How New Capabilities Enter Mainstream GI
By Dr. Charles Accurso and Praveen Suthrum
This chapter began with a question from Charley that is deceptively simple: if gastroenterology is changing so quickly, how does a real GI practice turn that change into something it can actually implement?
For years, the conversation about the future of GI has focused on what is coming. Scope Forward explored forces that could reorganize the specialty. The Shift went further into artificial intelligence, non-invasive diagnostics, virtual care, robotics, new care models, and companies already building around them. The technologies were no longer theoretical. The future had begun to show up.
What remained much less clear was where all of this belonged inside a functioning gastroenterology practice.
That distinction matters because a GI practice does not operate on a blank sheet of paper. There is already a full schedule, an ASC, an EHR, referring physicians, payer contracts, employees, partners, prior authorizations, recalls, phone calls, faxes, no-shows, and patients who need care today. When somebody arrives with a new diagnostic test, a virtual-care model, an AI platform, a metabolic program, or another promising service line, the practical question cannot simply be whether the innovation is exciting.
The question is how it enters the practice without destabilizing what already works.
That is why this chapter is the bridge in this book. The earlier chapters examined where GI creates value, what every practice should fix first, which assets it already owns, how procedures can become pathways, and what foundational infrastructure is necessary. Now we have to connect those ideas to the economic model.
How does innovation become care? How does care become a workflow? How does the workflow become sustainable? And how does a practice do all of this without simply carrying the operating model of GI 1.0 into the future?
Those questions are at the heart of GI 2.0.
Form Follows Function
As our discussions evolved, Charley repeatedly brought us back to a basic principle: form follows function.
Before a practice chooses a technology, hires a dietitian, contracts with a virtual-care company, creates a chronic-care program, or searches for a reimbursement mechanism, it should be able to answer a more fundamental question.
What are we trying to accomplish?
Does the proposed model improve patient care? Does it improve quality or outcomes? Does it make the patient experience better? Does it expand access? Does it help diagnose disease earlier, prevent progression, or keep a patient from falling through a gap? Does it reduce unnecessary testing, emergency visits, hospitalization, administrative work, or other avoidable costs?
If it does none of these things, the practice should ask why it is building the program at all.
This sounds obvious, but healthcare frequently develops in the opposite direction. A reimbursement opportunity appears and organizations construct a program around it. A technology becomes popular and practices search for places to deploy it. Another group launches an ancillary and competitors begin asking whether they should do the same.
The form begins determining the function.
That becomes especially dangerous now because there are so many forms available. A longitudinal-care program could involve a gastroenterologist, an APP, a dietitian, a behavioral-health professional, a care manager, a virtual company, an AI agent, remote monitoring, asynchronous communication, or some combination of them. The model could be financed through traditional visits, an existing care-management mechanism, a payer arrangement, an employer contract, a bundled model, a patient-paid component, or something new that has not yet become common.
Those are design choices. They are not the purpose of care.
The sequence should begin with the patient problem. Once the function is clear, the practice can design the care model, determine how it should operate, decide which capabilities belong inside the practice and which can come from elsewhere, identify who receives economic value when the pathway works, and then build the financial architecture that can sustain it.
The economics matter enormously. A program that consistently loses money will eventually disappear, however clinically elegant it may be. But the economics should support the function rather than dictate it.
A billing code can finance an innovation. It should not define the innovation.
That distinction became particularly important as we studied longitudinal-care models. Chronic Care Management, Principal Care Management, remote monitoring, traditional visits, and related mechanisms may have legitimate roles for appropriately eligible patients. They also come with clinical, documentation, staffing, payer, and compliance requirements that evolve over time.
The durable asset is therefore not the code.
It is the care model.
Clinical architecture comes first. Economic architecture follows.
Choosing Where to Play
GI practices are no longer short of opportunities. They have the opposite problem.
A practice could pursue colorectal cancer screening completion, fatty liver, metabolic care, virtual GI, IBS and IBD programs, clinical trials, advanced diagnostics, AI-enabled operations, remote monitoring, data partnerships, gut-brain care, new procedural technologies, microbiome-related services, employer programs, or several other ideas that will appear after this book is published.
Trying to pursue all of them would be a mistake.
A practice needs a way to decide.
Years ago, we used a simple framework in GI Mastermind called the Demand Matrix. The horizontal axis represented customer demand and the vertical axis represented business impact. Opportunities fell into four quadrants: Broken, Thankless, High End, and Golden Goose.
The framework remains useful, but it needs to evolve for GI 2.0.
Before an opportunity even enters the matrix, it should pass what we call the Patient-Purpose Gate.
Does this create legitimate value for the patient?
If the answer is weak, projected revenue should not rescue the idea.
