contemporary standards for working with AI are also regulated by the "Guiding Principles of Good AI Practice," published by the FDA and EMA in January 2026. These 10 principles (including human-centeredness by design, risk-based approach, compliance with standards, clear context of use, multidisciplinary expertise, data management, best practices in model development, risk-based performance evaluation, lifecycle management, and presentation of understandable information) correlate with ethical standards, creating a unified space of trust in technology [9, 10].

The development of the regulatory environment in 2026 is characterized by a transition to risk-based oversight: regulators determine the depth of control depending on the degree of AI's impact on patient safety and data reliability. The formation of this regulatory framework, including the Russian Code of Ethics, transforms AI into a source of new standardized quality requirements. Regulatory transparency ensures the sustainability of digital transformation processes and guarantees the ethical integrity of interactions with research participants in the global market [13, 15].

Digital twins. One of the most promising areas of AI application is patient digital twins. This technology, trained on extensive retrospective data, enables the creation of "virtual avatars" for each participant, allowing for reliable assessment of therapy effectiveness by comparing the actual effect with the predicted outcome of the digital twin, thereby eliminating the need for extensive placebo groups. This approach, known as TwinRCTs, has already received preliminary approval from the European Medicines Agency (EMA) for use as a control in late-phase clinical trials.

However, widespread implementation of digital twins is associated with challenges: the quality and representativeness of source data, algorithm interpretability (the "black box" problem), and the absence of unified validation standards. Overcoming these barriers requires the development of explainable AI (XAI) methods and the creation of extensive, unbiased training datasets.

Another notable trend is the proliferation of hybrid and decentralized clinical trials (DCTs). By 2025–2026, such models have ceased to be a rare solution and are increasingly perceived as an expected element of contemporary research practice, particularly in situations where frequent site visits are difficult due to distance, participant health status, high procedure frequency, or limited research site resources. The sources we reviewed describe models in which pre-screening, electronic consent acquisition, and subsequent follow-up are organized remotely [1, 2, 5, 7].

The practical meaning of hybrid models lies in the redistribution of individual procedures between in-person and remote formats. However, this does not imply a complete replacement of the research site with digital tools. The site retains a key role in medical assessment, safety monitoring, primary documentation management, and data quality assurance. Remote elements are used where they can enhance protocol feasibility without compromising scientific validity and regulatory traceability.

Advantages and challenges of decentralized trials. The development of decentralized trials is associated with the use of validated remote assessment methods and a more flexible approach to data in certain types of studies. This is particularly important for rare diseases, highly personalized therapies, and areas where in-person pre-screening is too costly or results in a high proportion of failed enrollments. Experts emphasize that this format does not waive data quality requirements: remote procedures must be pre-specified, standardized, validated when necessary, and the results obtained must be comparable to in-person visit data [2, 5, 7].

At the same time, more rigorous risk assessment is intensifying. Patient-centered and decentralized models may reduce participant burden but create new vulnerabilities: remote consent errors, technological failures, inconsistent data collection, delays in recognizing safety signals, and difficulties in confirming the validity of remote procedures. Therefore, contemporary logic consists not of maximizing the expansion of remote elements but of their selective implementation based on risk assessment, procedure significance, and impact on participant safety and data quality [5, 7].

Site readiness as a success factor. Of particular importance is the question of technological and operational readiness of research sites. The hybrid model requires a stable digital infrastructure, clear distribution of responsibilities, staff training, participant support, and consistency between systems. In 2025–2026 sources, site readiness is increasingly viewed as a study sustainability factor, as growing protocol complexity, number of procedures, and data volume increase the burden on the research environment [1, 4, 6, 8]. Therefore, the implementation of decentralized trial elements should be assessed not by the mere presence of digital solutions but by the site's ability to stably use them within a GCP-compliant process.

