628547
research-article2016
DSTXXX10.1177/1932296816628547Journal of Diabetes Science and TechnologyFonda et al
Original Article
The Cost-Effectiveness of Real-Time Continuous Glucose Monitoring (RT-CGM) in Type 2 Diabetes
Journal of Diabetes Science and Technology 2016, Vol. 10(4) 898–904 © 2016 Diabetes Technology Society Reprints and permissions: sagepub.com/journalsPermissions.nav DOI: 10.1177/1932296816628547 dst.sagepub.com
Stephanie J. Fonda, PhD1, Claudia Graham, PhD2, Julie Munakata, MS3, Julia M. Powers, MHS4, David Price, MD2, and Robert A. Vigersky, MD1
Abstract Background: This analysis models the cost-effectiveness of real-time continuous glucose monitoring (RT-CGM) using evidence from a randomized controlled trial (RCT) that demonstrated RT-CGM reduced A1C, for up to 9 months after using the technology, among patients with type 2 diabetes not on prandial insulin. RT-CGM was offered short-term and intermittently as a self-care tool to inform patients’ behavior. Method: The analyses projected lifetime clinical and economic outcomes for RT-CGM versus self-monitoring of blood glucose by fingerstick only. The base-case analysis was consistent with the RCT (RT-CGM for 2 weeks on/1 week off over 3 months). A scenario analysis simulated outcomes of an RT-CGM “refresher” after the active intervention of the RCT. Analyses used the IMS CORE Diabetes Model and were conducted from a US third-party payer perspective, including direct costs obtained from published sources and inflated to 2011 US dollars. Costs and health outcomes were discounted at 3% per annum. Results: Life expectancy (LE) and quality-adjusted life expectancy (QALE) from RT-CGM were 0.10 and 0.07, with a cost of $653/patient over a lifetime. Incremental LE and QALE from a “refresher” were 0.14 and 0.10, with a cost of $1312/patient over a lifetime, and incremental cost-effectiveness ratios were $9319 and $13 030 per LY and QALY gained. Conclusions: RT-CGM, as a self-care tool, is a cost-effective disease management option in the US for people with type 2 diabetes not on prandial insulin. Repeated use of RT-CGM may result in additional cost-effectiveness. Keywords behavior, continuous glucose monitoring, cost-effectiveness, health economics, self-care, self-monitoring
Real-time continuous glucose monitoring (RT-CGM) has been shown to improve glycemic control and/or reduce the frequency of mild to moderate hypoglycemic episodes in people with type 1 diabetes and people with type 2 diabetes taking prandial insulin.1-8 Similarly, in a randomized controlled trial (RCT), Vigersky and colleagues demonstrated various glycemic benefits of a 3-month course of RT-CGM in people with type 2 diabetes not taking prandial insulin.9-11 Cost-effectiveness analyses of RT-CGM and glycemic control in people with type 1 diabetes have shown that lowering A1C through the use of RT-CGM could result in quality of life benefits within the range of what is deemed to be costeffective.12-13 Given the evidence that RT-CGM has clinical benefit for people with type 2 diabetes not taking prandial insulin (about 72.6% people with type 2 diabetes do not take
insulin of any type),14 it is important to evaluate whether extending RT-CGM use to this large cohort is potentially cost-effective. Since there are no studies of cost-effectiveness of RT-CGM in people with type 2 diabetes, we examined the potential impact of RT-CGM on lifetime clinical and
1
Walter Reed National Military Medical Center, Bethesda, MD, USA Dexcom, Inc, San Diego, CA, USA 3 IMS Health, San Francisco, CA, USA 4 University of Colorado Denver, Aurora, CO, USA 2
Corresponding Author: Stephanie J. Fonda, PhD, Walter Reed National Military Medical Center, 8901 Wisconsin Ave, Bethesda, MD 20889-5600, USA. Email: [email protected]
899
Fonda et al economic outcomes, using results from Vigersky and colleagues’ RCT.9,10
Methods CORE Diabetes Model (CDM) We used the validated IMS CORE Diabetes Model (CDM), which has been broadly applied.15,16 The CDM is a nonproduct-specific computer simulation model. The simulations predict the long-term health outcomes and costs of particular interventions for type 1 and/or type 2 diabetes. In this case, the intervention is short-term, intermittent use of RT-CGM versus self-monitoring of blood glucose (SMBG) by fingerstick among people with type 2 diabetes not taking prandial insulin. Data that inform the probabilities of events (such as hypoglycemia, amputation, a myocardial infarction, etc.), the progression of A1C, systolic blood pressure, lipids, and appropriate risk adjustments are derived from the United Kingdom Prospective Diabetes Study (UKPDS), the Diabetes Control and Complications Trial (DCCT), the Framingham Heart Study, and other sources. The CDM simulates disease progression in both type 1 and type 2 diabetes, and is widely used to estimate the impact of interventions on clinical and cost outcomes, as well as a range of economic analyses (cost-effectiveness, cost-utility, cost-benefit or cost of disease). It can also be used to identify potential high-risk patient profiles. The outputs of the model include life expectancy (LE), quality-adjusted life expectancy (QALE), direct and indirect costs, cumulative incidence and time to onset of complications, and incremental cost-effectiveness ratios per additional life year (LY) or quality-adjusted life year (QALY) gained.
Cohort Inputs—Interventional Study The characteristics of the cohort simulated by the CDM to predict the long-term health outcomes and costs of RT-CGM for people with type 2 diabetes not taking prandial insulin are based on the study by Vigersky and colleagues.9-11 Briefly, it was a 52-week, prospective, 2-arm RCT in 100 adult subjects comparing the short- (12-week) and long-term (40week) effectiveness of RT-CGM (n = 50 subjects) and SMBG (n = 50 subjects). Those randomized to RT-CGM used the Dexcom™ SEVEN® (Dexcom, Inc, San Diego, CA). The RT-CGM use occurred in 4 periods (2 weeks on/1 week off) over 12 weeks, for a total of 8 weeks of usage. Those randomized to SMBG were asked to test their glucose before meals and at bedtime for 12 weeks, as well as times associated with symptoms of hypo- or hyperglycemia. After the initial 12 weeks, all participants were asked to perform SMBG for the duration of the study as recommended by their usual provider. Follow-up study visits were performed at 3-week intervals during the first 12 weeks and then every 3 months during the remaining 9 months of the study.
RT-CGM was provided as a self-care tool to inform the participants’ behavior choices. The framework for the RCT was guided by self-determination theory, which is concerned with the processes through which people become motivated for initiating new self-beneficial behaviors and maintaining them over time.17,18 People feel autonomous and competent when they regulate their behavior through personal choice and self-endorsement. Support for competence occurs when people are afforded the skills and tools for change. The study staff did not provide any care management and the study participants’ providers did not have access to the RT-CGM data. Also, patients in the setting where this study took place are discouraged from taking part in more than 1 interventional study at a time, so to our knowledge, they had no other behavior interventions. The participants in the intervention were, on average 57.8 years old (Table 1). To be eligible for the RCT, they had a diagnosis of type 2 diabetes for at least 3 months, had an initial A1C of between 7% (53 mmol/mol) and 12% (108 mmol/mol), and were treated with diet/exercise alone or other glucose lowering therapies except prandial insulin. There was no difference in the use of antidiabetic medications between the RT-CGM and SMBG groups. By the end of year 1, the participants who had used RT-CGM had a net reduction in A1C of 1.1 percentage points ± 1.5 percentage points (12 mmol/mol ± 16.4 mmol/mol), and the participants who had used SMBG alone had a net reduction of 0.5 percentage points ± 1.3 percentage points (5.5 mmol/mol ± 14.2 mmol/mol). There was no statistically significant impact of the intervention on other physiological parameters. These data, as well as additional background data needed for the CDM, provided the necessary inputs.
