Diabetes care’s most dangerous gap: what happens after the visit

Diabetes care’s most dangerous gap: what happens after the visit
Most clinicians don’t need another statistic to know diabetes care is hard. But a few numbers are impossible to ignore:

• In the U.S., fewer than 1 in 10 people with diabetes receive formal self-management education in their first year.
• Roughly 50% of patients don’t take their medications as prescribed, and only a small minority sustain diet, monitoring, and activity goals over time.

Major academic centers, including Duke, have consistently demonstrated that when patients are properly educated and engaged in their care, adherence and outcomes improve. Yet the majority of patients never receive that level of support—especially once they walk out of the clinic.

The result is a dangerous blind spot in the post-visit period, where gaps in education, monitoring, and follow-up quietly undermine everything that was carefully decided in the exam room.

This is exactly where AI-enhanced continuity workflows can help.

The post-visit gap: where adherence quietly falls apart

Inside the clinic, diabetes care can look well-structured: guideline-driven decision support, medication reconciliation, multidisciplinary teams, and clear care plans.

Outside the clinic, it often looks very different:

• Patients leave with limited recall of what was said.
• Written handouts rarely match literacy level, language, or health literacy needs.
• Remote patient monitoring devices may be deployed, but data is noisy and under-utilized.
• Follow-up is episodic and reactive—triggered by an A1c result, an ED visit, or a patient finally reaching out.
  
  

This isn’t a failure of intent. Clinicians and care teams are doing everything they can within the constraints of a system that still treats education and adherence as “one-time events” instead of continuous processes.

Three structural problems show up again and again:

1. Education isn’t continuous.
   Patients may get a brief explanation of diagnosis and medications at diagnosis, or a one-time DSME session if they’re lucky. But diabetes is a chronic condition; understanding and behavior need reinforcement over weeks, months, and years.
   
   
2. RPM data isn’t integrated into workflow.
   Glucose readings, weight, and blood pressure can stream in from home devices—but without clear triage rules and automation, it becomes background noise. Teams don’t have the capacity to manually sift through it day after day.
   
   
3. Signals arrive too late.
   By the time a clinician sees an elevated A1c, multiple missed doses, lifestyle drift, and silent clinical deterioration may have been going on for months.
   
   

To close this gap, we don’t just need “more data” or “another dashboard.” We need a workflow that extends the care plan into the patient’s daily life and scales without burning out the team.

From point solutions to continuity workflows
Over the past decade, healthcare has tried to fix pieces of this problem with:

• Patient portals
• Educational handouts and videos
• Diabetes apps
• Standalone RPM platforms
• Outreach calls and care management programs
  
  

Each of these helps, but they’re often fragmented, poorly integrated into clinical workflow, and dependent on manual effort. They don’t consistently:

• Reinforce the specific diagnosis and care plan of that patient from that clinician
• Interpret home data against clinician-defined thresholds
• Provide patient-facing guidance when readings drift, before a crisis
• Inform the care team in a structured, actionable way
  
  

That’s the shift from “tools” to continuity of care workflows: instead of adding another point solution, you embed a continuous, structured process that connects the clinic, the patient, and RPM data in one loop.

This is where Tali Health lives.

How Tali Health closes the post-visit gap
Tali Health is designed as a clinician-guided continuity of care workflow, enhanced by AI, specifically built to address these post-visit gaps in diabetes care.

At a high level, the workflow looks like this:

1. Clinician-defined plan, encoded once
   The clinician documents the diagnosis, goals, and individualized care plan—including desired monitoring, medication regimen, and thresholds for concern (e.g., glucose ranges that should trigger action).
   
   
2. Extension into the patient’s daily life
   Tali extends that plan into a 24/7, multilingual, patient-facing experience. Instead of relying on memory of a 15-minute visit, patients have continuous access to clear, consistent explanations of:
   
   
   • Their diagnosis
   • Their individualized care plan
   • Why specific behaviors and medications matter
     
     
3. RPM device connectivity and auto-sync
   Tali connects to RPM devices (such as glucose meters and continuous glucose monitors) and auto-syncs readings into the workflow. Patients don’t have to manually log values; their real-world data simply becomes part of the care loop.
   
   
4. AI-enhanced triage using clinician-defined thresholds
   Rather than simply displaying numbers, Tali applies clinician-defined thresholds to triage readings:
   
   
   • Values in range: the system supports patient self-management and reinforces adherence.
     
     
   • Values out of range or trending in the wrong direction: Tali flags them, following rules set by the care team.
     
     
5. AI is used to prioritize, pattern-recognize, and summarize, but the clinical logic comes from the clinicians themselves.
   
   
6. Patient education aligned to the clinician’s plan
   When readings or behaviors shift, Tali provides patient education aligned with the original plan—in language the patient can understand and in their preferred language. This includes:
   
   
   • Reinforcing how their diagnosis relates to current readings
     
     
   • Clarifying what the care plan expects (e.g., timing of medications, monitoring frequency)
     
     
   • Helping patients distinguish when self-management is appropriate vs. when to escalate
     
     
7. Timely notifications to the care team
   When thresholds are breached or risk patterns emerge, Tali generates structured, clinically relevant notifications to the care team—allowing them to focus on patients who truly need intervention, rather than scanning every single reading.
   
   

The key: the workflow is the product. AI is there to enhance scale, consistency, and triage—not to replace the clinician or dictate care.

Clinical and operational impact
When you engineer continuity of care into the workflow, several outcomes become possible:

• Improved adherence
  Patients are repeatedly exposed to the same, clinician-authored plan in plain language—while their real-time data is linked to that plan. That combination of education + feedback loop is exactly what research shows improves adherence and outcomes.
  
  
• Higher patient satisfaction and confidence
  Instead of feeling “on their own” between visits, patients have a 24/7, always-available resource that reflects their doctor’s intent—not generic search results. That reduces anxiety, confusion, and unnecessary outreach.
  
  
• Fewer avoidable calls and low-value appointments
  When patients understand their plan and have clear thresholds for when to self-manage vs. reach out, they make more appropriate use of the care team. Meanwhile, RPM-driven triage helps teams proactively reach out to the right patients at the right time.
  
  
• Scalable post-visit support without adding staff
  Because the workflow is encoded once and enhanced by AI, Tali can scale across large patient panels and multiple sites without requiring a proportional increase in nurse or educator FTEs. Clinicians stay in control; the workflow does the heavy lifting.
  
  

What this means for healthcare leaders
For health systems, medical groups, and value-based care organizations, the question is no longer whether diabetes care needs better post-visit support—it’s how to deliver it consistently and sustainably.

Tali Health offers one path forward:

• Start with clinician-guided, evidence-based care plans
• Extend them into the patient’s daily life through continuous, multilingual education
• Connect real-world data through RPM integration
• Use AI to triage, summarize, and prioritize, not to replace clinical judgment
• Deliver scalable continuity of care without adding staffing overhead
  
  

In a space where under-education and non-adherence silently drive avoidable complications and cost, closing the post-visit gap isn’t a “nice to have.” It’s central to high-quality diabetes care.

And now, with AI-enhanced continuity workflows, it’s finally operationally realistic.