After that comes demand. Some demand is visible. Practices can see it in referrals, patient volume, access problems, physician conversations, wait lists, and the conditions taking up more of the schedule. A group may notice rising numbers of patients with elevated liver enzymes, obesity-related digestive disease, functional GI disorders, or another clinical category. It may also see traditional referral patterns changing as non-invasive testing, consumer behavior, or new technologies alter the way patients enter gastroenterology.
Those are important signals.
But demand can also exist before the patient knows to ask for anything. A patient cannot request a coordinated colorectal cancer screening pathway if such a pathway does not exist. A patient with fatty liver may not ask for longitudinal liver management because the patient has been told only that an ultrasound showed “fatty liver.” An IBS patient may not know that structured nutrition and gut-brain behavioral care could be available.
GI 2.0 therefore has to look at both current demand and emerging demand.
The impact axis also needs to become broader than immediate revenue. A capability may generate margin, but it may also reduce cost, create physician capacity, strengthen referral relationships, improve outcomes, protect existing assets, create useful data, improve payer relationships, differentiate the practice, or position the organization for where GI is heading.
For GI 2.0, the vertical axis is better understood as clinical and strategic practice impact.
The old quadrants still tell us something valuable. Broken opportunities have little meaningful demand and little impact. Thankless opportunities have demand but an unattractive or inefficient delivery model. High-End opportunities may have lower current demand but significant clinical or strategic value. Golden Goose opportunities combine substantial demand with strong impact.
The important difference is that opportunities are no longer fixed in one quadrant.
They move.
IBS is a good example. Under the traditional physician-centric model, a patient may repeatedly return for short episodic visits while much of the actual need lies in nutrition, behavioral support, education, reassurance, and longitudinal symptom management. Demand is high, but the model can be frustrating for the physician, the patient, and the business. That is a classic Thankless category.
Change the operating model and the economics can change with it.
Fatty liver is another example. What was once frequently an evaluation followed by advice to lose weight is becoming a much larger clinical opportunity as disease burden, diagnostic capability, treatment options, metabolic medicine, and referral volume evolve.
The opposite can also occur. A Golden Goose today can move downward as technology, payment models, or patient behavior changes. Screening colonoscopy remains central to gastroenterology, but the pathway through which patients arrive for screening is changing. A practice that assumes yesterday’s demand pattern will remain permanent is taking a strategic risk.
That is why the GI 2.0 Demand Matrix needs arrows, not just dots.
[Figure 6.1: The GI 2.0 Demand Matrix — Patient-Purpose Gate → Demand × Clinical/Strategic Impact → Implementation Readiness]
The first visual should show the framework itself. A second can show an illustrative GI practice with different opportunities plotted across the matrix and arrows indicating how they might move over the next several years.
[Figure 6.2: Illustrative GI 2.0 Opportunity Map — example only, not a universal ranking]
That second figure is deliberately illustrative. A six-physician independent group should not make the same choices as a national platform. A rural group with limited specialist access may evaluate virtual care differently from a practice next to a major academic center. A group with a strong research infrastructure may see clinical trials as a major opportunity, while another may struggle to support them. A practice operating under value-based arrangements may evaluate longitudinal care differently from a purely fee-for-service group.
There is no universal GI 2.0 portfolio.
There is a disciplined way to create your own.
Readiness Before Ambition
An attractive position on the Demand Matrix does not mean a practice is ready to act.
This is where good ideas often die.
A practice may have strong patient demand, a compelling clinical rationale, and an attractive economic hypothesis but lack the capacity to implement the program. The physicians may have no time. Nobody may truly own the initiative. Staff may already be overwhelmed. The partner may not be ready. The workflow may cross three departments that rarely communicate.
Implementation readiness therefore needs to sit on top of the matrix.
Some opportunities are ready now. Some deserve a selective pilot. Others should be watched and prepared for. A few should be rejected.
There is an important distinction between rejection and deferral. An opportunity should be rejected when patient value is weak, clinical risk is inappropriate, utilization would be unnecessary, or the model creates unacceptable ethical or compliance concerns. An opportunity should be deferred when the idea is sound but the practice lacks time, leadership, capability, partner support, or organizational momentum.
The greatest implementation mistake, however, is starting too much.
Innovation meetings easily produce a long list of attractive projects. Every physician has another idea. Every vendor has another use case. Soon the group has four or five pilots, no clear owner, irregular meetings, and no one with enough time to make any single initiative work.
We would rather see a practice complete one meaningful project than announce five.
The first project needs a physician champion, but not merely a physician who likes the idea. It needs someone prepared to work through clinical disagreements, exceptions, staff concerns, workflow problems, partner issues, and the inevitable friction of implementation.
Lip service is not leadership.
The project also has to be small enough to finish. The workflow does not need to be perfect before launch; that is one of the purposes of a pilot. But the patient population, responsible people, measures, milestones, and recurring review need to be explicit.