Cost Inputs The cost inputs for the CDM included the costs of RT-CGM, SMBG, antidiabetic oral medications, insulin, routine management such as recommended screening, exams, and treatment for depression, and treatment of diabetes complications. No behavioral interventions that might be available to certain populations are included in the model, as these are not universal. SMBG is required to calibrate the Dexcom SEVEN (Dexcom, Inc, San Diego, CA) (twice daily), and the RT-CGM intervention was short-term and intermittent, so SMBG costs applied to both treatment groups (Table 2). Our assumptions regarding the frequency of SMBG, and therefore the cumulative costs of lancets and strips, were based on the observed frequency of SMBG in the interventional study, as previously reported,10 as well as published sources.19,20 RT-CGM costs applied to the RT-CGM group only, however, and the specific amounts were provided by Dexcom (Dexcom, Inc, San Diego, CA). Our assumptions regarding the annual cost of oral antidiabetic medications (Table 2) were based on the medications prescribed to the participants in the interventional study. The
900
Journal of Diabetes Science and Technology 10(4)
Table 1. Characteristics of the 100 Study Participants in RCT of RT-CGM in People With Type 2 Diabetes Not Taking Prandial Insulin. Variable
Base-case value
Demographics and clinical characteristics Age, mean years ± SD Proportion male (0-1) Proportion African American (0-1) A1C, mean %-points ± SD (mean mmol/mol ± SD) Duration of diabetes, mean years Systolic blood pressure, mean mm/Hg ± SD Total cholesterol, mean mg/dL ± SD HDL, mean mg/dL ± SD LDL, mean mg/dL ± SD Triglycerides, mean mg/dL ± SD Body mass index, mean kg/m2 ± SD Proportion with selected complications (0-1) Myocardial infarction Angina Peripheral vascular disease Stroke Heart failure Microalbuminuria Gross proteinuria Background diabetic retinopathy Proliferative diabetic retinopathy Macular edema Neuropathy Depression Impact of RT-CGM intervention Change in A1C from baseline at the end of year 1, mean percentage points ± SD (mean mmol/mol ± SD)
57.8 ± 10.8 0.55 0.52 8.3 ± 1.2 (67 ± 13.1) 9.0 ± 6.8 131.7 ± 17.8 166.1 ± 39.7 46.9 ± 13.8 94.1 ± 30.4 153.5 ± 110.7 32.3 ± 6.8 0.02 0.01 0.05 0.06 0.02 0.14 0.11 0.12 0.01 0.02 0.34 0.20 RT-CGM: –1.10 ± 1.50 (–12.0 ± 16.4) SMBG: –0.50 ± 1.30 (–5.5 ± 14.2)
medications were metformin (69% of the participants were taking this), sulfonylurea (48%), basal insulin (33%), pioglitazone (20%), exenatide (11%), sitagliptin (11%), glimepiride (5%), glyburide (4%), and repaglinide (3%). For each type of drug, we identified the lowest-cost products and the defined daily doses as reported in MediSpan PriceRx. The remainder of the direct costs, including costs for diabetes management (ie, exams, screening visits, etc.), cardiovascular disease complications, renal complications, acute events, eye disease, and neuropathy were obtained from published sources and inflated to 2011 US dollars.21-25 Costs and outcomes were discounted at 3% annually.
Analyses We explored 2 scenarios—the base case and a hypothetical refresher scenario. The base-case assumed no further use of RT-CGM after year 1, meaning the effect of the treatments on A1C observed in the RCT was applied for 1 year only. The refresher scenario assumed a second course of RT-CGM use, 1 year after the first course, in year 2. The second course
Table 2. Combined Costs by Treatment Group and CostEffectiveness Model. Costs Year 1, base-case scenario RT-CGM equipment SMBG strips and lancets Antidiabetic medications Total Year 2, base-case scenario RT-CGM equipment SMBG strips and lancets Antidiabetic medications Total Refresher scenario RT-CGM equipment SMBG strips and lancets Antidiabetic medications Total
RT-CGM group
SMBG group
$631.72 $369.80 $2083.00 $3084.52
$0.00 $369.80 $2083.00 $2452.80
$0.00 $174.24 $2083.00 $2257.24
$0.00 $174.24 $2083.00 $2257.24
$631.72 $174.24 $2083.00 $2888.96
$0.00 $174.24 $2083.00 $2257.24
Note: Costs of SMBG strips and lancets were higher in the year 1, basecase scenario because study participants in the RCT used for inputs to this analysis were asked to test 4 times daily. The costs are based on the frequency of testing that the study participants actually did.
would be identical to the first (4 periods of 2 weeks on/1 week off, distributed over 12 weeks, for a total of 8 weeks of usage). The refresher scenario also assumed that the second course of RT-CGM would maintain A1C at the level achieved from the first course. Convergence in A1C between the RT-CGM and SMBG treatment groups only would occur in the subsequent year of no RT-CGM use. This prediction was based on the observation that both groups from the original interventional study had an average A1C of 7.9% (63 mmol/ mol) at the end of year 2 (unpublished data from a follow-up analysis). This assumption may be conservative. The base-case and refresher scenarios assumed that patients in the simulated cohorts would continue on their diabetes medications until full uptake of insulin rescue therapy. There are no good data on the time to full uptake of insulin that can be applied to this cohort, so we used clinical judgment regarding the likely sequence of treatment and assumed time to full uptake of insulin to be by year 5 from the beginning of the most recent course of RT-CGM. To test the robustness of the cost-effectiveness, univariate and probabilistic sensitivity analyses were performed. All analyses were performed from a US third-party payer perspective, including direct costs obtained from published sources and inflated to 2011 US dollars. Outcomes were discounted at 3% per annum.
Results Base-Case Scenario The LE and QALE over a lifetime horizon are predicted to be 10.62 and 6.03 when RT-CGM is used, versus 10.52
901
Fonda et al Table 3. Lifetime Horizon Cost-Effectiveness of the Base-Case and Refresher Scenarios. RT-CGM Base-case scenario Life expectancy (LE) Quality-adjusted life expectancy (QALE) Total costs Incremental cost-effectiveness ratio per life-year (LY) gained Incremental cost-effectiveness ratio per quality-adjusted life-year (QALY) gained Refresher scenario Life expectancy (LE) Quality-adjusted life expectancy (QALE) Total costs Incremental cost-effectiveness ratio per life-year (LY) gained Incremental cost-effectiveness ratio per quality-adjusted life-year (QALY) gained
SMBG
Incremental
10.62 6.03 $66 094
10.52 0.10 (1.25 months) 5.96 0.07 (0.88 months) $65 441 $653 $6293 $8898
10.64 6.05 $66 734
10.50 0.14 (1.69 months) 5.95 0.10 (1.20 months) $65 423 $1312 $9319 $13 030
Notes: The outcomes in this table are rounded.
and 5.96 when SMBG alone is used (Table 3). The incremental LE and QALE were 0.10 and 0.07, respectively, meaning those who use RT-CGM could expect a gain of 1.25 months in LE or 0.88 quality-adjusted life months compared with using SMBG alone. The total lifetime costs in the RT-CGM and SMBG treatment arms were $66 094 and $65 441 respectively, and the incremental cost was $653 per patient over a lifetime. The incremental cost-effectiveness ratios were $6293 per LY gained and $8898 per QALY gained.
Refresher Scenario In the refresher scenario analysis, the LE and QALE over a lifetime horizon are estimated to be 10.64 and 6.05 when RT-CGM is used compared with 10.50 and 5.95 for SMBG alone (Table 3). The incremental LE and QALE were 0.14 and 0.10, or 1.69 months and 1.20 quality-adjusted life months. The total lifetime costs in the RT-CGM and SMBG treatment arms were anticipated to be $66 734 and $65 423, respectively, and the incremental cost was $1312 per patient over a lifetime. The incremental cost is approximately double that of the base-case scenario, reflecting the added cost of more RT-CGM-related supplies. The incremental costeffectiveness ratios were $9319 per LY gained and $13 030 per QALY gained.
Both Scenarios The use of RT-CGM translated into a reduction of cumulative rates of diabetes complications and deaths in cardiovascular disease, ulcers and amputations, and renal disease (Table 4). The one exception was stroke (death and event), which, paradoxically, is predicted to be higher for patients who use RT-CGM. This oddity is known as the “survival paradox” and has been observed for other health conditions and interventions.26
Sensitivity Analyses Probabilistic cost-effectiveness analysis suggests that the likelihood of the intervention being cost-effective is 70% at the willingness-to-pay threshold of $100 000 per QALY.
