Practical Implementation of Diabetes Technology: Real‐World Use

Introduction

The purpose of the practical implementation article of the ATTD Yearbook is to characterize the current diabetes technology landscape around the globe, focusing on emerging trends in device uptake, clinical utility and guidance, access to devices, educational strategies, and perceived benefit and burden of device use (1). Articles selected for this review were published between July 2018 and June 2019 and contained original research related to diabetes technology use in the real world rather than a structured research setting. Keyword searches included “education, access, economics, cost effectiveness, real world, behavior, barriers, registry, outcomes, implementation, decision support, and technology.” A total of 42 articles were retrieved, and 10 were selected for review here.

This year's article starts with highlighting new data from the Type 1 Diabetes (T1D) Exchange Registry from 2016 to 2018, which provides relevant insight into diabetes technology trends in the United States (2). As highlighted below, technology uptake has rapidly expanded in the past decade. This is most notable in use of continuous glucose monitoring (CGM), with a five‐fold increase in reported use compared with 8 years prior, and a 10‐fold increase in use for children <12 years old. Insulin pump use has slightly increased as well. Technophiles can celebrate this trend.

The second major point is more sobering: despite this increase in diabetes technology utilization, glycemic control has not improved overall. Glycemic data mirror the data from 8 years prior, with an actual worsening of glycemic control in adolescents. This paradox suggests a more nuanced picture of technology effectiveness, a picture that must include individual factors, clinical factors, and system factors all contributing to the “real‐world” glycemic benefits of technology.

This year's article therefore focuses on extra‐technological factors that support or hinder efforts of individuals using diabetes technology to manage diabetes. First, access, insurance coverage, and cost‐effectiveness data on sensor augmented pump (SAP) and hybrid closed loop (HCL) are reviewed. Next, new data from educational randomized control trials (RCT) are presented to emphasize the essential place for education in all diabetes self‐management. Finally, two studies of automated decision support are reviewed in order to illustrate some of the potential benefits and challenges of bringing this technology to the clinical environment.

Key Articles Reviewed for the Article.

State of type 1 diabetes management and outcomes from the T1D Exchange in 2016–2018

Foster NC, Beck RW, Miller KM, Clements MA, Rickels MR, DiMeglio LA, Maahs DM, Tamborlane WV, Bergenstal R, Smith E, Olson BA, Garg SK; for the T1D Exchange Clinic Network

Diabetes Technol Ther 2019;21: 66–72. Erratum in: Diabetes Technol Ther 2019;21: 230

Persistent heterogeneity in diabetes technology reimbursement for children with type 1 diabetes: the SWEET perspective

Sumnik Z, Szypowska A, Iotova V, Bratina N, Cherubini V, Forsander G, Jali S, Raposo JF, Stipančic G, Vazeou A, Veeze H, Lange K; on behalf of the SWEET study group

Pediatr Diabetes 2019;20: 434–443

Impact of Medicare continuous subcutaneous insulin infusion policies in patients with type 1 diabetes

Argento NB, Liu J, Hughes AS, McAuliffe‐Fogarty AH

J Diabetes Sci Technol 2019 Mar 31; 1932296819838292. [Epub ahead of print] DOI: 10.1177/1932296819838292

Cost‐effectiveness of sensor‐augmented insulin pump therapy vs continuous subcutaneous insulin infusion in patients with type 1 diabetes in the Netherlands

Roze S, Smith‐Palmer J, de Portu S, Delbaere A, de Brouwer B, de Valk HW

Clinicoecon Outcomes Res 2019;11: 73–82

Cost‐effectiveness analysis of the MiniMed 670G hybrid closed‐loop system versus continuous subcutaneous insulin infusion for treatment of type 1 diabetes

Jendle J, Pöhlmann J, de Portu S, Smith‐Palmer J, Roze S

Diabetes Technol Ther 2019;21: 110–118

Efficacy of an education program for people with diabetes and insulin pump treatment (INPUT): results from a randomized controlled trial

Ehrmann D, Kulzer B, Schipfer M, Lippmann‐Grob B, Haak T, Hermanns N

Diabetes Care 2018;41: 2453–2462

The impact of a structured education and treatment programme (FLASH) for people with diabetes using a flash sensor‐based glucose monitoring system: results of a randomized controlled trial

Hermanns N, Ehrmann D, Schipfer M, Kröger J, Haak T, Kulzer B

Diabetes Res Clin Pract 2019;150: 111–121

Pediatric endocrinology trainees' education and knowledge about insulin pumps and continuous glucose monitors

Marks BE, Wolfsdorf JI, Waldman G, Stafford DE, Garvey KC

Diabetes Technol Ther 2019;21: 105–109

Adjusting insulin doses in patients with type 1 diabetes who use insulin pump and continuous glucose monitoring: variations among countries and physicians

Nimri R, Dassau E, Segall T, Muller I, Bratina N, Kordonouri O, Bello R, Biester T, Dovc K, Tenenbaum A, Brener A, Šimunović M, Sakka SD, Nevo Shenker M, Passone CG, Rutigliano I, Tinti D, Bonura C, Caiulo S, Ruszala A, Piccini B, Giri D, Stein R, Rabbone I, Bruzzi P, Omladič JŠ, Steele C, Beccuti G, Yackobovitch‐Gavan M, Battelino T, Danne T, Atlas E, Phillip M

Diabetes Obes Metab 2018;20: 2458–2466

Automated insulin dosing guidance to optimise insulin management in patients with type 2 diabetes: a multicentre, randomised controlled trial

Bergenstal RM, Johnson M, Passi R, Bhargava A, Young N, Kruger DF, Bashan E, Bisgaier SG, Isaman DJM, Hodish I

