Best of LinkedIn: Health Tech CW 30/ 31

00:00:00: This episode is provided by Thomas Allgaier and Frennus based on the most relevant LinkedIn posts about health tech from CW-Thirtyandthirtyone.

00:00:08: Frenness equips HealthTech providers with The Market Intelligence to identify which hospitals to target, how to reach decision makers for hospital digitalization.

00:00:17: as a result of the Kronkenhaus-Sukenskizetz You can find more info in the description.

00:00:21: So we are cutting through the hype today.

00:00:23: writer.

00:00:24: For all of you digital transformation and tech professionals tuning in, welcome to our deep dive.

00:00:29: We are looking at the top health-tech trends making waves across LinkedIn And really we're mapping out the actual operational realities hitting the space.

00:00:37: Yeah let's unpack this because across imaging surgery cardiac care I mean The entire center of gravity and HealthTech is shifting right now?

00:00:47: proving a shiny new AI algorithm works in some isolated lab environment?

00:00:52: Right, the current challenge is proving it actually fits into the brutal day-to-day operational grind of a hospital.

00:00:58: Exactly!

00:00:59: But you know I have to ask this and push back on this whole transformation narrative.

00:01:04: isn't that true?

00:01:04: most these shiny AI pilots just die because Hospital IT departments simply aren't ready to absorb them.

00:01:11: Oh totally That's the multi million dollar bottleneck right there.

00:01:15: Yeah Joshua Liu posted Brilliant breakdown of this exact pilot problem.

00:01:20: He compared free health tech pilots to Costco samples.

00:01:26: Oh, Costco samples.

00:01:27: I like that right because everyone loves to walk down the aisle and try a free bite, but they rarely convert to actual buyers.

00:01:33: Implementing health tech isn't a free sample.

00:01:36: To actually turn on a pilot A hospital has to allocate hard IT resources You know spin up project managers And redirect data engineering teams

00:01:44: Which is incredibly expensive.

00:01:45: Exactly So if the hospital doesn't have financial skin in game.

00:01:49: The second they hit friction with legacy systems The pilot just stalls out.

00:01:53: Yeah, that makes perfect sense.

00:01:54: And it really connects to the point Matthew Davis brought up.

00:01:57: He was arguing that standardization actually has to precede optimization.

00:02:01: Oh absolutely

00:02:01: Like.

00:02:02: if you are building software for a hospital You literally cannot optimize a workflow That isn't consistent across their own departments.

00:02:08: If floor three handles patient intake differently than floor four, well your algorithm's gonna choke.

00:02:14: Spot on and Julia Strandberg echoed this too she emphasized that the most complex challenges we're facing aren't really tech itself.

00:02:21: it is operational reality of bridging these disconnected systems without adding cognitive friction for user.

00:02:28: Right, there is this massive readiness gap.

00:02:30: A few experts Kevin Bernard and Baroth Seisha pointed out how AI capabilities have completely outpaced health system readiness.

00:02:39: Yeah, Seisha specifically noted that shaping healthcare means we had to rethink force AI onto broken legacy processes.

00:02:48: And if we look at the end users, the clinicians their demands are getting incredibly specific.

00:02:53: Carla Goulart-Perron and Amy Martin both dug into some findings from The Future Health Index.

00:02:58: .The message from the clinical floor is crystal clear they want AI to reduce their cognitive burden.

00:03:03: Yeah ,they definitely do not want another dashboard.

00:03:05: No, no more alert fatigue.

00:03:08: Brandon Ferdig drove this home by reminding the tech community that AI is a tool not a replacement for human clinical judgment.

00:03:14: It

00:03:15: really isn't.

00:03:16: and you know The problem isn't a lack of data either.

00:03:19: P.J.

00:03:19: Lombardi outlined this beautifully.

00:03:21: if You just feed a clinician a raw heart rate That data point is functionally useless.

00:03:26: on its own

00:03:26: right context Is everything

00:03:28: exactly.

00:03:29: but A heart rate mathematically correlated in real time to respiratory status and device settings in medications, that becomes contextual data.

00:03:37: That actually drives action because right now as Miriam Fernandez-Martin highlighted, ninety seven percent of global health data is trapped in unstructured formats.

00:03:46: nobody can act

00:03:47: on.

00:03:47: Wait!

00:03:48: Ninety Seven Percent Is Trapped?

