Trends & Insights
Here is where we get to ponder trends and share insights about clinical information delivery in the pharmaceutical sector. What you find here can come from conferences or client work that has been completed. Whatever the source, this is where real value can be added, especially in a time of such technological upheaval as we have now.
PHUSE EU Connect 2025
Much is changing in clinical data science, with a clear shift away from fragmented legacy setups towards unified, governed environments that support reproducible analysis, end to end auditability and faster delivery under growing regulatory scrutiny and expanding data volumes. Modernisation efforts are increasingly cloud enabled and multi-language, bringing together SAS with R and Python, underpinned by version control and automated pipelines, shared metadata and stronger data governance aligned with FAIR principles. Open source adoption is accelerating, but the more instructive lessons are organisational, since effective change depends on sustained stakeholder engagement, training, communities of practice and dedicated expert teams that can embed new ways of working over the long term. Collaboration across industry is also maturing through reusable, validated tooling and governance frameworks that support hybrid and all R regulatory submissions while keeping validation and traceability central. Automation and AI are advancing from assisted coding to workflow generation and clinical querying, yet the emphasis remains on explainability, audit trails and a human in the middle approach for critical decisions. Alongside this, language agnostic infrastructure, synthetic data for early testing and privacy and more robust file and spreadsheet governance are emerging as practical enablers of scalable, compliant digital trial delivery.
SAS Innovate 2025
Held in Orlando, and later in London and other places, the 2025 event showcased developments with clear relevance to pharmaceuticals across clinical development, regulatory work and commercial analytics, with a strong focus on governed use of AI and modern data architecture. Highlights included a next generation clinical environment on SAS Viya that combines secure trial data management with regulated statistical computing, supporting audit trails, electronic signatures, standards alignment and multi-language programming while integrating with capture and validation systems and extending into protocol planning. A recurring theme was keeping generative capabilities inside controlled environments to protect sensitive data, support change control and preserve reproducibility, alongside growing use of natural language interfaces that broaden access for non programmers while retaining logging and access controls. Sessions also emphasised running analytics closer to where data resides to reduce data movement, address sovereignty constraints and improve performance, shared lessons from managed cloud migrations aimed at stability and compliance and explored modern engineering practices such as CI driven validation for clinical code, structured approaches to legacy migration and data lineage for auditability and ALCOA plus integrity. Trustworthy AI featured heavily, framing bias as a reflection of organisational practice and stressing explainability, validation, governance and human oversight, while change management discussions underlined that adoption depends as much on community support and engagement as it does on technical delivery.
CDISC European Interchange 2025
Across two days in Geneva, discussions pointed to clinical data standards shifting from static compliance artefacts towards connected, machine-executable infrastructure that supports end to end automation. Demonstrations showed how structured study definitions, semantic concepts and executable validation rules can cut study start up from weeks to minutes, improve cross study consistency and embed lineage directly in data flow, while early work continues on standardising analysis approaches and newer data types such as digital health, imaging and multiomics. Artificial intelligence was presented mainly as a practical aid for searching standards, drafting mappings and accelerating concept development, with strong emphasis on explainability, auditability and human review in regulated contexts. Real world data integration emerged as both necessary and difficult, driving renewed focus on provenance, harmonisation and bridging health record formats with submission structures, while regulatory initiatives such as ICH M11 and closer alignment between agencies and industry suggested a move towards global harmonisation rather than regional divergence. Repeatedly, speakers stressed that the hardest part is organisational change, with skills gaps, governance and sustained training determining whether automation efforts succeed and with pre competitive collaboration increasingly treated as essential shared infrastructure rather than optional goodwill.
SAS Innovate 2024
Across the event there was a clear emphasis on analytics becoming more industrial in how it supports pharmaceutical research, development, regulatory work and operations, with strong focus on governance, validation and reproducibility. Real world evidence work highlighted the value of standardising disparate data into a common model and capturing an end to end, auditable record of analytical decisions and outputs to strengthen credibility with regulators while speeding delivery. Manufacturing sessions showed how digital twins can use historical batch data to run large numbers of simulated experiments, helping identify optimal process settings with lower cost, faster iteration and better control of risk and documentation than physical trials alone. Platform discussions raised the practical trade-off between flexibility and revalidation when moving from legacy environments, including the strategic appeal of rewriting code into open languages to widen the talent pool and reduce dependency on a single vendor. A further theme was unifying low code and code first workflows in one governed environment, so operational users can produce routine analyses while statisticians refine models for higher performance, alongside renewed attention to modern development interfaces as a productivity and recruitment lever in tightly resourced clinical and statistical teams.