Gastroenterologist | Medical Director, Oshi Health | LinkedIn Top Healthcare Management Voice | Forbes Health Contributor

I’m a board-certified gastroenterologist and the Founding Medical Director at Oshi Health, where I’ve had the privilege of helping shape and expand a virtual-first GI care model now serving patients in all 50 states. Working alongside an incredible team, I’ve contributed to building scalable clinical pathways for IBS, IBD, and other gut conditions—anchored in team-based care, empathy, and measurable outcomes.

With 28,000 + followers, I’m among the most followed gastroenterologists on LinkedIn, where I regularly share insights on topics including:

• IBS & the gut-brain connection — emotional triggers, stress, and personalized care
• Digital health innovation — virtual-first models, AI in GI, and wearable integration
• Colorectal cancer screening — rethinking pathways and improving early detection
• Reducing unnecessary testing — CTs, labs, and low-yield procedures in IBS
• Value-based gastroenterology — lower cost, better outcomes, patient experience
• Health plan integration — improving patient access to covered virtual GI care
• Microbiome & nutrition science — evidence-based takes on fiber, probiotics, and gut myths
• Healthcare system reform — tackling disparities, burnout, and reimbursement models
• Behavioral science in GI — the role of habit loops, psychology, and coaching in digestive health
• Virtual care leadership — insights from helping scale a 50-state GI operation

As a LinkedIn Top Voice, I collaborate across digital health, clinical care, and health plans to drive forward-thinking solutions in gastroenterology—without waste, and with humanity.

Let’s connect if you’re working at the intersection of GI, tech-enabled care, or value-based delivery models.

Michael Bass, M.D.

Most of us have taken antibiotics and moved on. This new Sweedish study says your gut hasn't.

A new Nature Medicine paper linked 8 years of prescription data to stool metagenomes from nearly 15,000 adults in Sweden.

The headline finding: a single course of clindamycin, fluoroquinolones, or flucloxacillin was still associated with changes in gut bacteria 4 to 8 years later. Not months. Years.

Clindamycin use within a year of sampling was tied to 47 fewer detected species on average. That's a massive hit to an ecosystem most patients assume bounces right back.

The species that increased after antibiotics included bacteria previously linked to higher BMI, elevated triglycerides, and type 2 diabetes risk. The ones that decreased were tied to lower inflammation and better metabolic health.

This matters because the gut microbiome isn't a bystander in human health. It regulates metabolism, immune function, and inflammation. When you wipe out key species with a broad-spectrum antibiotic, you're not just treating an infection. You're reshaping an organ system.

And the recovery isn't automatic. The fastest rebound happened in the first two years. After that, diversity plateaued well below baseline. Some species may never fully return.

We prescribe a 10-day course for a dental abscess and tell the patient they're done. But the disruption we triggered may still be measurable half a decade later.

If you've had a patient who "just hasn't been the same" since antibiotics a few years ago, this paper gives you a biological reason to take that seriously.
We can't undo the antibiotics we've already prescribed. But we can stop pretending the consequences end when the pill bottle runs out.

undefined undefined

gastroenterology-specialists-of-delaware-llc

The Gut doesn't Forget 

LinkedIn Top Voice | Gastroenterologist I Global Medical Director @ ViomeDx

Dr Erika Tajra

The most dangerous phrase in cancer screening right now might be “no cancer signal detected.”

Multi-cancer blood tests are having a moment. A single blood draw that claims to detect 50+ cancers sounds revolutionary. And maybe someday it will be. But the evidence is still catching up to the excitement.

There are now roughly 40 multi-cancer early detection tests in development or already on the market. Yet very few have been tested in randomized trials. And the first large trial results? They didn’t show a clear reduction in late-stage cancers. Which is the whole point. Finding cancer only matters if you’re finding it early enough to change outcomes.

Why is this so hard? Because early cancers are biologically quiet. Most current tests search for fragments of tumor DNA circulating in the bloodstream. The problem is that very small tumors often shed almost no DNA at all. The signal can be incredibly faint. You’re essentially trying to hear a whisper in a stadium.