Lancet 2019;393: 1138–1148

State of type 1 diabetes management and outcomes from the T1D Exchange in 2016–2018

Foster NC¹, Beck RW¹, Miller KM¹, Clements MA², Rickels MR³, DiMeglio LA⁴, Maahs DM⁵, Tamborlane WV⁶, Bergenstal R⁷, Smith E¹, Olson BA⁷, Garg SK⁸; for the T1D Exchange Clinic Network

¹Jaeb Center for Health Research, Tampa, FLA; ²Endocrine/Diabetes Department, Children's Mercy Hospital, Kansas City, MO; ³Rodebaugh Diabetes Center, University of Pennsylvania Perelman School of Medicine, Philadelphia, PA; ⁴Department of Pediatrics, Indiana University School of Medicine, Indianapolis, IN; ⁵Department of Pediatrics–Endocrinology, Stanford University, Stanford, CA; ⁶Pediatric Endocrinology and Diabetes, Yale University School of Medicine, New Haven, CT; ⁷International Diabetes Center Park Nicollet, Minneapolis, MN; ⁸Barbara Davis Center for Childhood Diabetes, University of Colorado Denver, Aurora, CO

Diabetes Technol Ther 2019;21: 66–72. Erratum in: Diabetes Technol Ther 2019;21: 230

Background

Data from the T1D Exchange Registry are published every few years to provide longitudinal metrics by which to assess trends in T1D management and glycemic control in the United States, and allow comparison with other worldwide registries.

Methods

T1D Exchange data was collected from 81 pediatric and adult endocrinology practices in 35 states. The current report includes data from 22,697 registry participants (ages 1–93 years), collected between 2016–2018. Data on diabetes management (including diabetes technology usage) and metrics on glycemic outcomes were collected and compared with data collected by the registry in 2010–2012.

Results

Mean glycated hemoglobin (HbA1c) levels changed little between 2010–2012 and 2016–2018, though adolescents demonstrated a higher mean HbA1c in 2016–2018 compared with previous data. Mean HbA1c was 8.1% (65 mmol/mol) at age 5 years, increasing to 9.3% (78 mmol/mol) between the ages of 15 to 18 years, decreasing to 8.0% (65 mmol/mol) by age 28 and remaining stable at 7.5%–7.9% (58–63 mmol/mol) after age 30. Only 17% of youth achieved the goal HbA1c of <7.5% (58 mmol/mol), and only 21% of adults met the <7.0% (53 mmol/mol) goal HbA1c, set forth by the American Diabetes Association. Insulin pump use increased slightly, from 57% in the earlier cohort to 63% in the recent data. CGM use dramatically increased from 7% in 2010–2012 to 30% in 2016–2018, with the most significant increase in uptake being in ages <12 years old. Individuals who used CGM had better glycemic control (lower HbA1c) than non‐CGM users regardless of insulin modality (injections or insulin pump). There were significant differences in diabetes technology use by race/ethnicity, with lower utilization of both pumps and sensors in non‐Hispanic Black individuals than in both Hispanic and non‐Hispanic white persons across all age groups.

Conclusion

Despite a significant increase in diabetes technology adoption, glycemic control has not improved in the United States for individuals with T1D and has actually worsened in the adolescent age range.

Comment.

Registry data are important for benchmarking the practical march of diabetes technology from research to real world settings. These data represent only registry participants in the United States, and new data from Europe and worldwide are needed to present a more comprehensive picture of global utilization. As mentioned in the introduction to this article, two notable points from these data are that (a) diabetes technology uptake is increasing in the United States; and (b) despite these increases, glycemic control is not improving. Similar data have been previously published for Europe (3) and other regions of the world (4).

An additional takeaway from these data is that CGM appears to offer greater glycemic improvement compared with insulin pumps (see Fig. 9.1). CGM use was categorically associated with lower HbA1c levels regardless of whether the individual was using insulin pump or injections. These are important data to consider within a healthcare environment of constricting and restricting reimbursements, as CGM appears to offer the greater “bang for the buck” and suggest that clinicians should be encouraging CGM use even in those patients who are committed to injection regimens.

Access and Cost‐Effectiveness

Persistent heterogeneity in diabetes technology reimbursement for children with type 1 diabetes: the SWEET perspective

Sumnik Z¹, Szypowska A², Iotova V³, Bratina N⁴, Cherubini V⁵, Forsander G⁶, Jali S⁷, Raposo JF⁸, Stipančic G⁹, Vazeou A¹⁰, Veeze H¹¹, Lange K¹²; on behalf of the SWEET study group

¹Department of Pediatrics, Motol University Hospital, Prague, Czech Republic; ²Department of Pediatrics, Medical University of Warsaw, Warsaw, Poland; ³Department of Pediatrics, Medical University–Varna, UMHAT “Sv. Marina,” Varna, Bulgaria; ⁴Department of Endocrinology, Diabetes and Metabolism, UMC, University Children's Hospital, Ljubljana, Slovenia; ⁵Department of Women's and Children's Health, Azienda Ospedaliero‐Universitaria, Ospedali Riuniti di Ancona,“G. Salesi” Hospital, Ancona, Italy; ⁶Institute for Clinical Sciences, Dept of Ped, University of Gothenburg and the Queen Silvia Children's Hospital, Sahlgrenska Univ. Hospital, Gothenburg, Sweden; ⁷Department of Pediatrics, KLE University's Jawaharlal Nehru Medical College Belgaum, Belgaum, India; ⁸APDP‐Diabetes Portugal and Nova Medical School, Lisbon, Portugal; ⁹Department of Pediatrics, University Hospital Center “Sestre milosrdnice,” School of Dental Medicine, University of Zagreb, Zagreb, Croatia; ¹⁰Department of Pediatrics, Diabetes Center, P & A Kyriakou Children's Hospital, Athens, Greece; ¹¹Diabeter, Rotterdam, The Netherlands; ¹²Department of Medical Psychology, Hannover Medical School, Hannover, Germany

Pediatr Diabetes 2019;20: 434–443

Background

Developments in diabetes technology have continued to advance the care of children with T1D, enabling improved metabolic control without increasing the risk of hypoglycemia. However, access to these technologies for children with T1D can be severely limited by socioeconomic status and type of health insurance of the family, and the structure of the healthcare system and regulatory decisions of private and public payers. The purpose of this study was to evaluate the current reimbursement situation for insulin, glucose meters (and strips), insulin pumps, and CGM among the European countries participating in the SWEET project, an international network of pediatric diabetes centers established to reduce inequalities in access to and quality of pediatric diabetes care.

