Best of LinkedIn: Health Tech CW 16/ 17

00:00:00: This episode is provided by Thomas Allgeier and Freeness, based on the most relevant LinkedIn posts of health tech in CW-XVI and XVII.

00:00:09: Freenes equips HealthTech providers with market intelligence to identify which hospitals target and how to reach decision makers for hospital digitalization as a result of the Kronkenhaus-Skazetz.

00:00:21: You can find more info in description!

00:00:24: you know, when you think of a major medical breakthrough.

00:00:26: You probably picture something incredibly cinematic right?

00:00:28: Oh yeah absolutely like a gleaming robotic arm doing microscopic surgery or I don't know...a supercomputer just instantly curing a disease

00:00:36: Exactly!

00:00:36: You expect this very clean highly visible kind of magical moment.

00:00:41: But looking at the top health tech trends seen across Lincoln over the last two weeks, that expectation just completely shattered.

00:00:48: It really does!

00:00:49: The actual digital transformation happening in healthcare right now is well it's murky...it's messy and honestly the most revolutionary technologies aren't the loud or flashy ones.

00:00:57: they're these invisible systems just quietly untangling a web of deeply broken workflows

00:01:03: which is such a massive operational puzzle.

00:01:05: so Our mission for this deep dive is to basically piece together how that digital transformation is hitting the real world.

00:01:12: For you, we've clustered the insights into three distinct themes

00:01:16: right?

00:01:16: So first We are looking at how AI adoption is finally moving from just pure hype To actual organizational readiness.

00:01:24: Yep Then well unpack How clinical ai's becoming incredibly specialized

00:01:28: and Finally will dive in to how robotics Is expanding way beyond Just basic automation In The operating room.

00:01:34: So kicking off with that first theme, we are officially past the era of AI hype.

00:01:38: I mean... The foundational tech is built?

00:01:40: The algorithms exist?

00:01:41: The

00:01:41: challenge now is like how do we actually force that pristine technology to function inside really stubborn complex human systems?

00:01:50: exactly?

00:01:51: and Roshni Dada posted a brilliant breakdown on this.

00:01:53: she calls it the reality gap in health tech.

00:01:56: She points out four highly predictable places where this tech just hits a brick wall.

00:02:00: Oh right!

00:02:00: The adoption hurdles.

00:02:01: what were they again?

00:02:02: So

00:02:02: first is infrastructure.

00:02:04: You know, you can design this world-class software that needs high speed internet to process imaging but if you drop it into a basement clinic with the spotty connection...

00:02:13: It's completely useless!

00:02:14: Right.

00:02:15: and The second is training.

00:02:17: A hospital might actually buy the cutting edge tool But If the nurses aren't giving them hours To learn it just sits there

00:02:24: Like very expensive paperweight

00:02:25: Exactly.

00:02:26: Then theres awareness..if doctors don't have the tools they won't refer patients.

00:02:31: And the final one, which is huge...is reimbursement.

00:02:35: Oh!

00:02:35: The billing codes?

00:02:36: Yeah If insurance companies don't have a code for an AI-assisted procedure.

00:02:41: Adoption just stocks dead in its tracks.

00:02:43: I mean it's like designing state of art sports car but expecting to run flawlessly on dirt road The technology being fabulous as necessary But never sufficient.

00:02:54: That is a great analogy.

00:02:55: And Robert Slippin and Joshua Lew actually built on this exact dynamic, they pointed out that if you just take an inefficient healthcare process and slap AI on top of it... You haven't revolutionized anything!

00:03:05: You've just made a faster bad-processed.

00:03:07: Exactly —you create less inefficient but still totally ineffective workflow.

00:03:12: The true promise of AI requires us to completely redesign roles in workflows first.

00:03:17: Well, and you have to look at the foundation of all this right?

00:03:19: The data Kamal Nisralli shared.

00:03:21: This stat that is just wild.

00:03:23: Ninety-seven percent hospital data goes entirely unused

00:03:28: with ninety seven percent yes

00:03:30: an eighty percent it's totally unstructured.

00:03:32: we're talking messy faxes scan PDFs hasty notes

00:03:37: oh wow yeah.

00:03:38: Sigrid Bergman Rujan actually put out a warning about.

00:03:41: She argues that unless we fix those exact data silos, clean up the biases and find a clear ROI.

00:03:48: AI is never going to revolutionize health care.

00:03:50: because of the data as fragmented The AIs output will be flawed

00:03:53: exactly.

00:03:53: but when our hospital does get it right When they fixed the workflow?

00:04:02: It doesn't do surgery.

00:04:06: Doesn't diagnose anything, it just runs errands.

