Coverage for Continuous Glucose Monitoring for Individuals with Type 2 Diabetes Treated with Nonintensive Therapies: An Evidence-Based Approach to Policymaking

Numerous studies have demonstrated the clinical benefits of continuous glucose monitoring (CGM) in individuals with type 1 diabetes (T1D) and type 2 diabetes (T2D) who are treated with intensive insulin regimens. Based on this evidence, CGM is now a standard of care for individuals within these diabetes populations and widely covered by commercial and public insurers. Moreover, recent clinical guidelines from the American Diabetes Association and American Association of Clinical Endocrinology now endorse CGM use in individuals treated with nonintensive insulin regimens. However, despite increasing evidence supporting CGM use for individuals treated with less-intensive insulin therapy or noninsulin medications, insurance coverage is limited or nonexistent. This narrative review reports key findings from recent randomized, observational, and retrospective studies investigating use of CGM in T2D individuals treated with basal insulin only and/or noninsulin therapies and presents an evidence-based rationale for expanding access to CGM within this population.

spective studies investigating use of CGM in T2D individuals treated with basal insulin only and/or noninsulin therapies and presents an evidence-based rationale for expanding access to CGM within this population.

Introduction
A n estimated 34.5 million people in the United States have type 2 diabetes (T2D), 1 the majority of whom are treated with noninsulin medications. 22][13][14] Although often not considered, severe hypoglycemia is a significant burden for T2D individuals treated with insulin 15 and oral antihyperglycemic medications, 16 particularly for older individuals, many of whom have underlying heart disease. 17requent monitoring of glucose levels is a critical component of safe and effective diabetes self-management.9][20][21][22] As observed in a large cross-sectional study of 5104 individuals, adherence rates for prescribed BGM were as low as 44% for type 1 diabetes (T1D) adults and 24% for T2D adults. 22In a large, 12-month retrospective analysis of 1,329,061 Medicare beneficiaries with T1D and T2D treated with intensive insulin therapy (IIT), investigators reported that 38.14% of beneficiaries did not follow their prescribed glucose monitoring and an additional 35.42% had no record of ever obtaining glucose monitoring supplies. 20In a Swedish cross-sectional survey, the most common reasons for nonadherence reported were lack of time, not remembering to test, self-consciousness when testing in public, and the pain/discomfort of fingersticks. 19n increasing number of individuals with T1D and intensively managed T2D have transitioned from BGM to continuous glucose monitoring (CGM).Unlike BGM, which provides only point-in-time glucose values, current CGM devices automatically transmit real-time glucose data to users' smartphone or handheld receiver.Users have immediate access to viewing their current glucose values, recent trends, and rate-of-change trend arrows that indicate the direction and velocity of changing glucose, which can be easily and inconspicuously done in public.For added safety, these devices feature alarms and alerts that warn users about current and/or impending severe hypoglycemic (SH) and hyperglycemic events.This is particularly important to individuals with recurrent nocturnal hypoglycemia.
Retrospective analysis of overnight data enables users and their clinicians to adjust therapy to avoid SH events.CGM sensors last 7-14 days (and up to 180 days for implantable sensors), greatly reducing the patient burden associated with capturing these data relative to BGM.
][39] Despite a growing body of evidence supporting CGM use among people with T1D and T2D treated with IIT, commercial and public insurance coverage for CGM use in individuals using basal insulin only, or no insulin at all is limited, particularly within populations with lower socioeconomic status and racial/ethnic minority communities, which are disproportionately affected by diabetes. 40,41For example, among the 40 state in which Medicaid programs provide coverage, a 2022 report found that 27 currently cover CGMs only for individuals with T1D and T2D treated with IIT. 42In this narrative review, we report findings from current randomized and observational/retrospective studies investigating use of CGM in T2D individuals treated with basal insulin only and/or noninsulin therapies and present a rationale for expanding access to CGM within this population.