The three teaching cases that follow demonstrate different forms of innovation. Colorectal cancer screening shows ecosystem orchestration. Fatty liver shows longitudinal specialty ownership. IBS and IBD show how the delivery architecture itself can change.
Together, they demonstrate what we mean by integration.
Teaching Case 1: Colorectal Cancer Screening Completion
Stop Fighting the Modality. Complete the Journey
Colorectal cancer screening sits directly at the intersection of the traditional GI business model and a changing patient journey.
We need to establish the clinical purpose first. The purpose of colorectal cancer screening is to prevent colorectal cancer and save lives. It is not to preserve colonoscopy volume.
That does not mean the economics are irrelevant. Screening colonoscopy and ASC ownership remain important to many GI practices, and changes in screening patterns can have real financial consequences. But if the practice begins by defending one modality against another, it can miss the larger strategic opportunity.
The better question is: how do we help every appropriate patient complete the right screening pathway?
Once we ask that question, the fragmentation becomes obvious.
Primary care may recommend screening, but a recommendation is not completion. A patient may agree to a non-invasive test and never return it. A positive stool-based test may not lead reliably to colonoscopy. A GI referral may arrive but never become a scheduled appointment. A colonoscopy may be scheduled and then cancelled because of transportation, preparation, medication, insurance, or medical-clearance problems. A patient with prior polyps, family history, symptoms, or other risk factors may enter an inappropriate pathway because important information was not available when the screening decision was made.
Every participant sees a different part of the journey. The PCP sees the recommendation. The diagnostic company sees whether a test is completed. The GI practice sees referrals and procedures. The ASC sees the scheduled case. The payer sees claims.
The patient is often the only person moving through the entire system.
That is the opportunity.
Imagine that a GI practice and several of its important primary-care referral partners explicitly define the local screening journey. The physicians agree on the clinical logic: when specialist evaluation or colonoscopy should take priority, when non-invasive screening is reasonable, what information matters, what constitutes a red flag, and what should happen after an abnormal result.
Technology can then help operationalize the agreed pathway.
An AI-supported workflow could identify an incomplete step, reach out to the patient using physician-approved language, explain why follow-up matters, identify logistical barriers, support scheduling, and return unresolved clinical questions to the appropriate human. If a non-invasive test was selected but never completed, the system could follow up. If the result requires colonoscopy, the pathway could help close the loop. If a procedure is scheduled, the same infrastructure could help surface preparation, transportation, authorization, or clearance problems before they become cancellations.
AI does not decide which screening modality a patient should receive.
Physicians define the pathway.
AI provides persistence.
Over time, the patient-facing layer could become even more valuable. Instead of receiving disconnected communications from several organizations, the patient could have a longitudinal navigator that understands where the patient is in the screening journey and what should happen next.
The technology becomes the thread through a process that is currently fragmented.
This model also reframes the relationship between GI and non-invasive diagnostics. The strategic question no longer has to be whether one modality competes with another. The practice can help organize the architecture within which the appropriate patient reaches the appropriate test and, most importantly, completes the journey.
The same principle applies to the payer. Payers and physicians have legitimate points of friction, but colorectal cancer prevention creates meaningful alignment. Everyone benefits when appropriate screening is completed and abnormal findings are followed appropriately.
Historically, one difficulty with these relationships has been that the data lag behind the care. Claims arrive later. Outcomes are difficult to see in real time. Care gaps are known in the abstract but difficult to close patient by patient.
A better intelligence layer makes a different conversation possible. Instead of merely telling a payer that a program should work, the practice can increasingly show who was eligible, where patients dropped out, who was re-engaged, who completed screening, who required follow-up, and where the remaining gaps are.
The opportunity is therefore larger than any individual test.
It is the completed patient journey.
For most practices, the appropriate starting point is local and narrow. One referring group. One cohort. One or two points of failure. Learn where the friction lives, prove that the pathway can be closed, and expand from there.
That is how an innovation becomes part of mainstream GI.
Teaching Case 2: Fatty Liver and Metabolic GI
Turn an Existing Referral Into a Longitudinal Pathway.
Fatty liver presents a different opportunity because the patients are already entering gastroenterology.
A common journey begins with abnormal liver enzymes discovered during routine testing. Primary care may obtain imaging that suggests fatty liver, and the patient is referred to GI. The gastroenterologist evaluates the abnormal tests, considers other causes of liver disease, assesses risk, and determines whether more advanced disease may be present.
Most GI practices know how to do this part.
The uncertainty often begins after the diagnosis.
Who owns the patient longitudinally? Who determines when liver risk should be reassessed? Who manages liver-directed treatment? Who helps the patient address weight and metabolic risk? Who provides nutrition support? Who decides when a lower-risk patient can be seen less frequently? Who recognizes progression to advanced fibrosis or cirrhosis? How does all of this remain connected with a primary-care physician who may already be managing diabetes, hypertension, lipids, obesity, and cardiovascular risk?
The real problem is not simply diagnosis.