00:04:08: It carries blood bags and lab samples between floors

00:04:11: which sounds so basic but its actually huge!

00:04:14: It is up to forty percent of a nurses shift is just logistics.

00:04:18: they're basically walking miles as highly paid careers.

00:04:21: Healthcare rewards innovation that is invisible and actually relieves that burden.

00:04:26: And we saw a massive appetite for the exact kind of administrative relief at their recent DMEA, twenty-twenty six event?

00:04:33: Oh

00:04:33: absolutely!

00:04:34: Philip Martins and Nikolai Kolev were there sharing how Dr.

00:04:38: Libb's AI assistants are working in German practices right now.

00:04:41: Yeah

00:04:41: handling phone calls on the practice software rate.

00:04:43: Yep it's delivering actual lived reality relief.

00:04:47: It even caught attention from Health Minister Nina Worken.

00:04:50: Okay, but let me push back here for a second.

00:04:52: Because there's a risk!

00:04:53: What happens when an AI makes mistakes because of bad data rule?

00:04:57: Oh for sure...

00:04:57: Like Kayubi Ishiyama suggests that this Push For Efficiency is actually creating the need to create a totally new workforce role He calls it The Machine Manager.

00:05:06: Right someone whose entire job is just audit and supervise the AI To ensure patient safety.

00:05:12: But think about it.

00:05:13: hospitals are already strapped for cash.

00:05:15: Are they really gonna pay human just to babysit in

00:05:18: algorithm?!

00:05:19: It sounds counterintuitive, I know.

00:05:21: But think about the liability.

00:05:23: if an AI misreads a dosage across a thousand patients The harm is catastrophic.

00:05:28: You scale the tech but you can't scale accountability.

00:05:31: Yeah that makes sense.

00:05:32: So thats back office sorted.

00:05:34: Moving to our second theme What happens when we step onto front lines of clinical

00:05:38: care?

00:05:39: Thats where things shift.

00:05:40: We go from fixing workflows To fierce demand for specialization.

00:05:44: General AI chatbots are totally out.

00:05:47: Highly specific AI agents are in.

00:05:49: Right, because the stakes are literally life and death.

00:05:52: Myung-chah shared some great insights on Verily's new AI companion called Violet.

00:05:57: Why do they build a new one?

00:05:58: Well A recent JAMA study looked at off-the-shelf large language models And The failure rate for differential diagnosis was between ninety and hundred percent.

00:06:06: Wait

00:06:07: really?!

00:06:07: Almost a hundred percent failure rate.

00:06:09: Yeah, they fail completely on complex cases.

00:06:11: so Verily didn't just build one massive model for Violet.

00:06:15: They built it as a coordinated team of specialized agents.

00:06:18: How interesting!

00:06:19: It functions more like conservative medical board doing safe handoffs rather than One brain trying to guess everything.

00:06:26: Which

00:06:26: is so urgent because Steven G reported that sixty-six million Americans have already used AI chatbots for health advice and fourteen million of them actually skipped a doctor visit.

00:06:36: That

00:06:36: is terrifying, just trusting the general chatbot with your health?

00:06:39: It IS!

00:06:40: But hospitals are stepping up.

00:06:42: Anka Bashlari noted that Hartford HealthCare's patient GPT makes ninety percent fewer errors than standard chatGPT

00:06:50: Just by fencing it in

00:06:51: Exactly.

00:06:51: By fencing the AI into very specific verified clinical contexts.

00:06:56: Epic is doing the same thing, launching their own patient chat bot, Emmy.

00:07:00: And we are seeing this clinical intelligence get incredibly specialized.

00:07:04: Yesir Khan built an AI decision support system that brings pharmacogenomics right into the prescribing workflow.

00:07:10: Okay break that down for me.

00:07:11: Pharmacogenomics

00:07:12: Yeah.

00:07:13: so it cross references a patients DNA to prevent adverse drug reactions.

00:07:17: For example some people have a genetic variation That makes them poor metabolizers of a clopadocrel which has blood thinner.

00:07:24: Oh, got it.

00:07:24: So the drug wouldn't even work for them?

00:07:26: Exactly!

00:07:27: And CPIC guidelines are very clear on this.

00:07:29: but doctors rarely have time or data to check your DNA in an emergency.

00:07:35: Kahn's AI does that instantly and flags the doctor before they even prescribe him.

00:07:40: Wow... That is incredible.

00:07:43: But you know as these specialized systems get smarter Do humans kind of stop thinking?

00:07:47: That

00:07:47: is the exact pushback Matthias Goyen from Siemens Health and Ears brought up.

