Evaluation of the Cost-Effectiveness of Insulin Glargine Versus NPH Insulin for the Treatment of Type 2 Diabetes in the UK
By McEwan, Phil; Poole, Chris D; Tetlow, Tony; Holmes, Paul; Currie, Craig J
Key words: Cost – Economic – Effectiveness – Hypoglycaemia – Insulin glargine – NPH insulin – Type 2 diabetes
ABSTRACT
Background: People with type 2 diabetes typically require insulin treatment following failure on hypoglycaemic agents. The purpose of this study was to evaluate the relative cost-effectiveness of insulin glargine versus NPH insulin for the treatment of type 2 diabetes in the UK. Insulin glargine has been shown to reduce the likelihood of hypoglycaemia and improve HbA^sub 1c^ in type 2 diabetes by comparison with NPH insulin.
Methods: The study used a discrete event simulation model designed to forecast the costs and health outcome of a cohort of 1000 subjects over 40 years. The two main scenarios involved a difference in the likelihood of hypoglycaemia or a difference in HbA^sub 1c^. Prices were in UK 2005 costs. Costs and benefits were discounted at 3.5% per year. Effectiveness data were pooled data from randomised clinical trials.
Results: The incremental cost-effectiveness ratio of insulin glargine versus NPH insulin was 10 027 per quality adjusted life year under a differential hypoglycaemia-only scenario. The incremental cost-effectiveness ratio of insulin glargine versus NPH insulin was 13 921 per quality adjusted life year under a differential HbA^sub 1c^-only scenario. In wide-ranging sensitivity analysis, the ICER was consistently below 20 000 per quality adjusted life year gained.
Conclusion: This study was the first to evaluate the relative cost-effectiveness of insulin glargine versus NPH insulin. Insulin glargine resulted in significant health benefits and represents excellent value for money for the treatment of type 2 diabetes in the UK.
Introduction
Insulin glargine (glargine) is a long-acting insulin analogue intended for use in people with either type 1 or type 2 diabetes. Following a positive evaluation of the suitability of glargine by the National Institute for Health and Clinical Excellence (NICE) during its product launch in 2002, glargine has become established as basal insulin in the UK1. In 2005, a volume of 350000 glargine prescriptions were dispensed throughout England2.
Patients with type 2 diabetes treated with insulin typically have advanced disease and have gone through progressively more aggressive monotherapy and combination therapy with the oral hypoglycaemic agent (OHA) classes – metformin and sulphonylureas. Once subjects have ‘failed’ on OHAs there is usually a need for exogenous insulin alone or insulin in combination with OHAs3.
The general objectives of diabetes-specific treatments are widely described and relate to the achievement of good glycaemic control, typically measured as percent glycosylated haemoglobin (HbA^sub 1c^). There is irrefutable evidence that good glycaemic control is associated with a reduced likelihood of progressing to microvascular disease and macrovascular disease end points4 5. Furthermore, the use of exogenous insulin and some OHAs can result in hypoglycaemia. Hypoglycaemia occurs in differing severities, at differing frequency, and can sometimes require hospitalisation. Traditionally, there is a recognised trade-off between glycaemic control and the risk of hypoglycaemia6. Hypoglycaemia of any severity has a profound effect on patients’ quality of life7 and is regarded as the single greatest obstacle to achieving normoglycaemia8.
Many studies have found that patients treated with glargine achieved similar HbA^sub 1c^ levels to patients treated with NPH but with a lower rate of hypoglycaemia in both type 1 and type 2 diabetes9″17, whereas four trials in type 1 diabetes and one trial in type 2 diabetes supported an improvement in HbA^sub 1c^18-22. Even though the design of these randomised trials was only intended to show non-inferiority, the purpose of this study was to evaluate the relative cost-effectiveness of glargine versus NPH insulin using pooled data from these clinical trials in people with type 2 diabetes in the UK.
