There are limited data on the cost-effectiveness of insulin treatments in Bulgaria. With the increasing number of insulin treatment options, decision making based on economic evidence is essential to maximise health outcomes while effectively managing restrained budgets.
In this simple, short-term cost-effectiveness analysis, degludec was demonstrated to be cost-effective versus biosimilar glargine U100 in people with T1DM (ICER 4,498.68 BGN), T2DMBOT (ICER 399.11 BGN) and T2DMB/B (ICER 7,365.22 BGN). In all three patient groups insulin costs with degludec were higher than with glargine U100. However, these costs were offset by a reduction in the number of hypoglycaemic events with degludec in all three patient groups.
Cost-effectiveness analyses usually model the long-term impact of diabetes interventions on disease-related complications as a function of the differences in glycaemic control. However the data used in this model were derived from treat-to-target trials and glycaemic control was similar across both arms, thus the use of a long-term model based on differences in HbA1c was not appropriate. Therefore, this simple, transparent short-term model focuses on other important aspects associated with insulin therapy, including hypoglycaemia and insulin dosing. Although this model only reflects a 1-year time horizon, it not only represents the cost-effectiveness of degludec vs glargine U100 within the first year of treatment, but it can also be replicated for subsequent years, representing annual cost-effectiveness. This is supported by the insensitivity of the ICER to changes in the time horizon in the sensitivity analyses. The model has previously been used to evaluate the cost-effectiveness of degludec versus glargine U100 in patients with T1DM and T2DM in different settings [33, 34, 42, 44-46]. In the United Kingdom (UK) and Denmark, degludec was found dominant versus glargine U100 in T1DM and T2DMBOT and highly cost-effective in T2DMB/B [33, 44]. Similarly, in Serbia and Sweden, degludec was cost-effective versus glargine U100 in all three patient groups [34, 42]. These results are consistent with those observed in the current study.
Hypoglycaemia can have a major impact on patient’s quality of life and lead to significant psychological and physical morbidity and in severe cases death [47, 48]. Additionally, people with recurrent hypoglycaemic episodes are at risk of developing impaired awareness of hypoglycaemia which can cause patients to miss early symptoms and treat hypoglycaemia, increasing the risk of severe hypoglycaemic events [49]. A recent real world multi-national, non-interventional study assessed the prevalence of hypoglycaemia worldwide, including Bulgaria. In Eastern Europe the estimated overall annual rate of hypoglycaemic events was 66.9 per person per year in T1DM and 23.7 per person per year in T2DM [50]. The prevalence of nocturnal and severe hypoglycaemia were 9.8 and 4.5 per person per year in T1DM and 4.0 and 2.2 per person per year in T2DM. The nocturnal hypoglycaemic event rates in T1DM and severe events in T1DM and T2DM in Bulgaria are higher than those used in this model which, based on the impact of hypoglycaemic event rates demonstrated in the sensitivity analyses, suggests that degludec may be even more cost-effective in real world clinical practice in Bulgaria.
The real-world cost effectiveness of switching patients with T1DM from other basal insulins (glargine U100, insulin detemir and NPH insulin) to degludec has been investigated using the IQVIA CORE diabetes model from the perspective of the United Kingdom and Sweden [30, 51]. The CORE diabetes model is a lifetime Markov model predicting diabetes complications over time in patient populations representative of clinical practice and calculates the resulting economic impact. From both the UK and Swedish perspective, degludec was dominant versus other basal insulins, which was mainly driven by the significant reduction in HbA1c and lower rates of hypoglycaemia with degludec [30, 51].
Hypoglycaemia and the fear of hypoglycaemia are key barriers to the initiation and intensification of insulin regimens [52]. As a result of fear of hypoglycaemia, approximately 52% of people with T1DM and 41% of people with T2DM reduce their insulin dose following a hypoglycaemic event [53]. This compromises glycaemic control and puts patients at risk of serious long-term complications, such as cardiovascular disease, renal disease, retinopathy, neuropathy and amputations [54]. The unique pharmacological profile of degludec with a flat and stable action profile is associated with lower rates of hypoglycaemia [20, 22]. In phase 3a clinical trials comparing degludec with glargine U100, degludec demonstrated equivalent reductions in HbA1c with significantly lower rates of hypoglycaemia. In T1DM, degludec was associated with a 17% lower rate of non-severe nocturnal events, while in T2DMBOT and T2DMB/B, rates of non-severe nocturnal hypoglycaemia were decreased by 36% and 25%, respectively [27]. Additionally, the rate of non-severe daytime events was 17% lower with degludec in T2DMB/B, and the rate of severe event hypoglycaemia was 86% lower in T2DMBOT [27, 28].
The SWITCH 1 and SWITCH 2 trials, two phase 3b studies, sought to confirm the efficacy and safety of degludec in people with T1DM and T2DM with an increased risk of hypoglycaemia [55, 56]. In these studies, degludec achieved equivalent reductions in HbA1c with a significantly lower rate of overall symptomatic and severe hypoglycaemia versus glargine U100. These trials are more representative of patients in regular clinical practice than the phase 3a clinical trials, which excluded patients with recurrent hypoglycaemia. Data from SWITCH trials have recently been used to demonstrate that degludec is a cost-effective alternative to glargine U100 in the UK [57]. Furthermore, the hypoglycaemia benefit of degludec has been confirmed in real-world studies which demonstrate that switching to degludec from other basal insulin regimens is associated with significantly improved glycaemic control and a reduction in the rate of non-severe and severe hypoglycaemic events in T1DM and T2DM [29, 30].
In Bulgaria, SMBG tests are not paid by the healthcare payer but are paid out-of-pocket by patients, and were therefore not considered in this analysis. Due to the long duration of action with a flat and stable action profile and the lower day-to-day variability versus glargine U100, fewer SMBG tests are needed for titration and maintenance with degludec [20, 22, 58]. Fewer SMBG tests would be cost saving for the patient, although this was not assessed in the present analysis. The long and stable action profile of degludec also allows for flexibility of dosing time without compromising efficacy or risk of hypoglycaemia [25]. This flexible dosing option may provide an additional benefit especially for those people who have difficulties adhering to their treatment regimens (e.g. shift workers, frequent travellers, patients requiring help with insulin injections). Here, an estimate of the utility benefit of flexible dosing with degludec was included in the analysis and a utility gain of 0.006 derived from the study by Boye et al. [40] was applied for degludec. This can be considered a conservative estimate as a large time trade-off study identified a utility gain of 0.016 associated with flexible dosing of basal insulin [43].
It is important to acknowledge the limitations associated with the current analysis. With biosimilar glargine U100 being a relatively new-to-market insulin, no head-to-head trial data is available comparing degludec vs biosimilar glargine U100 and the analysis was based on currently available data and plausible assumptions, and the results should be interpreted accordingly. The analysis used hypoglycaemic event rates derived from published economic analyses using this model [33]. This could have resulted in an underestimation of the cost-effectiveness of degludec vs glargine U100 in Bulgaria, as a recent study suggests that rates of hypoglycaemia in real-world clinical practice in Bulgaria may be higher [50]. Additionally, the data used in this model to inform hypoglycaemia rate ratios originate from meta-analyses of phase 3a clinical trials [27, 28] to increase the statistical power of the analyses and increase the reliability of the data, and assumes replication of such rates in clinical practice. The clinical trials which informed this analysis used a treat-to-target approach in which patients were titrated until glycaemic targets were reached. In clinical practice, glycaemic targets are often not met for a variety of reasons, including non-adherence, and missed follow-up appointments. However, sensitivity analyses demonstrate that the results are robust to a wide variation in parameters, supporting the validity of the results.