That’s why researchers are exploring another layer of biology: RNA. 

RNA reflects what cells are actively doing, not just what mutations exist. Signals can appear earlier, even when tumor DNA is scarce. It can capture the body’s response to cancer, not just the tumor itself. And it provides insight into active biological pathways rather than static genetic changes.

None of this means DNA-based tests are wrong. But it highlights something important: detecting cancer early may require listening to more than one biological signal. DNA is probably part of the answer. It’s not the whole story.

Patients deserve innovation. But they also deserve evidence that it actually improves outcomes. Because “no cancer signal detected” only means something if the test was capable of detecting the signal in the first place.

The AI prescribing debate feels very different after you’ve cleared 40 refill requests in your inbox at the end of a clinic day.

Most people think prescribing is the clinical decision. Clinicians know the reality is everything wrapped around it. Prior auths, quantity limits, step therapy rules, plans that only cover 90 pills, missing labs, drug interaction checks, and the quiet liability of deciding what can safely go through and what needs a visit.

That’s why a recent The Washington Post  opinion criticizing Utah’s AI refill pilot with Doctronic caught my attention. Joseph Sakran , Johns Hopkins trauma surgeon and a medical researcher, argues that autonomous AI prescribing is moving forward on dangerously thin evidence.

His case is straightforward. The safety data rests largely on a single preprint on medRxiv that hasn’t been peer reviewed. The study authors hold equity in the company. The underlying data hasn’t been publicly shared.
They also point out that the widely cited 99.2% accuracy figure comes from urgent care encounters, not chronic medication renewals, which is what the Utah pilot is actually doing. Different context, different risk profile.

And they raise a fair concern about scale. Even a small error rate applied across thousands of prescriptions could affect vulnerable groups. Elderly patients, people on complex medication regimens, patients whose conditions have quietly changed since their last visit.

Those are legitimate concerns. But the authors also acknowledge something that gets lost in the headline debate. Medication renewals for stable chronic conditions are among the lowest-risk prescribing tasks in medicine.

If AI could safely handle routine renewals, it could free physicians to focus on the stuff that actually requires a doctor.

That tension is where the real conversation should be. Because the system clinicians are currently working in is already broken. 

Practices average nearly 40 prior authorizations per physician per week. Physicians and staff spend more than 13 hours weekly on them. Nearly 90% say prior auth contributes to burnout.

So I find myself in the middle on this one. I don’t want AI replacing clinical judgment. But I’m very open to AI reducing the administrative friction wrapped around routine prescribing. Stable medication renewals with guardrails and escalation paths? That’s not AI practicing medicine. That’s AI doing paperwork.

Those are two very different things. And until we can talk about them separately, this debate is going to keep going in circles.

Byron Crowe

graphical user interface, text, application

Everyone says AI will replace doctors. But the first thing physicians are actually using AI for? Reading the medical literature.

A new AMA survey shows more than 80% of physicians now use AI professionally. And the most common use isn’t diagnosis. It’s summarizing medical research and standards of care. That tells you something important about where medicine actually is right now.

The biggest bottleneck in healthcare isn’t intelligence. It’s bandwidth. Thousands of new papers publish every week. Guidelines constantly evolve. Biology keeps getting more complex. No physician can realistically read everything. 

So knowledge diffuses slowly through the system. Conferences, guidelines, word of mouth. Sometimes it takes years for a new insight to reach everyday clinical practice. I’ve seen it firsthand. A study changes how we should think about a condition, and five years later half the clinicians managing it haven’t read it yet.

AI changes that dynamic. Instead of reading 20 papers, a physician can read a two-minute synthesis. Instead of discovering a new mechanism years later at a conference, it can surface in clinical workflow the same week. That doesn’t replace physicians. It amplifies them.

Which is why most real-world physician AI use today looks nothing like the headlines,despite what the Silicon Valley tech- bros might be tweeting. 

It’s summarizing research, generating chart summaries, creating discharge instructions, reducing documentation burden. Not autonomous diagnosis. Not robot doctors.