Discussion We have shown that intermittent, short-term use of RT-CGM is a clinically effective approach for lowering A1C for people with type 2 diabetes who are not taking prandial insulin.9,10 The present analysis demonstrates that, for this cohort, it is also a cost-effective disease management adjunct. RT-CGM is not an intervention per se for this cohort, but could be considered as part of a diabetes management strategy that includes lifestyle advice, as well as the choice of optimal management of diabetes using medications. A refresher course of RT-CGM may result in longer-term cost-effective health benefits, due to an additional year of impact on A1C and the consequent long-term health outcomes. Much of the cost-effectiveness of RT-CGM for this cohort is due to the low cost of the intervention. Specifically, the incremental cost-effectiveness ratios per LY gained and QALY gained in both the base-case and refresher scenarios were substantially below the commonly discussed willingness-to-pay range of $50 000-$100 000 per QALY gained. Recent studies suggest that the acceptable threshold in the US setting could actually be much higher—in the range of $109 000-$297 000 per QALY gained.27 By comparison, a cost-effectiveness analysis using inputs from the Juvenile Diabetes Research–Continuous Glucose Monitoring (CGM) Trials in people with type 1 diabetes found that CGM reduced the expected lifetime incidence of complications while also increasing costs; that is, the incremental cost-effectives ratio for their base-case was $98 679 per quality-adjusted life-year gained.11 Much of the costs were attributed to health care provider time used for training in the use of CGM and diabetes management (which were not costs relevant
902
Journal of Diabetes Science and Technology 10(4)
Table 4. Cumulative Incidence of Diabetes Complications for the Base-Case and Refresher Scenarios, by Treatment Group and as Percentage of the Cohort. Base-case scenario Congestive heart failure death (%) Congestive heart failure event (%) Peripheral vascular disease onset (%) Angina (%) Stroke death (%) Stroke event (%) Myocardial infarction death (%) Myocardial infarction event (%) Ulcer (%) Recurrent ulcer (%) Amputation ulcer (%) Amputation recurrent ulcer (%) Neuropathy (%) Microalbuminuria (%) Gross proteinuria (%) End-stage renal disease (%) Nephropathy death (%)
Refresher scenario
RT-CGM
SMBG
RT-CGM
SMBG
21.32 23.43 7.74 8.96 13.00 18.18 11.63 14.88 26.86 33.31 8.27 2.61 52.81 24.61 10.23 5.52 3.10
21.41 23.50 7.91 9.07 12.93 17.94 11.78 15.26 27.83 34.70 8.68 2.62 54.78 26.27 11.24 6.08 3.45
21.45 23.49 7.82 9.03 13.11 18.30 10.93 14.12 27.02 33.87 8.36 2.54 53.24 24.96 10.46 5.64 3.21
21.48 23.67 7.92 9.03 12.85 17.82 11.82 15.39 27.89 34.64 8.65 2.66 55.00 26.42 11.35 6.12 3.52
to Vigersky and colleagues’ intervention), health service utilization as well as supply costs for the CGM and confirmatory SMBG testing. Another cost-effectiveness examination in people with type 1 diabetes, comparing CGM plus intensive insulin therapy versus SMBG plus intensive insulin therapy, found that CGM plus intensive insulin therapy resulted in an expected improvement in cost-effectiveness of 0.52 QALYs and an incremental cost-effectiveness ratio of $45 033 per QALY.13 In addition, our findings for short-term RT-CGM use for the simulated cohort compared favorably to previous costeffectiveness results for SMBG in people with type 2 diabetes. In an examination of SMBG in Canadian patients with type 2 diabetes who were not using insulin, Tunis and the Canadian Optimal Prescribing and Utilization Service (COMPUS) estimated incremental cost-effectiveness ratios to be Can$39 231 and Can$54 349 per QALY gained for adherence rates of 66% and 87%, respectively.28 In the Swiss context, 1, 2, or 3 times daily SMBG in patients with type 2 diabetes treated with oral antidiabetic agents resulted in incremental cost-effectiveness ratios of CHF9177, CHF12 928 and CHF17 342 per QALY gained, respectively, which are below the commonly accepted willingness-to-pay thresholds in Switzerland.29 In the US context, Tunis and Minshall estimated that the cost per QALY gained for SMBG performed 1 or 3 times per day among people with type 2 diabetes on oral antidiabetic agents was $7856 and $6601.30 A later study of this US cohort estimated that SMBG performed 1, 2, or 3 times daily increased QALY by 0.05, 0.12, and 0.13, and the incremental cost-effectiveness ratios were $26 206, $18 572, and $25 436 per QALY gained, increasing with SMBG frequency.31
A fundamental difference between this analysis and previous cost-effectiveness studies of monitoring in people with type 2 diabetes is the length of time to which the cost of the intervention was applied. In the interventional study on which this cost-effectiveness analysis was based, the use of RT-CGM in year 1 resulted in statistically significant improvement in A1C in the first 3 months of the study, and this was sustained until the end of the year, 9 months after the completion of the active intervention. This is a relatively short-term intervention, yet it had long-term health consequences and long-term cost-effective health benefits. The cost-effectiveness analysis predicted a small effect on quantity and quality of life. This small effect is typical of behavioral interventions, where no prolonged action of the intervention is assumed. RT-CGM, as used in the original clinical trial, is a tool to support learning and subsequent self-directed, behavioral choices because no interpretation of RT-CGM data by a clinician was provided, nor did study staff adjust diabetes therapies; physiologic effects were due to participants’ actions. Clinician interpretation might increase costs, but might also increase the beneficial effects of this technology, and thereby offset costs. A limitation of this analysis is the small size of the original study (n = 100) that provided many of the clinical inputs for the CDM. However, the original study was the longest and largest randomized clinical trial to date of RT-CGM use in people with type 2 diabetes not taking prandial insulin. Also, the original, interventional study used an older RT-CGM system (Dexcom SEVEN). Newer systems (eg, Dexcom G4 Platinum®) have features that may make them more user-friendly and, therefore, better tools to support
903
Fonda et al behavior than older systems. Specifically, they are more accurate, patients wearing a sensor can travel a greater distance from the handheld receiver and still transmit data, and they have smaller handheld receivers which may make them easier to carry in a pocket or purse. The aforementioned improved accuracy may result in greater trust of the information, which may enhance use of the RT-CGM device and of the RT-CGM information.32 Another potential limitation is that this analysis (and the original intervention) did not compare RT-CGM with structured SMBG, similar to that used in the Structured Testing Program (STeP).33 STeP had patients test their blood glucose 7 times in the 3 days prior to a health care visit (fasting, preprandial, 2 hours postprandial at each meal, and bedtime). The testing data were used to guide health care providers’ decision making, rather than to guide patient decision making, which was the focus of the RT-CGM RCT. Nevertheless, structured testing in the manner prescribed by STeP may be helpful for patient decision making, and this might also result in quality of life benefits considered to be cost effective. However, structured testing is different from RT-CGM in that it may not identify the true peak or nadir glucose, does not show the duration of hyperglycemia or hypoglycemia, does not show the speed of glucose rise or fall, does not alert a user to highs or lows, does not frame the results (with high or low thresholds), and is not necessarily structured in such a way as to show glucose responses to behaviors other than what one ate (eg, exercise, timing of medication, not getting enough sleep, etc). A strength of this study is that it is the first to date to evaluate the cost-effectiveness of RT-CGM in people with type 2 diabetes (on any treatment). Another is that it considers the cost-effectiveness of a refresher course; although, admittedly, this consideration is speculative since it is unclear at this time if a refresher course would be effective in maintaining (or perhaps further improving) glycemic control and, if so, how often and how long such a refresher course should be.