Methods

Surveys comprising questions about reimbursement (including conditions or limits) for glucose meters/strips, insulin pumps, and CGM devices were distributed to 24 members of the SWEET consortium (one per country), five representatives from non‐SWEET countries, and representatives of diabetes technology manufacturers. Data were collected over a 6‐month period between March and August 2017 and were compared with a previous survey of 27 European Union countries from 2009.

Results

Insulins are almost completely covered, with small exceptions for analogs in some countries, and analogs have become cheaper with the growth of biosimilars. There is great variability in coverage for glucose meter strips, although the general trend is toward unlimited strip reimbursement. Unlimited insulin pump reimbursement for children has increased since 2009, with some restrictions observed in a few countries based on age. Complete or partial reimbursement has been steadily increasing as this technology has matured, although some conditional restrictions continue to be in place (e.g., recurrent or nocturnal hypoglycemia). Most countries do not have compulsory outcome requirements for continued access to pumps or sensors. Reimbursement for “flash” glucose monitoring (FGM) was noted to be evolving so quickly that results could not be considered accurate; interestingly, there is a trend for reimbursement for FGM to negatively impact concomitant reimbursement for glucose test strips.

Conclusions

There is great variability among European countries participating in the SWEET consortium in reimbursement for diabetes‐related technologies, although the general trend is toward improved access for children with T1D. It is noted that reimbursement for test strips is insufficient to meet recommended testing frequency, and there is a trend for some conflict between reimbursement for strips and sensors. It is critical that children with T1D continue to have access to these important technologies to reduce inequities in care.

Impact of Medicare continuous subcutaneous insulin infusion policies in patients with type 1 diabetes

Argento NB¹, Liu J², Hughes AS², McAuliffe‐Fogarty AH²

¹Maryland Endocrine and Diabetes, Columbia, MD; ²T1D Exchange, Boston, MA

J Diabetes Sci Technol 2019 Mar 31; 1932296819838292. [Epub ahead of print] DOI: 10.1177/1932296819838292

Background

The Centers for Medicare and Medicaid Services (CMS) provide healthcare to over 15% of the United States population, and over 20% of CMS recipients have diabetes. With improved life expectancy and technology developments, insulin pumps are being used with increasing frequency in the Medicare population, but specific policies and regulations of CMS regarding qualification to initiate or maintain pump use may negatively impact access and outcomes in Medicare users. The purpose of this study was to evaluate how current CMS pump policies affect diabetes self‐management behaviors and health outcomes in Medicare users utilizing insulin pump therapy.

Methods

A 59‐question multiple‐choice/open‐ended question survey regarding experience with pump use while under Medicare coverage was conducted in adults with T1D identified through participation in the T1D Exchange Glu online community. Eligibility criteria included age >18 years, T1D >1 year, currently or previously on pump therapy >6 months, and under Medicare coverage >6 months.

Results

Among 241 adults with median age of 67 years, mean diabetes duration of 38 years, mean A1c 7.0%, and mean pump use 15 years, most individuals experienced problems in obtaining or maintaining coverage. Half of the patients new to pumps while on Medicare encountered problems trying to obtain coverage due to cost of pumps and supplies as well as difficulties finding Medicare providers who prescribed pumps. Almost three‐quarters of patients already using pump therapy before going on Medicare reported problems such as having to switch pump brand, insulin, or supplier. New and experienced pump users encountered delays in obtaining supplies due to paperwork issues and difficulty in meeting provider quarterly visit requirements. These delays frequently resulted in adverse health behaviors, including prolonging use of infusion sets, temporarily discontinuing pump use, and reducing basal rates to prolong use, which resulted in increased hyperglycemia, blood glucose variability, and pain/scarring at infusion sites. Over half of patients reported that their pump experience was worse on Medicare coverage than with their prior insurance.

Conclusions

Problems with obtaining and maintaining coverage and supplies for insulin pump therapy are common in patients with T1D on Medicare. Current CMS policies and regulations such as paperwork burden, required blood testing, quarterly provider visits, and warranty coverage frequently impact patients and resulted in compensatory nonrecommended behaviors and adverse clinical outcomes. Reevaluation of Medicare policies and regulations to allow for individualized patient requirements are needed to improve patient experience and outcomes.

Cost‐effectiveness of sensor‐augmented insulin pump therapy vs continuous subcutaneous insulin infusion in patients with type 1 diabetes in the Netherlands

Roze S¹, Smith‐Palmer J², de Portu S³, Delbaere A³, de Brouwer B⁴, de Valk HW⁵

¹HEVA HEOR, Lyon, France; ²Ossian Health Economics and Communications, Basel, Switzerland; ³Medtronic International Trading Sàrl, Tolochenaz, Switzerland; ⁴Medtronic Trading NL, Heerlen, the Netherlands; ⁵Leiden University Medical Center, Leiden, the Netherlands

Clinicoecon Outcomes Res 2019;11: 73–82

Background

The use of a sensor‐augmented pump (SAP) greatly extends the functionality of an insulin pump by providing additional information from the CGM and the potential for advanced features such as low glucose suspend or predicted low glucose suspend. The cost‐effectiveness of SAP compared with insulin pump alone is unknown; this study was undertaken to determine cost‐effectiveness in the Netherlands.

