Event and Cost Offsets of Switching 20% of the Type 1 Diabetes Population in Germany From Multiple Daily Injections to Continuous Subcutaneous Insulin Infusion

Abstract

Background:

Most patients with type 1 diabetes (T1D) administer insulin by multiple daily injections (MDI). However, continuous subcutaneous insulin infusion (CSII) therapy has been shown to improve glycemic control compared with MDI.

Objective:

The objective was to determine the key medical event and cost offsets generated over a 4-year period by introducing CSII to T1D patients who have inadequately controlled glucose metabolism on MDI in Germany.

Methods:

A decision-analytic budget impact model, simulating a treatment switch scenario, was developed. In the base case, all T1D patients received MDI, while in the switch scenario, 20% of the eligible T1D population, randomly selected, moved to CSII. The model focused on 2 medical endpoints and their corresponding cost offsets: severe hypoglycemic events requiring hospitalization (SHEH) and complication-borne diabetic events (CDEs) avoided. Event rates and costs were taken from the literature and official sources, adopting a health insurance perspective.

Results:

Compared with the base case, treating 20% of patients with CSII in the switch scenario resulted in 47 864 fewer SHEH and 5543 fewer CDEs. This led to total cost offsets of €183 085 281 within the 4-year time horizon. Of these, 92% were driven by avoided SHEH. Compared to an expected budget impact (cost increase) of 83%, only treatment costs considered, the total impact of the switch scenario amounted merely to a 24.5% increase in costs (reduction by 58.5% points; a factor of 3.4).

Conclusion:

The use of CSII resulted in fewer SHEH and CDEs compared to MDI. The incurred CSII implementation costs are hence offset to a substantial degree by cost savings in complication treatment.

Keywords

budget impact type 1 diabetes continuous subcutaneous insulin infusion complication-borne diabetic events severe hypoglycemic events German costs data glucose monitoring

It is estimated that there are 387 million patients with diabetes worldwide, of whom 5-10% in developed countries have type 1 diabetes mellitus (T1D).¹ There are estimated to be 350 000 patients with T1D in Germany.² Patients with T1D require insulin to maintain their glucose levels within a near normal range, usually measured by self-monitoring of blood glucose (SMBG).³ This is often done by multiple daily self-injections (MDI) of insulin.³ Maintaining a near normal blood glucose range is essential to the health and daily well-being of the patient, whereas an inadequately controlled glucose level can result in acute complications such as hypoglycemia or diabetic ketoacidosis, and to chronic complications over time (microvascular and macrovascular complications).^4-6 A near normal blood glucose range, hence, being adequately controlled, implies to maintain a defined A1C level of <7.5% without the emergence of SHEs, referred to German guidelines;⁷ patients who do not meet this definition are considered to be inadequately controlled. This, always under consideration of the individual’s needs, resources and preferences in terms of the course of their diabetes. Improved control of glucose levels can therefore help patients avoid complications. It has been shown that a 1-percentage-point improvement in A1C over time can lead to significant reductions in micro- and macro-vascular complications.^8,9 As 53% of the medical management costs associated with diabetes in Germany are a result of complications,¹⁰ there are consequently potential cost savings to be made, primarily as a result of improved management of glycemic control.

Continuous subcutaneous insulin infusion (CSII) has been shown to provide significant reductions in A1C, compared with MDI.^11-13 However, the costs of CSII are higher than the costs of MDI, given the higher level of technology and the consumables required. The overall budget impact of using CSII rather than MDI is therefore of interest when considering how to treat patients with T1D.

