Roadmap to the Effective Use of Continuous Glucose Monitoring in Primary Care Free

Diabetes has become the great epidemic of the 21st century and is arguably one of the greatest epidemics in human history (1,2). With an estimated prevalence of 37.3 million, or 11.3% of the population (3), diabetes and its complications consume one of every four health care dollars spent in the United States (4).

Yet, the specialty resources available to help battle this epidemic remain severely limited. Endocrinology as a subspecialty is significantly under-resourced to manage the diabetes epidemic; an estimated 8,000–9,000 board-certified endocrinologists currently practice in the United States (5). Even with a larger cohort of specialty-focused advanced practice clinicians to augment MD-credentialed endocrinologists, the capacity to guide over 10% of the American population in managing diabetes is inadequate.

Primary care providers are the obvious group to address the diabetes epidemic. With both established connectedness with people with diabetes and the capacity to see and help manage these individuals (6), primary care not only can be the pathway to optimized glycemic management, but also needs to be. The task, then, is to provide primary care providers with both the tools they need to optimize glycemic management and the systems of support they need to use those tools adequately.

Primary care, by its nature, is tasked with the holistic management of a wide spectrum of health conditions extending far beyond diabetes. For this reason, “bandwidth” can be a major limitation to attempts to optimize diabetes management in primary care settings. Primary care simply lacks the singular focus that specialty care can bring to diabetes management. Additionally, a lack of provider expertise, especially with regard to insulin therapy, is often another significant barrier. Finally, primary care clinics and clinicians often lack the systems of support that are typically available within endocrinology practices. Structured support for obtaining glycemic data for use during clinical interactions is typically available as part of standard workflows in endocrinology practices. However, in primary care settings, where workflows often involve capturing data from across the broader spectrum of human health, this level of support typically is not available.

These factors, combined with relatively infrequent clinic visits, lead to suboptimal glycemic management through the widely recognized phenomenon of clinical inertia in primary care settings (7). Optimizing the support of people with diabetes, and especially individuals with type 2 diabetes, must therefore, of necessity, require empowering primary care clinicians to improve how they manage diabetes. Empowering primary care providers in this regard requires providing them with the tools and data they need to optimize care and, beyond that, with the knowledge and support systems necessary to move diabetes care forward.

A1C is the standard diabetes quality measure in primary care and beyond, for compelling reasons. It was the primary outcome measure in the major outcomes trials conducted from the 1980s through the first decade of the 21st century (8–10). Correlations between A1C and the development and progression of diabetes complications have been established beyond doubt. Based on these studies, A1C targets for quality assessment on a population basis, and with some limitations (11–13) on an individual basis, are established, widely known, and widely used. Yet, this measure can only paint a broad picture of glycemic management. It is less useful for day-to-day management because it is insensitive to the daily excursions into hyperglycemia and sometimes hypoglycemia that are the target of glycemic management.

The established standard for day-to-day glycemic management, especially in primary care settings, has been fingerstick blood glucose monitoring (BGM). The value of BGM has been well proven in type 1 diabetes and insulin-treated type 2 diabetes (14,15). It is widely available and widely prescribed. Yet, BGM has been less effective in typical real-world settings than in clinical trials. Its limitations involve the burden of its use: the inconvenience of testing, discomfort, and payer-imposed limitations on the number of tests performed per day. These barriers limit BGM’s ability to reveal a person’s full glycemic picture throughout the day. Additionally, a lack of availability of glycemic data at the time of clinical interactions, especially in primary care settings, further limits the effectiveness of BGM. The promise of BGM has never been fulfilled, and today, many clinical practice guidelines recommend against its routine use in non–insulin-treated individuals with diabetes (16–18).

Continuous glucose monitoring (CGM) has been available since the early 2000s, but use has dramatically increased in just the past 5 years, driven by the availability of newer systems with improved ease of use (i.e., no calibration requirement, greater accuracy, and longer sensor wear times), increased availability at lower costs, and evolving data reporting and sharing to support their use in broad populations, including those with type 2 diabetes (19). Driven by randomized controlled trials (RCTs) showing the superiority of CGM versus BGM in improving A1C or reducing hypoglycemia in individuals with type 2 diabetes using a multiple daily injection (MDI) or basal-only insulin regimen (20–22), and further supported by observational data suggesting benefit in type 2 diabetes more broadly (23–29), practice recommendations have now evolved to include larger populations for whom CGM offers potential benefits (30).

