Across three sessions, the patient track at IQVIA TechIQ 2025 told one cohesive story: start with what matters to patients (Intelligence), turn that into workable solutions (Innovation), and prove measurable results for patients in the real world (Impact). The through line was “precision with purpose”— use patient data safely and responsibly, embed it in everyday clinical workflows, and show concrete benefits (driven by trustworthy AI).

Session 1: [Panel] Streamlining patient experience & evidence collection with technology & AI

Moderator: Ian Bonzani (IQVIA); Panelists: Jessica Braid (Roche), Tom Keeley (GSK), and Elan Josielewski (IQVIA).

Session take-home: Patient relevance first; let technology and AI serve that purpose, not define it.

What we heard. The panel widened the lens on Patient Experience Data (PED): not just traditional COAs (Clinical Outcomes Assessments), but a broader picture that also includes clinician and observer-reported insights, passive digital signals and novel composite measures leveraging wearables and sensors. The point is not more data—it’s the right data, anchored in patient relevance, that regulators, payers, clinicians and patients can act on.

Key insights. Teams are moving from broad, general ideas and concepts to specific, observable traits that digital measures can quantify with less burden (think sleep continuity or activity-based function). The panel cautioned against “precision first, meaning later”: define the concept of interest with patients up front, then select digital measures and appropriate tools that capture it. Meanwhile, regulators are leaning in (e.g., FDA expectations for PED content and growing EMA attention), which elevates the bar for quality, documentation, and presentation.

Where AI fits. AI is already useful for sifting time dense signals, assisting with qualitative analysis (e.g., LLMs to surface themes from patient interviews), and building individual baselines from continuous data—all under a privacy by design umbrella. The practical takeaway: when AI is applied to clearly defined, patient relevant concepts, it can highlight the most important information and reduce unnecessary noise, without adding extra work.

Session 2: [Presentation] Delivering pharma last-mile solutions via electronic medical record systems

Presenters: Vieshal Raja Gopal (AstraZeneca), John Rigg (IQVIA), Calum Yacoubian (IQVIA).

Session take-home: When patient-based analytics is integrated into clinical workflows, care gaps close and outcomes improve.

What we heard. Pharma companies experience substantial diagnostic, prognostic and care management hurdles which seriously constrain product sales; some of these hurdles are often referred to as “last-mile” challenges. Successful “last-mile” solutions put Clinical Decision Support Tools (CDSTs) directly inside the EMR (Electronic Medical Record), surfacing the right patient, at the right time, with the right next action. The pattern: extract EMR data → transform with NLP (Natural Language Processing) or other AI as needed → present prioritized worklists or point of care prompts that align with guidelines for prospective review by care coordinators.

Customer success stories:

Aortic Stenosis (AS):

• Using NLP to read echocardiogram text in the EMR found 50% more patients with AS (narrowing of a heart valve) than billing codes, and ~20% of them merited rapid specialist referral.
• AI “read what clinicians wrote” and surfaced people who needed attention sooner.
• Why it matters: faster, more accurate routing to specialty care prevents worsening disease and improves service planning.

Atrial Fibrillation (AF):

• A machine learning model flagged people at high risk of undiagnosed AF (an irregular heart rhythm that raises stroke risk).
• About 1 in 4 people flagged by the model had an abnormal, undiagnosed heart finding on the quick ECG test. The approach delivered roughly a 9× improvement in finding undiagnosed AF versus prior systematic or opportunistic benchmarks.
• Why it matters: earlier AF detection helps prevent strokes and reduces avoidable healthcare costs.

Asthma Guidelines in Practice:

• A healthcare initiative used electronic medical records (EMR) to identify patients who were frequently using short-acting inhalers and transitioned appropriate individuals to a single-inhaler therapy that combines both maintenance and relief functions.
• This approach led to:
• A significant increase in prescribing of the combined therapy
• A substantial reduction in severe asthma exacerbations
• A notable decrease in greenhouse gas emissions associated with asthma treatment

Why it matters: By finding the right patients in the EMR and supporting guideline-based care, the system reduced severe attacks and improved both health and environmental impact—showing how data can make everyday care safer and greener.

What makes these three customer examples work? They fit naturally into the way clinicians already work (right inside the electronic medical record), are managed responsibly, and are designed to be easy to use and measure—rolling out carefully to avoid overwhelming staff with too many alerts.

Session 3: [Panel] Best practices for scaling safe and innovative use of patient data

Moderator: Jonathan Green (IQVIA); Panelists: Vanessa Carter (The AMR Narrative), Sharon Lamb (McDermott Will & Schulte), Gurleen Singh Jhuti (Genentech).

Session take-home: To safely and effectively scale the use of patient data, organizations must make privacy and transparency a core part of every solution and involve patients in the process. Building trust through clear communication and responsible data practices is essential for unlocking real innovation in healthcare.

What we heard. There is an explainability gap: people want their data used ethically, but they often do not understand how modern data and AI pipelines work. The panel argued for plain language transparency, co-design with patients (not just consultation), and consent journeys that let people see how their contributions lead to tangible improvements in safety, equity, and access.

Navigating the rules without losing momentum. With Europe’s evolving frameworks (GDPR, EU AI Act) and the U.S. “patchwork” of governance, the remedy is first principles governance: privacy by design, clear model lifecycle controls, and risk-based safeguards that avoid “cost layering without value” - especially where software already falls under medical device regulation.

Center the patient, not only the policy. Involve patients in co-design and governance; consider consent journeys that inform people how their data contributes to societal benefit; close health equity gaps by responsibly combining underserved population attributes (e.g., race, income, education) with clinical data - balancing privacy risks with high value insights.

The Bottom Line – Precision with Purpose

The Patient Track’s unifying message is simple: start with the patient, ship solutions that fit the work, and prove the difference they make. Session One reset our measurement mindset around patient anchored concepts supported (not led) by AI; Session Two showed that EMR embedded CDSTs can deliver tangible gains—from earlier identification (AS, AF) to fewer severe asthma attacks and even lower emissions; Session Three made clear that explainability, privacy by design, and patient partnership are the price of admission for scale.

Final take-home: Intelligence. Innovation. Impact is not a tagline; it is the operating model for turning high quality patient data and Healthcare grade AI^® into earlier action, better decisions, and measurably better outcomes.