The Economic Burden of Insulin-Related Hypoglycemia in Spain

This study highlights the substantial cost burden of hypoglycemia in Spain for insulin-treated adults with diabetes. The cost of non-severe hypoglycemia in particular is often overlooked and underestimated; however, this study estimates the overall cost of non-severe episodes to be greater than the cost of severe episodes. This is due to the much higher rate of non-severe hypoglycemia compared with severe, despite the considerably lower cost of an average non-severe episode compared with severe. The costs associated with people with T1DM are higher than those with T2DM, despite the fact that there are approximately 50% more people with T2DM receiving insulin than people with T1DM. This is largely due to a higher rate of both severe and non-severe hypoglycemic episodes in people with T1DM, but also to the fact that a slightly higher percentage of severe episodes are treated in hospital rather than at home/work in T1DM. In addition, people with T1DM are 36% less likely to need hospitalization for a severe episode; however, if hospitalization is required, the average length of stay is nearly twice as long as for people with T2DM.

These results enable healthcare budget holders to estimate the cost of hypoglycemia in Spain or in a specific hospital or region, for example. In addition, the results enable decision makers to offset the cost of different insulin therapies with a more favorable hypoglycemia profile, as recommended in Spanish guidelines for hypoglycemia management [5], and patient education programs with the decrease in costs associated with an expected reduction in hypoglycemia rates.

Varying by ±20% the cost of a GP visit, the percentage of severe episodes treated in hospital for T1DM, and the percentage utilization for GP visits for a non-severe episode in T2DM had the largest effects on the overall results and therefore were the inputs that showed the greatest uncertainty. The cost and utilization of HCP visits (GP and diabetes specialist) were also among the main drivers of the model in an analysis conducted in the UK and Denmark [9, 10]. In addition, a study (SECCAID) that estimated the costs of diabetes in Spain reported that 28% of the total corresponds to primary care [24], further supporting the fact that GP costs are an important part of the cost of diabetes and its associated complications.

There are a number of limitations with the analysis; however, all assumptions are considered conservative and therefore the overall results are unlikely to be overestimated. General limitations of the approach are that only insulin-related hypoglycemia, symptomatic hypoglycemia, and an adult population are considered. While hypoglycemia is most commonly associated with insulin therapy, it can also be experienced with other antidiabetic medicines, e.g., sulfonylureas and glinides [3]. It is difficult to estimate whether asymptomatic hypoglycemia incurs additional costs and, if so, how much; however, not including any cost associated with asymptomatic episodes in the default case means that these costs will not be overestimated and that this analysis will, if anything, be conservative. Asymptomatic hypoglycemia is commonly associated with impaired awareness of hypoglycemia, which was reported in 55% of subjects with T1DM and 39% of subjects with T2DM in a questionnaire-based study in Spain [8]. Based on these figures, scenario analysis showed that asymptomatic patients were assumed only to suffer a non-severe hypoglycemic event, resulting in an additional SMBG test, but not incurring the additional cost of a GP appointment as in the default case. In addition, while there are much fewer children with diabetes than adults [12] and while it is impossible to speculate as to the utilization of healthcare resources in children compared with adults, there will at least be some cost for hypoglycemia in children that will add to the economic burden of hypoglycemia calculated in this study.

There are some necessary assumptions in the model. Firstly, the use of glucose and glucagon treatment is not included as a conservative approach. For severe episodes treated in hospital, it is assumed that these costs are included in the hospitalization tariff, A&E cost, and ambulance service cost, to avoid double counting. For non-severe episodes and severe episodes treated at home/work, no data were identified on the use of glucose or glucagon, and so these costs are not included; however, it is likely that glucose or glucagon would be used in the treatment of some (unknown) proportion of these hypoglycemic episodes. In an estimate of the cost of hypoglycemia in Denmark, data were available for glucose and glucagon use for severe episodes treated at home; 3.2% of episodes required glucagon, 6.8% required a glucose infusion, and 2.6% required a glucose tablet [10].