Conclusions In sum, RT-CGM was excellent value under the conditions of the original, interventional study, namely short-term, intermittent use in people with type 2 diabetes who were not taking prandial insulin at the time they began using RT-CGM and who used the technology as a tool to inform behavioral choices without clinician guidance. Future research might address the value of RT-CGM in a larger cohort with type 2 diabetes, including those on prandial insulin, and the costeffectiveness of using newer RT-CGM systems and/or clinician input into the behavior decisions that patients make in response to their RT-CGM data. Abbreviations CDM, CORE Diabetes Model; COMPUS, Canadian Optimal Prescribing and Utilization Service; LE, life expectancy; LY, life year; QALE, quality-adjusted life expectancy; RCT, randomized
controlled trial; RT-CGM, real-time continuous glucose monitoring; SMBG, self-monitoring of blood glucose; STeP, Structured Testing Program; UKPDS, United Kingdom Prospective Diabetes Study.
Declaration of Conflicting Interests The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: CG and DP are employees of Dexcom, Inc. No other competing financial interests exist.
Funding The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The project was sponsored by an investigator-initiated grant from Dexcom, Inc to SJF and RAV. The opinions expressed in this article reflect the personal views of the authors and not the official views of the US Army or the Department of Defense.
References 1. Juvenile Diabetes Research Foundation Continuous Glucose Monitoring Study Group. Continuous glucose monitoring and intensive treatment of type 1 diabetes. N Engl J Med. 2008;359:1464-1476. 2. Juvenile Diabetes Research Foundation Continuous Glucose Monitoring Study Group. The effect of continuous glucose monitoring in well-controlled type 1 diabetes. Diabetes Care. 2009;32:1378-1383. 3. Deiss D, Bolinder J, Riveline JP, et al. Improved glycemic control in poorly controlled patients with type 1 diabetes using real-time continuous glucose monitoring. Diabetes Care. 2006;29:2730-2732. 4. O’Connell MA, Donath S, O’Neal DN, et al. Glycaemic impact of patient-led use of sensor-guided pump therapy in type 1 diabetes: a randomised controlled trial. Diabetologia. 2009;52:1250-1257. 5. Battelino T, Phillip M, Bratina N, et al. Effect of continuous glucose monitoring on hypoglycemia in type 1 diabetes. Diabetes Care. 2011;34:795-800. 6. Garg S, Zisser H, Schwartz S, et al. Improvement in glycemic excursions with a transcutaneous, real-time continuous glucose sensor. Diabetes Care. 2006;29:44-50. 7. Bailey TS, Zisser HC, Garg SK. Reduction in hemoglobin A1C with RT-CGM: results from a 12-week observational study. Diabetes Technol Ther. 2007;9:203-210. 8. Yoo HJ, An HG, Park SY, et al. Use of a real time continuous monitoring system as a motivational device for poorly controlled type 2 diabetes. Diabetes Res Clin Pract. 2008;82:73-79. 9. Ehrhardt NM, Chellappa M, Walker MS, et al. The effect of real time continuous glucose monitoring on glycemic control in patients with type 2 diabetes mellitus. J Diabetes Sci Technol. 2011;5:668-675. 10. Vigersky RA, Fonda SJ, Chellappa M, et al. Short and long-term effects of real-time continuous glucose monitoring in patients with type 2 diabetes mellitus. Diabetes Care. 2012;35:32-38. 11. Fonda SJ, Salkind SJ, Walker MS, et al. Heterogeneity of responses to real-time continuous glucose monitoring
904 (RT-CGM) in patients with type 2 diabetes and its implications for application. Diabetes Care. 2013;36:786-792. 12. Huang ES, O’Grady M, Basu A, et al. The cost-effectiveness of continuous glucose monitoring in type 1 diabetes. Diabetes Care. 2010:33:1269-1274. 13. McQueen RB, Ellis SL, Campbell JD, et al. Cost-effectiveness of continuous glucose monitoring and intensive insulin therapy for type 1 diabetes. Cost Eff Resour Alloc. 2011;9:13-20. 14. Koro CE, Bowlin SJ, Bourgeois N, et al. Glycemic control from 1988 to 2000 among U.S. adults diagnosed with type 2 diabetes: a preliminary report. Diabetes Care. 2004;27:17-20. 15. Palmer AJ, Rose S, Valentine WJ, et al. Validation of the CORE Diabetes Model against epidemiological and clinical studies. Curr Med Res Opin. 2004;20(suppl 1):S27-S40. 16. Palmer AJ, Roze S, Valentine WJ, et al. The CORE Diabetes Model: projecting long-term clinical outcomes, costs and costeffectiveness of interventions in diabetes mellitus (types 1 and 2) to support clinical and reimbursement. Curr Med Res Opin. 2004;20(suppl 1):S5-S26. 17. Ryan RM, Deci EL. Intrinsic and extrinsic motivations: classic definitions and new directions. Contemp Educ Psychol. 2000;25;54-67. 18. Deci EL, Ryan RM. The “what” and “why” of goal pursuits: human needs and the self-determination of behavior. Psychol Inquiry. 2000;11:227-268. 19. Yeaw J, Lee WC, Aagren M, et al. Cost of self-monitoring of blood glucose in the United States among patients on an insulin regimen for diabetes. J Manag Care Pharm. 2012;18:21-32. 20. Lee WC, Conner C, Hammer M. Cost-effectiveness of liraglutide versus rosiglitazone, both in combination with glimepiride in treatment of type 2 diabetes in the U.S. Curr Med Res Opin. 2011;27:897-906. 21. Tunis SL, Willis WD, Foos V. Self-monitoring of blood glucose (SMBG) in patients with type 2 diabetes on oral anti-diabetic drugs: cost-effectiveness in France, Germany, Italy, and Spain. Curr Med Res Opin. 2010;26:163-75. 22. UK Prospective Diabetes Study (UKPDS) Group. Intensive blood-glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes (UKPDS 33). Lancet. 1998;352;837-853.
Journal of Diabetes Science and Technology 10(4) 23. Minshall ME, Oglesby AK, Wintle ME, et al. Estimating the long-term cost-effectiveness of exenatide in the United States: an adjunctive treatment for type 2 diabetes mellitus. Value Health. 2008;11:22-33. 24. Simon GE, Katon WJ, Lin EH, et al. Cost-effectiveness of systematic depression treatment among people with diabetes mellitus. Arch Gen Psychiatry. 2007;64:65-72. 25. Pelletier EM, Smith PJ, Boye KS, et al. Direct medical costs for type 2 diabetes mellitus complications in the US commercial payer setting: a resource for economic research. Appl Health Econ Health Policy. 2008;6:103-112. 26. Gruberg L, Weissman NJ, Waksman R, et al. The impact of obesity on the short-term and long-term outcomes after percutaneous coronary intervention: the obesity paradox? J Am Coll Cardiol. 2002;39:578-584. 27. Braithwaite RS, Meltzer DO, King JT Jr, et al. What does the value of modern medicine say about the $50,000 per quality-adjusted life-year decision rule? Med Care. 2008;46:349-356. 28. Tunis SL. Cost-effectiveness of self-monitoring of blood glucose (SMBG) for patients with type 2 diabetes and not on insulin: impact of modelling assumptions on recent Canadian findings. Appl Health Econ Health Policy. 2011;9:351-365. 29. Pollock RF, Valentine WJ, Goodall G, et al. Evaluating the cost-effectiveness of self-monitoring of blood glucose in type 2 diabetes patients on oral anti-diabetic agents. Swiss Med Wkly. 2010;140:w13103. 30. Tunis SL, Minshall ME. Self-monitoring of blood glucose in type 2 diabetes: cost-effectiveness in the United States. Am J Manag Care. 2008;14:131-140. 31. Tunis SL, Minshall ME. Self-monitoring of blood glucose (SMBG) for type 2 diabetes patients treated with oral anti-diabetic drugs and with a recent history of monitoring: cost-effectiveness in the U.S. Curr Med Res Opin. 2010;26:151-162. 32. Polonsky WH, Hessler D. Perceived accuracy in continuous glucose monitoring: understanding the impact on patients. J Diabetes Sci Technol. 2015;9:339-341. 33. Polonsky WH, Fisher L, Schikman CH, et al. Structured selfmonitoring of blood glucose significantly reduces A1C levels in poorly controlled, non-insulin treated type 2 diabetes. Diabetes Care. 2011;34:262-267.