Methods

This cost‐effectiveness analysis was performed in two T1D cohorts: individuals with suboptimal glycemic control (mean age 27 years, 51% female, baseline HbA1c 8.0% [64 mmol/mol]) and individuals with impaired hypoglycemia awareness (mean age 18.6 years, 51% female), indicating a higher risk for hypoglycemia. This study used the IQVIA CORE Diabetes Model with clinical data inputs from published literature on T1D morbidity and mortality. Analysis was performed using the societal perspective (i.e., incorporating both direct and indirect costs).

Results

In the suboptimal glycemic control cohort, SAP improved quality‐adjusted life expectancy by 1.77 quality‐adjusted life years (QALYs) when compared with insulin pump (15.54 QALYs vs 13.77 QALYs respectively). The lifetime cost of SAP was higher (€189,855 [EUR] vs €150,366 for insulin pump), and the resulting incremental cost‐effectiveness ratio (ICER) for SAP was €22,325 per QALY gained. SAP was associated with a 1‐year delay in onset of diabetes related complications. Sensitivity analysis indicated that the results were sensitive to quality‐of‐life assumptions around fear of hypoglycemia and also baseline HbA1c.

For the impaired hypoglycemia awareness cohort, SAP led to 2.16 QALYs gain compared with insulin pump (16.70 QALYs vs 14.53 QALYs) with higher lifetime costs (€204,013 vs €171,032). Therefore, the ICER for SAP was 15,243 per QALY gained. Sensitivity analysis indicated that analysis was strongly influenced by assumptions on rates of severe hypoglycemia for individuals on insulin pump alone.

Conclusion

From a societal perspective, SAP is likely to be cost‐effective both for individuals with suboptimally controlled T1D and individuals with impaired hypoglycemia awareness when compared with insulin pump alone.

Cost‐effectiveness analysis of the MiniMed 670G hybrid closed‐loop system versus continuous subcutaneous insulin infusion for treatment of type 1 diabetes

Jendle J¹, Pöhlmann J², de Portu S³, Smith‐Palmer J², Roze S⁴

¹Faculty of Medical Sciences, Örebro University, Örebro, Sweden; ²Ossian Health Economics and Communications GmbH, Basel, Switzerland; ³Medtronic International Trading Sàrl, Tolochenaz, Switzerland; ⁴HEVA HEOR, Lyon, France

Diabetes Technol Ther 2019;21: 110–118

Background

Hybrid closed loop (HCL) therapy provides additional benefits over insulin pump therapy in the ability to automatically increase or decrease insulin delivery based on sensor‐reported glucose levels. Although numerous studies have demonstrated the glycemic benefit of HCL, the cost‐effectiveness of the system is not known. This study was undertaken to determine cost‐effectiveness of HCL in Sweden.

Methods

This study used data from the 3‐month pre–post Minimed 670G safety study performed in the United States (T1D participants mean age 37.8, 56% female, 21.7 years duration of diabetes), and clinical data in Sweden on rates of severe hypoglycemia and diabetic ketoacidosis. Cost‐effectiveness was assessed over a lifetime using the IQVIA CORE Diabetes Model and calculated from a societal perspective (assessing both direct and indirect costs). Cost data were collected on Swedish references prices and literature, and a willingness‐to‐pay threshold of 300,000 Swedish krona (approx. €29,100) was used to determine cost‐effectiveness.

Results

Use of HCL was associated with quality‐adjusted life expectancy improvement of 1.90 QALYs, with higher associated costs, leading to an incremental cost‐effectiveness ratio (ICER) of 164,236 Swedish krona (approx. €15,930) per QALY gained. The QALY improvement was related to reduced incidence and delayed onset of diabetes complications and reduction in fear of hypoglycemia. In individuals with poorly controlled T1D [HbA1c ≥7.5%, (58 mmol/mol)], the HCL was associated with a 2.25 QALY increase over insulin pump, yielding an ICER of 15,830 Swedish krona (approx. €1,535).

Conclusion

HCL is likely to be a cost‐effective treatment compared with insulin pump alone in individuals with T1D, particularly in individuals who have suboptimal glycemic control on current therapy.

Comment.

The first two articles for this theme highlight the current realities faced by clinicians and patients wishing to prescribe and use diabetes technologies and demonstrate that as technologies mature, obstacles to their uptake persist but change in nature. For example, early trials of CGM demonstrated clinical benefit with consistent use, but technical barriers such as short wear time, inaccuracy, and need for frequent calibration limited their successful use (5,6).

A decade later, with improved accuracy and functionality, the T1D Exchange Registry reports that the rate CGM use is nearly 40% and continuing to climb (2). The current challenge is how to ensure that patients have access to these potentially life‐changing devices. No matter how effective the technologies become, they will not help people unless patients can get their hands on them: limitations today are primarily based on cost. The SWEET project found that for children with T1D, access to most current diabetes technologies is improving, particularly with respect to more mature technologies such as insulin pumps (7). Coverage for sensors also seems to be improving, although the SWEET consortium observed that improvements in CGM and FGM coverage were accompanied by a reciprocal decrease in coverage for blood glucose test strips. Could this be the harbinger of the decline and eventual extinction of capillary glucose monitoring?

