2025 – Top 5 RWE Regulatory Trends (with what’s changing) 

Source: RWR Regulatory Updates (Jan – Dec 2025)

UK-MHRA-AI-Consultancy-Dec-2025

AI moves from “innovation topic” to “regulated evidence lifecycle”

Consolidation creates execution pressure

  • Trend 1 = AI moves from “innovation topic” to “regulated evidence lifecycle”
  • Trend 2 = Health-data access and interoperability become regulatory infrastructure
  • Trend 3 = Harmonisation of NIS/RWE methodological expectations accelerates
  • Trend 4 = Data protection and cybersecurity directly shape RWE feasibility
  • Trend 5 = Operational standardisation increases predictability—but reduces flexibility

Introduction

Throughout 2025, real-world research (RWR) continued to move from a supportive role into the core of regulatory, safety, and reimbursement decision-making. Across regions, regulators and health systems placed less emphasis on whether real-world data (RWD) and real-world evidence (RWE) can be used, and far more on how they are generated, governed, accessed, and maintained over time.

This shift is visible in practical regulatory actions rather than high-level statements. New legislation, revised guidance, formal data-access mechanisms, mandated templates, and system changes are now directly shaping how non-interventional studies (NIS) are designed, approved, and operated. In many cases, these developments affect feasibility, timelines, and inspection risk as much as scientific design choices.

This brief is based on the RWR Regulatory Updates published between January and December 2025 and reflects observed regulatory activity at country, regional, and global levels. The five trends set out below highlight where expectations are becoming clearer, more structured, and more enforceable in practice.

The purpose is straightforward: to help sponsors, CROs, and real-world research teams understand where regulatory pressure is increasing, where operational risk is emerging, and what needs to be in place to deliver RWE that is credible, defensible, and fit for regulatory and reimbursement use.

Context: Acceptance Pressures and Practical Challenges Observed in 2025

While 2025 shows clear progress in the formalisation and use of RWE, the same period also highlights persistent challenges in acceptance and execution that are reflected across the monthly RWR Regulatory Updates.

Across regions, regulators and health systems have raised expectations around methodology, governance, data access, and operational control. In practice, this has made RWE harder to deliver at scale. Increased scrutiny, longer feasibility timelines, data-access constraints, and higher compliance burdens have reduced the number of situations where RWE can be deployed quickly or flexibly.

Several updates during 2025 note a growing gap between strategic ambition for RWE and the resources required to meet regulatory expectations in practice. For some organisations, this has resulted in reprioritisation or restructuring of internal RWE functions, particularly where evidence generation could not be clearly linked to defined regulatory or reimbursement decisions.

These challenges do not contradict the positive regulatory trajectory described below. Rather, they reflect a transition phase in which RWE is increasingly expected to meet regulatory-style standards, while operating models and resourcing have not always adjusted at the same pace.

  • Trend 1 = AI moves from “innovation topic” to “regulated evidence lifecycle”

    What’s changing: Regulators and health systems are shifting from generic AI enthusiasm to governance, accountability, and real-world performance monitoring. Expectations increasingly extend beyond development and validation to include transparency, bias management, privacy risk controls, and ongoing oversight.

    Seen in: Canada’s AI in HTA positioning and horizon scanning; EU AI governance and AI-in-regulation initiatives; UK MHRA AI Airlock and NHS AI acceleration programmes; France GDPR–AI compliance frameworks; Japan PMDA AI action plan.

    So what? RWE programmes using AI increasingly need explicit controls around dataset provenance, bias assessment, model lifecycle management, and ongoing performance monitoring. Acceptance depends less on novelty and more on demonstrable fitness of both data and methods for the intended use.

  • Trend 2 = Health-data access and interoperability become regulatory infrastructure

    What’s changing: Health-data access is being formalised through national and regional platforms, interoperability standards, and structured governance for secondary use of data, including cross-border access. System transitions are increasingly influencing study start-up and submission pathways.

    Seen in: European Health Data Space (EHDS) implementation and HealthData@EU releases; EMRN data strategy; TEHDAS2 draft Health Data Access Body guidance; FDA interest in HL7 FHIR-based submissions; UK IRAS replacement (Plan & Manage).

    So what? Data readiness is now inspectable. Metadata quality, linkage governance, permissions, auditability, and alignment with mandated technical standards are becoming key differentiators for feasibility, timelines, and approval risk.