Additional benefits of CGM include a decreased burden in obtaining robust glycemic data both day and night; the ability to access glycemic data through smartphone-linked, Cloud-based data repositories; and the availability of a standardized and remarkably intuitive data presentation format. That format, known as the ambulatory glucose profile (AGP) report (31), allows the rapid review of glycemic data in a structured and predictable manner by specialists, nonspecialist medical professionals, and people with diabetes (Figure 1).

The convergence of conclusive evidence from research, improved availability and ease of use of CGM systems, and availability of robust CGM-driven glycemic data in an intuitive and accessible format (the AGP report) all suggest that we now really do have a better vehicle to take us further down the road toward optimized glycemic management. Through the broader resource of primary care, we have the opportunity to reach the vast population of people with diabetes in the United States. Thus, we have a vehicle, and we have a driver for the vehicle. So, how do we create a roadmap to success? How do we avoid the same types of pitfalls that left BGM technology with its potential unfulfilled?

Our objective in creating a roadmap is to minimize barriers and maximize benefit. Minimizing barriers means making the best path also the easiest and most direct path, and making the roadmap clear and easy to follow for both primary care clinicians and people with diabetes. Maximizing benefit means using CGM and the data it provides to its fullest extent. We need to optimally use both point-in-time data to optimize the impact of lifestyle behaviors, nutrition, and pharmacotherapies, and retrospective data to pursue pattern-based glycemic management and improve shared decision-making via visualization of CGM data at the time of visits and clinical interactions (Figure 2).

A fundamental reality in creating a roadmap for the optimized use of CGM in primary care settings is that nobody benefits if CGM systems are unavailable to the people they might help. The journey starts with access, and access to technology typically starts with compelling trial data and, beyond that, real-world data to identify populations that might benefit from it.

RCT data are available to support the use of CGM for individuals with type 2 diabetes who are on insulin therapy regardless of regimen (20–22). These studies typically involved insulin titration by researchers and diabetes experts, with both BGM and CGM arms titrated using best-practice care. Whether the benefits of CGM over BGM exist to a greater or lesser extent in real-world settings and specifically in populations managed in primary care has not yet been shown definitively. However, compelling observational data are available (23–29), and real-world, fundamentally pragmatic studies are currently underway (32). Early studies and compelling observational data suggest a potential benefit of CGM use for individuals with type 2 diabetes who do not take insulin (33,34), but, again, definitive data are not yet available.

Beyond compelling evidence of efficacy, a long-range perspective on costs versus benefits and on the impact of improved glycemic status on total costs of care over time helps to support CGM use across broad populations and earlier in the course of illness, when optimized glycemic status has been shown to decrease long-term risks of complications (35). The vast majority of the $237 billion spent on direct medical costs of diabetes in 2017 went to inpatient hospitalizations and emergency department visits, nursing home care, and medications for diabetes complications. Shifting even a fraction of that cost to the 2% of the total spent on diabetes supplies (4) has the potential to dramatically decrease the total cost of care by improving care upstream (i.e., optimizing glycemic status earlier in the course of illness and for a greater portion of the overall population with diabetes).

If data support benefits, we need to make CGM systems available with minimal hassles and barriers to acquisition. For commercially insured individuals, availability and access, although variable, has improved in the past 5 years. A key component of availability for individuals managed in primary care settings is to minimize prior authorization hoops and hassles, which can be a fundamental barrier limiting access (36,37). In primary care clinics, which often have less developed processes for paperwork completion and less sophistication with regard to knowing the steps necessary to obtain coverage approvals, prior authorization requirements can be a common cause of abandonment of attempts to provide CGM for patients. Short-term workarounds in primary care settings may include process management for prior authorizations, but a far more cost-effective route would be to decrease the burden by decreasing prior authorization requirements.

For individuals covered by Medicare and Medicaid, the purchase of a CGM system is typically handled as a durable medical equipment (DME) acquisition. There have been significant improvements in Medicare-based DME access to CGM. Even with the sunsetting of short-term improved access during the coronavirus disease 19 (COVID-19) pandemic (38), requirements for onerous documentation of BGM-based testing for approval of CGM were dropped in 2021 (39), and the requirement of an MDI insulin regimen are currently under review and likely will be dropped in 2023 (40).