Methods

A decision-analytic budget impact model was developed in Microsoft® Excel® 2010 to consider the impact of switching patients with T1D from MDI to CSII. The considered CSII technology does not incorporate continuous glucose measurement. The model considers the key costs of treatment with each form of therapy, and examines 2 medical endpoints and their corresponding cost offsets: severe hypoglycemic events requiring hospitalization (SHEH) and complication-borne diabetic events (CDEs) (micro- and macrovascular) avoided. The model considers the German T1D population (350 000 patients),² and assumes a proportion of MDI patients of 85%, of which 88.2% have currently inadequate glucose control according to UK population data¹⁴ (no data for Germany available); and hence determine the eligible MDI-population for switch to CSII of 262 395. Two scenarios were considered: the base case, where all patients were treated with MDI; and a switch scenario, where—based on a practicable assumption—20% of the eligible patients were switched to CSII and 80% of the eligible patients remained on MDI. The model inputs were chosen to represent a patient population with T1D, and eligible patients were randomly selected for switching. However, in clinical practice, CSII has been used in patients with difficulties in managing T1D whose risk of hypoglycemia is higher than other patients. Therefore, a scenario with twice the risk of SHEH was modeled within the switch scenario, in addition.

Table 1 shows the frequency of severe hypoglycemic events associated with MDI and CSII (1.38 vs 0.31 patient/year), sourced from a long-term US clinical study in which patients with ≥12 months of treatment with MDI were switched to CSII.¹⁵ The annual frequency of SHEH with CSII was assumed to be the average of the rate over each of the 4 years considered in the study.¹⁶ In the absence of German data, the proportion of the events requiring hospitalization was obtained from a Scottish study on the frequency of severe hypoglycemia requiring hospital treatment (21.31%).¹⁷

Table 1.

Input Variables: Estimated Rates of SHEH and CDEs With MDI and CSII Over 4 Years, and Costs per Event.

Outcome	MDI	CSII	Costs per event^a
Severe hypoglycemic events per patient-year	1.38¹⁵	0.31¹⁵	—
Proportion of above requiring hospitalization (%SHEH)	21.31%¹⁷		€3533⁴¹
CDE incidence (cumulative over 4 years) of:
Microvascular events^b
Nephropathy^b	7.12%^8,36	6.34%	€5361^10,25
Neuropathy^b	7.92%⁹	6.17%	€4404^25,41
Retinopathy^b	17.13%⁹	12.70%	€902^10,25
Diabetic foot ulcer^b	8.51%⁹	6.62%	€2654^25,41
Macrovascular events^b
Angina^b	3.46%^25,38	3.01%	€2417^10,25
Myocardial infarction^b	3.92%^25,39	3.42%	€3893^10,25
Stroke^b	1.61%^25,37	1.41%	€5460^10,25
All-cause mortality^b	5.44%^19,41	4.89%	—

Cost figures inflated from base years 2005 and 2007 to 2014 values using the Consumer Price Index.

All values for MDI calculated from source values by standardizing to a 10-year time horizon, then reestimating over a 4-year time horizon using equations presented in Fleurence and Hollenbeak.¹⁸ Values for CSII estimated using the impact of a 1% reduction in A1C on these outcomes and the reported reduction in A1C with CSII vs MDI (−0.55).^8,19,20,42

The risk of each of the CDEs associated with MDI was estimated based on reported risks for patients with T1D, or, where data for T1D-only were not available, patients with either type 1 or type 2 diabetes. The risk of CDEs with CSII was estimated based on 3 factors: the risk of CDEs associated with MDI, the effect of reductions in A1C on the risk of each event, and the relative reduction in A1C provided with CSII, compared with MDI.

The risk for CDEs with MDI were standardized to a 10-year time horizon to allow comparability and a degree of consistency. The 10-year risks were then adjusted to provide the risk over the 4-year time horizon used in the model (Table 1). The risks have been adjusted using equations presented in Fleurence and Hollenbeak.¹⁸

The impact of A1C on CDEs was drawn from the Diabetes Control and Complications Trial⁹ and from a long-term study on the impact of intensive treatment on mortality.¹⁹ Based on these, a 1-percentage-point reduction in A1C over the 4-year time horizon was estimated to result in a 22.78% reduction in angina pectoris,⁹ an 8.89% reduction in nephropathy,⁹ a 38.33% reduction in neuropathy⁹ (assumed to apply to foot ulcers), a 42.22% reduction in retinopathy,⁹ a 22.78% reduction in cardiovascular disease⁹ (assumed to represent myocardial infarction and stroke), and a 17.66% reduction in all-cause mortality.¹⁹ A summary of the clinical inputs used in the model and its associated costs is provided in Table 1.

