Abstract
Background:
Continuous glucose monitoring (CGM) has been increasingly shown to be beneficial in patients with both types 1 and 2 diabetes using insulin. Despite this, challenges remain in obtaining coverage for these devices. We sought to define the process of initiation of CGM and better understand factors associated with successful initiation.
Methods:
A single-center retrospective cohort study of 271 patients seen over a 3-year period from 2017 to 2020 in the adult endocrinology clinic at Boston Medical Center who were prescribed CGM was performed. The primary outcome was time to CGM initiation. Secondary outcomes included factors associated with initiation and continued use of CGMs and glycemic control.
Results:
Obtaining CGM through pharmacy benefit was significantly faster than through durable medical equipment companies (78 days vs. 152 days, P < 0.0001). Factors associated with initiation of CGM were younger age, private insurance, and education with a clinical diabetes educator. Identifying as black or Hispanic was significantly associated with decreased initiation of CGM. Glycemic control as represented by hemoglobin A1c improved in patients initiated on CGM from 9.06% to 8.22% (P < 0.001).
Conclusion:
Prescribing CGM as a pharmacy benefit significantly reduces the time to initiation, but on average, still takes several months, delaying potentially life-saving care for patients living with diabetes. Barriers to CGM initiation must be addressed to ensure timely delivery of optimal care to our patients.
Introduction
Continuous glucose monitoring (CGM) is widely regarded as the gold standard glucose monitoring technology in type 1 diabetes (T1D), with increasing evidence of value in individuals living with type 2 diabetes (T2D) who are treated with insulin. 1 –3 CGM technology has advanced considerably since the first CGM device was approved in the United States over two decades ago. These advancements have resulted in significant increases in CGM usage in children and adults with T1D over the last decade. 4 However, despite supportive evidence for CGM as the standard in modern glucose monitoring, there are many challenges faced by patients and clinical practices in implementing personal CGM into practice. Recent surveys in the United States have documented that lack of insurance coverage and cost of supplies are the major reasons for not initiating or continuing CGM, respectively, and global access to CGM technology varies widely. 5,6
The United States Food and Drug Administration regards CGM as a prescription product that largely relies on private insurance coverage to ensure access. 7 Furthermore, insurance eligibility criteria vary widely and change frequently, and the process of obtaining a CGM device is complex, often consisting of multiple steps that are time-consuming and restrictive. 8 The most common methods of obtaining a CGM device in the United States are through the pharmacy prescription benefit of a health insurance plan (e.g., a prescription for CGM is sent to a local or mail-order pharmacy) or through a durable medical equipment (DME) company.
Given this complexity, we were interested in defining the logistical process of obtaining a personal CGM device in our endocrinology clinic population at Boston Medical Center (BMC). BMC is an academic safety-net hospital in Boston, Massachusetts, that cares for a predominantly publicly insured and underserved minority population. Despite inequities in diabetes technology use in our endocrine clinic 9,10 in 2017, we employed a dedicated diabetes technology navigator who is responsible for processing prescriptions and DME requests for CGM to help clinicians and patients navigate this complex process. In addition, Massachusetts residents enjoy wide access to health insurance through a combination of private and/or public insurance, with <5% of the population being uninsured. 11 Given these potential advantages to support CGM uptake, we sought to understand the logistical factors and challenges associated with successful CGM initiation in our clinic, particularly surrounding time to initiation of CGM.
Methods and Measures
Study design
In this retrospective cohort study, we collected data for individuals with T1D and T2D seen in the adult endocrinology clinic at BMC between June 1, 2017, and June 1, 2020. Patients were included in the study population if they were seen for a diabetes care visit and received a CGM prescription or DME request from clinicians in the endocrine clinic during the 3-year period. Patients who had used CGM before this period or who had CGM prescribed by other clinicians outside of the endocrine clinic were excluded. The study was reviewed and approved by the Boston University Medical Center Institutional Review Board.