Less optimistically, access to diabetes technologies for older American continues to lag behind that for children, even for established devices. As people with T1D live longer, healthier lives and rightfully expect to continue benefitting from technological advances, limitations to access can have serious consequences. In the survey conducted by Argento and colleagues, most Medicare patients utilizing insulin pump therapy experienced such disruptions to access to their supplies as to result in potentially unsafe self‐management practices, such as using infusion sites longer than recommended, cutting back on basal rates to “save” on insulin costs, and temporarily discontinuing insulin pumps altogether (8). This sobering study proves that even experienced pump users may suffer unnecessary adverse health outcomes due to lack of access.

The second two articles explore the cost‐effectiveness of SAP and HCL, both CGM sensor‐dependent technologies (9,10). Cost‐effectiveness is by nature a relative measure, both influenced by the comparator category and by how investigators predict future morbidity and mortality. Although different countries and healthcare systems define different cost‐effective thresholds, diabetes related interventions are generally considered to be very cost effective at ≤$25,000 (€ ∼22,000) per QALY, cost effective at $25,000–$50,000 (€ ∼22,000–45,000), marginally cost‐effective at $50,000–100,000 (€ ∼45,000–90,000) (11). Both of these studies indicate that adding a CGM to insulin pump therapy is highly cost effective, especially for individuals with suboptimal glycemic control.

These cost‐effectiveness studies are excellent companions for each other because they used the same simulation model, the IQVIA CORE Diabetes Model, the same societal perspective that includes both direct and indirect costs of diabetes, and the same comparator group, insulin pump users. They cannot be used to directly compare SAP with HCL, however, because they used different baseline populations and different country specific costs, expenditures, and mortality and morbidity trajectories. What can be gleaned from the pair of articles, however, is that there are highly cost‐effective benefits to adding a sensor to insulin pump therapy, regardless of whether the sensor is used for HCL or simply low glucose suspend. The sensor may be considered the true hero of the diabetes technology story, with algorithms only providing incremental benefit beyond.

Technology Education

Efficacy of an education program for people with diabetes and insulin pump treatment (INPUT): results from a randomized controlled trial

Ehrmann D^1,2, Kulzer B^1,2,3, Schipfer M¹, Lippmann‐Grob B³, Haak T³, Hermanns N^1,2,3

¹Research Institute Diabetes Academy Mergentheim, Bad Mergentheim, Germany; ²Department of Clinical Psychology and Psychotherapy, University of Bamberg, Bamberg, Germany; ³Diabetes Clinic Mergentheim, Bad Mergentheim, Germany

Diabetes Care 2018;41: 2453–2462

This manuscript is also discussed in the article on Insulin Pumps, page S‐17.

Background

While structured education programs and continuous subcutaneous insulin infusion (CSII) are both mainstays of intensive diabetes self‐management, there is no information about employing structured education specifically targeted to the use of CSII, or insulin pump. This study examines whether a targeted education program for CSII users impacts glycemic control, behavior, or psychosocial status.

Methods

Individuals currently using CSII were enrolled in a 6‐month multicenter, randomized trial of a biweekly structured education program (INPUT) compared with usual care. INPUT consisted of 12 sessions (90 minutes, groups of 3–8 participants) based on self‐management/empowerment and provided information on utilizing advanced pump features (temporary basal rates, basal patterns, etc.), recognizing glucose patterns, and setting goals. Additionally, participants' emotions, attitudes, and motivations related to CSII were explored. The primary outcome was change in HbA1c over 6 months.

Results

In total, 268 participants with mean CSII use of 9.5 years were randomized to INPUT (n=135) or usual care (n=133). At 6 months, INPUT group HbA1c decreased significantly from baseline (8.3±0.8 to 8.04±0.9; P<0.001) while the usual care group remained the same (8.33±1.0 vs 8.27±1.0; P=0.11), and the between group difference favored INPUT (Δ=−0.22; P>0.001). Incident rate ratio for severe hypoglycemia was 3.55 times higher in the control group compared with INPUT. There were significant improvements in diabetes distress, depression, treatment satisfaction, hypoglycemia awareness, and key attitudes about CSII after 6 months in the INPUT group compared with control group. The INPUT group also reported more use of temporary basal rates.

Conclusion

A structured education program targeting CSII use improved both glycemic outcomes and psychosocial outcomes compared with usual care for existing CSII users.

The impact of a structured education and treatment programme (FLASH) for people with diabetes using a flash sensor‐based glucose monitoring system: results of a randomized controlled trial

Hermanns N^1,2,3, Ehrmann D^1,2, Schipfer M⁵, Kröger J⁴, Haak T³, Kulzer B^1,2,3

¹Research Institute Diabetes Academy Mergentheim, Bad Mergentheim, Germany; ²Department of Clinical Psychology and Psychotherapy, University of Bamberg, Bamberg, Germany; ³Diabetes Clinic Mergentheim, Bad Mergentheim, Germany; ⁴Centre of Diabetology Hamburg Bergedorf, Hamburg, Germany; ⁵Profusa, Inc., South San Francisco, CA

Diabetes Res Clin Pract 2019;150: 111–121

Background

FGM produces real‐time glucose values up to every 15 minutes by scanning an interstitial glucose sensor with a reader. Similar to CGM, FGM systems provide substantial retrospective data for review and pattern management. There is little known about how to best assist users to contend with the abundance of glucose data available to them compared with self‐monitoring of blood glucose.

Methods

This is a multicenter, 1:1 randomized trial of a structured education program (FLASH) over a 6‐month period. It was conducted at 26 outpatient centers across Germany with at least one endocrinologist and one diabetes educator. Participants were individuals on multiple daily injections or CSII, ages 16–75 years, who were already using or intending to use a FGM. The FLASH intervention consisted of four 90‐minute group education sessions over 6 weeks that focused on recognizing glycemic patterns in retrospective FGM reports, contending with emotional and motivational obstacles toward FGM, and individual goal setting. The primary outcome was change in HbA1c from baseline to 6 months, and secondary outcomes included FGM glucose metrics and changes in behavioral/psychosocial measures.