The reduction in A1C with CSII compared with MDI was taken from Raccah et al, who report a −0.55-point change (improvement) in A1C from baseline MDI treatment after 6 months of treatment with CSII.²⁰ This reduction was then applied to the rate reduction for each CDE reported in the literature with a 1% change in A1C (ie, 55% of this rate reduction). The estimated rate reduction calculated for each CDE was then applied to the rate of each CDE reported in the literature for MDI to provide an estimate of the rate of each CDE with CSII (Table 1).

The costs data and sources used in the model for treatment are summarized in Table 2. For MDI therapy only insulin costs have been considered.²¹ CSII costs were based on the Animas® Vibe® insulin pump (Johnson & Johnson Medical GmbH, Norderstedt, Germany) specifications, and the corresponding official costs of each component (pump, pump utilities, insulin) sourced from the German reimbursement schedule 2014.^21-24 The estimated costs do not claim to be exhaustive regarding the respective therapies. In fact, for example, the costs for needles and pens for MDI as well as costs for pump training, additional visits and blood ketone measurements for CSII are not reflected. This prioritization was consistently pursued throughout the model, both for costs as well as for outcomes (SHEH and CDE).

Table 2.

Treatment Costs Associated With MDI and CSII Included in the Analysis.

Item	Units	Costs per unit	Annual costs
MDI
Basal insulin^a,b,e	25 per day^22,40	€0.06²¹	€527.15
Bolus insulin^a,b,e	25 per day^22,40	€0.04²³	€363.76
Total			€891
CSII
Insulin pump costs			920^f
Insulin cartridges (200 units each)^c,d	92 per year^e	€12.40²⁴	€1140.80
Vial (1000 units)^c,d	19 per year^e	€39.86²⁴	€757.42
Reservoir cap	2 per year^f	€30.70^f	€61.40
Battery	16 per year^f	€6.30^f	€100.80
Battery cap	2 per year^f	€36.79^f	€73.58
Subcutaneous infusion sets (10 units each set)	12 per year^f	€127.18^f	€1526.20
Total			€4580

Based on Walsh et al,²² a mean use of 49.4 units of insulin per day are assumed; rounded up and divided by 2 due to b.

It is assumed that basal and bolus account for 50% of the daily uptake rate of insulin, respectively.⁴⁰

Average costs per unit takes the average value of the insulins recommended for Animas Vibe insulin pump.

Insulin cartridges for the Animas Vibe insulin pump are sold as empty single-use consumables that are filled from insulin vials.

It is assumed that a MDI user will need the same amount of daily insulin units as an CSII user.

Johnson & Johnson data on file for insulin pump costs per year.

The costs of SHEH and CDE were obtained from published sources and inflated from the reported costs year to 2014 costs using the consumer price index (Table 2).²⁵

Results

Compared with remaining on MDI (base case), switching 20% of eligible T1D patients from MDI to CSII (switch scenario) was found to result in 224 610 fewer severe hypoglycemic events. The latter, including 47 864 fewer hospitalizations (Figure 1), at cost savings of €169 104 984 over a 4-year period. Furthermore, there were 5543 fewer CDEs over 4 years (Figure 2), including 288 avoided deaths, with additional cost savings of €13 980 297. Total cost savings amounted to €183 085 281, of which 92% are realizable in the short term as consequence of avoided SHEH.

Figure 1.

Severe hypoglycemic events (SHEs) and severe hypoglycemic events requiring hospitalization (SHEH) cumulatively avoided over 4 years when 20% of German T1D patients were switched from MDI to CSII.

Figure 2.

Number of CDEs avoided over 4 years when 20% of German T1D patients were switched from MDI to CSII.

Total treatment costs over 4 years were €935 081 580 for the base case and €1 709 521 688 for the switch scenario, with a difference of €774 440 108 (Table 3), representing an increase of 83% in treatment costs. However, once the cost reductions associated with avoiding SHEH and CDEs are considered, the total budget impact of CSII is €591 354 827, representing a 24.5% increase in costs compared with the base case. A significant proportion of the additional costs, as expected from a narrower procurement-only viewpoint, are therefore offset in the holistic perspective through the avoidance of complications.