It is ownership.
The first principle of a GI 2.0 fatty-liver pathway is therefore straightforward: GI owns the liver, not the entire patient.
That means the GI practice takes responsibility for liver-specific evaluation, risk assessment, treatment decisions, reassessment, and escalation. Primary care continues to manage the broader cardiometabolic patient unless the clinicians explicitly decide otherwise.
That boundary matters because duplication creates confusion. A patient may already be receiving a metabolic therapy through primary care, endocrinology, or obesity medicine. It makes little sense for GI to create a competing treatment process simply because fatty liver has been identified. At the same time, when liver-specific treatment originates within the GI pathway, responsibility for monitoring and follow-up needs to be clear.
The level of GI ownership should also change with risk.
A lower-risk patient may need a defined plan and periodic reassessment while most broader metabolic management remains with primary care. An intermediate-risk patient may benefit from a more structured APP-supported longitudinal pathway inside GI. A patient with advanced fibrosis, suspected cirrhosis, progressive disease, or greater uncertainty needs substantially closer specialist involvement.
The program should therefore be risk-stratified rather than one-size-fits-all.
This is what makes fatty liver a particularly interesting GI 2.0 opportunity. The practice does not have to create demand. The patients are already arriving. The first program can be constructed around work that is already being performed, but performed more consistently and longitudinally.
That also means the practice should resist beginning with the most elaborate version.
It is easy to imagine a comprehensive metabolic center with extensive testing, nutrition, exercise support, behavioral care, digital monitoring, several professionals, and optional patient-paid services. Such a program may eventually make sense in some settings. It can also create a high barrier to entry and overwhelm the organization before it understands what patients actually need.
The better question is: what is the smallest useful program?
For many practices, the answer may be surprisingly modest. One physician champion. A clearly defined pathway. Appropriate risk assessment. APP-supported follow-up. A process for reassessment and escalation. Explicit communication with primary care.
Additional capabilities can be added when the need becomes clear.
One of the most important moments in the pathway is what we call the risk-clarity moment. Fatty liver can sound abstract to a patient. The ultrasound says steatosis. The enzymes are mildly abnormal. The patient feels fine.
The physician has an opportunity to translate the diagnosis into something actionable without turning the conversation into fear. Where does the liver appear to be today? What could progression mean? What can the patient do? What should be monitored? When should reassessment occur?
When clinically appropriate, non-invasive fibrosis assessment can help make that discussion tangible. The point is not to sell a test. It is to make an otherwise abstract risk understandable enough that the patient can engage with the pathway.
The longitudinal work begins after that.
This is where AI can be useful without pretending to be the doctor. A first version could support enrollment, approved education, voice or text check-ins, simple patient-reported information, weight tracking where appropriate, reminders, and identification of barriers. More sophisticated versions can help staff see which patients are stable, which are disengaging, which have not completed an expected step, and which need human attention.
The same infrastructure can provide three different experiences. The patient feels that somebody is still paying attention. The staff sees an actionable worklist instead of an undifferentiated population. The physician receives a concise longitudinal picture rather than another stream of raw data.
The physician continues to make the clinical decisions.
AI keeps the pathway alive between those decisions.
The economic model should remain flexible. Some care will be supported through ordinary specialist visits and testing. Selected patients may appropriately fit existing care-management arrangements. Some practices may develop nutrition or metabolic capabilities internally. Others may use an external partner. In some markets, payer-supported arrangements may eventually become attractive. Optional enhanced services may also exist, provided they remain clearly distinct from medically necessary covered care.
There is no reason to decide in advance that one financing mechanism is universally correct.
The practice should first determine what care its patients need and what it is capable of delivering.
It should also talk with primary care before building too much.
This becomes particularly important in markets where primary-care organizations are responsible for total cost of care. A program that appears to create value from the GI perspective may look duplicative to a PCP if both sides have not clarified their responsibilities.
Even outside value-based arrangements, referral relationships benefit from explicit communication. Many groups may discover that a relatively small number of referring practices account for a substantial share of fatty-liver referrals. A brief conversation with those physicians can clarify what they want GI to own and what information they want back.
Who will manage existing metabolic medications? When should the patient return to primary care? Which patients should remain under GI surveillance? What should trigger earlier re-evaluation? What does primary care actually need to know from a dietitian, virtual partner, or longitudinal program?
The solution on day one may simply be a better progress-note summary.
If physicians later decide they want automated communication or shared data, technology can support that.
The technology should follow the clinical relationship, not define it.
The lesson from fatty liver is not that every GI practice should open a metabolic center.
It is that a common referral can become a structured longitudinal pathway when the practice defines ownership, risk, follow-up, communication, and escalation.
The practice owns the pathway.
It does not need to perform every task inside it.
Teaching Case 3: IBS, IBD, and the Distributed GI Practice
Make the Patient Visible Between Visits.