Persistent CGM Use in Nonintensively Treated T2D
Both randomized and prospective/retrospective studies have demonstrated significant glycemic improvement, reductions in diabetes-related events and hospitalization rates, and cost benefits of persistent CGM use by individuals with T2D who are treated with basal insulin only, basal plus noninsulin medications, and noninsulin medications without insulin (Table 1).The following is a summary of the major findings from these studies.
Bao et al., 45 in a sub-analysis of the MOBILE study, assessed the impact of rtCGM use in older adults age ‡65 years (n = 42) compared with younger patients (n = 133). 45he mean change in HbA1c from baseline was -1.08% in the older CGM group compared with -0.38% among older BGM users, with an adjusted mean difference of 0.65%.Significant increases in TIR (70-180 mg/dL) were observed in the older rtCGM and BGM groups (19%, P = 0.01 and 12%, P = 0.003, respectively).The mean difference in HbA1c between treatment groups was -0.35% in the younger study groups.
Martens et al. 46 in the recent randomized MOBILE study, assessed the effects of CGM use in 175 T2D adults who were treated with basal insulin with or without noninsulin medications. 46Patients were randomized 2:1 rtCGM (n = 116) or traditional BGM (n = 59) and followed for 8 months.Baseline HbA1c values in the rtCGM and BGM groups were 9.1% and 9.0%, respectively.
At study end, mean change in HbA1c among rtCGM users was -1.1% compared with -0.5% in the BGM group (P = 0.02), with significant increases in percentage of TIR (%TIR) in the rtCGM group (from 40% to 56%) compared with decreases in the BGM group (from 59% to 43%, P < 0.001), and significant decreases in percentage of TAR (%TAR >250 mg/dL) compared with the increases in the BGM group (-15 vs. 2, P < 0.001).Importantly, exploratory subgroup analyses suggested that the HbA1c improvements were present across the age range of 33 to 79 years and the baseline HbA1c range of 7.1% to 11.6%.Aleppo et al. 47 in this multicenter study, evaluated the effect of discontinuing rtCGM after 8 months of use. 47In Phase 1, patients had initially been randomized to either rtCGM or BGM for 8 months.In Phase 2, the BGM group continued to use BGM (n = 57) and rtCGM users were randomized to continue (n = 53) or discontinue use (n = 53).
All study groups were then followed for another 6 months.In the group that discontinued CGM, mean %TIR had increased from 38% to 62% during the 8 months using rtCGM.Six months after discontinuing rtCGM, %TIR decreased to 50% at 14 months (P = 0.01).Mean baseline HbA1c values starting Phase 2 were 7.9% in the discontinued