Table 3.

Total Population Costs per Scenario Estimated by the Decision-Analytic Model.

Costs outcomes in population (annual)	Base case^a	Switch scenario^a	Difference	Difference over 4 years
Treatment costs	€233 770 395	€427 380 422	€193 610 027	€774 440 108
SHEH costs	€272 622 521	€230 346 275	−€42 276 246	−€169 104 984
CDEs costs	€97 021 539	€93 526 465	−€3 495 074	−€13 980 297
Total costs	€603 414 455	€751 253 162	€147 838 707	€591 354 828
Relative budget impact (in any given counting period)			24.5%

The base case is where all patients are on MDI therapy; the switch scenario is where 20% of patients switch from MDI to CSII.

As CSII may be used in patients with demonstrated difficulty in managing their diabetes, who are consequently at greater risk of severe hypoglycemia, a subscenario was also considered where the rates of SHE for MDI and CSII were doubled (2.76 and 0.62, respectively; see Table 1), while the rate of CDE was unchanged. In this scenario, 95 729 SHEH were avoided, 47 864 more than with the standard SHEH rates. This resulted in increased cost savings of €338 209 967 relating to SHEH avoided, and incremental total costs of CSII over MDI of €422 249 843, a 12.1% increase over the base case.

Discussion

The current model demonstrates that while switching 20% of eligible patients from MDI to CSII is associated with increased treatment costs, the latter are substantially offset by cost savings as a result of avoided hypoglycemic and diabetic events. The switch scenario was associated with an increase in treatment costs of 83%, compared with the base case; however, once SHES and CDEs avoided were considered, the increase in budget impact was only 24.5%. The expected incremental budget impact of CSII is therefore significantly offset by savings as a result of improved glycemic control, and is approximately 30% only of what would be expected based on treatment procurement costs alone. If this model is applied to an example clinic of 500 eligible MDI patients, of whom 20% are shifted to CSII, there would be 428 fewer SHEs, 91 fewer SHEH, and 11 fewer CDEs over a 4-year period (Table 4). Almost 1 avoidable event per patient is expected.

Table 4.

Outcomes Avoided Over 4 Years in an Example Population of 500 T1D Patients.

	SHEs over 4 years	SHEH over 4 years	CDEs over 4 years
Base case	2760	588	276
Switch	2332	497	265
Difference (outcomes avoided)	428	91	11

The base case is where all patients are on MDI therapy; the switch scenario is where 20% of patients switch from MDI to CSII.

It should also be noted that patients were selected for switch completely at random in the model; in clinical practice, this is improbable to happen. It is likely that physicians would elect to switch specific patients from MDI to CSII according to medical guidelines,⁷ for example patients who are at high risk of hypoglycemic hospitalizations or complications related to a raised A1C, given that these patients could be expected to benefit more from CSII than others. An analysis was therefore performed which considered a higher (doubled) rate of SHEH for both MDI and CSII (the rate of CDEs was unchanged). This analysis found that the net increase in costs with the switch scenario in this group was down to 12.1%, compared to 24.5% from the original switch scenario versus the base case. This analysis ignores any increase in risk of CDEs in such high-risk patients in the base cases. It is consequently conceivable that appropriate selection of patients for CSII may allow the increased costs of treatment with CSII, compared to MDI, to be largely offset by cost savings as a result of avoided SHEs and CDEs.

As with any model, there are limitations in the current analysis. Diabetes modeling is generally complex, and cost-effectiveness models typically examine long-term outcomes based on complex risk equations.^26,27 However, the current model uses a simplified approach examining SHEH and CDEs over a 4-year horizon.

Due to the absence of data from Germany on the prevalence of inadequately controlled patients the model assumptions refer to recent UK data.⁸ Here, adequately controlled is defined as an A1C <6.5%, resulting in a relatively high prevalence of inadequately controlled patients (88.2%). However, only 20% of these were considered to switch from MDI to CSII, hence, it is not expected to lead to inappropriate conclusions.