Our discussion of IBS changed this chapter more than any other case.
At first, the opportunity appeared relatively straightforward. GI practices see large numbers of IBS patients. Many need nutrition or gut-brain behavioral care. Perhaps the solution was to add a dietitian, establish access to a psychologist, use a virtual partner, find a reimbursement mechanism, and create another service line.
That framing turned out to be too narrow.
Charley brought us back to first principles. What are we trying to improve?
The patient should understand the diagnosis and the condition. Symptoms and quality of life should improve. In more severe cases, better longitudinal care should help reduce the recurring cycle of urgent-care visits, emergency-department evaluations, repeated imaging, and repeated diagnostic workups that can occur when symptoms flare and the patient has nowhere else to turn.
Traditional GI already performs part of this journey reasonably well. The patient is evaluated. Serious disease is considered. Appropriate testing is performed. A diagnosis is made. Medical treatment is offered.
What often disappears is what happens afterward.
Nutrition. Gut-brain behavioral care. Psychological support. Education. Reassurance. Symptom tracking. Medication response. Adherence. Questions that arise between appointments.
And, more fundamentally, visibility.
The visit is an event.
The disease is continuous.
Farid Razavi, a practicing gastroenterologist who founded Etra Health, described the problem as the large amount of clinically meaningful activity that occurs between GI visits. A patient may spend a short period with a gastroenterologist and then not return for months. During that interval, bowel habits change, medications work or fail, diet changes, stress increases, sleep deteriorates, symptoms fluctuate, and portal messages accumulate.
At the next visit, the physician may spend much of the encounter reconstructing what happened.
What if the patient did not become invisible?
Etra’s model is one example of trying to solve that problem. Patients can report symptoms and other longitudinal information between visits, while technology helps organize the accumulated activity into something the physician can actually use. The important idea is not the company or a particular billing mechanism. It is that the next physician encounter can begin with a longitudinal picture of what happened rather than a memory exercise.
That changes the unit of care from the visit to the journey.
Access Has to Be More Than an Appointment
This is also where virtual GI becomes much more than telehealth.
In our discussion, Sameer Berry of Oshi Health described access using three useful concepts: care needs to be fast, frequent, and complete.
Traditional gastroenterology is not designed to provide all three for every chronic GI patient.
Fast access is difficult when specialist waits are long. Frequent access is difficult when every interaction requires another physician appointment. Complete access is difficult because many patients need expertise beyond the gastroenterologist, particularly nutrition and gut-brain behavioral care.
That is not a failure of gastroenterologists.
It is a limitation of the operating model.
A specialist whose highest-value work includes diagnosis, complex judgment, medication decisions, serious disease management, and procedures should not necessarily spend every hour personally delivering every component of longitudinal support.
The better question is how to construct the right care team around the patient.
The 2026 assessment from the Peterson Health Technology Institute provides a helpful way to think about this. It broadly distinguishes between virtual models that wrap services such as nutrition, behavioral health, navigation, and symptom support around an existing clinical relationship, and clinician-led models that integrate GI clinicians more directly into the virtual multidisciplinary team.
More importantly, the report reinforces that different GI diseases require different levels of care intensity.
For IBS, structured nutrition and gut-brain behavioral interventions can be important parts of effective care. Many patients do not require continuous gastroenterologist involvement once the diagnosis and appropriate medical plan are established.
IBD is different.
Crohn’s disease and ulcerative colitis can require objective monitoring of inflammation, complex medications, procedures, infusion or specialty-pharmacy coordination, and rapid specialist decisions when disease activity changes. Nutrition and behavioral care may add real value, but they do not replace management of the underlying inflammatory disease.
This leads to another principle for GI 2.0:
Do not overbuild simple care, and do not underbuild complex care.
The goal is not to surround every patient with the largest possible team.
It is to create the minimum sufficient care architecture that reliably produces the outcome.
For one IBS patient, that might mean appropriate GI evaluation followed by structured nutrition support and a gut-brain intervention, with the gastroenterologist available for escalation. Another patient may need an APP, behavioral-health clinician, dietitian, and more frequent medical input. An IBD patient may need the full specialist-led longitudinal pathway, with multidisciplinary services supporting rather than replacing gastroenterology care.
The form follows the function.
The Practice Does Not Have to Hire Everyone
The virtual-GI market also teaches an important lesson about partnership.
There is not one virtual-care model.
One approach is to send the patient to a separate organization that provides additional support around the existing medical relationship.
Another is a clinician-led virtual model in which the outside organization manages a larger portion of GI care.
A third approach is increasingly important for community practices: embed the capability around the local GI relationship rather than bypassing it.
Oshi’s evolution illustrates one version of this. Sameer described an earlier model in which patients could move between the virtual organization and local GI practices. Although the clinical rationale was understandable, the local practice could still worry about losing visibility, losing the patient relationship, and receiving limited operational or economic benefit.