Methods
Modelling approach and model
This evaluation was undertaken within the context of the UK National Health Service (NHS) and used the NHS as its perspective. The method used was a cost-utility analysis (CUA) intended to determine the cost per quality-adjusted life year gained (QALY gained) using glargine versus NPH insulin, and determine if this is within the threshold for treatments that are considered to represent good value for money for general use within the NHS (200000 -30000 per QALY)23.
The modelling approach here was a discrete event simulation (DES) model of people with type 2 diabetes using Microsoft Excel. The DES ‘engine’ was written in embedded C++ programming language to minimise computation time. A flow diagram of the model is illustrated in Figure 1. The model had the ability to assess the economic impact of either a reduction in hypoglycaemia or an improvement in glycaemia or both of these factors at the same time. The time increment was 1 year and the model was designed to simulate a cohort of subjects (up to 10000 people) over a maximum time horizon of 40 years. In the main, the model used evidence from the UKPDS study to simulate disease progression in type 2 subjects. For the base-case analysis the characteristics of patients switching to insulin glargine used were obtained via the THIN database (Table I)24.
Modelling events for the type 2 diabetes component of the model was undertaken by coding the UKPDS 68 outcomes model25. As described in the original study, a series of seven Weibull proportional hazards regression equations were used to predict diabetes-related complications with UK life tables used to predict allcause mortality.
The complications predicted by these equations were:
* Ischaemic heart disease (IHD)
* Acute myocardial infarction (AMI)
* Congestive heart failure (CHF)
* Stroke
* Amputation (neuropathy)
* Renal failure (nephropathy)
* Blindness (retinopathy)
* Hypoglycaemic events (mild, moderate and severe)
* Non-cardiovascular death
Regarding the processing of simulated patients within the model, in the first instance a patient cohort was generated based upon the user-specified number of subjects to simulate and whether 1st- or 2nd-order uncertainty was selected. With Ist-order uncertainty selected, all simulation subjects had exactly the same patient profile, with 2nd order selected each patient differed (the degree or variability a function of the standard deviation or other distributional parameters entered by the user). Each individual within the cohort progressed through the model in yearly increments. Checks were made at the beginning of each time period for specific fatal or non-fatal events. The order in which these events occurred was randomised. If a fatal event occurred, all costs, life years and quality-adjusted life years were recorded and the simulation ended for that individual. The simulation model then selected the next individual and the simulation process began again. Assuming a subject did not die in any specific year, then following the ‘check for events’ stage, a simulated subject’s disease state was updated. Any appropriate decrement in health-related utility was then applied, together with any associated costs. The simulation clock was advanced and, if the end of the simulation time horizon had been reached, the simulation ended for that individual and the process resumed for the next simulated individual. Once all subjects had been simulated, the process ended and all summary statistics were generated for that particular run of the model.
Table 1. Base-case baseline patient characteristics
Figure 1. Flow diagram of the discrete event simulation model
In this instance, the second run of the model for the alternative treatment utilised exactly the same patient cohort data as the first, but applied the alternative treatment effects to the relevant variables (typically HbA^sub 1c^ and/or hypoglycaemic events). Alternative therapy costs were also applied. Insulin costs were derived per subject based on their respective weight. After applying any differential effects to the patient data, the model was then re- initialised and run through in exactly the same manner as for the first run. The model could be run flexibly under a wide variety of simulated scenarios.
Effectiveness
When using glargine, clinical trials have shown consistently a significant reduction in the risk of nocturnal hypoglycaemia and a trend towards a reduction in the risk of severe and symptomatic hypoglycaemia compared with NPH insulin9-22. Two meta-analyses have been carried out. The first reported a significant reduction in severe, nocturnal and symptomatic hypoglycaemia26 and this was supported more recently in a further meta-analysis that included three additional studies completed27. Relative risks and relative risk reductions are reported in Table 2. A meta-regression analysis investigating glargine’s ability to reduce HbA^sub 1c^ in type 2 diabetes has also been reported16. This also demonstrated that for a given rate of confirmed symptomatic hypoglycaemia, there was a 0.44% lower HbA^sub 1c^ with glargine versus NPH. For confirmed nocturnal hypoglycaemia this same model showed a 0.87% reduction in HbA^sub 1c^ (these values were used in sensitivity analysis in this study). The relative risk reduc\tions (RRRs) listed in Table 2 for the various categories of hypoglycaemia constitute model scenarios.