Another limitation of this model relates to CDEs, which were based on the reported risks of these in the literature, the impact of reductions in A1C on these outcomes, and the reduction in A1C with CSII compared with MDI. While it was possible to source event risks for T1D patients for most outcomes, it was necessary to use data from studies in patients with both T1D and T2D for diabetic foot ulcers (Table 1, contained in neuropathy).⁹ It is therefore possible that the risk of these events is different in the T1D population; however, in the absence of specific data for the T1D population alone, the risks shown here represent a pragmatic estimate. The impact of reductions in A1C on these CDEs (with the exception of all-cause mortality) was obtained from the 1993 report on The Diabetes Control and Complications Trial.⁹ While this publication is more than 20 years old, the effect of lowering A1C on diabetic events is not expected to have changed much during this time. The impact of CSII on A1C was sourced from a single trial in 132 adults and children with inadequately controlled T1D.²⁰ While this is the population of interest, this may be viewed as a selective use of data. However, there are no large-scale trials of CSII versus MDI, and 3 meta-analyses have reported a similar impact on A1C: Pickup et al reported a −0.51% difference in favor of CSII over MDI in 2002¹² and a −0.61% difference in 2008;¹³ and Jeitler et al reported a −0.4% reduction in studies under 6 months long and a −0.8% reduction in studies over 6 months long.¹¹ The value used in the model (−0.55%) may therefore be conservative, given the 4-year time frame. In spite of these limitations, they should be viewed in the context of the overall cost offsets, as CDEs avoided represent only 8% of these cost offsets. Therefore, inclusion of these does not significantly bias the overall results, and provides decision-makers and budget holders with an estimate of the impact of these outcomes on the overall budget impact of CSII. Above all, latest published data on cardiovascular mortality within T1D in Sweden was not yet considered. Steineck et al conclude that, among people with T1D, using CSII is associated with lower cardiovascular mortality than treatment with MDI.²⁸

SHE-driven hospitalizations avoided represent the large majority of the cost offsets in the model (92%). The rate for SHEs per patient-year was obtained from a US-based study;¹⁶ this outcome has not been widely reported in the literature, but a 2008 meta-analysis by Pickup et al found that the rate of severe hypoglycemia was substantially higher in patients using MDI, compared with patients using CSII (rate ratio 4.19, 95% CI 2.86-6.13),¹³ substantiating the rates used in the current analysis. The proportion of SHEs requiring hospitalization was based on a Scottish study (21.31%).¹⁷ It is unclear if hospitalization rates in Germany would be expected to be different from those reported in Scotland, and moreover, how this would impact on the cost estimates derived from our analysis. However, Germany is known to have some of the highest levels of hospital activity across the OECD. Hence, it might be likely that the actual proportion of hospitalization due to, or following, a SHE is at least on the level of the Scottish.

While there are limitations in the current model, it should also be noted that it is conservative in many of its assumptions. There is evidence that lowering A1C results in reduced mortality, and the model demonstrates the impact of CSII on mortality. While reducing mortality should be regarded as one of the primary goals of treatment, monetary costs cannot easily be assigned to a death. The model also only considers costs to the health care system due to SHEs resulting in hospitalizations and CDEs. However, SHEs which do not result in hospitalization may still have associated costs, for example emergency admission or outpatient care (eg, office visits and GlucaGen HypoKit®) costs, or lost productivity at work or missing work. Non–health care payers, such as employers, pension funds and society as a whole may therefore benefit from the improved glycemic control offered by CSII over MDI.

Furthermore, only insulin has been considered as consumed resource for MDI therapy; however, other consumables such as syringes (or insulin pen-needles) are needed. According to relevant guidelines, a consumption of 5 syringes daily is recommended in MDI therapy,⁷ corresponding to an uptake rate of 1825 syringes per MDI patient annually. Assuming the costs of 100 syringes to be €25, MDI costs would increase from €891 to €1347 per year per patient, representing an increase of 51%. Related to the total budget impact of the model, this increase in MDI costs would lead to only 17.1% rather than 24.5%, where syringes are not considered.