00:03:49: Yeah!

00:03:49: Ninty Seven Per Cent!

00:03:50: That's

00:03:50: a wild systemic failure.

00:03:52: I mean if almost all our data are locked up how do we even benchmark if an autonomous clinical AI is doing its job?

00:03:59: Well you have to change how your measure success.

00:04:02: Jules Friedman brought up the new Nature Medicine, Matt S. Framework.

00:04:05: it devaluates AI not against perfection but against reality.

00:04:09: okay

00:04:09: how so?

00:04:10: So an AI model with seventy-five percent diagnostic accuracy might sound terrifying to a software engineer who wants six nines of reliability.

00:04:18: But if the real world alternative is a patient waiting six months To see a specialist where their accuracy during that wait Is literally zero percent then that seventy five percent is A massive upgrade in care.

00:04:30: Oh wow, that makes total sense.

00:04:31: You compare the algorithm to the baseline status quo not a theoretical ideal.

00:04:36: and you know if we look at where AI is actually scaling first it isn't the sexy diagnostic stuff No

00:04:41: its the admin side

00:04:42: Right.

00:04:43: Christian Heim broke down this huge contract sales force just signed with U.S Department of Veterans Affairs.

00:04:48: It's worth upto one point six billion dollars deploying agentic AI to tackle unglamorous admin work answering calls, verifying benefits that sort of thing.

00:04:58: They are cutting scheduling cycle times from an average of twenty eight days down to minutes.

00:05:02: That is huge!

00:05:03: Thats where the immediate ROI lives.

00:05:05: But boundaries between consumer tech and clinical tech Are totally blurring now.

00:05:10: Pooja Apatel & Stephanie Steinek were discussing open AI rolling out chat GPT health Allowing users connect their actual medical records

00:05:19: Which raises massive architectural questions right?

00:05:21: About Data privacy and who owns the insights?

00:05:24: Exactly.

00:05:25: Meanwhile, inside the highly regulated clinical walled garden Epic is taking a dual approach.

00:05:31: Jackie Gerhardt explained how they've deployed an assistant called Emmy for patients And one called Art For Clinicians.

00:05:37: So prepping both sides for better conversations.

00:05:39: Yeah synthesizing the chart beforehand so The actual physical visit Is highly informed.

00:05:44: It's all about reducing friction.

00:05:47: Robert Sleppin praised Doctronic for offering low-friction virtual care for just thirty nine dollars and the scale is crazy.

00:05:53: Kylie Marks noted, Doximity has hit sixty three million telemedicine visits.

00:05:57: Wow!

00:05:58: Sixty Three Million.

00:05:59: Yeah Stephen Chalup kind of wrapped this theme up by arguing a dedicated AI clinical reasoning layer like background processing engine synthesizing all patient info.

00:06:08: That's incredible, but you know AI is just software.

00:06:12: It can analyze it can route But it cannot physically intervene.

00:06:15: if we want that intelligence to act in the physical world We have to look at the hardware which brings us to robotic and digital surgery.

00:06:22: Yeah The ai is the brain robotics of the hands

00:06:24: Exactly.

00:06:25: And by the way If you are finding this deep dive helpful for keeping up with tech friends make sure You hit subscribe so you catch our future ambitions.

00:06:34: Going back to robotics, I have an analogy for this.

00:06:36: A robotic surgery program is like a Formula One team.

00:06:39: it absolutely doesn't matter how advanced the car Is if the pit crew doesn't have the right tires ready?

00:06:44: Oh that's perfect because perception really diverges from reality here.

00:06:48: Austin Chiang and Rajit Kamal pointed out The gap between what patients expect Like sci-fi cyborgs And the reality of commercializing medical AI.

00:06:58: We were at the intersection Of digital planning in robotic execution.

00:07:02: We're

00:07:02: seeing huge global milestones, too.

00:07:04: But Gumbhan Yolgesen and Amr al-Mohammadi highlighted the first cases for the Medtronic Hugo system in Turkey and Saudi Arabia And Anastasia Andreenova Kendra Wotruba and Derek Minnick all celebrated Medtronik being named one of times one hundred most influential companies for twenty twenty six.

00:07:22: It's on moving to an ecosystem approach now.

00:07:24: Christian Valowider detailed the Stealth ECHI SS system which unifies planning, navigation and GE's BK active intraoperative ultrasound for cranial surgery.