Second, no data were identified for the use of extra SMBG tests following a severe episode, and so the number used following a non-severe episode is assumed to also apply to severe episodes. This is also likely to be conservative, since patients are likely to monitor their blood glucose more closely following a severe episode than following a non-severe episode.

Third, the percentage of patients visiting a GP and diabetes specialist following a severe episode is assumed to be the same for T1DM and T2DM. The source for these data is a trial that included both patients with T1DM and patients with T2DM, in a 47:53 ratio. It has previously been reported that people with T2DM have a higher healthcare resource utilization than people with T1DM for a severe hypoglycemic episode [16], and data used in the present analysis show that a higher percentage of people with T2DM contact a healthcare professional following a non-severe hypoglycemic episode than people with T1DM [8]. Therefore, using data from a study in which approximately 50% of patients have T1DM is likely to misrepresent the true mix of patients, which is made up of only about 10% T1DM. This means that the utilization value obtained from this study is likely to be underestimated.

Fourth, a visit to a healthcare professional following a non-severe episode is assumed to be with a GP rather than a diabetes specialist, which is conservative as this is the less expensive of the two options. In addition, the cost of a hospital stay is assumed to be in a general ward rather than intensive care, which also costs the least of the two options.

Finally, it is possible that the hypoglycemia rates used in the analysis are under-reported. The study that the rates are obtained from is a questionnaire-based study that relies on patient recall. In addition, the occurrence and recurrence of hypoglycemia (particularly severe) can prevent people with diabetes from keeping or obtaining a driver’s license, particularly a commercial license [25], and so patients may not report all hypoglycemic episodes experienced for fear of losing their license.

There have been a limited number of studies reporting the cost of hypoglycemia in Spain. A survey-based study of insulin-related severe hypoglycemia in Spain in 2006–2007 [14], which informed the percentage of severe episodes treated in hospital for this analysis, estimated a cost per episode of €30 for family/domestic treatment, €246 for community HCP treatment, and €1396 for hospital HCP treatment in patients with T1DM, and €52, €371, and €1370 in patients with T2DM, respectively. An average was reported using the split between family/domestic treatment and community/hospital treatment that was also used in the present analysis, and different splits between community treatment and hospital treatment. The base case split for community/hospital of 50%/50% for T1DM and 35%/65% for T2DM gave an average cost per severe episode of €577 for T1DM and €691 for T2DM. These costs are for 2007; inflating to 2016 costs using the Spanish consumer price index [26] gives €621 for T1DM and €744 for T2DM, which are comparable to those estimated in the present analysis (€717 for T1DM and €680 for T2DM). Two other retrospective studies that calculated the cost of severe hypoglycemia, one only in patients with T1DM [27] and one only in Andalusia [28], reported a cost per episode of €239 and €702, respectively. The cost from the study in patients with T1DM is noticeably lower than that estimated in the present analysis (€808), which appears to be due to the fact that severe episodes were assumed to be treated by a family member in 75% of cases, compared with 33% in the present analysis. Another study that estimated the cost of severe hypoglycemia using experts’ opinions reported a cost per episode of just €172.33 [29] for all severe hypoglycemic episodes (not just insulin-related). However, there are a number of reasons for this difference. Again, a higher percentage of episodes were assumed to be treated by a family member (68%) than by a medical professional. In addition, for episodes that resulted in hospitalization, a stay of only 1 day was assumed, and only one extra SMBG test was assumed to be used following a hypoglycemic episode. There were also a number of acknowledged limitations with this study: experts may have non-representative experience of hypoglycemia; experts may find it difficult to correctly estimate average resource consumption; experts do not treat family-treated episodes (by definition); and data were not always available for resource use in Spain, so data from Greece or an average from other countries was used instead.

In contrast to the two studies reporting a lower cost per severe hypoglycemic episode [14, 29], a review article was also identified that reported the cost for a severe episode as €3597 and estimated the cost of a non-severe episode as between €30 and €35. These costs are both higher than the estimates in the present analysis, with the severe cost substantially so. In fact, the cost reported for a severe episode is in the same range as the highest cost of a severe episode calculated in the present analysis using maximum resource utilization (€4378 for T1DM and €3006 for T2DM).