Despite these changes, DME-based acquisition of CGM can be problematic because of the paperwork requirements of individual DME suppliers. For individuals with Medicaid coverage, access is even more variable and remains extremely limited in many states. The fact remains that CGM technology can only benefit the broader population of people with type 2 diabetes if they have access to the technology. This is especially true for populations with the highest burden of diabetes complications, who often, because of demographic, social, and financial barriers related to social determinants of health, have the most limited access to newer technology and pharmacotherapies (41–43). The role of health equity advocacy in reaching these often-marginalized populations cannot be overstated. You simply cannot make the journey if you don’t have a vehicle.

Access to CGM technology is necessary but not sufficient to improve care in primary care settings through the use of CGM. The CGM system itself ultimately is not the agent of glycemic optimization; rather, it is the data produced by the CGM system that drive favorable change.

In this regard, it should be noted that CGM produces two types of data: point-in-time data and retrospective data. Point-in-time data include the current glucose value, a trend arrow, and a trend line, allowing people with diabetes to see the impact of dietary, exercise, and medication factors on an immediate basis and thereby take steps to significantly improve glycemic management on a purely heuristic basis. Aggregated retrospective data, presented as glycemic metrics and visualized via the AGP report, allows for rapid review and interpretation of the percentage of time spent in various glycemic ranges, other key glycemic metrics, as well as a composite “modal day” graph and individual daily view graphs (31,44,45). Glycemic metrics summarized from retrospective CGM data allow for the rapid identification of problematic glycemic patterns and enhancing discussion of therapeutic options using shared decision-making.

An optimized approach to CGM involves both helping people with diabetes to use point-in-time data optimally and reviewing retrospective data consistently at the time of clinical interactions to drive care improvement. Typically, point-in-time data are available to CGM users via a handheld reader or smartphone app without further logistical requirements. Reviewing retrospective data requires access to the data, which can be a key challenge and barrier in primary care settings.

Aggregate retrospective CGM data can be obtained in several ways. Reader devices can be downloaded during clinic visits if they are available. People with diabetes can upload data from their reader at home, allowing access via industry Cloud-based repositories. Finally, CGM data can be linked to industry Cloud-based repositories by smartphone, allowing web-based visualization by clinicians in real time. Although all three of these mechanisms can be used and are helpful for clinical interactions, they all have limitations that can present barriers to use in primary care settings. Downloading readers during clinic visits can be limited by organizational firewalls or a lack of access to drivers, preventing linking with Cloud-based repositories. Additionally, this mechanism requires that the CGM reader be physically present at the time of clinical encounters, which limits its usefulness during telehealth encounters, and relies on device users to remember to bring their reader to appointments.

Uploading device data to a Cloud-based repository via a home computer is also feasible for people with diabetes, but operating system and driver incompatibility, lack of appropriate cables, or simply the need to remember to upload the data can all limit access through this mechanism. Home uploading of data requires levels of computer sophistication, Internet access, and engagement that can pose a significant limitation.

Ultimately, access via industry Cloud-based resources allows the broadest access to primary care teams, whether for clinic visits, telehealth encounters, or quick clinical touchpoints to optimize insulin titration. For individuals with compatible smartphones, app-based Cloud access can allow for the most consistent and smoothest data delivery to clinicians. Once smartphones have been connected to receive sensor data and to communicate the data to a Cloud-based platform, users simply accept a sharing invitation via e-mail to allow their care team to access their CGM data. Lack of access to a smartphone can be a limitation to this mechanism of data acquisition in resource-poor environments, although the technology gap does show signs of narrowing, at least with regard to smartphones (46). Gaining Web-based access to data are currently the option that has become the best-practice alternative to obtaining retrospective CGM data by other means.

Even with Web-based access, barriers remain to obtaining data when they are needed during clinic visits. Computer systems used in clinical practice are typically highly protected by firewalls to prevent intrusion. Institutional and organizational firewalls can limit access to industry data, either by directly blocking that access or by blocking access to drivers necessary to upload the data. Engagement of the clinic or health system Information Technology team can be critical when working toward organizational solutions to gain broader access to CGM data; whereas smaller endocrinology departments can often use “one-off” solutions for data access, the much larger world of primary care will require more global access.