00:07:34: So tying everything into one platform?

00:07:36: Right!

00:07:36: And George Murgatroyd covered touch surgery integrating in to the ABLE ENT ecosystem.

00:07:41: it no longer just standalone hardware.

00:07:43: Tying back to the F-One analogy, Jonathan Sayle, FDSC pointed out that optimizing robotic utilization requires understanding really mundane workflows like making sure the sterilization department actually has enough stuff to prep the instruments.

00:07:55: Right otherwise a ten million dollar robot just sits there and training is evolving too!

00:07:59: Andrea Hagenfeldt in Muhitten Copan discussed scaling training via live surgery broadcast from Keralinsk and hands on hydrogel prostate models for your course.

00:08:08: And sometimes the breakthroughs are physiological.

00:08:11: Antonio Casino Marlanda shared this clinical breakthrough, an intraoperative autonomic field block that targets visceral pain and collectemies.

00:08:20: It cuts post-operative nausea so much they can do.

00:08:23: twenty four hour discharges.

00:08:25: Twenty Four Hour Discharge for a Collecteme is wild.

00:08:28: but to guide those advanced robots we need actually see what were operating on.

00:08:33: Diagnostics are shifting from predicting probabilities.

00:08:37: I

00:08:40: get that we want to find disease earlier, but is finding a microscopic fragment always helpful or do you risk overloading clinicians with noise?

00:08:48: That's the core tension right now.

00:08:50: You have to separate this signal from the noise!

00:08:52: Gia Tamar explained The Unite AI blood test using really brilliant analogy.

00:08:56: Finding early cancer DNA in normal blood Is like finding single contact lens dropped on beige carpet.

00:09:01: Oh wow...that sounds impossible.

00:09:03: Right but the AI model processes these subtle patterns and detected cancer in forty-seven percent of cases from just a standard blood draw.

00:09:12: And Guido Matthews noted, Phenomics is making extra chromosomal DNA visible to AI target discovery models.

00:09:18: finding this loose highly active fragments that drive tumors.

00:09:22: That's fascinating!

00:09:23: We are also seeing the sheer speed work flows accelerate.

00:09:26: Dan Xu and Alexei Skutarev highlighted how spectral CT now completes cardiac evaluations in under fifteen minutes.

00:09:34: Which is a game changer for emergency departments?

00:09:36: Absolutely!

00:09:37: And Niels Vandeverf explained how AI deblurring on spectral CT increases calcium detectability by twenty-three percent, plus.

00:09:44: Dr Daniel Strohmer shared insights from the Neotom Connect meeting on maximizing photon counting CT technology.

00:09:49: It's happening globally too For Esal Hassan, noted India's first AI-able electaeval lanac for precision radiotherapy.

00:09:57: And Kim Nissen mentioned Siemens etelica launching capillary testing.

00:10:00: Oh from just finger pricks

00:10:02: Exactly!

00:10:02: Twenty four assays form a fingerprint instead of massive blood draw.

00:10:06: All this is shifting the paradigm.

00:10:08: Dr.

00:10:09: Filippo Cadmartiri discussed the Preven SSEVD model reclassifying cardiovascular risk to one in five adults.

00:10:17: We are moving from predicting disease based on stats to actually looking for it with imaging.

00:10:22: And sometimes the innovations help in unexpected ways.

00:10:25: Switching Wang shared an amazing story about SmartHeart's automated free-breathing cardiac MR planning.

00:10:31: It allowed a successful scan For a deaf patient who couldn't hear the breath hold instructions.

00:10:35: The AI just adapted.

00:10:37: Good design creates accessibility, even in vet med.

00:10:40: Andrea Petraca noted MSKCT positioning is Just as critical of a dog named Flash

00:10:46: Yeah, Anja Helwitz reflected on Siemens' forty-year history of MRI gradient engineering.

00:10:52: Decades of foundational engineering got us here!

00:10:54: But you know none of this.

00:10:55: AI robotics or imaging scales without the unglamorous plumbing... The underlying infrastructure, cardiac pathways and regulations keeping everything safe.

00:11:03: It's wild to think we're applying health tech space travel but still struggle with basic security questionnaires and regulatory alignment Here On Earth.

00:11:11: Oh,

00:11:11: the plumbing is EVERYTHING.

00:11:13: Let's look at Cardiac AI At Scale.