Data collection
Demographic characteristics, including age, gender, race, ethnicity, primary language spoken, insurance status, diabetes type, diabetes complications, and diabetes treatment information, were collected from the electronic medical record (EMR). We also extracted narrative data related to the process of initiating a patient on the CGM. This included CGM type, patient education on CGM use by certified diabetes care and education specialists (CDCES), and documented contact between clinic staff and DME companies, insurance companies, and pharmacies.
Outcomes and measures
Our primary outcome of interest was to compare the time it took for a patient to initiate CGM depending on whether they received the device through insurance pharmacy benefit or DME benefit. Secondary outcomes included a descriptive analysis of the logistical factors that influenced this process and glycemic control of patients started on CGM as represented by the hemoglobin A1c (HbA1c).
Covariates
The process for initiating CGM use was grouped into three categories: (1) pharmacy benefit, (2) DME, or (3) out-of-pocket payment. The time to initiation of CGM was measured in days after the initial prescription order was placed and treated as a categorical variable. Initiation was defined as follows: EMR documentation confirming CGM use in clinician notes or through access of CGM download software when applicable. The time to initiation was grouped as <30, 30 to <90, 90 to <180, and ≥180 days.
Race and ethnicity data were categorized according to the United States Office of Management and Budget standards classifications. Languages spoken were categorized as English, Spanish, and other languages. Insurance status was categorized as private, public (Medicare, Medicaid), or none. HbA1c results included the most recent HbA1c available before the initial CGM prescription and again within 3–6 months after the start of CGM (if there was an HbA1c available at both 3 and 6 months, the latter was used).
Analyses
Continuous variables were reported as arithmetic means and standard deviations. Categorical variables were reported as numbers of patients and percentages. Univariate comparisons were performed using independent sample t-test for normally distributed continuous data, Mann–Whitney U-test for non-normally distributed continuous data and Chi-square or Fischer exact tests for categorical variables. Multivariable logistic regression models, including all covariates were then created for the primary outcome. From these models, we reported adjusted odds ratios (ORs) and 95% confidence intervals (CIs) for the association between each covariate and the primary outcome. All statistical tests were two-sided with level of statistical significance <0.05. SPSS version 23 (SPSS Inc.) was used to perform all statistical analyses.
Results
There were 325 patients who met the study entry criteria, and of these, 271 patients were included for analysis. Twenty patients were excluded who were lost to follow-up after the initial CGM prescription/DME request, 23 patients were excluded due to insurance changes occurring between the period of CGM prescription/request and starting to use the device, and 11 were excluded due to insufficient documentation. At the time of data extraction, 208 (78%) had successfully initiated CGM and 63 had not (22%). Over the course of data review, 16 (8%) patients discontinued or intermittently used their CGM.
Demographics and glycemic control
As shown in Table 1, patients started on CGM were younger (49.07 ± 14.8 years) compared with patients who were not started on CGM (56.41 ± 15.14 years, P = 0.001). There was no statistically significant difference in gender distribution between these two groups. Many patients spoke English as their primary language. There was no statistically significant difference in race or ethnicity distribution among the groups. All patients were covered by health insurance and most patients had private health insurance, with a statistically significantly higher percentage of those initiated on CGM having private insurance compared with patients who did not start on CGM (74% vs. 49.2%, P < 0.001).
Characteristics by Continuous Glucose Monitoring Status
Data are presented as mean ± SD or n (%). Fisher's exact P values are reported.
Other languages included Albanian, Amharic, Cape Verdean, Chinese, Ethiopian, German, Haitian Creole, Portuguese, Tigrinya, and Vietnamese.
Other race/ethnicity included Asian, American Indian or Alaska Native, Native Hawaiian or other Pacific Islander, multiple, or unknown.
Public health insurance included Medicare and Medicaid.
Education is referred to education/training for CGM by clinical diabetes educator.
CGM, continuous glucose monitoring; CSII, continuous subcutaneous insulin infusion; HbA1c, hemoglobin A1c; MDI, multiple daily injections; SD, standard deviation; T1D, type 1 diabetes; T2D, type2 diabetes.
