Post-market Review Products Used in the Management of Diabetes Report to Government Stage 2: Insulin Pumps February 2015

5.5 Literature review: cost-effectiveness studies of insulin pumps

The review assessed cost-effectiveness studies published between 2007 and October 2014 comparing insulin pump therapy to multiple insulin injections in patients with type 1 diabetes.

Four cost-effectiveness studies met the inclusion criteria: one undertaken in Australia (Cohen 2007), one in the United Kingdom (UK) (Cummins 2010), one in Canada (Charles 2009) and one in the United States (US) (St Charles 2009). Reported incremental cost-effectiveness ratios (ICERs) and selected sensitivity analyses results were converted to $AUD prices for the year 2013 to facilitate comparison across studies.

All studies used the computer-simulation CORE model, which has been developed to estimate the long term clinical and economic consequences of interventions for type 1 and type 2 diabetes (Palmer 2004). The CORE model may overestimate the incidence of macrovascular complications in people with type 1 diabetes (Cummins 2010). In all studies, the perspective was from a public or private payer, and costs were taken from national sources. In the base cases, the time horizon was 50 or 60 years and the discount rate varied between 3% and 5% per annum applied to both costs and clinical outcomes.

The ICER per quality-adjusted life year (QALY) adjusted for purchasing power parity and converted to AUD$ 2013 prices ranged between AUD$28,874 in the US study and AUD$96,220 in the UK study assuming a reduction in HbA1c level of 0.9% in one study and 1.2% in three studies. As all studies used the CORE model, with similar perspectives and time horizons, differences in results appeared to rise from differences in input variables for the benefits of insulin pump therapy, which included reductions in HbA_1c level and severe hypoglycaemia episodes, and costs of diabetes complications.

Reductions in HbA_1c level included in the base cases of all cost-effectiveness studies were higher than the reductions that were reported in the systematic review of RCTs (0.10–0.30%) and statistically significant results from many of the observational studies. Sensitivity analyses undertaken in the cost-effectiveness assessments demonstrated that the ICER was most sensitive to varying reductions in HbA_1c levels with insulin pump therapy compared with multiple daily injections. Large increases in the ICER were observed with lower reductions in HbA_1c. In the UK study, decreasing the reduction in HbA_1c level from 0.9% in the base case to 0.6% increased the ICER by 43%. In the three other studies, decreasing the reduction in HbA_1c level from 1.2% in the base case to 0.51% increased the ICER by 68%, 113% and 132% in the Australian, Canadian and US studies, respectively. Assuming reductions in HbA_1c levels of 0.51% to 0.675% with insulin pump therapy compared to multiple insulin injections, the ICER per QALY converted to AUD$ 2013 prices ranged from AUD$63,274 in the US study to AUD$292,952 in the UK study.

Studies used different assumptions (0% or 50%) for reductions in severe hypoglycaemic events avoided with insulin pump therapy compared to multiple insulin injections. All undertook sensitivity analyses varying hypoglycaemic events avoided up to 75%. ICERs were sensitive to changes in reductions in severe hypoglycaemic rates only when the cost of severe hypoglycaemia was valued in the high range (2007 USD$1,234).

With the exception of the UK study, all studies appear to underestimate the ICER by assuming a pump life span longer than the warranty period of 4 years for most insulin pumps currently marketed in Australia. Pump life span used in the studies was 7 years in the US study, and 8 years in the Australian and Canadian studies. In the US study, reducing the pump life from 7 years in the base case to 4 years in the sensitivity analysis increased the ICER by 54%.

With the exception of changes to quality of life associated with fear of severe hypoglycaemia, no study examined improvements in quality of life associated with insulin pump therapy. Stakeholders emphasised quality of life improvements as a key benefit of insulin pump therapy. The literature review of RCTs identified that reported quality of life was generally better with insulin pump therapy than multiple insulin injections in adults, adolescents and children.

In summary, the cost-effectiveness studies may have overestimated the cost-effectiveness of insulin pump therapy compared to multiple insulin injections by assuming larger decreases in HbA_1c than the results observed in RCTs and many of the observational studies, and longer pump life spans than the standard warranty period for insulin pumps. However, improvements in quality of life associated with insulin pump therapy were not assessed in the models with the exception of fear of severe hypoglycaemia. The main limitation of all the cost-effectiveness studies related to the assumption that improvements in HbA_1c observed in short term trials will generate long term morbidity benefits resulting in a decrease in diabetes complications. There are no long term trials of insulin pump therapy to support this assumption.

5.6 Insulin pump use under the Programme

The following figures are derived from the data collected and supplied by JDRF on the Insulin Pump Programme. The data collected includes:

• 
  age/sex of the recipient;
  
• 
  date of receipt;
  
• 
  pump brand;
  
• 
  household income;
  
• 
  co-payment amount; and
  
• 
  co-payment assistance indicator.
  

Up to 31 December 2012, 439 insulin pumps had been provided to eligible patients under the Programme (JDRF 2013). Figure 2 shows the number of insulin pumps provided by brand each year under the Programme. The Medtronic Paradigm Veo 754 is the most commonly supplied pump at 47%.

Figure 3 shows the total number of pumps supplied each year under the Programme. The uptake of the Programme increased significantly due to an agreement in October 2011 between the manufacturers and JDRF to pay the co-payment amount for families in the lowest income threshold of Family Tax Benefit Part A. This resulted in the lowest income families receiving a pump at no cost. Subsequently, in the 2012-13 financial year, Programme funding was exhausted by September 2012, and no more pumps were provided under the Programme beyond 31 December 2012. This explains the drop in the number of pumps supplied in 2012-13 seen in Figure 3.

Figure 4 shows the age distribution of insulin pump recipients under the Programme. Further information on the breakdown of insulin pump brand utilisation is detailed in Appendix C.

Figure 2. Number of insulin pumps administered under the Programme by brand and financial year.

Figure 3. Total number of insulin pumps administered under the Programme by financial year.

Figure 4. Number of insulin pumps administered under the Programme by age group.

Figure 5. Programme recipient income ranges for 2008-09 to 2012-13.