00:11:16: Prisa Ashar detailed domain-adapted ECG foundation models screening for structural heart diseases.

00:11:22: But Javier Cisvivo pointed out that the clinical value of AI AFib prediction depends entirely on the horizon, meaning

00:11:29: what exactly?

00:11:30: Well predicting AFib one hour out versus five years out are two completely different clinical workflows.

00:11:36: if it's five years High accuracy doesn't help an ER doctor today.

00:11:40: Right, utility is the metric.

00:11:42: Ricardo Moralea highlighted the CardioSmart Assist AI framework for echocardiography and Stacey Kelleher, Dionne Harrison & Che Trivias all talked about the necessity of connected cardiac enterprise workflows in three-D echo imaging

00:11:55: because if data's fragmented you're operating blind.

00:11:58: Jonathan Morgan noted Metronics Triage HF Algorithm is helping NHS Wales detect worsening heart failure early using existing implanted devices.

00:12:06: To build that, you need the tech infrastructure.

00:12:09: Yossi Matias announced Google is transferring the open health stack to The Linux Foundation creating a vendor-neutral home for digital health building blocks

00:12:18: In Scott's sale explained Oracle OCI strategy of separating clinical ops from enterprise APIs so the local hospital keeps running even if internet goes down.

00:12:27: Then you hit the regulatory engine, which is a beast.

00:12:30: Asim Khan and Veronique Tordov contrasted the US total lifecycle approach with EU's AI Act and MDR periodic reports building data pipelines to satisfy both as massive engineering headache

00:12:42: And barrier entry getting steeper.

00:12:44: Joel Kent pointed out FDA five ten K fees are rising nearly ten percent.

00:12:48: for twenty-twenty seven plus.

00:12:50: you need reimbursement.

00:12:52: Claudia Vasco advocates for DG like reimbursement virtual human twins

00:12:56: because if a hospital can't bill for it, they won't buy it.

00:12:59: Emily Woodward highlighted that UAE adoption of BPH tech depends heavily on local reimbursement

00:13:04: context.".

00:13:05: That's why Danielle Niprash stresses the need for early-stage market access consulting.

00:13:10: Jim Blueston actually launched MedDeviceIQ just to teach this commercial intelligence.

00:13:14: But even with clearance, you have to pass security.

00:13:17: Larry Trotter II warned how enterprise security questionnaires expose weak AI governance and scattered policies.

00:13:23: Yeah it adds massive delays And Christina Martinez noted infusion pump safety relies on human factors engineering during procurement not just design.

00:13:32: If the purchasing department buys a pump that clashes with nursing workflows You introduce massive risk.

00:13:38: It really takes a broader ecosystem.

00:13:40: Ajay Park champions cross-disciplinary collaboration.

00:13:44: Liz Carnabucci discussed value based med tech in APAC, Paul Mearns covered Ireland's digital health ambitions and Jolera Alessandra highlighted Italy & Germany data sovereignty ecosystem.

00:13:55: Manish Chand is pushing this forward with the Digital Health Hub Pavilion at HLTHUSA And Robin Goldsmith rounds out care access with Dr.

00:14:03: Sarah Matz five tillers for borderless healthcare.

00:14:05: Sustainability is key too.

00:14:07: Gussman Young's Jacob highlighted Philips Blue Seal Helium Reducing MRI, which is huge for remote installs.

00:14:13: And Alexander Devanshun showed consumer infrastructure via Philips Sonicare AI toothbrushes.

00:14:18: It all comes back to accessibility.

00:14:20: A tool group shared a visit to NASA Mission Control with Major General Joseph Schmid.

00:14:25: They proved that the constraints of space medicine like remote monitoring and latency heavily overlap with earth-based telemedicine in imaging.

00:14:33: The same tech for astronauts can fix rural health care, it's incredible.

00:14:37: But you know we've talked a lot today about AI reclaiming time and standardizing workflows.

00:14:42: but here is the question to take back your teams.

00:14:45: once AI successfully optimizes all our clinical and administrative bottlenecks what do we actually plan with save?

00:14:52: Will we use it to see more patients, or will we finally use it?

00:14:56: listen the ones that have?

00:14:57: If you enjoyed this episode new episodes drop every two weeks.

00:15:00: Also check out our other editions on cloud insights and sovereignty digital products in services AI and agentic systems green ICT and sustainable AI ICT INTECH Insights and defense tech.