Results

In total, 216 individuals were randomized to either FLASH education (n=108, mean age 44.8 years, 50% female, mean HbA1c 8.4±0.9%) or usual care (n=108, mean age 47.0 years, 53.7% female, mean HbA1c 8.4±0.9%), with all participants on multiple daily injections. At 6 months, the FLASH education group demonstrated a larger decrease in HbA1c (−0.28% [95% CI −0.16% to −0.40%]) compared with the usual care group (−0.11% [CI 0.00% to −0.22%]), with the overall between‐group change favoring the FLASH education group (Δ −0.17% [CI −0.01 to −0.33], P>0.001). The FLASH group had a 3.8% increase in time in range compared with usual care ([CI −7.0 to −0.5], P=0.027), and no significant reduction in hypoglycemia. FLASH group reported lower diabetes distress and higher satisfaction at 6 months compared with usual care group, and no differences were observed in depression, treatment satisfaction, or fear of hypoglycemia.

Conclusion

This is the first randomized trial to report on the effect of structured education on FGM use for individuals with diabetes. Structured education may contribute to a small improvement in glycemic control and levels of diabetes distress and satisfaction.

Pediatric endocrinology trainees' education and knowledge about insulin pumps and continuous glucose monitors

Marks BE, Wolfsdorf JI, Waldman G, Stafford DE, Garvey KC

Division of Endocrinology, Boston Children's Hospital, Boston, MA

Diabetes Technol Ther 2019;21: 105–109

Background

Although there has been a rapid increase in the use of CSII and CGM in diabetes care, glycemic control has not improved to the levels expected, especially in the pediatric population. Provider understanding and competency with these devices may be a contributing factor to this phenomenon. Little is known about training and experience with CSII and CGM among pediatric endocrinology trainees.

Methods

Pediatric endocrinology fellows and attendings supervising fellows were included in this study, all members of the international Pediatric Endocrinology Society in the United States and Canada. Participants were surveyed via Likert scale (1 to 5) on perceptions of knowledge of CSII and CGM, clinical management using these devices, and fellows education on devices. There were additional knowledge questions on CSII (e.g., insulin dose setting configurations, temporary basal rates, extended boluses, features of different systems) and CGM (e.g., interpretation of arrows, calibration, troubleshooting, interpretation of reports). Descriptive statistics were calculated as well as qualitative analysis of free text answers.

Results

In total, 42 pediatric endocrinology fellows (50% first year, 23.6% second year, 26.5% third year) and 44 supervising attending physicians (87.2% with >5 years of experience) reported low incidence of formal training on CSII and CGM at their institutions (14.7% and 25.6% respectively). Overall combined (n=86) perception of knowledge of devices were moderate at 3.6±1.0 for CSII and 3.6±0.9 for CGM. Two concepts (knowledge of unique pump features and ability to adjust insulin doses from CGM reports) differed according to year of training. Fellows who reported receiving formal education on devices reported higher perceived knowledge on CSII (P<0.04) and CGM (P<0.05). Fellows' perceived knowledge was highest for CSII and CGM benefits and limitations, assessing basal rates/carbohydrate ratios/correction factors, and was lowest for extended boluses, infusion sets, and counseling patients on practical use and unique features of devices. Only 51.2% of respondents agreed that their education was adequate in CSII and CGM, with 89.7% reporting that formal curriculum related to device use would enhance their fellowship experience and improve patient care.

Conclusion

Pediatric endocrinology fellows report moderate familiarity with CSII and CGM and could benefit from increased formal education on diabetes devices in their fellowship training.

Comment.

While structured education is not as flashy as automated dosing algorithms, it is arguably as effective for improving glycemic control. The two randomized clinical trials from Germany show the efficacy of education in a compelling way. The INPUT study by Ehrmann et al. reports a clinically and statistically significant decrease in HbA1c of 0.26% from baseline to 6 months, which is nearly the same glycemic improvement seen with initiation of CSII therapy as reported in meta‐analyses (mean reduction 0.3%) (12,13). This means that the addition of structured education could nearly double the benefits of diabetes technology initiation. A subsequent report comparing the INPUT intervention effect in the randomized control trial compared with an implementation trial (routine care), and reported even stronger results in the implementation trial after adjusting for baseline HbA1c (14), indicating great promise for real‐world implementation of programs like INPUT or FLASH.

Education interventions rely on competent and skilled educators and are also costly due to the expert hours involved. The study by Marks et al. makes a compelling case that endocrinologists are not necessarily trained to be the diabetes technology experts in the field (15). This report is brief and the sample is limited to endocrinologists in the United States and Canada, but it nonetheless highlights the fact that endocrinologists are woefully underprepared to contend with specialty devices for diabetes self‐management, especially in the role of expert. This trend will likely continue as the technological field continues to expand and economic pressures continue to diminish the time available for physicians to spend with their patients.

Understanding the effectiveness of structured education and the technological undertraining of endocrinologists should yield a systematic conclusion: diabetes educators should be trained as the technology experts in the room. Adding additional training on diabetes technology and technology‐focused structured education could greatly improve a clinic's ability to provide competent care for technology users. Educators are a more cost‐effective resource when compared with endocrinologists and should be better leveraged to implement wide‐scale device‐related structured education.

In order to do this, payer systems must shift to reimburse educator time in meaningful and consistent ways. Diabetes educators should not be relegated to time‐fillers in a busy endocrinology practice or a bonus service available to larger medical practices; rather, they should be an indispensable expert resource proactively leveraged to improve outcomes with diabetes self‐management. This era of diabetes technology—which is insufficient to move outcomes alone—may be the right moment to validate the critical role of diabetes educators. Are we ready for a diabetes technology educator certification? Patients are ready—can payers and practices get on board?