Caitlin Owens highlighted this in a recent Axios Healthcare newsletter. AI’s earliest clinical adoption is largely about helping physicians manage information and administrative load, not replacing clinical judgment. That distinction matters more than most people realize.

Medicine isn’t moving toward autonomous doctors. It’s moving toward augmented cognition. Doctors still make the decisions. But AI helps them see patterns faster in an ocean of information that no human brain can process alone. And as biology gets more complex, that ability won’t just be nice to have. It’ll be essential.

No alternative text description for this image

Novo Nordisk  just agreed to let hims & hers  distribute Wegovy and Ozempic through its platform and dropped the lawsuit it filed last month over compounded semaglutide. This isn’t a GLP-1 headline. It’s a preview of where telehealth is going next.

The first era of telehealth was about convenience. Fast visits, easy prescribing, consumer-friendly access. The next era will be judged on something much harder: can telehealth become a trusted distribution and care layer for complex chronic disease?

For years, telehealth has been underestimated as a lightweight front-end. Good for access, good for demand capture, good for meeting patients where they are. But this deal signals something bigger. Large pharma now sees digital care platforms as legitimate channels for branded therapy distribution. Not lead generators. 

Because once telehealth becomes a true distribution channel, the business model gets serious fast. It’s no longer about customer acquisition and conversion. It’s about clinical credibility, regulatory discipline, patient safety, adherence, supply continuity, and what happens after the prescription is written. 

Especially in obesity care, where anyone can build demand but very few can build durable care models. The companies writing scripts the fastest aren’t necessarily the ones that will survive this next phase.

There’s a second signal buried in this story too. Just days after U.S. regulators warned telehealth firms over misleading compounded GLP-1 advertising, one of the most visible players struck a deal around approved branded drugs instead. 

Telehealth is no longer disrupting healthcare from the outside. It’s being absorbed into the system itself. That creates more legitimacy and more accountability in the same move.

Giovanni Di Napoli is an accomplished Medical Devices Executive, with over 20 years of experience in Life Sciences of both entrepreneurial and corporate backgrounds.

After spending 12 years as President of Medtronic Gastrointestinal, Di Napoli now leads as Chief Executive Officer of Cosmo Pharmaceuticals, to further their mission of improving people's lives by developing products that address unmet medical needs in the fields of gastroenterology, dermatology, and healthtech.

When leading Medtronics Gastrointestinal (GI) Operating Unit (OU) Di Napoli achieved growth of a niche OU of two products to a global OU with ~20 products achieving significant contribution through implementing a direct sales model, restructuring teams;innovation as well as acquisitions.

As President of the GI OU, Di Napoli pioneered the implementation of AI in the GI field. Recognized by Forbes Magazine for innovative efforts in Mar '21, Di Napoli was the co-author of the article "The brave new world of artificial intelligence: dawn of a new era". 
Integral to the acquisition of Covidien by Medtronic, Di Napoli led the acquisition of the PillCam business worth ~$1BN.

With full responsibility over all GI operations, P&L; (Sales, Marketing, Quality, BD, R&D, Clinical, Regulatory; others) Di Napoli ensured equity of opportunity with 42.4% of diverse representation, leading consistently above corporate standards.
Within and outside of his current role, Di Napoli continues to raise awareness of colon and esophageal cancer, through collaboration with the Colon Cancer Alliance and Esophageal Cancer Association, giving back to the public community as part of Medtronics STEM project and personally providing support to children's education.
Di Napoli also has strong professional followership through both in-person and virtual contributions, with over 30,000 LinkedIn followers and actively publishing articles. 

Di Napoli won the ASGE Crystal Award for contributions to the field of GI and is also a frequent speaker at conferences in Silicon Valley and medical device and tech summits globally. 
Currently, Di Napoli is an executive external board member of company Palliare and has led longstanding scientific and business development collaborations with Mayo Clinic, Amsterdam UMC, Northwestern Chicago, Rush Chicago, Stanford, Johns Hopkins Maryland, NHS UK, Amazon, AWS, NVIDIA, SFDC.