CGM COVERAGE FOR NONINTENSIVELY TREATED T2D
rtCGM group, 8.2% in the continued rtCGM group, and 8.4% in the BGM group.Mean HbA1c increased to 8.2% after discontinuing CGM (P = 0.06), whereas HbA1c in patients who continued rtCGM decreased to 8.1% HbA1c, and in the BGM group increased to 8.5% at 14 months.
Wada et al. 48in this 24-week, multicenter, open-label, randomized parallel-group study, randomly assigned 93 adult patients with noninsulin-treated T2D (1:1) to isCGM (n = 49) or BGM (n = 51) use. 48Change in HbA1c in each group was assessed.Baseline HbA1c levels were similar between the isCGM and BGM groups (7.83% and 7.84%, respectively).HbA1c was significantly decreased from baseline in the isCGM group at 24 weeks (-0.46%,P < 0.001), but not the BGM group (-0.17%,P = 0.124); a significant between-group difference was also observed (-0.29%,P = 0.022).Treatment satisfaction scores obtained using the Diabetes Treatment Satisfaction Questionnaire were also significantly improved, along with the mean glucose levels, glycemic variability, and time in hyperglycemia compared with the BGM group.
Large proportions of both T2D and T1D patients enabled and customized their glucose alerts.However, a higher proportion of T1D patients utilized the data SHARE feature compared with T2D patients (38.0% vs. 10.0%,respectively) and had more followers compared with T2D patients.Similar proportions of T2D (53.8%) and T1D (59.1%) patients used the CLARITY software throughout the observation period.
Chesser et al. 51 in this pilot, single-arm, interventional study assessed the usability and feasibility of 12 weeks' use of CGM in nine adolescents and young adults (13-21 years old) with T2D for ‡6 months, HbA1c >7.0%, and treated with basal insulin and/or noninsulin therapy (NIT); seven participants completed. 51At 12 weeks, participants reported statistically significant improvement in diabetes-related quality of life (QoL), with the mean Pediatric Quality of Life inventory (PedsQL) 52 diabetes score increasing from 70 to 75 after using CGM (P = 0.026).
Norman et al., 54 in the retrospective analysis of administrative claims data in the Optum Research Database (ORD), investigated the impact of rtCGM on diabetes-related medical costs within the T2D population following ‡6 months of rtCGM use. 54Changes in diabetes-related health care resource utilization costs were expressed as per-patientper-month (PPPM) costs.Within the cohort, 454 (80%) patients were treated with IIT, 58 (10%) were treated with NIIT, and 59 (10%) were treated with NIT.Results showed that the average PPPM for diabetes-related medical costs decreased by $424 (P = 0.035) after initiating rtCGM.Reductions in inpatient hospital admissions (-0.006PPPM, P = 0.057) and in total hospital admissions (-0.042PPPM, P = 0.139) were also observed.
Guerci et al. 55 conducted a retrospective analysis of the French national Syste `me National des Donne ´es de Sante ŕeimbursement claims database and identified 5933 individuals with T2D who had initiated isCGM between August 1, 2017, and December 31, 2018; 78.9% of subjects were treated with basal insulin and other antihyperglycemic medications. 55Claims data for the 12 months before, and up to 24 months after isCGM initiation, were analyzed to identify hospitalizations for adverse diabetes events (ADEs), including SH events, diabetic ketoacidosis (DKA), coma, and hyperglycemia-related admissions.Investigators reported that 2.01% of subjects had experienced at least one hospitalization for any ADE in the year before isCGM initiation, compared to 0.75% at 1 year and 0.60% at 2 years of isCGM use.During the first year of isCGM, there were 75% fewer DKA admissions and a 44% reduction in SH admissions.These improvements persisted after 2 years, with a further 43% reduction in DKA rates.
Similar findings were reported by Carlson et al. 57 in a U.S. chart review of 100 T2D adults managed with NIIT. 57After 3 to 6 months of isCGM use, the greatest reductions in HbA1c were observed in patients with >9.0%HbA1c at baseline (-1.7%, P < 0.0001).
Miller et al., 61 using a similar study design, analyzed data from the MarketScanÔ administrative claims database to assess the effects of isCGM on rates of all-cause hospitalizations (ACHs) and ADEs in a larger cohort of T2D adults (n = 10,282) treated with NIIT or noninsulin medications. 61uring the 6-month observation period, investigators reported significant reductions in ACHs, from 0.177 to 0.151 events/patient/year (P = 0.002).The rate of ADE decreased from 0.076 to 0.052 events/patient/year (P < 0.001).These decreases were significant in both the insulin-treated and noninsulin-treated patients (-0.040,P < 0.001 and -0.014, P = 0.015 events/patient/year, respectively).
Norman et al. 62 in an earlier analysis of the ORD, assessed the glycemic effects of CGM use (rtCGM or isCGM) on glycemic control compared with BGM within a cohort of 82,983 T2D patients treated with IIT, NIIT, or NIT. 62Following ‡6 months of CGM use, significant reductions in HbA1c were observed with any CGM and rtCGM versus BGM (-0.46 and -0.72 vs. -0.09,respectively, P < 0.001).HbA1c reductions were significant in CGM versus BGM users in the IIT group (any CGM, -0.38%; rtCGM, -0.68%, P < 0.001) and NIT group (any CGM, -0.67%, P < 0.001; and rtCGM, -0.87, P = 0.008).Investigators also reported significantly higher percentages of CGM versus BGM users achieved ‡1.0 HbA1c reductions in the IIT and NIT groups.
Intermittent CGM Use in Nonintensively Treated T2D In addition to the glycemic benefits observed with persistent CGM use, studies have also demonstrated the utility of intermittent use in modifying user behaviors and improving their understanding of their diabetes and importance of adherence to prescribed nonintensive treatment regimens (Table 2).The following is a summary of the major findings from these studies of intermittent CGM use.Many of these studies involved older CGM technologies.