Automated Decision Support

Adjusting insulin doses in patients with type 1 diabetes who use insulin pump and continuous glucose monitoring: variations among countries and physicians

Nimri R¹, Dassau E², Segall T³, Muller I³, Bratina N⁴, Kordonouri O⁵, Bello R¹, Biester T⁵, Dovc K⁴, Tenenbaum A^1,6, Brener A¹, Šimunović M⁷, Sakka SD⁸, Nevo Shenker M¹, Passone CG⁹, Rutigliano I¹⁰, Tinti D¹¹, Bonura C¹², Caiulo S¹², Ruszala A¹³, Piccini B¹⁴, Giri D¹⁵, Stein R¹⁶, Rabbone I¹¹, Bruzzi P¹⁷, Omladič JŠ⁴, Steele C¹⁸, Beccuti G¹⁹, Yackobovitch‐Gavan M^1,6, Battelino T^4,20, Danne T⁵, Atlas E³, Phillip M^1,6

¹Jesse Z and Sara Lea Shafer Institute for Endocrinology and Diabetes, National Center for Childhood Diabetes, Schneider Children's Medical Center of Israel, Petah Tikva, Israel; ²Harvard John A. Paulson School of Engineering and Applied Sciences, Harvard University, Cambridge, MA; ³DreaMed Diabetes Ltd, Petah Tikva, Israel; ⁴Department of Pediatric Endocrinology, Diabetes and Metabolism, University Medical Centre‐University Children's Hospital, Ljubljana, Slovenia; ⁵Diabetes Centre for Children and Adolescents, Auf der Bult, Kinder‐ und Jugendkrankenhaus, Hannover, Germany; ⁶Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel; ⁷Department of Pediatrics, University Hospital Centre Split, Split, Croatia; ⁸Department of Endocrinology and Diabetes, Evelina London Children's Hospital, London, UK; ⁹Instituto da Criança – HCFMUSP, University of Sao Paulo, Sao Paulo, Brazil; ¹⁰Pediatrics IRCCS Casa Sollievo della Sofferenza, San Giovanni Rotondo, Italy; ¹¹Centre of Pediatric Diabetes, Department of Pediatrics, University of Turin, Turin, Italy; ¹²San Raffaele Hospital, Vita‐Salute San Raffaele University, Milan, Italy; ¹³Department of Pediatric and Adolescent Endocrinology, Institute of Pediatrics, Jagiellonian University Medical College, Krakow, Poland; ¹⁴Diabetology Unit, Meyer Children's Hospital, Florence, Italy; ¹⁵Department of Paediatric Endocrinology and Diabetes, Bristol Royal Hospital for Children, Bristol, UK; ¹⁶Paediatric Endocrinology and Diabetes Unit, Dana‐Dwek Children's Hospital, Sourasky Medical Centre, Tel Aviv, Israel; ¹⁷Departments of Medical and Surgical Sciences of Mothers, Children and Adults, Azienda Ospedaliero—Univeristaria of Modena Policlinico, Paediatric Unit, Modena, Italy; ¹⁸Paediatric Endocrinology and Diabetes, Leeds Children's Hospital, Leeds, UK; ¹⁹Division of Endocrinology, Diabetes and Metabolism, Department of Medical Sciences, University of Turin, Turin, Italy; ²⁰Faculty of Medicine, University of Ljubljana, Ljubljana, Slovenia

Diabetes Obes Metab 2018;20: 2458–2466

Background

As CSII and CGM use continuing to increase, analyzing and translating the large amounts of data from patient uploads into useful recommendations to adjust diabetes therapy can be overwhelming and time‐consuming. Furthermore, there are limited data and no uniform guidelines to help providers and patients adjust insulin pump parameters to optimize diabetes control. Automated decision support systems (ADS) may help simplify and optimize diabetes management. This study evaluated the consistency in insulin dosage recommendations among a range of physicians treating patients with diabetes and compared physician recommendations with those of the Advisor Pro ADS.

Methods

Carelink Pro uploads of 15 patients with T1D (10 children and 5 young adults) were provided to 26 physicians (13 faculty, 13 fellows). Each physician was asked to recommend insulin dose and/or behavioral changes (if any) based on 3 weeks of pump, sensor, and meter data. Primary endpoints of the study were the percentage of comparison points in which there was full or partial agreement or disagreement on the trend of insulin adjustment among physicians and between physicians and the Advisor Pro. Additional endpoints explored the effect of experience level of physician, physicians at same or different centers, and centers with high versus low pump utilization on recommendation consistency.

Results

There was only fair consistency (41%–46% full agreement) among physicians regarding the three domains of insulin dose recommendations (carb ratio, correction factor, and basal rate). Agreement among physicians was similar to agreement between physicians and Advisor ADS. Level of experience of the physician, extent of center experience with insulin pumps, and whether physicians were from the same or different center did not affect the level of agreement. Magnitude of dose changes were similar between physicians and the Advisor Pro, but the Advisor tended to recommend significantly more basal rate, carb ratio, and correction factor time periods.

Conclusions

Insulin dosing recommendations to optimize diabetes management in patients using pumps and CGM greatly among physicians, regardless of physician level of training, experience with pumps, or working with other physicians. Recommendations provided by Advisor Pro did not differ significantly in directionality or magnitude from advice given by physicians. Larger, controlled studies comparing the efficacy and time saving of Advisor Pro and other ADS are needed.