The background rates of hypoglycaemia were taken from published data. The rates of severe, nocturnal and symptomatic hypoglycaemia were drawn from the DCCT trial4, the Cardiff Hypoglycaemia Survey7, Riddle and colleagues’ and Pampanelli and colleagues (Table 3)28. It has been shown that the risk of hypoglycaemia was similar in patients treated with insulin irrespective of whether they had type 1 or type 2 diabetes29.
Estimates of financial costs
The financial costs applied to the simulation model are detailed in Table 4. Costs of medical treatment were obtained from UK sources and were indexed to year 2005 with UK Treasury rates31. Costs were attributed as follows (Table 4):
Hypoglycaemia. Costs over the 12-month study period in one study of 244 admissions for hypoglycaemia totalled 92078 of Healthcare resources, some 377.37 per episode29. Given that this represents only a fifth of all severe episodes4, the average known cost across all episodes in 1998 (year of data collection) was estimated to be 88.69. Although other estimates include the cost of rescue medication administered in the community32, data are not available to confirm what proportion of severe hypoglycaemia was treated here. The model used a conservative weighted average based only on known healthcare resource usage.
Table 2. Hypoglycaemia risk reduction: glargine versus NPH insulin
Table 3. Estimated annual hypoglycaemia risks and number of expected events
Insulin. Unit costs and market shares of both glargine and NPH were obtained from the British National Formulary No. 50 and Department of Health Prescription Cost Analysis data for 2004. A 0.40 IU/kg/day dose regimen was assumed for both glargine and NPH when used in a basal-insulin only insulin regimen.
Macrovascular events. Macrovascular event costs are split into either fatal or non-fatal event costs and were applied in the year in which the event occurred. Maintenance costs for those subjects surviving were applied in all subsequent years until either the end of the simulation time horizon or until the subject died.
Retinopathy. Subjects were split probabilistically into those who were screened and those who were not. Screening costs were an annual cost applied to whichever retinopathy state a subject was in, except if blind (severe visual loss). Subjects who were not screened could transfer to the screened group.
Blindness (severe visual loss). Subjects were assumed to incur blindness in both eyes simultaneously and therefore the event of blindness occurred once only. The initial cost related to the event was assumed to be equal to zero, with subsequent maintenance costs applied annually thereafter from published data38.
Table 4. Event, maintenance and therapy costs (indexed to UK 2005)
Nephropathy. Patients who had microalbuminuria (MAU) incurred an annual diagnostic cost (test strip) and angiotensin-converting enzyme (ACE) inhibitor therapy charges. Dialysis costs were annual weighted mean costs for peritoneal and haemodialysis. Transplant costs were applied in the year of transplant, and maintenance costs applied in all subsequent years.
Peripheral vascular disease. Amputations had a single cost associated with the event and a subsequent annual maintenance cost. No other costs were included.
Health-related utility
Utility estimates were derived either from the UKPDS study25 or generated via the HODaR database3940. Table 5 details the utility decrements used for each specified event. Where possible, utility values reported in the UKPDS outcomes model were used; where these were not reported, utility values from the HODaR database were used. Subsequent events, by default, incurred the same utility decrement as in the initial event, although the model allowed for alternative values for subsequent events to be applied. Utility associated with hypoglycaemia events was handled somewhat differently. Statistical models were developed that related the frequency and severity of hypoglycaemia to fear of hypoglycaemia, and subsequently to changes in health-related utility7.
Table 5. Health-related utility decrements
Further model details
The base-case model followed a cohort of 1000 subjects in each arm over 40 years. All prices were adjusted to 2005 values (UK) and discount rates of 3.5% were applied to both costs and benefits. The base-case scenarios were evaluated firstly for a difference in only hypoglycaemia, and secondly for a difference in only HbA^sub 1c^.
Internal validation was conducted by comparing the model’s clinical event output data to the macrovascular and microvascular output reported in the UKPDS study25.