Randomized controlled trials
Moon et al. 63 conduced a multicenter, randomized prospective study to investigate the efficacy of intermittent,

CGM COVERAGE FOR NONINTENSIVELY TREATED T2D
short-term rtCGM use in 61 T2D adults who were poorly controlled with noninsulin medications. 63Patients were randomly assigned to one of three treatment groups: one session of rtCGM (group 1) or two sessions of rtCGM with a 3-month interval between sessions (group 2) and a control group.All participants used blinded rtCGM for up to 6 days before randomization.Among the 48 patients who completed the study (baseline HbA1c 8.2%), investigators observed a significant HbA1c reduction in treatment group 1 (-0.60%,P = 0.044) and treatment group 2 (-0.64%,P = 0.014) compared with the control group at 3 months.At 6 months, only group 2 achieved a significant HbA1c reduction (-0.68%,P = 0.018).Cox et al. 64 in this randomized clinical trial, compared conventional medication management to medication management in conjunction with a lifestyle intervention using CGM in 30 T2D adults treated with noninsulin therapies and mean 8.8% HbA1c at baseline. 64Participants were randomly assigned (1:2) to routine care (RC) or use of rtCGM with four discussion sessions about how to minimize glycemic excursions.At the 5-month follow-up, the rtCGM group showed significant improvements in HbA1c compared with the RC group (from 8.9% to 7.6% vs. 8.8% to 8.7%, respectively, P = 0.03).The rtCGM group also showed a reduced need for diabetes medication (P = 0.01), reduced carbohydrate consumption (P = 0.009), and improved diabetes knowledge (P = 0.001), QoL (P = 0.01), and diabetes distress (P = 0.02) and trended to more empowerment (P = 0.05) with no increase in hypoglycemia.
Vigersky et al. 65 and Fonda et al. 66 in this randomized controlled trial, assessed the glycemic short-and long-term effects of intermittent rtCGM use in a cohort of 100 T2D adults treated with nonintensive therapies. 65The majority of patients (n = 60) were treated with oral or noninsulin injectable medications only.Investigators compared the effects of 12 weeks of intermittent rtCGM use with BGM on glycemic control over a 40-week follow-up period.At 12 weeks, there was a significant difference in HbA1c that was sustained during the follow-up period.Investigators observed decreases in HbA1c of 1.0%, 1.2%, 0.8%, and 0.8% in the rtCGM group at weeks, 12, 24, 38, and 52, respectively, compared with reductions of 0.5%, 0.5%, 0.5%, and 0.2% in the BGM group (P = 0.04).
The improvements observed in the rtCGM group occurred without a greater intensification of medication compared with those in the BGM group.Within this cohort, Fonda et al. reported that intermittent rtCGM use was a cost-effective diabetes management option, and that frequent use may result in additional cost-effectiveness. 66oo et al., 67 in this early prospective, randomized trial, investigated the potential effects of intermittent rtCGM use compared with BGM on glycemic control, weight, and selfmanagement behaviors in a cohort of 65 T2D adults who were poorly controlled with basal insulin and/or oral hypoglycemic agent therapy (8.0% to 10% HbA1c range at baseline). 67Patients were randomly assigned to rtCGM or BGM use and followed for 3 months.rtCGM patients used their device once a month for 3 days (due to the wear time of the sensor).The BGM group continued to test glucose levels ‡4 times/week for 3 months.Medication dosages were not to be changed over the study period.Investigators reported significant HbA1c reductions with rtCGM use compared BGM ((-1.1% vs. -0.4%,respectively, P < 0.01).