Another review article [30] estimated the total direct cost of diabetes in Spain in 2009 to be €5.1 billion, meaning that the cost of hypoglycemia, as calculated in the present analysis, represents 13% of the total cost of diabetes, which is a substantial proportion.

For comparison, the cost of hypoglycemia has also been estimated using the LIHT for the UK [9] and for Denmark [10]. The results have been converted to Spanish Euros using purchasing power parities to directly compare and are shown alongside the results for the present analysis in Table 8.

The cost per severe episode is substantially higher in Spain than in Denmark and the UK, and the cost per non-severe episode is higher than in Denmark, but lower than the UK for patients with T1DM and comparable for patients with T2DM. The higher cost for severe episodes is due to a number of reasons. The main difference is that there are a much higher percentage of patients treated in hospital for severe episodes in Spain (66.7% and 64.3% for T1DM and T2DM, respectively) than in Denmark (20.3% for both T1DM and T2DM) or the UK (23% for T1DM and 53% for T2DM). The percentage of episodes resulting in hospitalization is lower in Denmark (4.4%) than in Spain (16% for T1DM and 25% for T2DM) but comparable to the UK (21%). Ambulance use and A&E treatment are similar in Spain and Denmark, but are slightly higher in the UK. The percentage of diabetes specialist visits following a severe episode treated in hospital is also higher in the UK and Denmark than in Spain. In terms of costs, ambulance, A&E, and hospitalization are all more expensive in Spain than in the UK and Denmark (when costs are converted into Spanish EUR). The cost of a diabetes specialist visit is comparable in Spain, the UK, and Denmark.

For non-severe episodes, the percentage of GP visits is higher in Spain than in Denmark, but is lower in Spain than the UK for T1DM and comparable to the UK for T2DM. The number of extra SMBG tests used following a hypoglycemic episode is similar in Spain and the UK, but lower in Denmark. Costs for a GP visit were also similar in Spain and the UK but lower in Denmark, and the cost of SMBG tests was similar in Spain and the UK but higher in Denmark.

The cost of severe hypoglycemia per 100,000 general population is nearly 20-fold higher in Spain than in Denmark, but is comparable to the UK (although slightly higher). The substantially higher cost in Spain than in Denmark is due to not only a higher cost per episode but also a higher prevalence of diabetes (7.8% vs. 5.7%) and higher rates of severe hypoglycemia for both patients with T1DM (0.9 vs. 0.6 episodes per year) and patients with T2DM (0.32 vs. 0.1 episodes per year). The cost per severe episode and the diabetes prevalence are both higher in Spain than in the UK (prevalence 7.8% vs. 6.0%), but this is offset by a lower severe hypoglycemia rate for both patients with T1DM (0.9 vs. 3.2 episodes per year) and patients with T2DM (0.32 vs. 0.7 episodes per year).

The cost of non-severe hypoglycemia per 100,000 is higher in Spain than in both Denmark (approximately 6-fold) and the UK (approximately 3-fold). The rate of non-severe hypoglycemia is similar in Spain and Denmark for both patients with T1DM (88.0 vs. 98.6 episodes per year) and patients with T2DM (27.2 vs. 27.4 episodes per year), but the diabetes prevalence and cost per non-severe episode are both higher in Spain. For the UK the cost per non-severe episode is lower in Spain for patients with T1DM and comparable for patients with T2DM, and the higher overall cost is due to a higher diabetes prevalence in Spain and a much higher non-severe hypoglycemia rate for both patients with T1DM (88.0 vs. 29.0 episodes per year) and patients with T2DM (27.2 vs. 10.2 episodes per year).

These differences in costs combine to mean that the total cost of hypoglycemia per 100,000 general population in Spain is calculated as 9.00 times the cost in Denmark and 2.03 times the cost in the UK, and is a substantial cost for people with insulin-treated diabetes that should not be ignored.