Sensitivity analysis
One-way sensitivity analysis was carried out on all key model inputs. This was conducted by running the model with the upper and lower limits specified in Tables 3-6. Modification to hypoglycaemia background rates and risk reduction, costs and utilities were applied both separately and combined. Sensitivity analysis on utility values also included applying half the event specific utility decrement in subsequent years following each respective event. The equations used to predict utility values associated with hypoglycaemia were modified by increasing/decreasing the published equation coefficients by 1.96 SE.
Sensitivity analysis was also conducted on discount rates (0-6%) and model time horizon (5, 10 and 20 years).
Results
Model validation
Internal validation of the model using the UKPDS 68 equations was performed using baseline UKPDS profiles along with data from the UKPDS itself. Findings are illustrated in Figure 2. Reasonable concordance was observed generally with the complications included in the model; however, the model slightly over-predicted total mortality when compared with UKPDS findings.
Table 6. The forecast frequency of vascular end points when using glargine or NPH insulin under the two main scenarios (see footnote to explain non-integer values)
Number of end points forecast in each arm
The model was designed to output the frequency in each of a range of pertinent vascular events and mortality that are listed in Table 6 for the two main scenarios. In the first scenario, where only hypoglycaemia was considered, there was an overall difference of 1999 events, 9392 events and 18786 events for severe, nocturnal and symptomatic hypoglycaemia, respectively. In the second scenario, where HbA|c differences were considered, there was a difference in combined events (excluding hypoglycaemia) of 33 events (Table 6). In this scenario there was less hypoglycaemia in the NPH group as a result of differential survival. The biggest difference was in the forecast frequency of acute myocardial infarction (16 events).
Total costs and QALYs forecast in each arm
Total treatment costs over the simulation period when using NPH insulin were less than glargine because of the difference in unit cost of the two insulin options. In the differential hypoglycaemia scenario, discounted total costs were 4892 and 6433 per patient for NPH insulin and glargine, respectively (Δ [asymptotically =] 1541). There was a total difference of 111 discounted qualityadjusted life years (QALYs) with this scenario across the 1000 subjects. Under this scenario, there was no difference in total life years because of the very small likelihood of dying from hypoglycaemia. Under the alternative scenario of a difference in HbA1, the rounded difference in discounted total costs, total life years and QALYs was 1.5 million, 150 years and 111 QALYs, respectively (Table 7).
Incremental cost-effectiveness ratios and sensitivity analysis
Under the hypoglycaemia scenario the mean incremental cost- effectiveness ratio (ICER) was 10027 per QALY gained. Under the HbA, scenario, the mean ICER was 13921 per QALY gained.
In detailed one-way sensitivity analysis intended to investigate the effects of uncertainty, the ICER varied within reasonable limits (Figure 3) and the majority of the mean ICER values were within 20000 per QALY gained and all scenarios were within 30000 per QALY. Various scenarios that were expected to change the ICER are shown in the ‘tornado plot’ illustrated in Figure 3. The ICER was most sensitive to the price of glargine, the utility decrement associated with hypoglycaemia, and the cohorts’ mean weight. Further sensitivity results are presented in Table 8 which indicate that reducing the hypoglycaemia risk and HbAlc treatment effects by 50% impact substantially on the results, with costs per QALY of 29 040 and 22420, respectively. Modifying the coefficients for the hypoglycaemia utility equations had the greatest impact upon the hypoglycaemia only scenario in which the cost per QALY was increased to 22557.
Figure 2. Twelve-year cumulative incidence observed in the UKPDS cohort and forecast using the model
Table 7. Total costs and benefits under the two main scenarios
Discussion
This economic evaluation found that insulin glargine when used to treat type 2 diabetes was cost-effective in the UK, i.e. it was within commonly accepted thresholds for treatments that are regarded as providing value for money by NICE23. Furthermore, this conclusion was upheld under a range of plausible scenarios.