Oshi has subsequently developed more embedded approaches that allow portions of its virtual clinical capacity and multidisciplinary support to function more closely with participating GI practices.
Salvo Health illustrates another version of the same larger idea. Its model was deliberately designed around the local gastroenterologist rather than around bypassing the community practice. The outside organization supplies longitudinal capabilities such as nutrition, behavioral-health support, care-management staff, and technology, while the local GI relationship and physician authority remain central.
That distinction is strategically important.
A GI practice already owns valuable assets that a digital company cannot simply manufacture: the trust of existing patients, relationships with local referring physicians, clinical credibility, access to procedures, knowledge of the local market, and years of accumulated presence in the community.
Virtual care can bypass that asset.
It can compete with it.
Or it can strengthen it.
A practice should know which of those three things is happening before it signs a partnership.
There is therefore a third option between “hire everybody ourselves” and “refer the patient somewhere else.”
The practice can extend itself.
This is what we mean by an embedded capability.
It also gives us another way of thinking about longitudinal care. The gastroenterologist does not personally need to become continuously available to every patient.
The practice does.
Longitudinal care can be provided through physicians, APPs, nurses, dietitians, behavioral-health professionals, embedded partners, asynchronous communication, and AI-supported workflows. The gastroenterologist remains available for the work that genuinely requires gastroenterology judgment.
That is a much more scalable definition of access.
The Economic Bridge Is Not the Destination
IBS also forced us to look more carefully at the economics.
At first, it was easy to assume that longitudinal-care companies were largely variations on the same reimbursement strategy. They are not.
Some provider-facing models use existing care-management or monitoring mechanisms to make longitudinal services financially feasible inside a GI practice. Salvo is one example. Etra currently approaches the problem from another angle but similarly views existing reimbursement as a way to finance more structured between-visit care.
Oshi represents a different path. Sameer described beginning with the desired clinical model, developing evidence around that model, and working with commercial payers to create economic arrangements that could support a more intensive multidisciplinary service.
These approaches should not be collapsed into a single category.
Instead, they help reveal the difference between an economic bridge and a strategic payment model.
An economic bridge answers a practical question: how do we finance the first useful version of this care model?
For one practice, ordinary specialist reimbursement may be enough. Another may appropriately use existing care-management mechanisms for qualifying patients. A partner may supply capabilities that the group cannot yet afford to build internally. The organization may invest some of its own capital. An existing payer benefit may provide another route.
These mechanisms can make experimentation possible.
The strategic payment question is different: if the pathway works, who receives enough economic value from the result to want to finance it at scale?
Farid described that distinction clearly. Existing care-management reimbursement can support the infrastructure required to improve between-visit care. But if the model later demonstrates better symptom control, fewer emergency visits, less unnecessary imaging, fewer hospitalizations, or lower total cost, the more important long-term relationship may be directly with the payer.
Oshi began closer to that end of the continuum by developing payer-supported arrangements around its broader clinical model.
Both perspectives are useful.
Starting with a billing mechanism and never looking beyond it can trap the practice inside reimbursement-based medicine.
Waiting indefinitely for a perfect value-based contract can prevent a useful program from ever getting started.
The practical answer is to know which phase you are in.
Use an economic bridge when it legitimately finances a useful clinical model. Measure what the model actually does. If the value becomes large enough, build a more durable economic architecture around it.
This leads to one of the most useful questions in this chapter:
Who benefits economically when the patient gets better?
For IBS, the payer may benefit when better longitudinal management reduces repeated evaluation, emergency care, and other avoidable utilization. An employer may benefit when symptoms interfere less with work.
For IBD, preventing even a small number of major flares or hospitalizations can create substantial economic value.
For colorectal cancer screening, value can accrue to the patient, GI practice, primary care organization, diagnostic company, payer, and health system in different ways.
For fatty liver, the value may be distributed across GI, primary care, the payer, and the broader metabolic-care ecosystem.
This is a more strategic question than asking only which code can be billed.
The future revenue opportunity may come from proving to the party carrying the economic burden that the pathway solves the problem better.
The Payer Needs Proof
This changes the payer conversation.
Traditional negotiations are often organized around rates. The practice argues that reimbursement should be higher. The payer argues about cost. Both sides are usually working with claims information that arrives well after the clinical activity.
A longitudinal pathway creates the possibility of a different conversation.
Here is the population.
Here is the gap.
Here is the clinical pathway.
Here is who completed care.
Here is what happened to symptoms or disease control.
Here is what happened to utilization.
Here is what the program cost.
Now the practice is not merely asking to receive more money for performing the same set of events. It is demonstrating that it can organize care differently.
The PHTI assessment is useful precisely because it asks more than whether patients enjoy virtual care. It evaluates clinical effectiveness and economic impact while also acknowledging that the evidence base remains young and that longer-term, higher-quality measurement is still needed.