Di Napoli is also a former professional basketball coach, who is dedicated to being an innovative and passionate leader, and determined to deliver value to shareholders, employees, and colleagues.

Giovanni Napoli

Before we accelerate, we align.

This week, we did exactly that.

We held our FY26 Leadership Team meeting and aligned on our North Star. Alignment at the top sharpens our direction and allows us to challenge assumptions and, most importantly, agree on what matters most. When leaders are not aligned, execution fragments. We do not allow that.

With our new mission of Building Health Confidence, the bar rises. Moving ahead, this will define how we assess every investment and priority.

Every decision must support patients.
Every investment must strengthen outcomes.
Every leader must own execution.

We have a direction; now we deliver. 

FY26 begins with this team!

Giovanni Di Napoli

LinkedIn Top Voice | CEO, Cosmo Health Confidence | Harnessing A.I. to Improve Patient Outcomes | Health Equity Advocate | Empowering Teams

This headline deserves more attention: Jeff Bezos is raising tens of billions to buy industrial companies disrupted by AI.

I think this says a lot about where AI is headed.

Most conversations we hear focus on models and productivity tools. Bezos is looking at it from another perspective: manufacturing and supply chains. That is where complexity sits and where large gains will come from.

His strategy is simple: buy companies and invest in AI to help them operate. Looking at the bigger picture, AI stops being a feature and becomes infrastructure.

I think we will see the same move across many industries. The next phase of AI will not sit on top of systems, but reshape how those systems run.

That is a much bigger opportunity. 

https://lnkd.in/e3w8pPUB

This headline caught my attention:

“Tesla plans $20 billion capital spending spree in push beyond human-driven car”.

Interestingly, Tesla plans to reduce Model S and Model X production in California and use that factory space to build Optimus humanoid robots. That decision stood out to me. Factory space is limited, and companies protect it carefully. When leadership reallocates it, to me, this reflects a clear priority.

A closer look tells us that Tesla plans to invest $20 billion in robotics and AI infrastructure.

This shows how the company is building for its next phase, not only improving existing products.

This raises an important question: is Tesla scaling robots as a supporting capability or positioning robotics as a core product category?

Open article: Tesla plans $20 billion capital spending spree in push beyond human-driven cars by reuters.com

Tesla plans $20 billion capital spending spree in push beyond human-driven cars

Something interesting appears to be happening in manufacturing right now is the push toward humanoid robots to fill labour gaps.

Companies like Agility Robotics are deploying their robot Digit in warehouses and factories to handle repetitive physical tasks that many companies struggle to staff. Large firms including Amazon and Toyota have already begun testing or deploying the systems. 

The driver behind this is fairly simple: many industrial economies are dealing with aging workforces and labour shortages, particularly for physically demanding roles. 

So the interesting question may not be whether robots replace workers.

It may be how robotics becomes part of the workforce in sectors where companies simply can’t find enough people to do the job.

Open article: These robots are coming for the jobs no one wants — and could fill workforce gaps by businessinsider.com

Agility Robotics is building humanoid bots to address a labor gap in the manufacturing industry, which is seeing vacancies and an aging workforce.

These robots are coming for the jobs no one wants — and could fill workforce gaps

taplio

The Battle

Michael Bass, M.D.

Giovanni Napoli

Followers

Posts

Engagements

Likes

Performance ScoreMichael Bass, M.D.'s LinkedIn Performance

Performance ScoreGiovanni Napoli's LinkedIn Performance

Ready to grow your LinkedIn presence?

taplio

The Battle

Michael Bass, M.D.

Giovanni Napoli

Followers

Posts

Engagements

Likes

Performance Score.css-goub6s{font-size:var(--chakra-fontSizes-md);font-weight:500;color:var(--chakra-colors-gray-600);display:block;margin-top:var(--chakra-space-1);}Michael Bass, M.D.'s LinkedIn Performance

Performance ScoreGiovanni Napoli's LinkedIn Performance

Ready to grow your LinkedIn presence?

Performance ScoreMichael Bass, M.D.'s LinkedIn Performance