Figure 3. Base-case and sensitivity analysis: a ‘tornado plot’ of the incremental cost-effectiveness ratios under alternative scenarios. Hypo only, refers to the hypoglycaemia base-case scenario; HbA^sub 1c^ only, refers to the HbA^sub 1c^ base-case scenario
Table 8. Sensitivity analysis
This study had a number of strengths and weaknesses. Firstly, it was a modelled simulation and, as such, is only as reliable as the predictive algorithms and other data that were used in its construction. However, the model used some of the most contemporary and widely regarded data and risk algorithms available, for example, the UKPDS and HODaR, wi\th baseline characteristics matched to the typical profile of patients switching to insulin. The internal validation exercise reported here provides evidence that the model is operating reasonably reliably and, while it is true that no modelled economic evaluation is going to provide point estimates that are demonstrably precise, the predicted ICERs generally remained under the 20000 per QALY threshold. Under most scenarios, this provides further confidence in the main conclusion of this study.
An important consideration here was that the underlying effectiveness data were derived from pivotal clinical trials that were designed to only show non-inferiority in clinical trials. The design of this trial is entirely justified; however, data elsewhere in this supplement demonstrate that glargine shows considerable benefits over-and-above those shown in the clinical trials24,41. Therefore, this evaluation could reasonably be regarded as being conservative.
This study has a number of implications for policy makers. It demonstrates that the original decision by NICE to allow glargine to be used in the UK has been vindicated, and there has likely to have been considerable health benefit as a consequence.
The main area for research that this study highlights – along with the sister studies reported in this journal Supplement24,41,42 – is the considerable differences in the findings from clinical trials and from ‘real-life’ or ‘naturalistic’ data. This should be investigated thoroughly.
Conclusion
When compared to NPH insulin, treatment with insulin glargine resulted in considerable health benefits and it was cost-effective when used to treat people with type 2 diabetes in the UK. Within the limitations of this study this analysis shows that insulin glargine represents good value for money.
Acknowledgements
This study was funded by sanofi-aventis UK.
References
1. NICE appraisal no. 53. [http://www.nice.org.uk/page. aspx?o=TA053] (accessed 28th April 2006)
2. Prescription cost analysis for England, (accessed 28th April 2006) [http://www.ic.nhs.uk/pubs/prescostanalysis2005]
3. Bagust A, Evans M, Beale et al. A model of long-term metabolic progression of type 2 diabetes mellitus for evaluating treatment strategies. Pharmacoeconomics 2006;24(Suppl 1):5-19
4. The 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;329:977-86
5. Turner RC, Milns H, Neil HA, et al. Risk factors for coronary artery disease in non-insulin dependent diabetes mellitus: United Kingdom Prospective Diabetes Study (UKPDS: 23). BMJ 1998;316:823-8
6. Davis EA, Keating B, Byrne GC, et al. Impact of improved glycaemic control on rates of hypoglycaemia in insulin dependent diabetes mellitus. Arch Dis Child 1998;78:111-15
7. Currie CJ, Morgan CLl, Poole CD, et al. Multivariate models of health-related utility and the fear of hypoglycaemia in people with diabetes. Curr Med Res Opin 2006;22:1523-34
8. Winocour PH. Effective diabetes care: a need for realistic targets. BMJ 2002;324:1577-80
9. Pieber TR, Eugene-Jolchine I, Derobert E. Efficacy and safety of HOE901 versus NPH insulin patients with type 1 diabetes. Diabetes Care 2000; 23:157-62