Hybrid studies also change data management requirements. With distributed data collection, the importance of unified data capture rules, completeness control, timeliness of transmission, provenance sources, and subsequent verification increases. This is particularly important when using digital biomarkers, wearable devices, remote monitoring, and real-world data. Experts note that participant follow-up becomes more continuous, and data collection becomes more distributed over time and less tied to individual visits [1, 2, 5, 8].

Thus, hybrid and decentralized elements are increasingly considered an expected part of a modern clinical trial, rather than an exception or additional option. Their significance lies in improving study accessibility and reducing burden on participants and sites without weakening data quality, safety, and GCP control requirements [1, 2, 7, 13].

Contemporary clinical trials are increasingly relying on data collected outside traditional in-person visits to the research site. The importance of digital biomarkers, wearable devices, and continuous monitoring models is growing, particularly in neurology, cardiology, pulmonology, and metabolic disorder research [2, 5].

Continuous monitoring refers to the regular or constant collection of data on participant status between visits to the research site. Such data may come from wearable devices, mobile applications, home medical devices, or other digital tools. Unlike the classical model, where participant status is assessed on pre-specified visit dates, continuous monitoring allows for capturing the dynamics of indicators in everyday life. This makes data collection more distributed over time and less dependent on a limited number of control points [2, 5].

Real-world data (RWD) and tokenization. The use of real-world data (RWD) is being increasingly implemented. Such data includes information from electronic medical records, registries, insurance databases, laboratory systems, pharmacy dispensing data, medical devices, and other sources reflecting healthcare delivery in routine practice. In clinical trials, RWD are increasingly used at the planning stage, for protocol feasibility assessment, identification of suitable populations, follow-up, and interpretation of result applicability to real-world medical practice [1, 2, 4, 7, 8].

Special attention is devoted to the tokenization method—a way of linking a participant's medical records to a de-identified identifier without revealing their identity. This model allows for matching data from different sources and conducting long-term follow-up, particularly in complex studies including gene therapy. According to the WCG Clinical review, among approximately 3,000 new protocols reviewed between May 2024 and May 2025, tokenization was mentioned in 21 cases. No cases of deferral or denial of protocol approval for reasons related to tokenization were recorded. These data indicate that when procedures are correctly described in the informed consent and confidentiality requirements are observed, tokenization can be used without significant objections from the ethics committee [1].

Risk-based data management. The implementation of the risk-based approach in clinical trial conduct also continues. Each year, we observe an increasing transition from the methodology level to routine operational practice. The Veeva Systems report explicitly states that the theory of risk-based data management (RBDM) must become a reality, with 57% of data management professionals expecting RBDM to influence the development of their role within the next two years [8]. Against the backdrop of growing data volume and diversity, the traditional model of comprehensive manual verification is becoming increasingly unsustainable, and focus is shifting toward early signal detection, centralized analytics, prioritization of critical data, and the connection between risk management, data quality, and current operational decisions [8].

Thus, by 2026, RWD, digital biomarkers, and continuous monitoring are becoming part of a broader evidence system in clinical trials. Their significance is associated with the ability to better assess the applicability of results to real-world practice, support long-term participant follow-up, and reduce excessive burden on participants and research sites. At the same time, the expansion of such approaches requires clear rules for data provenance, quality control, confidentiality protection, and preservation of the connection between the participant's digital model and the real clinical context [1, 2, 5, 8, 13].

One of the significant organizational trends is the growing importance of research center readiness to participate in clinical trials. Currently, site readiness is increasingly understood not as formal readiness to start but as the center's ability to stably execute the protocol under conditions of high procedural complexity, staffing burden, growing data volume, and expansion of digital tools [1, 4, 6, 8].

According to the 2026 WCG Clinical report, 45% of research sites indicate that persistent operational problems affect their ability to participate in new studies. Such problems include complex protocols, budget cuts, the variety of software required to conduct the study, administrative burden, and delays [1]. WCG separately notes that 35% of sites cite the complexity of clinical trials as their main problem [1].