research-article2016
DSTXXX10.1177/1932296816628547Journal of Diabetes Science and TechnologyFonda et al
Original Article
The Cost-Effectiveness of Real-Time Continuous Glucose Monitoring (RT-CGM) in Type 2 Diabetes
Journal of Diabetes Science and Technology 2016, Vol. 10(4) 898–904 © 2016 Diabetes Technology Society Reprints and permissions: sagepub.com/journalsPermissions.nav DOI: 10.1177/1932296816628547 dst.sagepub.com
Stephanie J. Fonda, PhD1, Claudia Graham, PhD2, Julie Munakata, MS3, Julia M. Powers, MHS4, David Price, MD2, and Robert A. Vigersky, MD1
Abstract Background: This analysis models the cost-effectiveness of real-time continuous glucose monitoring (RT-CGM) using evidence from a randomized controlled trial (RCT) that demonstrated RT-CGM reduced A1C, for up to 9 months after using the technology, among patients with type 2 diabetes not on prandial insulin. RT-CGM was offered short-term and intermittently as a self-care tool to inform patients’ behavior. Method: The analyses projected lifetime clinical and economic outcomes for RT-CGM versus self-monitoring of blood glucose by fingerstick only. The base-case analysis was consistent with the RCT (RT-CGM for 2 weeks on/1 week off over 3 months). A scenario analysis simulated outcomes of an RT-CGM “refresher” after the active intervention of the RCT. Analyses used the IMS CORE Diabetes Model and were conducted from a US third-party payer perspective, including direct costs obtained from published sources and inflated to 2011 US dollars. Costs and health outcomes were discounted at 3% per annum. Results: Life expectancy (LE) and quality-adjusted life expectancy (QALE) from RT-CGM were 0.10 and 0.07, with a cost of $653/patient over a lifetime. Incremental LE and QALE from a “refresher” were 0.14 and 0.10, with a cost of $1312/patient over a lifetime, and incremental cost-effectiveness ratios were $9319 and $13 030 per LY and QALY gained. Conclusions: RT-CGM, as a self-care tool, is a cost-effective disease management option in the US for people with type 2 diabetes not on prandial insulin. Repeated use of RT-CGM may result in additional cost-effectiveness. Keywords behavior, continuous glucose monitoring, cost-effectiveness, health economics, self-care, self-monitoring
Real-time continuous glucose monitoring (RT-CGM) has been shown to improve glycemic control and/or reduce the frequency of mild to moderate hypoglycemic episodes in people with type 1 diabetes and people with type 2 diabetes taking prandial insulin.1-8 Similarly, in a randomized controlled trial (RCT), Vigersky and colleagues demonstrated various glycemic benefits of a 3-month course of RT-CGM in people with type 2 diabetes not taking prandial insulin.9-11 Cost-effectiveness analyses of RT-CGM and glycemic control in people with type 1 diabetes have shown that lowering A1C through the use of RT-CGM could result in quality of life benefits within the range of what is deemed to be costeffective.12-13 Given the evidence that RT-CGM has clinical benefit for people with type 2 diabetes not taking prandial insulin (about 72.6% people with type 2 diabetes do not take
insulin of any type),14 it is important to evaluate whether extending RT-CGM use to this large cohort is potentially cost-effective. Since there are no studies of cost-effectiveness of RT-CGM in people with type 2 diabetes, we examined the potential impact of RT-CGM on lifetime clinical and
1
Walter Reed National Military Medical Center, Bethesda, MD, USA Dexcom, Inc, San Diego, CA, USA 3 IMS Health, San Francisco, CA, USA 4 University of Colorado Denver, Aurora, CO, USA 2
Corresponding Author: Stephanie J. Fonda, PhD, Walter Reed National Military Medical Center, 8901 Wisconsin Ave, Bethesda, MD 20889-5600, USA. Email: [email protected]
899
Fonda et al economic outcomes, using results from Vigersky and colleagues’ RCT.9,10
Methods CORE Diabetes Model (CDM) We used the validated IMS CORE Diabetes Model (CDM), which has been broadly applied.15,16 The CDM is a nonproduct-specific computer simulation model. The simulations predict the long-term health outcomes and costs of particular interventions for type 1 and/or type 2 diabetes. In this case, the intervention is short-term, intermittent use of RT-CGM versus self-monitoring of blood glucose (SMBG) by fingerstick among people with type 2 diabetes not taking prandial insulin. Data that inform the probabilities of events (such as hypoglycemia, amputation, a myocardial infarction, etc.), the progression of A1C, systolic blood pressure, lipids, and appropriate risk adjustments are derived from the United Kingdom Prospective Diabetes Study (UKPDS), the Diabetes Control and Complications Trial (DCCT), the Framingham Heart Study, and other sources. The CDM simulates disease progression in both type 1 and type 2 diabetes, and is widely used to estimate the impact of interventions on clinical and cost outcomes, as well as a range of economic analyses (cost-effectiveness, cost-utility, cost-benefit or cost of disease). It can also be used to identify potential high-risk patient profiles. The outputs of the model include life expectancy (LE), quality-adjusted life expectancy (QALE), direct and indirect costs, cumulative incidence and time to onset of complications, and incremental cost-effectiveness ratios per additional life year (LY) or quality-adjusted life year (QALY) gained.
Cohort Inputs—Interventional Study The characteristics of the cohort simulated by the CDM to predict the long-term health outcomes and costs of RT-CGM for people with type 2 diabetes not taking prandial insulin are based on the study by Vigersky and colleagues.9-11 Briefly, it was a 52-week, prospective, 2-arm RCT in 100 adult subjects comparing the short- (12-week) and long-term (40week) effectiveness of RT-CGM (n = 50 subjects) and SMBG (n = 50 subjects). Those randomized to RT-CGM used the Dexcom™ SEVEN® (Dexcom, Inc, San Diego, CA). The RT-CGM use occurred in 4 periods (2 weeks on/1 week off) over 12 weeks, for a total of 8 weeks of usage. Those randomized to SMBG were asked to test their glucose before meals and at bedtime for 12 weeks, as well as times associated with symptoms of hypo- or hyperglycemia. After the initial 12 weeks, all participants were asked to perform SMBG for the duration of the study as recommended by their usual provider. Follow-up study visits were performed at 3-week intervals during the first 12 weeks and then every 3 months during the remaining 9 months of the study.
RT-CGM was provided as a self-care tool to inform the participants’ behavior choices. The framework for the RCT was guided by self-determination theory, which is concerned with the processes through which people become motivated for initiating new self-beneficial behaviors and maintaining them over time.17,18 People feel autonomous and competent when they regulate their behavior through personal choice and self-endorsement. Support for competence occurs when people are afforded the skills and tools for change. The study staff did not provide any care management and the study participants’ providers did not have access to the RT-CGM data. Also, patients in the setting where this study took place are discouraged from taking part in more than 1 interventional study at a time, so to our knowledge, they had no other behavior interventions. The participants in the intervention were, on average 57.8 years old (Table 1). To be eligible for the RCT, they had a diagnosis of type 2 diabetes for at least 3 months, had an initial A1C of between 7% (53 mmol/mol) and 12% (108 mmol/mol), and were treated with diet/exercise alone or other glucose lowering therapies except prandial insulin. There was no difference in the use of antidiabetic medications between the RT-CGM and SMBG groups. By the end of year 1, the participants who had used RT-CGM had a net reduction in A1C of 1.1 percentage points ± 1.5 percentage points (12 mmol/mol ± 16.4 mmol/mol), and the participants who had used SMBG alone had a net reduction of 0.5 percentage points ± 1.3 percentage points (5.5 mmol/mol ± 14.2 mmol/mol). There was no statistically significant impact of the intervention on other physiological parameters. These data, as well as additional background data needed for the CDM, provided the necessary inputs.