10. Raskin P, Klaff L, Bergenstal R, et al. A 16-week comparison of the novel insulin analog Lantus (HOE901) and NPH human insulin with insulin lispro inpatients with type 1 diabetes. Diabetes Care 2000;23:1666-71
11. Home PD, Rosskamp R, Forjanic-Klapproth J, Dressier A. A randomized multicentre trial of insulin glargine compared with NPH insulin in people with type 1 diabetes. Diabetes Metab Res Rev 2005;21:545-53
12. Schober E, Schoenle E, Van Dyk J, Wernicke-Panten K. Comparative trial between insulin glargine and NPH insulin in children and adolescents with type 1 diabetes mellitus. J Pediatr Endocrinol Metab 2002; 15:369-76
13. Ratner R, Hirsch IB, Niefmg JL, et al. Less hypoglycaemia with Lantus in intensive insulin therapy for type 1 diabetes. Diabetes Care 2000;23:639-43
14. Rosenstock J, Park G, Zimmerman J. Basal Lantus (HOE901) versus NPH insulin in patients with type 1 diabetes or multiple daily insulin regimens. Diabetes Care 2000;23:1137-42
15. Rosenstock J, Schwartz SL, Clark CM, et al. Basal insulin therapy in Type 2 diabetes 28-week comparison of insulin glargine (HOE 901) and NPH insulin. Diabetes Care 2001;24:631-6
16. Yki-Jrvinen H, Dressier A, Ziemen M. Less nocturnal hypoglycaemia and better post-dinner glucose control with bedtime insulin glargine (HOE 901) versus NPH insulin during combination therapy in type 2 diabetes. Diabetes Care 2000;23:1130-6
17. Riddle MR, Rosenstock J, Gerich J. The Treat-to-Target Trial: randomized addition of glargine or human NPH insulin to oral therapy of type 2 patients. Diabetes Care 2003;26:3080-6
18. Fulcher GR, Gilbert RE, Yue DK. Glargine is superior to neutral protamine Hagedorn for improving glycated haemoglobin and fasting blood glucose levels during intensive insulin therapy. Intern Med J 2005;35:536-42
19. Rossetti P, Pampanelli S, Fanelli C, et al. Intensive replacement of basal insulin in patients with type 1 diabetes given rapidacting insulin analog at mealtime. Diabetes Care 2003;26: 1490- 6
20. Porcellatti F, Rossetti P, Pampanelli S, et al. Better long- term glycaemic control with the basal insulin glargine as compared with NPH in patients with Type 1 diabetes mellitus given mealtime lispro insulin. Diabet Med 2004;21:1213-20
21. Murphy N, Keane S, Ong K, et al. A randomized crossover trial comparing insulin glargine plus insulin lispro with NPH insulin plus soluble insulin in adolescents with type 1 diabetes. Diabetologia 2002;45(Suppl 2):871
22. Fritsche A, Schweitzer MA, Hring H. Improved glycaemic control and reduced nocturnal hypoglycaemia in patients with type 2 diabetes with morning administration of insulin glargine compared with NPH insulin. Diabetes Care 2002;51 (Suppl 2):149
23. Devlin N, Parkin D. Does NICE have a cost-effectiveness threshold and what other factors influence its decisions? A binary choice analysis. Health Economics 2004; 13:437-52
24. Poole CD, Tetlow T, McEwan P, et al. The prescription cost of managing people with type 1 and type 2 diabetes following initiation of treatment with either insulin glargine or insulin detemir in routine general practice in the UK: a retrospective database analysis. Curr Med Res Opin 2007;23(Suppl 1):S41-48
25. Clarke PM, Gray AM, Briggs A, et al. UK Prospective Diabetes Study (UKDPS) Group. A model to estimate the lifetime health outcomes of patients with type 2 diabetes: the UK Prospective Diabetes Study (UKPDS) Outcomes Model (UKPDS no. 68). Diabetologia 2004;47:1747-59
26. Rosenstock J, Park G, Zimmerman J. Basal Lantus (HOE901) versus NPH insulin in patients with type 1 diabetes or multiple daily insulin regimens. Diabetes Care 2000;23:1137-42
27. Medical Research Matters Limited. Insulin glargine and insulin detemir for type 1 and type 2 diabetes: systematic review of randomised controlled trials. Report for sanofi-aventis, January 2006 (data on file sanofi-aventis, Guildford).