Overall, 57.9% of the patients were living with T2D and 41.3% with T1D. The baseline HbA1c in the overall cohort was 9.05% ± 1.85%, and there was no statistically significant difference in baseline HbA1c between patients who initiated CGM and those who did not. For those who initiated CGM, glycemic control improved after starting on CGM with the mean HbA1c 3–6 months after initiation improving to 8.22% ± 1.67% (P < 0.001).
CGM pharmacy versus DME coverage and time to CGM initiation
The majority of CGM insurance coverage was through pharmacy benefit (58.9%), with a mean CGM initiation time for this group of 78 ± 138 days following the initial prescription (Fig. 1). Of these, 59% of patients started on CGM within 30 days, 15.6% between 30 and 90 days, 12.3% between 90 and 180 days, and 13.1%>180 days. In the group that received CGM via DME, the mean start time was 152 ± 142 days. In the DME group, only 2% of patients started on CGM within 30 days, 47.1% between 30 and 90 days, 21.6% between 90 and 180 days, and 29.4%> 180 days. For the remaining 15.8% who paid for CGM out of pocket, the mean start time was 82 ± 120 days, similar to the pharmacy benefit group.
The percentage of patients initiating CGM through DME or pharmacy insurance benefit per time period from initial prescription. CGM, continuous glucose monitoring; DME, durable medical equipment.
CGM coverage through pharmacy benefit was significantly faster than DME coverage (P < 0.0001). There was a higher percentage of patients in the DME group who were covered by public insurance compared to patients in the pharmacy group (40.4% vs. 10.8%, P < 0.001).
Factors and challenges associated with initiation of CGM were analyzed through EMR chart review. For the pharmacy benefit group, delays between 30 and 90 days were most frequently related to the timing of clinic follow-up appointments with prescribing clinicians, rather than delays in receiving the device from the pharmacy. For delays >90 days, there was more variation in factors identified, including delays in picking up prescriptions from the pharmacy, other health-related problems, missed clinic visits, and delays in insurance prior authorization approval. For the DME coverage group, delayed initiation most commonly was secondary to missing clinical information in the medical record required for insurance coverage determination resulting in delayed approval and shipment of devices, missed clinic visits, and other health-related problems. There were typically multiple communications documented in the EMR between our CGM patient navigator, the prescribers, and/or the insurance/DME companies in these cases.
Among the patients who were unable to initiate CGM (Table 2), lack of insurance approval was the major barrier identified (94%) and given the utilization of an experienced diabetes technology navigator, clinical documentation was typically submitted only when the patients were expected to meet insurance criteria for CGM coverage. For those who initiated and then discontinued or had interrupted CGM use, cost was the main reason in 61.5% and 66.7%, respectively.
Barriers to Continuous Glucose Monitoring Use
Data reported as n (%).
To better understand factors associated with successful initiation of CGM use, a multivariable logistic regression analysis was performed (Table 3). The odds of initiating CGM decreased with increasing age (OR 0.96, 95% CI 0.94–0.99). In addition, the odds of starting CGM were lower for patients who identified as black or Hispanic (OR 0.32, 95% CI 0.14–0.78, and OR 0.31, 95% CI 0.10–0.96). Factors associated with increasing odds of starting on CGM were private insurance and diabetes education before being prescribed a CGM. There was no significant difference based on gender, presence of diabetes complications, baseline HbA1c, or type of diabetes.
Multivariable Analyses Associated with Starting on Continuous Glucose Monitoring
Discussion
Despite near universal health insurance coverage in Massachusetts and widespread acknowledgment of the benefit of CGM technology, timely and straightforward access to CGM devices remains a challenge. In our specialty endocrinology clinic caring for an underserved urban minority population, ∼70% of the patients prescribed CGM were private health insurance beneficiaries. All patients were supported by a specialty diabetes care team, including a dedicated in-house diabetes technology navigator with experience working with insurers, DME companies, and pharmacies. Despite this, it still often took >3 months for a patient to obtain a CGM device. It is encouraging to note that nearly 60% of CGM requests were routed through the pharmacy benefit and the percentage of requests proceeding through the pharmacy increased over the 3-year study period.