Cost Inputs The cost inputs for the CDM included the costs of RT-CGM, SMBG, antidiabetic oral medications, insulin, routine management such as recommended screening, exams, and treatment for depression, and treatment of diabetes complications. No behavioral interventions that might be available to certain populations are included in the model, as these are not universal. SMBG is required to calibrate the Dexcom SEVEN (Dexcom, Inc, San Diego, CA) (twice daily), and the RT-CGM intervention was short-term and intermittent, so SMBG costs applied to both treatment groups (Table 2). Our assumptions regarding the frequency of SMBG, and therefore the cumulative costs of lancets and strips, were based on the observed frequency of SMBG in the interventional study, as previously reported,10 as well as published sources.19,20 RT-CGM costs applied to the RT-CGM group only, however, and the specific amounts were provided by Dexcom (Dexcom, Inc, San Diego, CA). Our assumptions regarding the annual cost of oral antidiabetic medications (Table 2) were based on the medications prescribed to the participants in the interventional study. The
900
Journal of Diabetes Science and Technology 10(4)
Table 1. Characteristics of the 100 Study Participants in RCT of RT-CGM in People With Type 2 Diabetes Not Taking Prandial Insulin. Variable
Base-case value
Demographics and clinical characteristics Age, mean years ± SD Proportion male (0-1) Proportion African American (0-1) A1C, mean %-points ± SD (mean mmol/mol ± SD) Duration of diabetes, mean years Systolic blood pressure, mean mm/Hg ± SD Total cholesterol, mean mg/dL ± SD HDL, mean mg/dL ± SD LDL, mean mg/dL ± SD Triglycerides, mean mg/dL ± SD Body mass index, mean kg/m2 ± SD Proportion with selected complications (0-1) Myocardial infarction Angina Peripheral vascular disease Stroke Heart failure Microalbuminuria Gross proteinuria Background diabetic retinopathy Proliferative diabetic retinopathy Macular edema Neuropathy Depression Impact of RT-CGM intervention Change in A1C from baseline at the end of year 1, mean percentage points ± SD (mean mmol/mol ± SD)
57.8 ± 10.8 0.55 0.52 8.3 ± 1.2 (67 ± 13.1) 9.0 ± 6.8 131.7 ± 17.8 166.1 ± 39.7 46.9 ± 13.8 94.1 ± 30.4 153.5 ± 110.7 32.3 ± 6.8 0.02 0.01 0.05 0.06 0.02 0.14 0.11 0.12 0.01 0.02 0.34 0.20 RT-CGM: –1.10 ± 1.50 (–12.0 ± 16.4) SMBG: –0.50 ± 1.30 (–5.5 ± 14.2)
medications were metformin (69% of the participants were taking this), sulfonylurea (48%), basal insulin (33%), pioglitazone (20%), exenatide (11%), sitagliptin (11%), glimepiride (5%), glyburide (4%), and repaglinide (3%). For each type of drug, we identified the lowest-cost products and the defined daily doses as reported in MediSpan PriceRx. The remainder of the direct costs, including costs for diabetes management (ie, exams, screening visits, etc.), cardiovascular disease complications, renal complications, acute events, eye disease, and neuropathy were obtained from published sources and inflated to 2011 US dollars.21-25 Costs and outcomes were discounted at 3% annually.
Analyses We explored 2 scenarios—the base case and a hypothetical refresher scenario. The base-case assumed no further use of RT-CGM after year 1, meaning the effect of the treatments on A1C observed in the RCT was applied for 1 year only. The refresher scenario assumed a second course of RT-CGM use, 1 year after the first course, in year 2. The second course
Table 2. Combined Costs by Treatment Group and CostEffectiveness Model. Costs Year 1, base-case scenario RT-CGM equipment SMBG strips and lancets Antidiabetic medications Total Year 2, base-case scenario RT-CGM equipment SMBG strips and lancets Antidiabetic medications Total Refresher scenario RT-CGM equipment SMBG strips and lancets Antidiabetic medications Total
RT-CGM group
SMBG group
$631.72 $369.80 $2083.00 $3084.52
$0.00 $369.80 $2083.00 $2452.80
$0.00 $174.24 $2083.00 $2257.24
$0.00 $174.24 $2083.00 $2257.24
$631.72 $174.24 $2083.00 $2888.96
$0.00 $174.24 $2083.00 $2257.24
Note: Costs of SMBG strips and lancets were higher in the year 1, basecase scenario because study participants in the RCT used for inputs to this analysis were asked to test 4 times daily. The costs are based on the frequency of testing that the study participants actually did.
would be identical to the first (4 periods of 2 weeks on/1 week off, distributed over 12 weeks, for a total of 8 weeks of usage). The refresher scenario also assumed that the second course of RT-CGM would maintain A1C at the level achieved from the first course. Convergence in A1C between the RT-CGM and SMBG treatment groups only would occur in the subsequent year of no RT-CGM use. This prediction was based on the observation that both groups from the original interventional study had an average A1C of 7.9% (63 mmol/ mol) at the end of year 2 (unpublished data from a follow-up analysis). This assumption may be conservative. The base-case and refresher scenarios assumed that patients in the simulated cohorts would continue on their diabetes medications until full uptake of insulin rescue therapy. There are no good data on the time to full uptake of insulin that can be applied to this cohort, so we used clinical judgment regarding the likely sequence of treatment and assumed time to full uptake of insulin to be by year 5 from the beginning of the most recent course of RT-CGM. To test the robustness of the cost-effectiveness, univariate and probabilistic sensitivity analyses were performed. All analyses were performed from a US third-party payer perspective, including direct costs obtained from published sources and inflated to 2011 US dollars. Outcomes were discounted at 3% per annum.
Results Base-Case Scenario The LE and QALE over a lifetime horizon are predicted to be 10.62 and 6.03 when RT-CGM is used, versus 10.52
901
Fonda et al Table 3. Lifetime Horizon Cost-Effectiveness of the Base-Case and Refresher Scenarios. RT-CGM Base-case scenario Life expectancy (LE) Quality-adjusted life expectancy (QALE) Total costs Incremental cost-effectiveness ratio per life-year (LY) gained Incremental cost-effectiveness ratio per quality-adjusted life-year (QALY) gained Refresher scenario Life expectancy (LE) Quality-adjusted life expectancy (QALE) Total costs Incremental cost-effectiveness ratio per life-year (LY) gained Incremental cost-effectiveness ratio per quality-adjusted life-year (QALY) gained
SMBG
Incremental
10.62 6.03 $66 094
10.52 0.10 (1.25 months) 5.96 0.07 (0.88 months) $65 441 $653 $6293 $8898
10.64 6.05 $66 734
10.50 0.14 (1.69 months) 5.95 0.10 (1.20 months) $65 423 $1312 $9319 $13 030
Notes: The outcomes in this table are rounded.
and 5.96 when SMBG alone is used (Table 3). The incremental LE and QALE were 0.10 and 0.07, respectively, meaning those who use RT-CGM could expect a gain of 1.25 months in LE or 0.88 quality-adjusted life months compared with using SMBG alone. The total lifetime costs in the RT-CGM and SMBG treatment arms were $66 094 and $65 441 respectively, and the incremental cost was $653 per patient over a lifetime. The incremental cost-effectiveness ratios were $6293 per LY gained and $8898 per QALY gained.
Refresher Scenario In the refresher scenario analysis, the LE and QALE over a lifetime horizon are estimated to be 10.64 and 6.05 when RT-CGM is used compared with 10.50 and 5.95 for SMBG alone (Table 3). The incremental LE and QALE were 0.14 and 0.10, or 1.69 months and 1.20 quality-adjusted life months. The total lifetime costs in the RT-CGM and SMBG treatment arms were anticipated to be $66 734 and $65 423, respectively, and the incremental cost was $1312 per patient over a lifetime. The incremental cost is approximately double that of the base-case scenario, reflecting the added cost of more RT-CGM-related supplies. The incremental costeffectiveness ratios were $9319 per LY gained and $13 030 per QALY gained.
Both Scenarios The use of RT-CGM translated into a reduction of cumulative rates of diabetes complications and deaths in cardiovascular disease, ulcers and amputations, and renal disease (Table 4). The one exception was stroke (death and event), which, paradoxically, is predicted to be higher for patients who use RT-CGM. This oddity is known as the “survival paradox” and has been observed for other health conditions and interventions.26
Sensitivity Analyses Probabilistic cost-effectiveness analysis suggests that the likelihood of the intervention being cost-effective is 70% at the willingness-to-pay threshold of $100 000 per QALY.
