The Resolution of the Congress of Clinical Pharmacologists, held on May 21–22, 2026 in Moscow within the framework of the IV Russian Congress "Pharmacotherapy Safety 360°: Noli nocere!", defines the priority objectives of the professional community for the period up to 2030. The document specifies the pathways for implementing the Healthcare Development Strategy of the Russian Federation (Presidential Decree No. 896 of 08.12.2025) by strengthening the role of clinical pharmacologists in ensuring effective, safe, and cost-effective pharmacotherapy, introducing domestic medicines, pharmacovigilance, curbing antimicrobial resistance, personalized medicine, and combating polypharmacy. The Resolution formulates 19 specific items with designated responsible parties and deadlines, covering the development of a specialty strategy, updating the regulatory framework, integrating clinical pharmacologist consultations into the clinical-statistical groups system, participation in working groups on clinical guidelines, establishing a national research center, advancing educational programs (undergraduate, residency, continuing professional education), organizing formulary committees, as well as activities for career guidance, accreditation, and enhancing the prestige of the specialty. Special emphasis is placed on integrating clinical pharmacological technologies (pharmacogenetic testing, therapeutic drug monitoring, deprescribing) into clinical guidelines and preparing for the 30th anniversary of the specialty in 2027.

Introduction. Every year, clinicians and researchers face a growing volume of scientific publications, necessitating a critical appraisal of the quality and validity of clinical data. On one hand, the pharmaceutical industry demands accelerated clinical trial timelines; on the other hand, it cannot accept reduced accuracy and reliability of the results. Striking a balance between speed and quality has become a key challenge in modern evidence-based medicine.

Objective. Based on the European Medicines Agency (EMA) document “Clinical Evidence 2030”, to systematize and analyze six key principles that enable high-quality, methodologically rigorous and clinically relevant design and conduct of clinical trials.

Main points. The article sequentially discusses the following principles: patient-centered trial design; leveraging existing data to identify knowledge gaps; clear formulation of research questions (including the use of the PICO framework); rational selection of data sources and methods (including real-world data and artificial intelligence); early alignment of approaches with all stakeholders; and ensuring transparency and openness as the foundation of public trust. Each principle is illustrated with practical examples and references to current methodological standards (CONSORT, ICH guidelines). 

Conclusion. Consistent application of the six principles improves the quality, efficiency and reproducibility of clinical research, while strengthening trust in research results among the scientific community, regulators and patients. This article may serve as a practical guide for researchers, clinical pharmacologists and healthcare organizers when planning projects of varying complexity.

Objective. To develop and validate a hybrid biostatistical methodology for generating Real-World Evidence (RWE) to assess the efficacy and safety of food supplements.

Methods. The proposed framework integrates causal inference techniques (Inverse Probability of Treatment Weighting, IPTW) with ensemble machine learning methods (CatBoost, Random Forest) and Multiple Imputation by Chained Equations (MICE). Validation was performed using two pilot non-interventional studies (N=249) and a high-dimensional synthetic dataset (N=6000) with missingness up to 83.33%. Multi-objective Pareto optimization was applied for benefit–risk assessment. 

Results. IPTW achieved a standardized mean difference (SMD) <0.10 across all baseline covariates. The efficacy regression model attained an R² of 0.3596. The CatBoost-based safety classification model, after threshold optimization, achieved a PRfor Magnesium 400 mg). Pareto optimization provided an objective comparison of supplement profiles without subjective AUC of 0.123 and an F1-score of 0.218, enabling the detection of latent adverse event signals (e.g., 12.47% indicator weighting.

Conclusion. The developed methodology improves the consistency and reliability of post-market supplement evaluation using RWD and can support regulatory activities of organizations such as Food and Drug Administration, European Food Safety Authority, and Roszdravnadzor RF.

Objective. To compare the effectiveness of achieving the target 24-hour area under the concentration-time curve (AUC₂₄) and the safety of two vancomycin dosing regimens in patients with bone and joint infections.