Significant reductions in weight (P = 0.014) and BMI (P = 0.008) were observed in the rtCGM group but not the BGM group.There was a significant reduction in total calorie intake in the rtCGM group (from 1858.7 to 1690.0 cal/day, P = 0.002).A significant increase in exercise time in rtCGM users compared with BGM users (P = 0.02) was also observed.
Prospective/retrospective studies Porter et al., 68 in this prospective pilot study, examined how use of rtCGM impacts glycemic metrics, weight loss, lifestyle, and patient perspectives on using the sensor among 37 T2D adults not treated with prandial insulin. 68Participants were randomized to either lifestyle counseling with two 10-day sessions of rtCGM use in blinded mode (n = 22) or four 10-day sessions of rtCGM use in unblinded mode but with no counseling (n = 15); 13 rtCGM participants completed the study.At 24 weeks, 6 (46.0%) reported weight loss of ‡10 lbs, 11 (84.6%) reported they were motivated to increase physical activity and excluded or eliminated certain foods as a result of rtCGM use, 12 (92.3%)reported they would wear rtCGM on a regular basis, and 13 (100%) reported that use of rtCGM contributed to self-care.
Bergenstal et al. 69 in this retrospective analysis, evaluated patient satisfaction with CGM use in 594 Onduo/Virtual Diabetes Clinic (VDC) participants. 69Satisfaction was assessed using the CGM satisfaction questionnaire. 70The reported CGM satisfaction score was 4.5 out of 5. Most respondents (94.7%) agreed/strongly agreed that CGM use improved their understanding of the impact of CGM on eating (97.0%), and that it increased their knowledge about diabetes (95.7%) and helped improve diabetes control when not wearing the sensor (79.4%).HbA1c decreased from 7.7% -1.6% to 7.1% -1.2% (P < 0.001; 10.2 months).These data suggest that it is feasible to provide CGM directly to individuals with T2D through a VDC without in-office training.Importantly, a subgroup analysis revealed significant reductions in HbA1c among both the insulin and noninsulin user groups with a baseline HbA1c ‡8.0%, -1.5% -2.1%, and -2.0%-1.7%, respectively (both P < 0.001).
Majithia et al. 71 conducted the Onduo VDC program, which is a telehealth program for people with T2D, and the eligibility criteria for participating in the VDC were age ‡18 years and ‡8% HbA1c at program entry.The program combines use of a mobile app, remote lifestyle coaching, connected devices, and live video consultations with boardcertified endocrinologists.In this prospective single-arm study, investigators evaluated glycemic outcomes associated with rtCGM use by program participants for 4 months in 55 program members. 71Participants were asked to use their rtCGM sensor intermittently over the course of 4 months, wearing a total of six 10-day sensors.At the end of the observation period, HbA1c levels decreased significantly (-1.6%, P < 0.001).
Polonsky et al. 72 used the 17-item Diabetes Distress Scale (DDS17) 73 to evaluate change in diabetes distress among 228 Onduo/VDC participants who reported moderate distress (score 2.0-2.9) or high distress (score ‡3.0) on the time of enrollment.Significant reductions in overall distress from 3.0 -0.8 at baseline to 2.5 -0.9 (P < 0.001) were reported by participants at an average of 6 months of follow-up.Participants who used intermittent rtCGM (n = 77) versus nonusers (n = 151) reported significantly greater reductions in overall distress (P = 0.012) and regimen-related distress (P < 0.001).

