Field Notes from the 2026 West Windsor PHUSE SDE
Author: Aga Rasinska, Director of Strategy
Margaret Wishart (of Bristol Myers Squibb) and I built the West Windsor PHUSE SDE around a premise: Clinical data analytics is changing faster than the systems and processes designed to support it. Not faster than our capability; faster than our foundations.
But what does innovation mean right now? Are we doing it, or are we calling something else by that name? Because the honest answer is that most of what gets labeled innovation is automation wearing a faster, cleaner outfit. Still the same decision, still the same schema, just with better-looking dashboards as outputs.
In the event’s fireside chat, Mike Stackhouse and I talked peer to peer about:
- What kills innovation in our industry more reliably than anything else
- Whether you can move fast in clinical development and what that costs
- What we’d tell someone in the room who wanted to be innovative
At one point he asked what agile meant to me, after all these years of using the word. My honest answer? Keep asking why and keep checking in on whether you’re delivering value.
That’s the tension the whole event was built to pressure-test, and it moved through the room all day.
The Executive Panel Opened It
Sangeeta Bhattacharya (Johnson & Johnson), Samar Noor (Bristol Myers Squibb), and Jeen Liu (Regeneron) followed with a real, honest talk on the long-standing assumptions they’d had to abandon and the moments incremental optimization stopped being enough. We moved through:
- Where disruption is real inside each of their organizations versus where it’s still a phrase that shows up on a slide
- Where the real friction sits, like technology, standards, regulation, or people
- What it costs to make the case for investment when the return is hard to quantify and the timeline runs past the next budget cycle
We also talked about the gap between what leadership believes is happening with the data and what’s happening on the ground, and what it takes to close it honestly rather than paper over it with better dashboards.
The question that stayed with me the longest wasn’t one I asked so much as one already sitting in the room: Are we spending years solving the same problems while AI is about to make the choice irrelevant, and if so, how did all that effort matter? No one has a tidy answer yet, including me.
The Day Kept Testing It
From there, the day kept testing the same premise from different angles.
Merck’s team walked through their proof-of-concept, taking CDISC ARS metadata to drill-down reporting—a specification for what a reviewer can trust. Stephen Hamburg (Jazz Pharmaceutical) presented a working prototype that takes the SAP and produces specs, shells, ADaM requirements, and code end to end, with a human reviewing at every stage. But the prototype wasn’t the only point he wanted to leave the audience with. He also focused on the mindset shift underneath it, from “I write the code that produces this table” to “I own the quality of this table, however it gets produced.”
It was also honest take that headcount demand is still unclear, as efficiency per programmer rises, but so might the volume of work sponsors ask for. He was just as honest that the skill mix isn’t ambiguous at all. Value moves upstream, toward specification, validation, and judgment, whether the industry is ready for that shift or not.
Michael Chow (Posit) made the case for Python inside a multilingual, regulated stack, arguing not R versus Python, but what it takes to hold both without the schema underneath them fracturing. Kanishk Singh’s (Johnson & Johnson) persona stack talk offered the clearest answer to the opening question all day: The model isn’t the variable, what you stack around it is, like identity, a skill, live evidence, task scope. He also posed that a general-purpose model starts citing the standard it used and naming what it won’t answer.
Rostislav Markov (AWS) pushed the same idea further into infrastructure, focusing on what it takes to move from AI-assisted coding to governed SDTM derivation teams can sign off on. Emma Cleary (BIP) wrapped up the talks session with the number that resonate with anyone writing an AI business case right now: 72% of life sciences CIOs report their organizations broke even or lost money on AI investment. AI doesn’t equal ROI on its own, and when it scales faster than adoption does, what we lose isn’t just productivity, it’s trust in the next initiative.
She traced it back to something more human than technical: Employee willingness to support organizational change has fallen from 74% to 43% in less than a decade, because people are absorbing more planned change than they can process. Value starts with people, not the technology stacked around them.
The Technical Panel Closed the Loop
By the time Michael Chow (Posit), Rostislav Markov (AWS), and Andrew Holz (Posit) closed things out on the technical panel, the pattern was hard to miss. The question was whether the multilingual reality we’re all living in is a strength or just a new layer of complexity to manage.
We pushed past the abstractions on purpose:
- Asking not whether cloud infrastructure belongs in clinical trials, but inspecting the gap between what it promises and what teams can implement this quarter
- Questioning not whether AI belongs in the pipeline, but where the real risk sits, in the model, the data, the validation layer, or the human who signs off at the end
“We built it fast” and “we can trust it” are not the same sentence, and none of the three let the room forget it. Breaking the schema means rebuilding it on purpose, instead of letting it calcify by default.
Are We Creating Real Value?
Name one thing the industry keeps calling innovation that isn’t, and one thing it’s doing that is. My answer, after a full day of conversations? We keep calling faster innovative, but it isn’t. The winners in innovation, as I said closing the day, won’t be the organizations with the most advanced technology. They’ll be the ones who can combine innovation with adoption, technology with expertise and people, and speed with critical thinking.
Which brings me back to the question we opened the day with, the one I don’t think we’re done answering: If the future of clinical analytics were built from first principles today, what would it look like?
Aga Rasinska
Director of Strategy, Atorus
With more than a decade of hands-on experience in bioinformatics, data science, and program leadership, Aga Rasinska brings a dual perspective that bridges business strategy and technical innovation, transforming complex clinical and omics data challenges into scalable, results-driven solutions. Her expertise spans project and product management, change leadership, and data-driven decision-making within regulated life sciences environments. She is also a frequent industry speaker, panelist, and contributor focused on the intersection of science, data, and business strategy.