In the majority of cases the switch from MDI to CSII treatment leads to a decrease in the patient’s demand on insulin due to a more continuous delivery of insulin in the course of a day;^29,30 however, the number of daily insulin units consumed was assumed to be equal for CSII and MDI in the current model.

As with any budget impact model, also the present one did not include the beneficial impact of CSII on patients’ quality of life (QoL), which may be improved as a consequence of the better health outcomes enabled with CSII, compared to MDI.³¹ In addition, patients with diabetes often have a fear of hypoglycemia,³² which can have a significant negative impact on diabetes management due to so-called “hypo-avoidance behaviour,”³³ metabolic control, and subsequent health outcomes.³⁴ Providing such patients with CSII may therefore help alleviate their fear and allow them to better manage their diabetes, leading to improved health outcomes.

Finally, the model uses only a 4-year time horizon; however, diabetes must be managed for the duration of the patient’s life. A1C is known to be linked to the rate of diabetic complications (micro- and macro vascular events),⁹ and it is possible that the true impact of improved glycemic control would become apparent over a longer time scale only. Future work should therefore include the impact of these factors on the overall budget impact of CSII. Furthermore, recently available data on costs and outcomes of sensor-augmented pump therapy (SAP) may be worth investigating, as recent data from the UK suggests SAP to be cost-effective.³⁵

The results of the current analysis use German costs and, as such, may not be applicable to other countries due to differences in health care systems and unit costs. However, the applied rates of outcome parameter differences in SHEH and CDE rates are unlikely to vary substantially between countries, and this model may therefore serve as an indication of the clinical and economic impact of CSII. As far as the authors are aware, this is the first publication on the budget impact of CSII in Germany. While cost-effectiveness analyses have been published for other countries, none have been published for Germany. The current study therefore represents a useful addition to the body of evidence that may assist decision makers and budget holders.

Conclusion

The use of CSII results in fewer SHEH, CDEs and deaths, which results in avoided costs. While the costs of CSII treatment is higher than MDI, these costs are offset to a substantial degree by cost savings as a result of avoided events, and the overall impact of switching 20% of an eligible diabetes population in Germany from MDI to CSII represents a 24.5% increase in costs (instead of 83% when complication costs are not considered). It is quite likely that the current model underestimates the potential cost savings, as the selection of switch patients is random, rather than targeted. Other benefits, such as QoL, are not considered. CSII may therefore represent wise spending of scarce health care resources.

Footnotes

Acknowledgements

The authors would like to thank Ralf Reintjes for reviewing an earlier version of this research article.

Abbreviations

A1C, glycated hemoglobin; CDE, complication-borne diabetic event; CSII, continuous subcutaneous insulin infusion; MDI, multiple daily injections; QoL, quality of life; SAP, sensor-augmented pump therapy; SHE, severe hypoglycemic event; SHEH, severe hypoglycemic event requiring hospitalization; SMBG, self-monitoring of blood glucose; T1D, type 1 diabetes mellitus.

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: Magnus Stüve and Marion Schauf are employees of Johnson & Johnson Medical GmbH.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: York Francis Zöllner, Ralph Ziegler, and Julia Krumreich have received consultancy honoraria from Johnson & Johnson Medical GmbH.

References

International Diabetes Federation. IDF Atlas Update Poster. 6th ed. Brussels, Belgium: IDF; 2014.

Danne

Deutscher Gesundheitsbericht Diabetes. 2015. Available at: http://www.diabetesde.org/fileadmin/users/Patientenseite/PDFs_und_TEXTE/Infomaterial/Gesundheitsbericht_2015.pdf. Accessed July 2015.

Deutsche Diabetes Gesellschaft. Therapie des Typ-1-Diabetes. 2011.

Cryer

PE.

The barrier of hypoglycemia in diabetes. Diabetes. 2008;57(12):3169-3176.

Fishbein

H PP

. Acute Metabolic Complications in Diabetes in America. 2nd ed. Bethesda, MD: National Institute of Diabetes, Digestive and Kidney Diseases, National Institute of Health; 1995.