That is the right standard.
Enrollment is not enough.
App usage is not enough.
Engagement by itself is not enough.
Eventually, the question becomes whether the patient became better and whether the system became more efficient.
GI practices that can answer those questions will approach payers from a very different position.
AI Is the Continuity Layer, Not the Care Model
AI appears naturally in every case in this chapter, but this is not the AI chapter.
That comes later.
Here, AI matters because it changes what is operationally possible.
In colorectal cancer screening, AI can provide persistence across a fragmented journey by conducting outreach, identifying barriers, helping with scheduling, tracking incomplete steps, and moving unresolved issues to the right human.
In fatty liver, it can support enrollment, education, longitudinal check-ins, simple monitoring, reminders, and summaries.
In IBS and IBD, it can help make the patient visible between visits by organizing longitudinal information, identifying patterns, and directing expensive human attention toward the moments where it is most valuable.
There is a common architecture underneath all three examples. The patient generates information. AI organizes the workflow-relevant parts of that information. The right human becomes involved when clinical judgment, empathy, behavior change, treatment, or escalation is required.
Farid’s description of Etra is useful here because AI is not presented as the care itself. It helps turn large amounts of longitudinal activity into something clinically useful.
Jeff Glueck’s description of Salvo adds another important nuance. In behavioral change, nutrition, chronic symptoms, and gut-brain care, the human relationship can itself matter. AI can scale education, personalization, tracking, coordination, and repetitive work, while clinicians provide judgment, trust, empathy, and the nuanced support required to help patients actually change behavior.
So the principle is broader than simply saying that AI handles volume and humans handle judgment.
AI handles volume and continuity. Humans provide judgment, trust, empathy, and behavior change.
That balance will evolve. Chapter 8 will examine what happens when these capabilities are applied across the entire practice.
For this chapter, the lesson is simpler.
Do not add AI to a poorly designed pathway and expect the pathway to become intelligent.
Design the care first.
Then use AI to make it scalable.
Fund the Future Before You Build It
There is another practical constraint on innovation: it consumes management attention and operating capacity before it creates anything.
That is why the foundational work in the previous chapter matters.
A practice struggling with unanswered phones, manual referral intake, inefficient scheduling, weak recalls, repetitive eligibility work, preventable denials, and overloaded staff will find it difficult to launch a sophisticated longitudinal-care program.
Fortunately, many of those operational problems are now among the most practical uses of AI and automation.
The reason to begin there is not simply to make the practice look modern. Better operations can create the financial and organizational capacity required to build the next model.
If automation reduces repetitive administrative work, capacity increases. If a better referral workflow converts patients who were previously lost, the practice captures revenue from assets it already possesses. If recall automation brings overdue patients back into care, both access and utilization improve. If eligibility, prior authorization, denials, appeals, and other revenue-cycle processes become more efficient, margin improves without creating a new clinical service line.
Those gains can finance more uncertain innovations.
AI-enabled operations therefore function as a funding engine for GI 2.0.
The first innovation a practice implements does not have to be the most visionary one. It often makes sense to begin where the friction is obvious and the return can be demonstrated quickly.
Successful execution creates confidence.
Confidence creates capacity for the next move.
Beyond the Three Cases
CRC, fatty liver, and IBS/IBD are teaching cases. They are not the entire GI 2.0 opportunity map.
Clinical trials may become highly attractive for practices with sufficient patient populations and the right research infrastructure or partner. Many community groups already possess the patient volume that life-sciences companies value but lack the operational structure required to participate consistently.
Remote monitoring can be useful inside selected disease pathways, but it should not automatically be treated as a standalone revenue stream. The important questions are what is being monitored, what clinical decision the information changes, and who becomes responsible when the data indicate deterioration.
Microbiome science remains strategically important, but scientific excitement should not be confused with clinical readiness. Broad commercial testing has moved faster than the evidence supporting routine clinical action. That makes microbiome a good example of an opportunity to watch and prepare for rather than force prematurely into mainstream practice.
Advanced diagnostics will continue to change the way patients enter GI. New stool-, blood-, imaging-, molecular-, and other non-invasive approaches may begin as relatively narrow opportunities and move toward broader adoption as evidence, reimbursement, and clinical familiarity develop.
Robotics and increasingly automated procedural technologies will eventually create similar questions inside the procedure room.
Data may also become an important strategic asset. But a practice should be cautious about imagining a lucrative data business before it has solved the more basic challenge of organizing useful, governed, interoperable clinical and operational information.
None of these opportunities should be pursued because they appear on a list in this book.
They should pass through the same discipline.
What Should You Do Monday Morning?
If another gastroenterologist finished this chapter on a flight and came to Charley the next morning asking what to do first, he would not begin by recommending fatty liver, virtual GI, an AI company, or a particular reimbursement strategy.
He would first tell the physician to understand the practice deeply.