Discussion We have shown that intermittent, short-term use of RT-CGM is a clinically effective approach for lowering A1C for people with type 2 diabetes who are not taking prandial insulin.9,10 The present analysis demonstrates that, for this cohort, it is also a cost-effective disease management adjunct. RT-CGM is not an intervention per se for this cohort, but could be considered as part of a diabetes management strategy that includes lifestyle advice, as well as the choice of optimal management of diabetes using medications. A refresher course of RT-CGM may result in longer-term cost-effective health benefits, due to an additional year of impact on A1C and the consequent long-term health outcomes. Much of the cost-effectiveness of RT-CGM for this cohort is due to the low cost of the intervention. Specifically, the incremental cost-effectiveness ratios per LY gained and QALY gained in both the base-case and refresher scenarios were substantially below the commonly discussed willingness-to-pay range of $50 000-$100 000 per QALY gained. Recent studies suggest that the acceptable threshold in the US setting could actually be much higher—in the range of $109 000-$297 000 per QALY gained.27 By comparison, a cost-effectiveness analysis using inputs from the Juvenile Diabetes Research–Continuous Glucose Monitoring (CGM) Trials in people with type 1 diabetes found that CGM reduced the expected lifetime incidence of complications while also increasing costs; that is, the incremental cost-effectives ratio for their base-case was $98 679 per quality-adjusted life-year gained.11 Much of the costs were attributed to health care provider time used for training in the use of CGM and diabetes management (which were not costs relevant
902
Journal of Diabetes Science and Technology 10(4)
Table 4. Cumulative Incidence of Diabetes Complications for the Base-Case and Refresher Scenarios, by Treatment Group and as Percentage of the Cohort. Base-case scenario Congestive heart failure death (%) Congestive heart failure event (%) Peripheral vascular disease onset (%) Angina (%) Stroke death (%) Stroke event (%) Myocardial infarction death (%) Myocardial infarction event (%) Ulcer (%) Recurrent ulcer (%) Amputation ulcer (%) Amputation recurrent ulcer (%) Neuropathy (%) Microalbuminuria (%) Gross proteinuria (%) End-stage renal disease (%) Nephropathy death (%)
Refresher scenario
RT-CGM
SMBG
RT-CGM
SMBG
21.32 23.43 7.74 8.96 13.00 18.18 11.63 14.88 26.86 33.31 8.27 2.61 52.81 24.61 10.23 5.52 3.10
21.41 23.50 7.91 9.07 12.93 17.94 11.78 15.26 27.83 34.70 8.68 2.62 54.78 26.27 11.24 6.08 3.45
21.45 23.49 7.82 9.03 13.11 18.30 10.93 14.12 27.02 33.87 8.36 2.54 53.24 24.96 10.46 5.64 3.21
21.48 23.67 7.92 9.03 12.85 17.82 11.82 15.39 27.89 34.64 8.65 2.66 55.00 26.42 11.35 6.12 3.52
to Vigersky and colleagues’ intervention), health service utilization as well as supply costs for the CGM and confirmatory SMBG testing. Another cost-effectiveness examination in people with type 1 diabetes, comparing CGM plus intensive insulin therapy versus SMBG plus intensive insulin therapy, found that CGM plus intensive insulin therapy resulted in an expected improvement in cost-effectiveness of 0.52 QALYs and an incremental cost-effectiveness ratio of $45 033 per QALY.13 In addition, our findings for short-term RT-CGM use for the simulated cohort compared favorably to previous costeffectiveness results for SMBG in people with type 2 diabetes. In an examination of SMBG in Canadian patients with type 2 diabetes who were not using insulin, Tunis and the Canadian Optimal Prescribing and Utilization Service (COMPUS) estimated incremental cost-effectiveness ratios to be Can$39 231 and Can$54 349 per QALY gained for adherence rates of 66% and 87%, respectively.28 In the Swiss context, 1, 2, or 3 times daily SMBG in patients with type 2 diabetes treated with oral antidiabetic agents resulted in incremental cost-effectiveness ratios of CHF9177, CHF12 928 and CHF17 342 per QALY gained, respectively, which are below the commonly accepted willingness-to-pay thresholds in Switzerland.29 In the US context, Tunis and Minshall estimated that the cost per QALY gained for SMBG performed 1 or 3 times per day among people with type 2 diabetes on oral antidiabetic agents was $7856 and $6601.30 A later study of this US cohort estimated that SMBG performed 1, 2, or 3 times daily increased QALY by 0.05, 0.12, and 0.13, and the incremental cost-effectiveness ratios were $26 206, $18 572, and $25 436 per QALY gained, increasing with SMBG frequency.31
A fundamental difference between this analysis and previous cost-effectiveness studies of monitoring in people with type 2 diabetes is the length of time to which the cost of the intervention was applied. In the interventional study on which this cost-effectiveness analysis was based, the use of RT-CGM in year 1 resulted in statistically significant improvement in A1C in the first 3 months of the study, and this was sustained until the end of the year, 9 months after the completion of the active intervention. This is a relatively short-term intervention, yet it had long-term health consequences and long-term cost-effective health benefits. The cost-effectiveness analysis predicted a small effect on quantity and quality of life. This small effect is typical of behavioral interventions, where no prolonged action of the intervention is assumed. RT-CGM, as used in the original clinical trial, is a tool to support learning and subsequent self-directed, behavioral choices because no interpretation of RT-CGM data by a clinician was provided, nor did study staff adjust diabetes therapies; physiologic effects were due to participants’ actions. Clinician interpretation might increase costs, but might also increase the beneficial effects of this technology, and thereby offset costs. A limitation of this analysis is the small size of the original study (n = 100) that provided many of the clinical inputs for the CDM. However, the original study was the longest and largest randomized clinical trial to date of RT-CGM use in people with type 2 diabetes not taking prandial insulin. Also, the original, interventional study used an older RT-CGM system (Dexcom SEVEN). Newer systems (eg, Dexcom G4 Platinum®) have features that may make them more user-friendly and, therefore, better tools to support
903
Fonda et al behavior than older systems. Specifically, they are more accurate, patients wearing a sensor can travel a greater distance from the handheld receiver and still transmit data, and they have smaller handheld receivers which may make them easier to carry in a pocket or purse. The aforementioned improved accuracy may result in greater trust of the information, which may enhance use of the RT-CGM device and of the RT-CGM information.32 Another potential limitation is that this analysis (and the original intervention) did not compare RT-CGM with structured SMBG, similar to that used in the Structured Testing Program (STeP).33 STeP had patients test their blood glucose 7 times in the 3 days prior to a health care visit (fasting, preprandial, 2 hours postprandial at each meal, and bedtime). The testing data were used to guide health care providers’ decision making, rather than to guide patient decision making, which was the focus of the RT-CGM RCT. Nevertheless, structured testing in the manner prescribed by STeP may be helpful for patient decision making, and this might also result in quality of life benefits considered to be cost effective. However, structured testing is different from RT-CGM in that it may not identify the true peak or nadir glucose, does not show the duration of hyperglycemia or hypoglycemia, does not show the speed of glucose rise or fall, does not alert a user to highs or lows, does not frame the results (with high or low thresholds), and is not necessarily structured in such a way as to show glucose responses to behaviors other than what one ate (eg, exercise, timing of medication, not getting enough sleep, etc). A strength of this study is that it is the first to date to evaluate the cost-effectiveness of RT-CGM in people with type 2 diabetes (on any treatment). Another is that it considers the cost-effectiveness of a refresher course; although, admittedly, this consideration is speculative since it is unclear at this time if a refresher course would be effective in maintaining (or perhaps further improving) glycemic control and, if so, how often and how long such a refresher course should be.