Protocol complexity growth and burden on research sites. In Phase III studies, since 2015, the number of mandatory procedures has increased by 42%, the number of endpoints by 37%, and nearly one-third of collected data relates to procedures that do not support the primary endpoints [1]. Additional burden is created by the fragmentation of the digital environment. Sites must work with electronic databases, medical information systems, electronic questionnaires, remote monitoring systems, electronic consent solutions, laboratory portals, and other tools. According to WCG Clinical data, 79% of research sites indicate that when using digital tools in clinical trials, it is site staff, rather than the sponsor, CRO, or technology provider, who most often become the first line of support for patients with technical problems [1].

Protocol amendments and operational sustainability. A significant indicator of vulnerability remains protocol amendments: Phase III studies now have on average 3.5 substantial amendments compared to 2.3 ten years ago [1]. Each such amendment requires document revision, staff retraining, informed consent updates, and additional communications with participants. According to WCG data, 47% of sites consider protocol simplification critically important for operational viability [1].

Site readiness is increasingly assessed through staffing resources, staff training, role distribution, quality of internal procedures, ability to work with digital systems, and resilience to protocol changes. For hybrid and decentralized elements, biobanking logistics, remote procedures, participant support, timely data capture, and deviation control are additionally important.

Ways to improve site readiness. As WCG experts note, one of the key directions is shifting operational risk assessment to the early stages of study design. In addition, process standardization, implementation of integrated platforms (CTMS, EDC, eCOA), technical support provision, and staff training are important. The need to reduce "white space"—delays between start-up stages associated with excessive approvals and manual data transfers—is separately emphasized [1].

Thus, organizational readiness of the research site is becoming one of the factors in clinical trial sustainability. It is determined by the site's ability to withstand protocol complexity, work with digital systems, maintain data quality, and adapt to changes without losing study manageability [1, 4, 6, 8].

One of the key recent trends is the transition from the classical patient-centered approach to a partnership model with the study participant. If previously patient-centeredness was more often described as a general principle of quality research, in 2025–2026 materials it is increasingly presented as a practical organizational task: reducing participant burden, improving communication, retention, and involvement in study design [1, 5–7, 13].

In the 2026 WCG report, this shift is described as a transition to studies in which the participant becomes an active source of feedback during study design. This approach affects recruitment speed, participant retention, and data quality. Therefore, the participant experience becomes an independent factor in the operational sustainability of the study [1, 4–7].

Practical measures to improve participant experience. Practically, this is expressed in more understandable informational materials, flexible visit schedules, the possibility of remote participation, the use of digital tools, transportation and organizational support, and more attentive work with participant retention in the study. Contract Pharma magazine separately emphasizes that the growth of obesity research increases the importance of designs that take into account participants' motivational and logistical barriers [4]. Cuttsy and Cuttsy Ltd. describes the same trend through the transition from a protocol-oriented approach to a model that takes into account the real-life circumstances of the study participant [6].

An important part of the new model is the recognition of the participant as an active partner in the research process. This implies earlier involvement of participants and patient organizations in discussion of study design, participant materials, and communication strategy. This approach helps to better account for language, expectations, concerns, and real limitations of participants, as well as increase trust in the study [1, 5, 6, 7].

Compensation for participant time and effort. Particular attention in 2026 is devoted to the issue of fair compensation for participants' time and effort. WCG analysis showed that 95% of informed consent forms do not offer compensation beyond direct expenses (transportation, parking). At the same time, analysis of nearly 25,000 ethics committee records revealed only 77 cases where payment issues were brought up for discussion, and in no case did the committee consider the compensation amount too high. On the contrary, significant variability was identified in participant payment for the same study across different sites (standard deviation from $9 to $170), raising questions of fairness and ethics [1].