Fowler

Microvascular and macrovascular complications of diabetes. Clin Diabetes. 2008;26(2):77-82.

Böhm

Dreyer

Fritsche

Füchtenbusch

Gölz

Martin

Therapie des Typ-1-Diabetes. Diabetologie und Stoffwechsel. 2014;9:S125-S135.

Stratton

Adler

Neil

. Association of glycaemia with macrovascular and microvascular complications of type 2 diabetes (UKPDS 35): prospective observational study. BMJ. 2000;321(7258):405-412.

Diabetes Control and Complications Trial Research Group. The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus. N Engl J Med. 1993;329(14):977-986.

10.

von Ferber

Koster

Hauner

. Medical costs of diabetic complications total costs and excess costs by age and type of treatment results of the German CoDiM Study. Exp Clin Endocrinol Diabetes. 2007;115(2):97-104.

11.

Jeitler

Horvath

Berghold

. Continuous subcutaneous insulin infusion versus multiple daily insulin injections in patients with diabetes mellitus: systematic review and meta-analysis. Diabetologia. 2008;51(6):941-951.

12.

Pickup

Mattock

Kerry

Glycaemic control with continuous subcutaneous insulin infusion compared with intensive insulin injections in patients with type 1 diabetes: meta-analysis of randomised controlled trials. BMJ. 2002;324(7339):705.

13.

Pickup

Sutton

AJ.

Severe hypoglycaemia and glycaemic control in type 1 diabetes: meta-analysis of multiple daily insulin injections compared with continuous subcutaneous insulin infusion. Diabet Med. 2008;25(7):765-774.

14.

Diabetes

UK.

Diabetes in the UK: key statistics on diabetes. 2014. Available at: https://www.diabetes.org.uk/Documents/About%20Us/Statistics/Diabetes-key-stats-guidelines-April2014.pdf. Accessed February 2015.

15.

Bode

Steed

Davidson

PC.

Reduction in severe hypoglycemia with long-term continuous subcutaneous insulin infusion in type I diabetes. Diabetes Care. 1996;19(4):324-327.

16.

Bode

Sabbah

Gross

Fredrickson

Davidson

PC.

Diabetes management in the new millennium using insulin pump therapy. Diabetes Metab Res Rev. 2002;18(suppl 1):S14-S20.

17.

Leese

Wang

Broomhall

. Frequency of severe hypoglycemia requiring emergency treatment in type 1 and type 2 diabetes: a population-based study of health service resource use. Diabetes Care. 2003;26(4):1176-1180.

18.

Fleurence

Hollenbeak

CS.

Rates and probabilities in economic modelling: transformation, translation and appropriate application. Pharmacoeconomics. 2007;25(1):3-6.

19.

Orchard

Nathan

Zinman

. Association between 7 years of intensive treatment of type 1 diabetes and long-term mortality. JAMA. 2015;313(1):45-53.

20.

Raccah

Sulmont

Reznik

. Incremental value of continuous glucose monitoring when starting pump therapy in patients with poorly controlled type 1 diabetes: the RealTrend study. Diabetes Care. 2009;32(12):2245-2250.

21.

Liste

Rote

. Cost of insulin glargine. 2015. Available at: http://online.rote-liste.de/. Accessed June 2015.

22.

Walsh

Roberts

Bailey

Guidelines for insulin dosing in continuous subcutaneous insulin infusion using new formulas from a retrospective study of individuals with optimal glucose levels. J Diabetes Sci Technol. 2010;4(5):1174-1181.

23.

Liste

Rote

. Value been derived from the average cost of Animas Vibe approved insulin vials (Apidra, Humalog and Novorapid) which contain 1,000 units. 2015. Available at: http://online.rote-liste.de/. Accessed June 2015.

24.

Liste

Rote

. Average cartridge price for the Animas Vibe approved insulins Apidra, Humalog and NovoRapid. 2015. Available at: http://online.rote-liste.de/. Accessed June 2015.

25.