  • Trend 3 = Harmonisation of NIS/RWE methodological expectations accelerates

    What’s changing: Regulators and standards bodies are converging on clearer expectations for NIS using RWD, particularly where evidence informs regulatory, safety, or reimbursement decisions. Emphasis continues to grow around transparency, pre-specification, and methodological justification.

    Seen in: EMA RWE roadmap and NIS reflection paper; CIOMS RWD/RWE guidance; ICH RWE-related guideline development including ICH M14; FDA emphasis on visibility of RWE in decision-making; Saudi Arabia draft RWE framework.

    So what? “Non-interventional” no longer implies low scrutiny. Acceptance depends on whether the underlying data are fit for use and whether the resulting evidence is fit for the specific purpose being proposed.

  • Trend 4 = Data protection and cybersecurity directly shape RWE feasibility

    What’s changing: Expanded obligations around consent, reuse limitations, cross-border transfers, breach reporting, vendor controls, and cybersecurity requirements are directly affecting what RWE is feasible in practice.

    Seen in: Peru and Mexico data-protection reforms; India Digital Personal Data Protection Rules; China Cybersecurity Law revisions; EU pharmacovigilance tightening with downstream RWE governance implications; Czech and Finnish NIS2 cybersecurity effects; France CNIL clarifications.

    So what? RWE success increasingly depends on clear controller–processor roles, defensible transfer mechanisms, and cybersecurity-ready partners. Weak governance can prevent studies from proceeding at all.

  • Trend 5 = Operational standardisation increases predictability—but reduces flexibility

    What’s changing: Regulators and health systems are standardising the operational “how” of research through single-review models, mandatory templates, model agreements, structured amendment processes, and defined post-market reporting formats.

    Seen in: Germany “one study, one vote”; UK GDPR wording templates, revised amendment processes, mCDA/mNISA updates, amendment history logs, and device PMSR format; Brazil SINEP decree implementation; EU coordinated assessment pilots; Spain IVDR national implementation affecting PMPF; Australia ethics statement updates.

    So what? Local interpretation is increasingly replaced by mandated artefacts. TMFs and operating procedures must track templates, agreements, tools, and effective dates by jurisdiction to remain inspection-ready.

2025 Country Trend Signals (at-a-glance): Consolidation and Execution Pressure

Signal key:

● Strong signal  ○ Emerging signal  – Not a key signal for that year

Trend codes:

T1 – AI and advanced analytics governance
T2 – Health-data access and interoperability infrastructure
T3 – Methodological expectations for NIS/RWE
T4 – Data protection, privacy, and cybersecurity
T5 – Operational standardisation

2025-RWE-Trends-Comparison-Table

2025 Country Trend Signals (at-a-glance) – By Country/Region

Trend codes:

T1 – AI and advanced analytics governance
T2 – Health-data access and interoperability infrastructure
T3 – Methodological expectations for NIS/RWE
T4 – Data protection, privacy, and cybersecurity
T5 – Operational standardisation

2025-Country-Trends

Terminology Note (fit-for-use and fit-for-purpose)

Throughout this brief, references to RWE quality or acceptability should be read in line with ICH M14 and related guidance. Regulators do not assess RWE as a fixed “grade”. Instead, they consider whether the underlying RWD are fit for use and whether the resulting RWE is fit for the specific regulatory, safety, or reimbursement purpose being proposed. Acceptance is conditional, contextual, and question-driven.

Looking Ahead to 2026

Based on the regulatory signals observed throughout 2025, the overall direction for RWE is now relatively clear. Regulators are unlikely to revisit fundamental questions about whether RWE can be used. Instead, expectations around data fitness, methodological rigour, governance, and operational control are likely to be applied more consistently across use cases.

For organisations generating RWE, the challenge moving into 2026 is therefore less about innovation and more about execution. Evidence will increasingly need to be designed against clearly defined regulatory, safety, or reimbursement questions, with justification that both the data and the methods are fit for that purpose.

Progress in 2026 is likely to remain uneven. Organisations able to align RWE activity tightly to specific decisions and demonstrate control across the full evidence lifecycle will be better positioned than those relying on exploratory or loosely governed approaches. This outlook reflects a continuation of trends already visible across 2025.

Bottom line: RWE is now regulated end-to-end—data, methods, technology, and operations. Acceptance increasingly depends on fitness for use and fitness for purpose, rather than on any notion of a universal evidence “grade.”