Health Insurance Portability and Accountability Act regulations have significant implications for accessing health information via the Web. Safe password practices and steps to avoid storing data on personal devices are crucial. Two-factor authentication is becoming the industry standard in accessing Cloud-based data repositories. Although necessary and appropriate, the work involved to maintain and ensure the connectivity of passwords and password protection while accessing multiple sites in the course of busy clinical practice can be a significant limitation to accessing data.

Although it is possible for individual clinicians to personally access CGM data via the Web at the time of clinical interactions, time constraints can be a significant barrier. Primary care clinicians are typically tasked with managing the broad spectrum of health in a 15- to 30-minute time slot. Diabetes visits comingle with visits for medical urgencies and emergencies, health maintenance activities, phone and other messaging, and various other health needs of the broader population. A key element of support, and therefore a key element of the roadmap to successful use of CGM in primary care, is the creation of a workflow, which typically involves office personnel and sets the protocol for a smooth, uniform process for obtaining glycemic data in advance or at the time of clinical interactions. A team-based, uniform, and consistent approach to managing chronic disease can be a key component of quality optimization in primary care settings (47–49) and is likewise a key component of optimized CGM use in primary care. Workflow and consistent processes facilitate the availability of key glycemic data during clinical interactions, enabling clinicians to focus on the patient interaction and shared decision-making, rather than on data acquisition, during visits.

Beyond workflow, optimized use of CGM in primary care starts with optimized set-up of devices at the time of CGM initiation. Designating a clinic diabetes champion who has expertise in setting up and troubleshooting CGM systems can be extremely helpful in ensuring success at the time of start-up and also in helping people with diabetes understand and use their CGM data. Populations with type 2 diabetes tend to be older and less technologically savvy than populations with type 1 diabetes; ensuring that steps as seemingly simple as having the correct language set on the device and having the date and time correctly set can prevent significant issues down the road when trying to access patients’ glycemic data. Diabetes educators, when available, can often be the optimal resource in this champion role, but other diabetes care team members can also assist with these tasks.

Clinicians at all levels typically work and live in an electronic medical record (EMR) system. These systems offer significant benefits with regard to data availability and data integration, and, with an additional push from American governmental incentives (50), have been widely adopted throughout the U.S. medical system. The near-universal use of EMR systems presents a unique opportunity for glycemic data integration: direct importation of CGM-based glycemic data and AGP reports into patients’ EMRs. For people with diabetes who are able to link smartphone CGM data to industry Cloud-based platforms, an opportunity exists for collaboration among industry, device manufacturers, and health care organizations to create processes to pull CGM data directly into the EMR system, removing significant barriers to data access.

Direct importation of CGM data into EMRs likely represents the single best solution for data accessibility during clinical visits. In this regard, significant progress has been made in creating this critical on-ramp to the information superhighway. Proof-of-concept trials of direct EMR-based access have been conducted at several institutions, including the International Diabetes Center in Minneapolis, MN; Northwestern University in Chicago, IL; and the Children’s Hospital of Los Angeles/University of Southern California (51–56). We await more formal publications on these integration projects. Although this mechanism of data access is not yet widely available, it is clear that it is both feasible and critical in facilitating the broader use of CGM glycemic data in primary care settings. Direct EMR-based access to glycemic data is a key element on the roadmap to using CGM successfully in primary care.

Just as access to devices is necessary but not sufficient to provide glycemic data at the time of clinical interactions, having data at the time of clinical interactions is necessary but not sufficient to optimize glycemic management. The final key aspects of improving care include improving primary care expertise in managing glucose-lowering therapies (especially insulin) and improving the cadence of titration to reduce clinical inertia.

Improving the cadence of titration becomes much easier in a CGM-based world, given that access to glycemic data are much easier using Cloud-based mechanisms. Glycemic data becomes available for multiple touchpoints and interactions beyond traditional clinic visits. The explosion of telehealth, or “virtual” care, driven by the COVID-19 pandemic (57–59) has provided new and unique opportunities. With the availability of Cloud-based CGM data, the opportunity for multiple and frequent diabetes titration touchpoints offers significant promise for reducing clinical inertia.