Materials and methods. This prospective observational study included 20 female patients (35–65 years old) with chronic osteomyelitis, implant-associated infection, or septic arthritis. Patients were distributed into two groups (n=10): Group 1 received a loading dose of 1500 mg, followed by 750 mg every 8 hours (daily dose 2250 mg); Group 2 received a loading dose of 2000 mg, followed by 1000 mg every 12 hours (daily dose 2000 mg). On day 3, peak (V₁) and trough (V₂) serum vancomycin concentrations were measured, and AUC₂₄ was calculated. The target range was 400–600 h·μg/mL. Safety was assessed based on serum creatinine dynamics, creatinine clearance, and adverse events.

Results. The groups were comparable in age, anthropometric parameters, baseline creatinine, and creatinine clearance (p > 0.05). Median AUC₂₄ was 698.8 (566.9–755.6) h·μg/mL in Group 1 and 636.9 (432.5–749.1) h·μg/mL in Group 2 (p = 0.684). Target AUC₂₄ was achieved in 2 (20 %) patients in Group 1 and 3 (30 %) in Group 2; excessive levels (>600) were observed in 7 (70 %) and 5 (50 %), respectively (p = 0.7). Trough concentration was significantly higher in Group 1 than in Group 2 (12.8 vs. 5.3 μg/mL, p = 0.002). One case of vancomycin-induced nephropathy was recorded in Group 1 (AUC₂₄ 1028 h·μg/mL, decline in creatinine clearance >50 %). A moderate correlation was found between vancomycin trough concentration and serum creatinine level on day 7 (r = 0.458, p = 0.042).

Conclusion. Both vancomycin dosing regimens (750 mg every 8 hours and 1000 mg every 12 hours) show comparable rates of achieving target AUC₂₄ in patients with orthopedic infections. The more intensive regimen (with a total daily dose of 2250 mg) is associated with higher trough concentrations and a trend toward exceeding the target range, potentially increasing the risk of nephrotoxicity. Monitoring of trough concentrations and AUC₂₄ calculation are necessary for therapy personalization, especially in patients with additional risk factors (blood loss, elderly age).

Background. Healthcare evidence generation is shifting beyond randomized controlled trials (RCTs) to embrace real-world data (RWD). Collected from diverse routine care settings, RWD underpins real-world evidence (RWE), offering insights into patient outcomes, treatment effectiveness, and healthcare delivery. While RWE enhances generalizability and cost-effectiveness compared with RCTs, its inherent complexities such as data quality issues, missingness, confounding, and selection bias demand rigorous analytical approaches.

Methods. This critical review synthesizes the spectrum of analytical methodologies applied to RWD, ranging from traditional statistical techniques and causal inference frameworks to machine learning and advanced methods including natural language processing (NLP), Bayesian modeling, and network analysis. Each approach is appraised in terms of strengths, limitations, and suitability for addressing the unique challenges of RWD.

Results. Traditional statistical models provide interpretability and control for observed confounding but are limited by strong assumptions and vulnerability to unmeasured confounders. Causal inference methods, such as instrumental variables and target trial emulation, strengthen causal interpretation but require strong assumptions and specialized expertise. Machine learning approaches excel in prediction and high-dimensional data analysis but face interpretability and generalizability challenges. Advanced methods, including NLP and Bayesian models, extend analytical capacity but demand significant expertise and computational resources. Across categories, triangulation of multiple methods enhances robustness and credibility.

Conclusions. RWE plays a critical role in drug development, post-market surveillance, comparative effectiveness research, health economics and outcomes research, and personalized medicine. Future progress hinges on emerging innovations such as federated learning, synthetic data generation, and explainable AI (XAI), alongside advances in data integration, harmonization, and reproducibility. Methodological transparency, rigorous validation, and interdisciplinary collaboration are essential to ensure that RWE delivers trustworthy, ethical, and clinically actionable insights capable of informing regulatory decisions and optimizing healthcare globally.