This is particularly important given that DME requests took almost twice as long as pharmacy prescriptions. Nonetheless, the average time between sending a prescription to a pharmacy and CGM initiation was still nearly 3 months and a patently disappointing 5 months in the DME group. These delays are unacceptable given the importance of this technology and its clearly established benefits. It is important to highlight that this 5-month delay was despite an active and established technology navigation process, which suggests that delays may be even longer in clinical environments without similar insurance coverage or without specialty expertise and navigator support. Documentation revealed that delays were attributed to a combination of factors, primarily the need to resubmit paperwork, including blood glucose logs, clinical notes, and HbA1c results, along with completion of certificates of medical necessity.
The significant improvement in time to CGM initiation through pharmacy benefit seen in our data provides strong support for increasing coverage of CGM through the pharmacy benefit and abandonment or widespread reform of the DME process. Multiple communications (calls with insurers, DME companies, and patients), faxes, and EMR chart note requests are required to direct the DME process. We believe that there are too many opportunities for failure built into this process and that universal pharmacy coverage for CGM should become the accepted standard in future. This is particularly important for patients with low health literacy or who do not speak English as their primary language as they often face more challenges in attempting to navigate this process, leading to further health care access disparities.
In addition, a higher percentage of patients in the DME group had public insurance compared to the pharmacy group, which suggests that there may be further discrepancies in access to CGM devices depending on insurance status. With changes to CGM device coverage requirements implemented by the Centers for Medicare and Medicaid Services in July 2021, particularly the elimination of the requirement to monitor blood glucose four times per day before approval, it is likely that this discrepancy will be reduced over time.
It is notable that there were no significant differences observed between individuals living with T1D and T2D in terms of successful CGM initiation, and nearly 60% of prescriptions were initiated for individuals living with T2D. Nearly 90% of our overall cohort were insulin users (multiple daily injections or basal insulin), which suggests that CGMs were prescribed for individuals who would likely derive the most benefit according to the current literature and evolving standards of care. Only 2% of CGM prescriptions were originated for patients who were already using continuous subcutaneous insulin infusion (CSII), which is also consistent with our clinic experience that most pump users have already started using CGM before or concurrent with CSII initiation.
Even when patients successfully initiate CGM, discontinuation or interrupted use is not uncommon (8%). In approximately two-thirds of cases, excess cost (despite the presence of health insurance) was cited as the reason in the EMR, suggesting that copays or out-of-pocket expenses remain an issue of concern for some individuals. While universal health care is widespread in Massachusetts, this is even greater cause for concern in areas with less widely available health insurance and remains a barrier to CGM initiation in some populations.
An additional finding of interest was that receiving CGM-focused diabetes education by a CDCES before CGM prescription was associated with a higher likelihood of successfully starting on CGM. Despite widespread availability of CDCES appointments in our clinic, only 60% of our patients received dedicated CGM-focused education before or after initiation, which argues for programmatic adjustments that encourage early and consistent engagement with diabetes educators, who are often intimately aware of the challenges we describe. This may be a significant challenge outside of dedicated endocrinology clinics where CDCES expertise may not be widely available, particularly as increasing pharmacy CGM coverage will likely result in increasing access to CGMs through primary care clinicians in the future.
Conclusion
In summary, our experience heralds a call to change. Would we accept a 3-month delay for a life-saving drug prescription? Or require multiple communications from clinicians to address these delays? There is overwhelming and compelling evidence of clinical benefit for CGM. 8 Widespread insurance coverage is still insufficient but continues to improve. However, individuals living with diabetes are still, unfortunately, indirectly denied the benefit because of the bureaucratic process that could be much simpler and more transparent. Pharmacy benefit CGM insurance coverage should be the goal to provide the best evidence-based care to the largest number of patients with elimination of unnecessary barriers.
Footnotes
Author Disclosure Statement
No competing financial interests exist.
Funding Information
No funding was received for this article.