Discussion
In 23 (79%) of the 29 studies reviewed here, investigators reported associations between CGM use and improvements in HbA1c 44-47,54,56-60,62,63,65,67,69,71,75 and/or key CGM metrics. 44,47,71These findings, alone, provide strong evidence that supports providing access to this technology to all lessintensively treated T2D patients who would benefit and are able to use it safely and effectively.Monitoring regimens (persistent or intermittent) should be based on each person's individual needs and modified when these needs change.
9][20][21][22] Although not well studied, one can reasonably assume that poor BGM adherence is a likely contributor to the high prevalence of poorly controlled diabetes in the United States, 2,6-10 which continues to worsen. 2 By design, current CGM systems address all of the most common reasons reported by patients for not performing BGM. 19As demonstrated in the Onduo T2D studies, satisfaction was extremely high, users found their experiences with CGM both helpful and enlightening, 69 and diabetesrelated distress was significantly reduced. 72Moreover, the rapidly increasing adoption of CGM, worldwide, is a strong indicator of the value patients with diabetes place on this technology.
While some may argue against expanding access to CGM for economic reasons, it is important to consider the costs of uncontrolled diabetes.In 2017, the estimated total cost of diabetes in the United States, included *$237 billion attributable to direct medical care and an additional $90 billion in reduced productivity. 9A substantial percentage of these costs results from hospitalizations and emergency department utilization subsequent to diabetes-related adverse events. 10Most of these events are avoidable.Moreover, as demonstrated in an early study by Fonda et al. 66 and more recently by Norman et al., 54 use of CGM by patients treated with nonintensive therapies is cost effective.
Both public and commercial health insurers are now recognizing the challenges and complexity of effective diabetes management and how current and evolving technologies such as CGM, sensor-augmented insulin pump devices and now, and automated insulin delivery systems can dramatically reduce patient burden and the increasing costs associated with uncontrolled diabetes.However, payers need to have a more comprehensive calculation of the actual cost:benefit ratio, recognizing that the anticipated cost savings of not utilizing these technologies is far outweighed by the unanticipated consequences, such as increased hospitalizations and emergency department utilizations, which are associated with suboptimal glycemic control.This will require greater communications between the various departments within each payer organization to fully assess how the coverage policy will impact the total costs of providing or not providing coverage.
One example of how working with incomplete data can obfuscate the true cost of a policy decision is how the Centers for Medicare & Medicaid Services (CMS) interpreted the ''cost savings'' attributed to its national launch of the competitive bidding program in 2013.A 2017 study by Puckrein et al., examined changes in accessing BGM supplies test strips between 2010 and 2014, before and after nationwide expansion of the competitive bidding program. 76They found that the proportion of beneficiaries filling their test strip prescription partially or not at all grew significantly during that period.
In a 2016 study that looked only at the test period for competitive bidding, a correlation between lower rates of BGM use and increases in mortality, inpatient admissions and costs were observed. 77This demonstrates that coverage decisions regarding CGM and other innovative technologies must involve all stakeholders from payer organizations as well as the broader diabetes community.
CMS took an important step on April 16, 2023, expanding coverage of CGM to users of any insulin as well as to individuals who do not use insulin but who experience recurrent Level 2 or a single Level 3 hypoglycemic event. 76These changes are consistent with current American Diabetes Association (ADA) clinical guidelines. 37In light of the evidence reviewed here, there is clearly a growing evidentiary basis for permitting others to access this important technology as well.
9][80] Overcoming these challenges can only be achieved when patients and clinicians have unfettered access to the tools and technologies that have been proven effective in improving diabetes management and engaging patients in their self-management regimens.The more individuals understand their diabetes, the more willing and better able they will be to perform the daily tasks required to achieve optimal diabetes self-management.

Conclusions
CGM should be made readily available to all individuals with diabetes who are able to use this technology safely and effectively.Large randomized controlled trials and prospective real-world studies assessing the benefits of CGM compared with BGM in real-world settings would be helpful.Payers and policy makers need to catch up to the current research, broadening coverage and eliminating preauthorization requirements.Researchers need to catch up to needs of most people with T2D who currently have limited access to CGM and its benefits.

Table 1 .
Improvements in Key Outcomes from Randomized and Prospective/Retrospective Studies of Persistent Continuous Glucose Monitoring Use

Table 2 .
Improvements in Key Outcomes from Randomized and Prospective/Retrospective of Intermittent Continuous Glucose Monitoring Use The table illustrates studies where CGM was only worn on an intermittent basis, not continuously.a Hypoglycemia, other acute events, hospitalizations/ER visits.-, Metric was not measured, or change was insignificant.X indicates improvement.BMI, body mass index.