Human context versus digital model. This trend is shared by experts in Russia as well. Thus, I.V. Karnaukhov, in his presentation at the Stereotax8 conference, adds an important limitation to the digital approach. He emphasizes that the digital patient model is constructed from laboratory, instrumental, molecular genetic, clinical, and questionnaire data; however, it does not fully reflect the fear, pain, fatigue, hope, family circumstances, and personal motivation of the individual [16]. Therefore, partnership with the participant requires maintaining a balance between data, protocol requirements, and the patient's real-life situation.

Of particular importance is the role of the research team. According to Karnaukhov's conclusion, it is the investigator and the site team that become the "bridge" between protocol requirements and the participant's real situation [16]. This is especially important in complex and lengthy studies where compliance is influenced by medical parameters, emotional state, family support, trust in the physician, and understanding of the meaning of participation.

Digital tools should enhance human interaction and maintain direct participant contact with the research team. WCG emphasizes the importance of trust, empathy, regular communication, and live contact with the research team [1]. A similar conclusion is found in Karnaukhov's report: digital data and algorithms cannot fully replace the physician's clinical judgment and direct contact with the patient [16]. Consequently, the development of digital solutions should support communication between the participant and the research team, preserving the human context throughout the participant's journey: from initial information to retention, support, completion of participation, and recognition of the individual's contribution to the acquisition of new medical knowledge [1, 5, 6, 7, 13, 16].

Another important direction is the development of advanced therapies, primarily cell and gene therapy. These approaches are gradually moving beyond the narrow experimental agenda and forming a separate class of studies with increased requirements for infrastructure, safety, logistics, and regulatory support. It is these requirements that may become practical constraints for conducting studies [1, 2, 5, 7].

According to WCG ClinSphere, as of the end of 2025, there were more than 3,600 active and planned studies in the field of cell and gene therapies (CGT) worldwide, with 66% in oncology [1]. In the United States, 46 CGT products have been approved, classified as umbilical cord blood derivatives, CAR-T therapies, other gene therapies, and other cell therapies [1].

Unique logistical and manufacturing requirements. Cell and gene products often require individualized manufacturing, strict control of chain of identity and chain of custody, special transportation conditions, and precise coordination between the research site, laboratory, manufacturing facility, and sponsor. For such studies, an error in labeling, storage, transportation, or material transfer can have direct implications for participant safety and result interpretation [1, 2, 5].

Advanced therapies may be accompanied by specific risks, including delayed adverse events, immunological reactions, risks associated with genetic intervention, and the need for long-term monitoring after product administration. Therefore, such studies require pre-specified monitoring procedures, response pathways for complications, and a robust follow-up system [1, 5, 7].

Staffing and infrastructure requirements. Conducting cell and gene therapy studies requires specialists familiar with the specifics of participant selection, biomaterial handling, biosafety, working with genetic data, specific information provision, and complication monitoring. In some cases, multidisciplinary team participation is required: clinicians, laboratory specialists, pharmacists, geneticists, quality specialists, and logistics and regulatory support experts [1, 2, 5].

High manufacturing costs, complex logistics, limited numbers of suitable participants, long-term follow-up, and increased infrastructure requirements make such programs financially and organizationally sensitive. Therefore, errors at the planning stage, site selection, or protocol feasibility assessment can lead to significant delays and cost increases [1, 2, 5].

Regulatory innovations. For gene and cell therapy, more flexible and modular clinical trial designs are relevant. This is associated with the high uncertainty of early data and the personalized nature of many programs. Such studies may involve refinement of selection criteria, dosing regimen, monitoring parameters, and individual endpoints. Such flexibility is permissible only with pre-specified rules, scientific justification, ethical acceptability, and regulatory traceability [7].

Thus, the development of advanced therapies shifts the focus from discussing scientific novelty to the question of the readiness of the entire research system. The main challenge is the ability to ensure safe manufacturing, correct logistics, qualified conduct, long-term follow-up, and sustainable quality control. Therefore, cell, gene, and other advanced therapies are becoming a test of the maturity of clinical research infrastructure [1, 2, 5, 7, 13].