Data acquisition, interpretation, and utilization all take time. The viability of using CGM to optimize care rests on the viability of reimbursement for time spent using CGM to optimize care. Fortunately, reimbursement for cognitive resources in managing CGM is growing. Currently, reimbursement is available for the start-up and application of CGM devices as well as the interpretation of data (60). Current trials of Medicare-based remote patient monitoring codes (61,62) open additional possibilities for care model innovation, making optimizing glycemic management more viable within a wider clinical practice spectrum.

Building expertise within primary care to use CGM to optimize glycemic management is the final stop on the roadmap to successful CGM use in primary care. Access to devices and data only helps if people with diabetes and the clinicians who help them manage their diabetes know how to use CGM data appropriately. The fingerstick BGM experience suggests that technology without training or appropriate use does not improve care (16). The question is not only “Can we do better with CGM-based monitoring?” but also “Will we do better with CGM-based monitoring?” CGM data are robust relative to BGM data, and standard data presentation formats like the AGP report make data interpretation straightforward. Concepts such as time in range (TIR; the percentage of time a person spends with glucose levels within a target range) are both intuitive and actionable at the individual level and create a framework for understanding in primary care.

Clinical programs offering training on the TIR and the related concepts of time above range and time below range, as well as the use of other CGM-based glycemic metrics to optimize therapy have emerged online through professional organizations and advocacy groups such as the American Diabetes Association, American College of Physicians, and American Academy of Family Physicians and at clinical conferences. Literature on using CGM and CGM-based metrics targeting primary care audiences is available and expanding (44,63,64). At its heart, this literature emphasizes a systematic approach to using CGM glycemic data. The knowledge base for using CGM and AGP data reports in primary care settings is improving, but work remains to be done. The diffusion of these approaches into training programs is ongoing and needs to continue. The potential payoff of this final stop on the roadmap to optimized use is huge.

An additional component of successful diabetes management in primary care settings is the concept of team-based care. Diabetes management using CGM integrates nicely into team-based models of care. The spectrum of diabetes management ranges from relatively straightforward noninsulin glycemic management to complex issues of polypharmacy and MDI insulin regimens. Primary care clinicians are in the truest sense “jacks of all trades, but masters of none.” That fact, combined with typically brief clinical visits and very busy schedules, makes team-based management of complex chronic medical conditions not a luxury, but rather a necessity.

Resources for team-based management vary. In some practices, a clinician acting as a diabetologist or an endocrinologist able to do outreach using a Project ECHO–type (or similar) model (65–67) can help to optimize complex care. Pharmacists who can perform medication management and registered nurses who can coordinate care can be tremendous assets in team-based management. Diabetes educators perhaps hold the most promise for teaming with primary care providers to optimize glycemic management. Having access to diabetes educators with expertise, patient focus, and more time to address the complexities of diabetes management has been shown to improve metabolic parameters (68), and the addition of CGM-based management to this role likely provides additive benefits. What remains to be developed is a reimbursement model to allow the expansion of this critical resource in managing 13% of the U.S. population.

The rapidly expanding availability of CGM in primary care settings has already increased awareness of pattern-based glycemic management, including the identification of problem areas and pitfalls rarely appreciated using only BGM data. The robust nature of CGM data allows for more nuanced glycemic insights with the potential to dramatically improve the efficacy and safety of glycemic management, especially for individuals using insulin. A primary care colleague made this analogy: “BGM is like looking at a room through a keyhole; CGM is like looking at the room with the door wide open.” The impact of CGM in primary care has been large, but the potential impact remains much larger.

The path forward involves building systems of support to allow the optimized use of CGM in the much larger world of primary care. By removing barriers to the availability of CGM technology, optimizing data-sharing so that CGM data and AGP reports are universally available, and giving primary care clinicians both the knowledge and data they need to advance care forward at an appropriate cadence, we have the opportunity to improve the care of broad populations of people with diabetes. It is time to move beyond the roadmap and create the roadway. It’s time to make the future of diabetes management a reality for primary care.

T.W.M. is employed by HealthPartners Institute, which has contracts with Abbott Diabetes Care, Dexcom, Eli Lilly, Insulet, Medscape, Medtronic, Novo Nordisk, Sanofi US Services, Inc., and Tandem for his services as a research investigator, speaker, and/or consultant. He is paid on salary and receives no personal income from any of these services. No other potential conflicts of interest relevant to this article were reported.

As the sole author of this article, T.W.M. researched the data and wrote and revised the manuscript and is the guarantor of this work.