28. Pampanelli S, Fanelli C, Lalli C, et al. Long-term intensive therapy in IDDM: effects on HbA1c, risk of severe and mild hypoglycaemia, status of counterregulation and awareness of hypoglycaemia. Diabetologia 1996;39:677-86
29. Leese GP, Wang J, Broomhall J, et al. Frequency of severe hypoglycaemia requiring emergency treatment in type 1 and type 2 diabetes: a population-based study of health service resource use. Diabetes Care 2003;26:1176-80
30. HOE-901/4002. Target glycaemic control and the incidence of symptomatic nocturnal hypoglycaemia in insulin naive subjects with Type 2 diabetes on oral hypoglycaemic agent(s) and treated with insulin glargine or NPH human insulin. Data on file, sanofi- aventis, Guildford, Surrey, UK: 2002
31. UK Treasury discount rates [http://www.hm-treasury.gov.uk/ economic_data_and_tools/gdp_deflators/data_gdp_index.cfm] (accessed 28th April 2006]
32. Clarke P, Gray A, Legood R, et al. The impact of diabetesrelated complications on healthcare costs: results from the United Kingdom Prospective Diabetes Study (UKPDS Study No. 65). Diabetic Med 2003;20:442-50
33. Palmer AJ, Weiss C, Sendi PP, et al. The cost-effectiveness of different management strategies for type 1 diabetes: a Swiss perspective. Diabetologia 2000;43:13-26
34. UK Drug tariff, December 2005
35. McEwan P, Dixon S, Baboolal K, et al. Evaluation of the cost effectiveness of sirolimus versus tacrolimus for immunosuppression following renal transplantation in the UK. Pharmacoeconomics 2006;24:67-79
36. NHS Reference costs [http://www.dh.gov.uk/PolicyAndGuicknce/ OrganisationPolicy/FinanceAndPlanning/NHSReferenceCosts/fs/ en] (accessed 2nd May, 2006)
37. Costs of glargine from the BNF [http://www.bnf.org/bnf/] (accessed 2nd May, 2006)
38. O’BrienJA, Patrick AR, Caro JJ. Cost of managing complications resulting from type 2 diabetes mellitus in Canada. BMC Health Serv Res 2003;3:7
39. Currie CJ, McEwan P, Peters JR, et al. The routine collation of health outcomes data from hospital treated subjects in the Health Outcomes Data Repository (HODaR): descriptive analysis from the first 20,000 subjects. Value Health 2005;8: 581-90
40. Morgan CLl, McEwan P, Morrissey M, et al. Characterisation and comparison of health-related utility in people with diabetes in various states of single and multiple vascular complications. Diabetic Med 2007 (in press)
41. Currie CJ, Poole CD, Tetlow T, et al. The outcome of care in people with type 1 and type 2 diabetes following switching to treatment with either insulin glargine or insulin detemir in routine general practice in the UK: a retrospective database analysis. Curr Med Res Opin 2007;23(Suppl l):S33-39
42.McEwan P, Poole CD, Tetlow T, et al. Evaluation of the costeffectiveness of insulin glargine versus NPH insulin for the treatment of type 1 diabetes in the UK. Curr Med Res Opin 2007;23(Suppl 1):S7-19
43. Lee AJ, Morgan CLl, Conway P and Currie CJ. Characterisation and comparison of health-related quality of life for patients with renal failure. Curr Med Res Opin 2005;21:1777-83
CrossRef links are available in the online published version of this paper: http://www.cmrojournal.com
Paper CMRO-3633_4, Accepted for publication: 11 January 2007
Published Online: 07 February 2007
doi: 10.1185/030079906X167570
Phil McEwan(a), Chris D. Poole(b), Tony Tetlow(b), Paul Holmes(c) and Craig J. Currie(d)
a School of Mathematics, Cardiff University, Cardiff, UK
b Cardiff Research Consortium, Heath Park, Cardiff, UK
c sanofi-aventis UK, Guildford, Surrey, UK
d Department of Medicine, School of Medicine, Cardiff University, Cardiff, UK
Address for correspondence: Dr Phil McEwan, Medicentre, Heath Park, Cardiff CF14 4UJ, UK.
Tel.: +44 (0) 29-2068-2048; Fax +44 (0) 29-2074-2049; email: mcewan@cardiff.ac.uk
Copyright Librapharm 2007
(c) 2007 Current Medical Research and Opinion. Provided by ProQuest Information and Learning. All rights Reserved.