Ask the Expert: CGM at Hospital Discharge

CGM at hospital discharge: Turning a vulnerable transition into actionable follow-up

Hospital discharge is one of the moments in diabetes care where I see the greatest disconnect between clinical intention and real-world execution. A patient may leave the hospital with a new diabetes diagnosis, a new insulin regimen, recent diabetic ketoacidosis, steroid-associated hyperglycemia, or major medication changes — but once they are home, the care team often loses visibility into what is happening day to day. That gap matters. The first days and weeks after discharge are when patients are adapting to new therapies and routines while waiting for follow-up that may not happen quickly enough, due to a variety of circumstances.

As a clinician who has worked across inpatient and outpatient diabetes care, I think about CGM at discharge less as a “device start” and more as a transition-of-care strategy. The American Diabetes Association (ADA) now recognizes that initiating CGM just before hospital discharge may be beneficial when clinically appropriate and paired with follow-up planning.¹ For clinicians, the practical question is not whether CGM has value in diabetes care broadly, but how to use it thoughtfully at this high-risk point in care.

How should clinicians think about CGM at hospital discharge for patients with diabetes who are leaving the hospital with new or changing treatment plans?

What the evidence suggests

The evidence base for CGM in diabetes care is well established across multiple populations, particularly for improving glycemic outcomes such as A1C, time in range, and hypoglycemia-related measures.²⁻⁴ In one randomized study of adults with type 2 diabetes after hospital discharge, patients using CGM had better glycemic control and lower glucose variability than those using point-of-care testing, including significantly lower mean glucose variability at 4 weeks.⁵ Real-world analyses have also suggested that CGM initiation after a diabetes-related health care encounter may be associated with lower emergency department use.⁶ Since recent studies have also shown an association between CGM use and emergency department visits and hospital admissions in people with type 1 and type 2 diabetes, the ADA states that it may be beneficial to initiate CGM just prior to discharge to facilitate follow-up and possibly prevent acute diabetes-related complications and readmission.¹

What I take from this evidence is not that CGM alone solves the challenges of care transitions, but that it provides objective data during a critical period when therapy is changing, risk is elevated, and traditional fingerstick logs often lack the full story. In my experience, that’s when CGM can be particularly valuable, helping identify emerging hypoglycemic or hyperglycemic patterns before they go unnoticed or contribute to preventable complications.

What I see in real-world practice

In practice, the patients who often need the most support after discharge are also the patients least likely to come back with clean, interpretable glucose data. I have seen patients return to follow-up without a meter, without enough readings, or with values written down but no clear sense of timing, meals, insulin doses, or symptoms. That makes it difficult to distinguish between fasting hyperglycemia, postprandial excursions, missed insulin, overcorrection, or hypoglycemia that may be driving over corrective eating.

This is especially challenging for patients discharged on insulin for the first time or after DKA, severe hyperglycemia, or a new diabetes diagnosis. In those situations, CGM can help shift the follow-up visit from “What do we think might be happening?” to “Here are the patterns we can act on.” The AJMC case examples highlight this same issue: clinicians at Regions Hospital described how patients using CGM arrived at follow-up with more actionable data, enabling more confident treatment adjustments than fingerstick monitoring alone.⁷

What I also see is how this changes the patient experience between visits. Instead of relying on intermittent fingersticks, patients have a clearer, more continuous understanding of their glucose during a time when routines are changing and confidence may be low. Rather than trying to interpret isolated readings, they can begin to recognize patterns and better understand how meals, activity, or medication changes affect their glucose. Alerts may also help patients recognize impending lows or rising glucose sooner, rather than reacting after the fact. In my experience, that added awareness can provide reassurance during a time when patients are often navigating uncertainty.

From a behavioral change perspective in my experience, the period immediately after discharge is a fragile window where new routines are being formed but are not yet stable. Early experiences during this time often determine whether behaviors are reinforced or abandoned—particularly if patients feel uncertain or encounter unexpected hypo- or hyperglycemia without clear feedback. CGM helps close that gap by providing continuous, real-time visibility that allows patients to connect their actions (meals, insulin, activity) to glucose trends. This more immediate feedback loop may support learning, build confidence, and reinforce adherence at a time when patients are translating hospital-based care into self-management. In my experience, that added visibility helps patients stay engaged and more proactively adjust behaviors before patterns escalate.

Patient Care Workflow

Inpatient phase	Transition phase	Outpatient phase
New onset T1D/T2D diagnosis Standard diabetes education CGM training 2-4 hours before discharge Sensor placed at discharge 3 sensors provider per patient Connected to clinic portal	Handoff to outpatient team via EHR Contact within 72 hours Weekly check-ins for first month Telemedicine visit at 1-2 weeks Medical assistant handles insurance	Remote monitoring via CGM data Treatment adjustments based on trends Alternating telemedicine appointments In-person visit at 1 month Gateway to insulin pump therapy

How this changes clinical conversations

What especially stands out to me is how differently the follow-up visit unfolds when CGM data is available. For example, a patient may have looked stable on a carbohydrate-controlled inpatient diet but experience significant post-meal hyperglycemia once they return home. Another patient may be discharged on basal insulin and report feeling “fine,” but CGM reveals overnight lows that would not have been captured with sporadic daytime checks.

Having access to these patterns makes counseling more specific and individualized. Instead of giving general advice about diet, insulin timing, or hypoglycemia prevention, clinicians can point to specific trends and ask, “What was happening here?” or “What changed on this day?” That kind of discussion helps patients connect glucose patterns to meals, activity, medication timing, illness, or fear of hypoglycemia. It also supports more confident treatment adjustment. In the AJMC hospital case study, one clinician described CGM as moving them from a “snapshot” to a “video” of glucose, which is exactly the kind of visibility that can improve clinical decision-making after discharge. ⁷

Another important consideration is patient safety during the early post-discharge period, particularly for patients starting insulin or recovering from DKA. When CGM is in place, alerts may help patients recognize concerning trends earlier and prompt appropriate action before those patterns potentially worsen. 

Practical use of CGM in this context

From what we’ve seen, successful implementation begins well before the patient leaves the hospital. Successful programs identify candidates early, assess coverage and supply needs, determine whether the patient needs a receiver or can use a smartphone, provide education, and connect the patient to follow-up or data-sharing workflows before discharge pressure peaks. The AJMC paper highlights real world implementation examples where CGM education and sensor placement were embedded into discharge planning, including sensor placement within a few hours of discharge and follow-up contact within days.⁷ 

CGM should not be treated as a standalone intervention. It works best when paired with early follow-up, clear team roles, patient education, and a plan for who will review the data and act on it. That may include endocrinology, diabetes educators, pharmacists, primary care, virtual care teams, or a bridge clinic model. Without that structure, clinicians risk generating data without creating a clear pathway for clinical response.