Introduction. Artificial intelligence (AI) has undergone rapid development in pharmacovigilance (PV), evolving from experimental application to being considered a key tool in daily practice. Relatively simple AI models, including statistical signal detection methods, have been used in PV for decades, while recent advances in Natural Language Processing (NLP) have significantly expanded the scope of potential applications.

Objective. This article provides a critical analysis of the potential of NLP systems to optimize routine PV tasks, taking into account data protection requirements. The application of semantic search AI models based on alternative architectural approaches, specifically embedding models and retrieval-augmented generation (RAG), is examined separately. 

Main points. The authors distinguish between processes that do not involve personal data and allow the use of open AI solutions (searching and systematizing scientific literature, generating publication summaries), and processes involving the handling of confidential information (e. g., data extraction from Individual Case Safety Reports (ICSRs), automated generation of clinical case descriptions, benefit-risk analysis, and compliance with regulatory requirements and reporting standards), which require the use of corporate AI systems deployed within a secure infrastructure. The article also discusses limitations, risks, practical implementation aspects, as well as issues of ensuring the reliability, reproducibility, and transparency of the solutions used.

Conclusion. AI models, particularly NLP models, have significant potential for integration into routine PV processes. However, successful integration of AI models is impossible without a systematic approach to managing associated risks. Subject to these conditions, AI can become a robust tool for increasing the efficiency of PV processes.

Actuality. Respiratory syncytial viral (RSV) is a most common reason for acute respiratory infections in infants and children. Prevalence of non-severe RSV infection forms in kids does not mean absent it's influence for country economics and society at all. Severe cases of RSV infections could require hospitalization, that increases budget impact.

Objective. Social-economic burden of RSV infection evaluation in children first two years of life in Russian Federation basing on real data of epidemiologic monitoring, experts’ opinion for patients’ separation in depends on severity of illness and modelling of expenditures.

Materials and methods. Data of non-personal base of traditional flu and acute respiratory infections laboratory monitoring of National institute of flu named after Smorodintsev (seasons 2022–2023, 2023–2024, 2024–2025 yy) were included into retrospective analysis for burden modelling for 1 year (season). Extrapolation of monitoring results of proven laboratory indicators of RSV infection prevalence in cities of detection (minimum 5,8 %, maximum 12,3 % out of prevalence of acute respiratory viral infections), were made on all kids 0–2 y. o. population in the country. Minimal calculated amount of RSV infection was 245 723, maximal 521 101 per season. Four models of patients based on illness severity and necessity in hospitalization, including ICU, complications have been defined by medical experts. Direct medical (ambulance, outand in-patient treatment) indirect medical (cost of parental seek live) as well as indirect costs (GDP loss in case of parental seek live and cost of life years loss due to premature death of child) were calculated. Horizon of analysis one epidemic season, and long-term economic consequences in case of death during hospitalization (discount rate 3.5 %).

Results. Total direct medical expenditures in hospitals can reach 12.9 bln RUR, and for out-patient treatment 1.01 bln RUR annually (season) of RSV infection. Direct non-medical expenditures were calculated as 9.6 bln RUR, and GDP loss due to parental seek live no less than 7.7 bln RUR. Total common expenditures were reached in 32 bln RUR. If we can calculate burden of RSV infection in case of premature death in severe illness total budget loss could reach 3.31 trillion RUR (minimal) or 7.01 trillion RUR (maximal). Sensitivity analysis confirmed stability of model to prevalence RSV infection parameter changing.

Conclusion. Social-economic burden of RSV infection in children 0–2 y. o. has a huge size and has progressive dynamic in case of prevalence increasing. Healthcare costs, economic and social losses dictate the need to improve prevention measures in the most vulnerable age group children aged 0–2 years.