00:07:52: He warns about this dangerous psychological shift when... ...the system recommends a treatment.

00:07:57: Yeah, automation bias!

00:07:59: Yes it carries this aura of neutrality.

00:08:02: It makes clinicians feel less fully responsible.

00:08:04: You know like-it makes it much heavier for the human to disagree with the machine.

00:08:08: Because if you disagree with AI in your wrong its entirely your fault

00:08:11: Exactly, and patient trust is so tied to this.

00:08:15: Carla Goulart-Perone from Philips highlighted that eighty seven percent of cardiac patients want their clinicians to explicitly explain how AI's being used in there care.

00:08:24: They wanna know a human ultimately responsible

00:08:26: Right!

00:08:27: they refuse be handed off the black box.

00:08:29: Well actually perfectly transition says our third theme.

00:08:33: Because while AI handles all that cognitive load, robotics and spatial computing are becoming the ultimate physical extension of the surgeon.

00:08:40: Moving way beyond just basic automation

00:08:42: Exactly!

00:08:43: Standardizing excellence in operating room.

00:08:45: But Dr Michael Menegini argues we aren't ready for a fully autonomous surgeon bot.

00:08:50: Why not?

00:08:51: because...a robot can do repetitive chores on fixed grid.

00:08:55: Sure But human biology has immense variants.

00:08:58: You know, varying bone density previous injuries soft tissue tension surgical judgment just cannot be automated.

00:09:04: Yeah biological variance is chaotic.

00:09:06: but Michael Arman highlighted a study that counters that of it.

00:09:09: the studies show That robots actually leveled playing field.

00:09:12: really how so well

00:09:13: low volume surgeons who used robotic assistance achieved outcomes that were comparable to high-volume expert manual surgeons?

00:09:20: It just raises the baseline.

00:09:22: oh wow That is huge for regional hospitals.

00:09:24: Yeah, and the momentum is clear.

00:09:26: Irene Zew noted Stryker unveiled The Mako RPS at the A.A.U.S convention And Medtronic's Hugo-R.E.S platform Is expanding fast!

00:09:35: Right Catherine Nelson Noted it being used At the Cleveland Clinic And Andrea Bianchi Saw It at the Cahors Hospital in France.

00:09:41: But you know its no longer just about the physical robot arm.

00:09:44: Its About what the surgeon can actually see.

00:09:46: Herkishov Narayan Analyzed Proprio's Paradigm Platform For Spinal Surgery.

00:09:50: What does that one do?

00:09:51: It replaces intermittent, flat X-rays with a continuous real time digital twin.

00:09:57: We're talking two hundred and fifty gigabytes of data per hour.

00:10:00: A live three D map during the surgery?

00:10:02: Exactly!

00:10:03: it totally eliminates that deadly translation error... ...of a surgeon trying to map a two d image onto a three D patient.

00:10:10: That is wild.

00:10:11: I mean digital twins are fundamentally changing spatial computing in healthcare.

00:10:15: Dipsiano-Passeridi discussed Siemens' health in ears using image based cardiac digital twins for VT ablation.

00:10:21: So, for guiding heart procedures?

00:10:22: Yeah!

00:10:22: For faster intraprocedural guidance.

00:10:24: and it's not even just biology.

00:10:26: Dr.

00:10:26: Jens first noted that Siemens replaced static KPI dashboards with an immersive three d digital twin at their HEP factory Just to massively speed up the reaction times.

00:10:36: It really is becoming the standard.

00:10:38: So if we look at the big picture here, We've talked about AI-fixing workflows and robots fixing bodies but The ultimate goal of all this health tech Is making sure that patient never reaches the operating room At

00:10:50: ALL.

00:10:50: A complete shift to prevention

00:10:52: Exactly!

00:10:53: Seema Verma Lucy Jones And Miko Radolotti All pointed This out.

00:10:57: Healthcare must pivot to prevention and value-based care.

00:11:01: And patients are moving

00:11:02: fast.

00:11:02: That's better than the clinics, right?

00:11:04: Oh

00:11:04: definitely!

00:11:05: Vicki Britton shared how UK patients are self organizing online to plan GLP One oral maintenance strategies entirely without official guidance.

00:11:14: They're

00:11:14: just figuring it out themselves.

00:11:15: Yeah patient demand is outpacing the establishment.

00:11:18: So, here is a final provocative thought for you to mull over.

00:11:22: Will AI

00:11:23: finally be the bridge that makes proactive preventative care scalable?

00:11:44: Yeah, thanks for tuning in and make sure to subscribe.

00:11:47: Catch

00:11:48: you on the next one!