Consumer Price Index. Historic inflation in Germany. 2015. Available at: http://www.inflation.eu/inflation-rates/germany/historic-inflation/cpi-inflation-germany.aspx. Accessed April 2015.

26.

Hayes

Leal

Gray

Holman

Clarke

PM.

UKPDS outcomes model 2: a new version of a model to simulate lifetime health outcomes of patients with type 2 diabetes mellitus using data from the 30 year United Kingdom Prospective Diabetes Study: UKPDS 82. Diabetologia. 2013;56(9):1925-1933.

27.

Palmer

Roze

Valentine

. The CORE Diabetes Model: projecting long-term clinical outcomes, costs and cost-effectiveness of interventions in diabetes mellitus (types 1 and 2) to support clinical and reimbursement decision-making. Curr Med Res Opin. 2004;20(suppl 1):S5-S26.

28.

Steineck

Cederholm

Eliasson

. Insulin pump therapy, multiple daily injections, and cardiovascular mortality in 18 168 people with type 1 diabetes: observational study. BMJ. 2015;350:h3234.

29.

David

Gill

Gunnarsson

Shafiroff

Edelman

Switching from multiple daily injections to CSII pump therapy: insulin expenditures in type 2 diabetes. Am J Managed Care. 2014;20(11):e490-e497.

30.

Lauritzen

Pramming

Deckert

Binder

Pharmacokinetics of continuous subcutaneous insulin infusion. Diabetologia. 1983;24(5):326-329.

31.

Nicolucci

Maione

Franciosi

. Quality of life and treatment satisfaction in adults with type 1 diabetes: a comparison between continuous subcutaneous insulin infusion and multiple daily injections. Diabet Med. 2008;25(2):213-220.

32.

Kamps

Roberts

Varela

RE.

Development of a new fear of hypoglycemia scale: preliminary results. J Pediatr Psychol. 2005;30(3):287-291.

33.

Barnard

Thomas

Royle

Noyes

Waugh

Fear of hypoglycaemia in parents of young children with type 1 diabetes: a systematic review. BMC Pediatr. 2010;10:50.

34.

Wild

von Maltzahn

Brohan

Christensen

Clauson

Gonder-Frederick

A critical review of the literature on fear of hypoglycemia in diabetes: Implications for diabetes management and patient education. Patient Educ Couns. 2007;68(1):10-15.

35.

Roze

Brandt

De Portu

Papo

Jendle

Health-economic analysis of real-time continuous glucose monitoring in people with type 1 diabetes. Diabet Med. 2015;32(5):618-626.

36.

Chaturvedi

Bandinelli

Mangili

Penno

Rottiers

Fuller

JH.

Microalbuminuria in type 1 diabetes: rates, risk factors and glycemic threshold. Kidney Int. 2001;60(1):219-227.

37.

Hagg

Thorn

Putaala

. Incidence of stroke according to presence of diabetic nephropathy and severe diabetic retinopathy in patients with type 1 diabetes. Diabetes Care. 2013;36(12):4140-4146.

38.

McAlpine

Morris

Emslie-Smith

James

Evans

JM.

The annual incidence of diabetic complications in a population of patients with type 1 and type 2 diabetes. Diabet Med. 2005;22(3):348-352.

39.

Mohammedi

Bellili-Munoz

Marklund

. Plasma extracellular superoxide dismutase concentration, allelic variations in the SOD3 gene and risk of myocardial infarction and all-cause mortality in people with type 1 and type 2 diabetes. Cardiovasc Diabetol. 2015;14:845.

40.

BMJ. Type 1 diabetes—treatment approach 2014. 2014. Available at: http://bestpractice.bmj.com/best-practice/monograph/25/treatment/step-by-step.html. Accessed June 2015.

41.

Hammer

Lammert

Mejias

Kern

Frier

BM.

Costs of managing severe hypoglycaemia in three European countries. J Med Econ. 2009;12(4):281-290.

42.

Bergenstal

Tamborlane

Ahmann

. Effectiveness of sensor-augmented insulin-pump therapy in type 1 diabetes. N Engl J Med. 2010;363(4):311-320.