The practical sequence is:
- Know where you are. Understand your referral patterns, access problems, patient mix, current sources of revenue, major assets, operational bottlenecks, and areas where volume or margin may already be changing.
- Identify one meaningful patient problem. Do not begin with a vendor. Find a problem that matters clinically and occurs often enough in your practice to justify attention.
- Educate yourself deeply. Study the evidence, economics, partners, technology, payer issues, and regulatory considerations. Do not copy a program merely because another GI group is doing it; what works in another market may not fit yours.
- Choose a small project and a real leader. Pick something narrow enough to implement and assign a physician champion who has both the interest and the capacity to lead it.
- Define success before launch. Decide what should become better for the patient, the practice, and the broader system. Put ownership, milestones, review meetings, and measurement on the calendar.
- Start, learn, and build. Some initiatives will fail. Some partners will disappoint. Some technologies will change. Leadership requires accepting that possibility. The larger risk is allowing uncertainty to become a reason to do nothing.
Charley uses an old expression: if you are not the lead dog, the view never changes.
There is truth in it.
Nobody outside the practice is going to define the future of the practice for you in a way that perfectly matches your interests. Vendors will naturally optimize around their products. Payers will optimize around their economics. EHR companies will optimize around their platforms.
Someone inside the GI practice has to decide what the practice itself should become.
Do Not Take GI 1.0 Into the Future
There is a subtle trap in all of this.
A gastroenterology practice can embrace innovation while changing almost nothing.
It can add an AI tool to the existing workflow. It can hire a dietitian and call it a metabolic program. It can contract with a virtual-care company while leaving the patient journey fragmented. It can add another ancillary because the reimbursement is attractive. It can use tomorrow’s technology to make yesterday’s operating model slightly more efficient.
That is not GI 2.0.
The larger opportunity is to reconsider the operating model itself.
Colorectal cancer screening illustrates one possibility. The practice does not need to defend one screening modality against another. It can help organize the pathway through which appropriate patients move from a screening decision to completion.
Fatty liver illustrates another. Patients are already entering GI practices. The opportunity is to turn a diagnosis that can easily become unowned into a clearly defined longitudinal liver pathway while preserving the appropriate role of primary care and other metabolic clinicians.
IBS and IBD show something more fundamental. The traditional physician visit does not have to remain the unit around which all care is organized. Nutrition, behavioral health, APPs, virtual clinicians, digital tools, embedded partners, and AI can create a distributed care team that provides different levels of support to different patients.
None of these models requires the practice to own every capability.
They do require the practice to think more broadly about what it owns.
The practice may own the patient relationship. It may own the clinical pathway. It may own the standards by which partners participate. It may own the longitudinal information. It may own the decision about when a physician, APP, dietitian, behavioral-health professional, outside partner, or AI system enters the journey.
That is a much broader definition of a gastroenterology practice than the one the specialty inherited.
Procedures remain central. Colonoscopy, upper endoscopy, advanced endoscopy, infusion, and ASCs will continue to create substantial clinical and economic value. GI 2.0 is not an argument to walk away from what gastroenterology has built extraordinarily well.
It is an argument not to be limited by it.
Many of the next opportunities in GI will not enter through the procedure room. Some will arise before a procedure is needed. Others will emerge after the procedure is over. Some will come from chronic conditions the traditional practice has not organized longitudinally. Some will come from data. Some will come through partners. Some will emerge because payers are prepared to finance a better way of managing an expensive patient population. Many will become practical because AI allows coordination and continuity to occur at a scale that previously required too much human labor.
A GI 1.0 practice looks at an innovation and asks, “How can we fit this into what we already built?”
A GI 2.0 practice asks a different question:
If we were designing the best way to care for this patient today, knowing what is now clinically, technologically, and operationally possible, what would that pathway look like?
Then it works backward toward the operating and economic model.
That requires courage because there will never be complete information. Reimbursement will change. Companies will evolve. Some will disappear. Technology will improve. A program that works well in one market may fail in another. The first version will almost certainly contain mistakes.
But waiting is also a decision.
The first move does not have to be grand. It has to be real.
Choose one patient problem worth solving. Understand it deeply. Design the smallest useful model. Assign someone who will lead it. Decide what success looks like. Start. Measure what happened.
If the model produces better care, better quality, better outcomes, a better patient experience, or lower unnecessary cost, then build the economic architecture that allows that value to grow.
Do not put the economics first.
Create the value first, and then make the economics support it.
That is how innovation moves from the conference stage into mainstream gastroenterology. The future does not enter a practice simply as a technology. It enters as a workflow, a care model, and eventually a different way of organizing the practice itself.
Innovation is not the strategy. Integration is the strategy.
Once that decision is made, the next question becomes practical: which parts of the new operating model should the practice perform itself, which should be automated, which should come through partners, and which should remain with someone else?
That is where we go next.