Automated insulin dosing guidance to optimise insulin management in patients with type 2 diabetes: a multicentre, randomised controlled trial

Bergenstal RM¹, Johnson M¹, Passi R¹, Bhargava A², Young N², Kruger DF³, Bashan E⁴, Bisgaier SG⁴, Isaman DJM⁵, Hodish I^4,6

¹International Diabetes Center, Minneapolis, MN; ²Iowa Diabetes and Endocrinology Research Center, Des Moines, IA; ³Henry Ford Medical Center Endocrinology, Detroit, MI; ⁴Hygieia Inc, Livonia, MI; ⁵Department of Biostatistics, School of Public Health, University of Michigan, Ann Arbor, MI; ⁶Department of Internal Medicine, Division of Metabolism, Endocrinology and Diabetes, University of Michigan Medical Center, Ann Arbor, MI

Lancet 2019;393: 1138–1148

Background

Successful management of insulin‐requiring type 2 diabetes (T2D) requires frequent and effective manual titrations of insulin, but restrictions in time and expertise among providers results in most patients being under‐dosed with insulin and only sporadically dose titrated. The aim of this study was to determine whether use of the d‐Nav system, a handheld device that contains a blood glucose meter and data management system that identifies glucose patterns and recommends insulin doses, was superior to provider support alone in improving glycemic control in insulin‐requiring T2D.

Methods

This was a 6‐month multicenter parallel‐design RCT in adults (age 21–70 years) with insulin‐requiring T2D, baseline A1c 7.5%–11%, comprising four treatment regimens: once‐daily basal insulin + daily blood glucose (BG) monitoring; twice‐daily biphasic/premixed insulin + twice daily glucose monitoring; basal‐bolus therapy with fixed meal and correction doses plus four times daily BG monitoring; or basal‐bolus therapy with flexible insulin–carbohydrate and correction factor ratios plus four times daily BG monitoring. Primary outcome was change in A1c over six months; additional outcomes included percentage of subjects reaching predefined A1c thresholds and hypoglycemia. Subjects had three in‐person visits and four telephone check‐ins over 6 months.

Results

Among 181 enrolled subjects (93 in the d‐Nav group, baseline A1c 8.7%, and 88 in the control group, A1c 8.5%), a significantly greater drop in A1c was noted in the d‐Nav group compared with the control group (1.0 vs 0.3%; P<0.0001). Significantly more subjects in the d‐Nav group achieved final A1c level of <7% or <8% compared with control group (22% vs 5% and 62% vs 33%, respectively). The d‐Nav group had significantly fewer BG readings in the <50 range. Increases in total daily insulin dose and weight were seen only in d‐Nav group. 70% of subjects in the d‐Nav group were comfortable or very comfortable with having their doses managed by the system.

Conclusions

Use of the d‐Nav decision support system plus provider support was more effective than provider support alone in reducing A1c without increasing the risk of hypoglycemia in a diverse group of insulin‐requiring subjects with T2D. It is particularly notable that this is a scalable intervention that does not require additional provider time or effort to maintain once subjects are trained.

Comment.

While intensive insulin therapy is already the cornerstone of management of T1D, there is increasing recognition that more intensive regimens may be required for optimal care of insulin‐requiring patients with T2D as well. However, the vast numbers of patients with T2D requiring intensive management presents a daunting challenge for a healthcare system already straining to meet increasing demand while at the same time struggling with shrinking visit times and reimbursements. Automated decision support (ADS) systems have the potential to assist clinicians and patients in optimizing insulin therapy; the selected articles highlight the progress and challenges in automating decision support in real world settings (16,17).

For T2D management, which involves fewer variables than T1D management, simple algorithms and devices may be effective solutions for automated dose titration based on blood glucose data. The Bergenstal study is particularly noteworthy for including subjects with a wide range of management schemes, from once‐daily basal insulin dosing only to more sophisticated multiple‐dose regimens, including both premixed and basal‐bolus options (17). Larger‐scale implementation of ADS such as the one evaluated in this study appears to be a viable option for T2D and has potential to significantly improve outcomes for large numbers of people without regular access to sophisticated endocrinological care.

In contrast, the sheer volume of data and multitude of choices available to clinicians attempting to adjust therapy in patients with T1D using pumps and sensors makes development of ADS systems for this population a much more complex and challenging project. This interesting study by Nimri highlights the fact that there is little agreement or consistency among physicians regarding recommendations for insulin dosing changes for patients using pumps and sensors, lending credence to the old adage of medicine being as much an art as a science (16). The good news is that this at a very broad level, the particular ADS system under study, the Advisor Pro, seemed to agree with doctors as much as doctors agree with doctors. Feasibility studies of ADS in T1D have been promising (18,19); if results from larger‐scale, longer‐duration studies demonstrate efficacy, the incorporation of these systems into diabetes practices promises to improve efficiency, care delivery, and outcomes.

Summary

We are currently experiencing a “golden age” of diabetes technology: small, accurate, easy‐to‐use CGMs may soon render blood glucose testing obsolete, while pumps are increasingly integrated with sensors to automate insulin delivery for better glycemic results. Nevertheless, the data clearly demonstrate that devices alone will not improve diabetes outcomes: for these technologies to truly, positively impact outcomes, people with diabetes must have access to the devices and supplies without undue financial burden, as well as the educational support necessary to effectively integrate them into their own self‐management practices. As these technologies are increasingly adopted into routine care for more people with T1D and T2D, the widespread use of automated decision support can potentially offset and reduce overall healthcare costs by improving outcomes with more efficient investment of clinical resources.

Author Disclosure Statement

Stuart Weinzimer has served as a speaker for Tandem Diabetes and Insulet Corporation and serves as a consultant for Eli Lilly and Company, Sanofi, and Zealand Pharma. He has received grants to his institution from Medtronic. Laurel H. Messer is a contracted product trainer for Medtronic Minimed, has received consultant/speaking fees from Tandem Diabetes and DexCom Inc., and has consulted for Clinical Sensors and Capillary Biomedical.