Notes from Scarlet's digital health summit

Thanks to Scarlet for the invite — I was genuinely impressed by the personal touch and by the quality of the event itself. Real debates, no clichés. Here are the bite-sized bullets from my notes, expanded with a bit more context.

AI scribes and the return of empathic time

One of the more optimistic threads of the day: AI scribes have the power to give clinicians back the joy of their job, by cutting into the time currently lost to administrative documentation in favour of their empathic, patient-facing role. The scale of the claimed benefit varies a lot depending on the study and the specialty — some large multi-site trials show more modest single-digit reductions in documentation time, while smaller studies in specific settings report much larger gains, and one specialty-specific estimate puts documentation load at roughly a third of a clinician's total time to begin with. Wherever the exact number lands for a given setting, the direction is consistent: less time on the keyboard, more time looking at the patient.

The risk of not acting

From MHRA's CEO: regulators are used to putting heavy weight on the risk of acting — approving a new device, allowing off-label use — and much less on the risk of not acting, such as preventing a solution from reaching a critical health gap. The benefit-risk ratio, as currently practiced, is unbalanced. It's a point MHRA's leadership has been making publicly in various forms recently, framing overly cautious decision-making as itself a form of harm, not a neutral default. Safety stays non-negotiable, but the argument is that regulation shouldn't function as a paternalistic barrier when patients have informed choices available to them.

Duty of care at consumer scale

Consumerised medical platforms reach tens of millions of users, and that reach brings a duty of care that goes beyond traditional efficacy metrics. A 99% success rate still leaves 1% facing clinical risk — and at consumer platform scale, 1% can mean millions of people. The visibility these platforms have into risk-benefit patterns and real user needs is uniquely powerful, and arguably underused relative to how much signal actually flows through them.

Continuous deployment breaks traditional audit logic

Digital health's shift to continuous deployment doesn't fit neatly into how regulatory review has traditionally worked. Scarlet's own promise — software change reviews measured in days, not months — is a direct response to that mismatch. The open question is what this looks like five years out: could the regulatory review barrier eventually be conceptually replaced by something closer to a simultaneous check, running alongside deployment rather than gating it beforehand? That's a genuinely open design question for regulators and platforms alike, not a solved one.

The controversial counterpoint

Is more and faster medtech actually good for consumers and health systems? Or does it just create more bottlenecks, more problems of choice, and more risk for users who aren't equipped to evaluate what they're choosing between? And does accelerating deployment this way effectively redefine the role of government in healthcare — shifting it from gatekeeper toward something closer to a real-time monitor? Nobody in the room had a tidy answer, and that's precisely what made it a real debate rather than a panel full of consensus.

The room itself

A wonderful afternoon, and a rare occasion to catch up with so many past and present colleagues and clients in one room — genuinely one of the better parts of days like this. Congratulations to the whole Scarlet team for pulling together an event with this much substance behind it.

Giulia Paggiola with colleagues from partner consultancy MDcompliance (https://www.mdcompliance.co.uk/)

Methodology note: This article is based on my original LinkedIn post (link), reflecting notes and reflections from Scarlet's digital health event. AI assisted in elaborating the topic into a broader article by adding background context on the themes discussed, including published research on AI scribe time savings and public statements from MHRA leadership on risk-proportionate regulation. All analysis and perspectives are my own, and all content has been reviewed by me for accuracy.