For the Russian context, electronic informed consent is of particular importance [14]. It cannot yet be considered a definitively established standard for all types of clinical trials; however, there are already grounds to speak of it as an emerging direction of development. This is associated with amendments to Article 43 of Federal Law No. 61-FZ "On the Circulation of Medicines," which establish the possibility of signing informed consent in electronic form and describe the types of electronic signatures used [14].

Thus, the electronic form of the patient information sheet and voluntary consent receives direct legal recognition in industry legislation. The significance of this change lies in the emergence of an independent legal basis for the application of electronic consent in clinical trials. Further development of this direction will depend on technological implementation, law enforcement practice, site readiness, and the ability of study organizers to ensure an understandable, safe, and verifiable consent process [14].

Improving clarity and accessibility. An important practical emphasis is contained in Karnaukhov's report at Stereotax8. The author emphasizes that the electronic form can improve the clarity of information provision through more flexible presentation of materials: video, infographics, interactive explanations, and other formats that help the participant better understand the content of the study [16]. At the same time, the digital format does not remove the key requirement for live contact with the research team: the participant must have the opportunity to ask questions, receive answers, and make a decision with an understanding of the essence of the study [16].

Consequently, electronic informed consent should be considered not as a simple replacement of a paper document, but as part of a broader process of interaction with the participant. Its value is determined by convenient access to materials, document version transparency, recording of the fact of review, the possibility of remote interaction, and preservation of action traceability. At the same time, process quality depends on the extent to which the electronic system supports communication between the participant and the research team, and also takes into account digital literacy, age, health status, and individual limitations of the patient [16].

Russian eConsent platforms. In this context, the development of a Russian eConsent platform can be considered as an example of a solution focused on the study participant and the research team. The platform is designed for convenient document exchange between investigators and patients, provision of materials in digital format, support for remote scenarios, recording of the review process, document version management, and preservation of audit traceability. With correct implementation, such solutions can increase the transparency of the consent process and reduce the administrative burden on the research site.

There are also other technological solutions in this area. For example, the DM365 team is developing the MainEDC ecosystem, which includes the capability for electronic signing of informed consents. This experience is important as confirmation that electronic consent in Russia is developing at the level of applied platforms for clinical trials. It can be separately noted that the experience of remote and digital solutions accumulated particularly actively during the COVID-19 period, when the need to continue studies and reduce in-person contacts accelerated interest in virtual and hybrid models.

Prospects for eConsent development in the Russian Federation. Thus, the legal and technological framework for the application of electronic informed consent in Russia has already been established. Further dissemination of this approach will depend on practical decisions by sponsors and contract research organizations. The applicability of the electronic form in a specific protocol, the readiness of the research team to use digital tools for working with patients, economic feasibility for the sponsor, and the possible establishment of an obligation to offer the patient an electronic version alongside the paper version will all be significant [14, 16].

The conducted analysis allows us to formulate several practical conclusions for the Russian clinical research system, including recommendations for investigators, sponsors, and regulatory authorities.

By 2026, the development of clinical trials is determined by a combination of technological, organizational, and patient-oriented changes. Artificial intelligence, hybrid models, real-world data, digital biomarkers, risk-based data management, and advanced therapies are forming new requirements for study design, conduct, and quality control [1, 2, 4–8, 13].

The main conclusion is the changing status of these directions. They are increasingly considered as practical conditions for a sustainable study: feasible for the site, understandable for the participant, manageable for the sponsor, and acceptable from a regulatory requirements perspective. Therefore, the quality of a modern clinical trial is determined by the balance between scientific validity, operational feasibility, digital traceability, participant safety, and the preservation of the human context of interaction.

For the Russian Federation, the key immediate steps are: regulatory consolidation of rules for conducting decentralized clinical trials, widespread implementation of eConsent using domestic platforms, monitoring of AI Code of Ethics application, and mandatory inclusion of a "Limitations" section in scientific publications to enhance transparency and compliance with international standards.