Conclusions In sum, RT-CGM was excellent value under the conditions of the original, interventional study, namely short-term, intermittent use in people with type 2 diabetes who were not taking prandial insulin at the time they began using RT-CGM and who used the technology as a tool to inform behavioral choices without clinician guidance. Future research might address the value of RT-CGM in a larger cohort with type 2 diabetes, including those on prandial insulin, and the costeffectiveness of using newer RT-CGM systems and/or clinician input into the behavior decisions that patients make in response to their RT-CGM data. Abbreviations CDM, CORE Diabetes Model; COMPUS, Canadian Optimal Prescribing and Utilization Service; LE, life expectancy; LY, life year; QALE, quality-adjusted life expectancy; RCT, randomized
controlled trial; RT-CGM, real-time continuous glucose monitoring; SMBG, self-monitoring of blood glucose; STeP, Structured Testing Program; UKPDS, United Kingdom Prospective Diabetes Study.
Declaration of Conflicting Interests The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: CG and DP are employees of Dexcom, Inc. No other competing financial interests exist.
Funding The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The project was sponsored by an investigator-initiated grant from Dexcom, Inc to SJF and RAV. The opinions expressed in this article reflect the personal views of the authors and not the official views of the US Army or the Department of Defense.
References 1. Juvenile Diabetes Research Foundation Continuous Glucose Monitoring Study Group. Continuous glucose monitoring and intensive treatment of type 1 diabetes. N Engl J Med. 2008;359:1464-1476. 2. Juvenile Diabetes Research Foundation Continuous Glucose Monitoring Study Group. The effect of continuous glucose monitoring in well-controlled type 1 diabetes. Diabetes Care. 2009;32:1378-1383. 3. Deiss D, Bolinder J, Riveline JP, et al. Improved glycemic control in poorly controlled patients with type 1 diabetes using real-time continuous glucose monitoring. Diabetes Care. 2006;29:2730-2732. 4. O’Connell MA, Donath S, O’Neal DN, et al. Glycaemic impact of patient-led use of sensor-guided pump therapy in type 1 diabetes: a randomised controlled trial. Diabetologia. 2009;52:1250-1257. 5. Battelino T, Phillip M, Bratina N, et al. Effect of continuous glucose monitoring on hypoglycemia in type 1 diabetes. Diabetes Care. 2011;34:795-800. 6. Garg S, Zisser H, Schwartz S, et al. Improvement in glycemic excursions with a transcutaneous, real-time continuous glucose sensor. Diabetes Care. 2006;29:44-50. 7. Bailey TS, Zisser HC, Garg SK. Reduction in hemoglobin A1C with RT-CGM: results from a 12-week observational study. Diabetes Technol Ther. 2007;9:203-210. 8. Yoo HJ, An HG, Park SY, et al. Use of a real time continuous monitoring system as a motivational device for poorly controlled type 2 diabetes. Diabetes Res Clin Pract. 2008;82:73-79. 9. Ehrhardt NM, Chellappa M, Walker MS, et al. The effect of real time continuous glucose monitoring on glycemic control in patients with type 2 diabetes mellitus. J Diabetes Sci Technol. 2011;5:668-675. 10. Vigersky RA, Fonda SJ, Chellappa M, et al. Short and long-term effects of real-time continuous glucose monitoring in patients with type 2 diabetes mellitus. Diabetes Care. 2012;35:32-38. 11. Fonda SJ, Salkind SJ, Walker MS, et al. Heterogeneity of responses to real-time continuous glucose monitoring
904 (RT-CGM) in patients with type 2 diabetes and its implications for application. Diabetes Care. 2013;36:786-792. 12. Huang ES, O’Grady M, Basu A, et al. The cost-effectiveness of continuous glucose monitoring in type 1 diabetes. Diabetes Care. 2010:33:1269-1274. 13. McQueen RB, Ellis SL, Campbell JD, et al. Cost-effectiveness of continuous glucose monitoring and intensive insulin therapy for type 1 diabetes. Cost Eff Resour Alloc. 2011;9:13-20. 14. Koro CE, Bowlin SJ, Bourgeois N, et al. Glycemic control from 1988 to 2000 among U.S. adults diagnosed with type 2 diabetes: a preliminary report. Diabetes Care. 2004;27:17-20. 15. Palmer AJ, Rose S, Valentine WJ, et al. Validation of the CORE Diabetes Model against epidemiological and clinical studies. Curr Med Res Opin. 2004;20(suppl 1):S27-S40. 16. Palmer AJ, Roze S, Valentine WJ, et al. The CORE Diabetes Model: projecting long-term clinical outcomes, costs and costeffectiveness of interventions in diabetes mellitus (types 1 and 2) to support clinical and reimbursement. Curr Med Res Opin. 2004;20(suppl 1):S5-S26. 17. Ryan RM, Deci EL. Intrinsic and extrinsic motivations: classic definitions and new directions. Contemp Educ Psychol. 2000;25;54-67. 18. Deci EL, Ryan RM. The “what” and “why” of goal pursuits: human needs and the self-determination of behavior. Psychol Inquiry. 2000;11:227-268. 19. Yeaw J, Lee WC, Aagren M, et al. Cost of self-monitoring of blood glucose in the United States among patients on an insulin regimen for diabetes. J Manag Care Pharm. 2012;18:21-32. 20. Lee WC, Conner C, Hammer M. Cost-effectiveness of liraglutide versus rosiglitazone, both in combination with glimepiride in treatment of type 2 diabetes in the U.S. Curr Med Res Opin. 2011;27:897-906. 21. Tunis SL, Willis WD, Foos V. Self-monitoring of blood glucose (SMBG) in patients with type 2 diabetes on oral anti-diabetic drugs: cost-effectiveness in France, Germany, Italy, and Spain. Curr Med Res Opin. 2010;26:163-75. 22. UK Prospective Diabetes Study (UKPDS) Group. Intensive blood-glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes (UKPDS 33). Lancet. 1998;352;837-853.
Journal of Diabetes Science and Technology 10(4) 23. Minshall ME, Oglesby AK, Wintle ME, et al. Estimating the long-term cost-effectiveness of exenatide in the United States: an adjunctive treatment for type 2 diabetes mellitus. Value Health. 2008;11:22-33. 24. Simon GE, Katon WJ, Lin EH, et al. Cost-effectiveness of systematic depression treatment among people with diabetes mellitus. Arch Gen Psychiatry. 2007;64:65-72. 25. Pelletier EM, Smith PJ, Boye KS, et al. Direct medical costs for type 2 diabetes mellitus complications in the US commercial payer setting: a resource for economic research. Appl Health Econ Health Policy. 2008;6:103-112. 26. Gruberg L, Weissman NJ, Waksman R, et al. The impact of obesity on the short-term and long-term outcomes after percutaneous coronary intervention: the obesity paradox? J Am Coll Cardiol. 2002;39:578-584. 27. Braithwaite RS, Meltzer DO, King JT Jr, et al. What does the value of modern medicine say about the $50,000 per quality-adjusted life-year decision rule? Med Care. 2008;46:349-356. 28. Tunis SL. Cost-effectiveness of self-monitoring of blood glucose (SMBG) for patients with type 2 diabetes and not on insulin: impact of modelling assumptions on recent Canadian findings. Appl Health Econ Health Policy. 2011;9:351-365. 29. Pollock RF, Valentine WJ, Goodall G, et al. Evaluating the cost-effectiveness of self-monitoring of blood glucose in type 2 diabetes patients on oral anti-diabetic agents. Swiss Med Wkly. 2010;140:w13103. 30. Tunis SL, Minshall ME. Self-monitoring of blood glucose in type 2 diabetes: cost-effectiveness in the United States. Am J Manag Care. 2008;14:131-140. 31. Tunis SL, Minshall ME. Self-monitoring of blood glucose (SMBG) for type 2 diabetes patients treated with oral anti-diabetic drugs and with a recent history of monitoring: cost-effectiveness in the U.S. Curr Med Res Opin. 2010;26:151-162. 32. Polonsky WH, Hessler D. Perceived accuracy in continuous glucose monitoring: understanding the impact on patients. J Diabetes Sci Technol. 2015;9:339-341. 33. Polonsky WH, Fisher L, Schikman CH, et al. Structured selfmonitoring of blood glucose significantly reduces A1C levels in poorly controlled, non-insulin treated type 2 diabetes. Diabetes Care. 2011;34:262-267.
