Cost-Effectiveness of Blood Glucose Test Strips in the Management of Adult Patients with Diabetes Mellitus (2024)

Introduction

Diabetes mellitus is a chronic disease characterized by the body’s inability to produce sufficient insulin and/or properly use insulin.1 Worldwide, an estimated 285 million people have diabetes, and this number is projected to increase to 438 million by 2030.2 In 2004/2005, approximately 1.8 million (5.5%) Canadians aged 20 years and older had diagnosed diabetes.1 However, it is estimated that 2.8% of the general adult population has undiagnosed type 2 diabetes mellitus,7 and the true prevalence of diabetes may approach 2.0 million.3

Approximately 10% of patients with diabetes have type 1 diabetes, which is the result of little or no insulin being produced by the body.4 The majority of patients with diabetes have type 2 diabetes, which is a metabolic disorder caused by varying degrees of insulin resistance, where the body usually produces insulin but is unable to use it properly.4 When inadequately managed, diabetes is likely to result in poor glycemic control.1 Impaired glycemic control, if prolonged, may result in diabetes-related complications such as ischemic heart disease, stroke, blindness, end-stage renal disease, and lower limb amputation.5^,6 Therefore, control of blood glucose levels to reduce a patient’s risk of developing these complications is an important component of diabetes management. Recommended approaches to improve glycemic control include lifestyle modifications such as weight control, proper nutrition, and adequate exercise; the use of medications such as insulin and oral antidiabetes drugs; and self-monitoring of blood glucose (SMBG).7

The purpose of SMBG is to collect detailed information about glucose levels across various time points each day and take appropriate action should those levels be outside the desired range.8^,9 SMBG requires that patients prick their fingers with a lancet device to obtain a small blood sample.8^,9 The blood is applied to a reagent strip or blood glucose test strip, and glucose concentration is determined by inserting the strip into a reflectance photometer or an electrochemical sensor.8 Results, based on an automated reading, are available from the photometer within five to 30 seconds.8 The results can be stored in a glucose meter’s electronic memory or recorded in the patient’s logbook. It has been suggested that patients can adjust food intake, physical activity, and pharmacotherapy in response to their blood glucose readings to better maintain optimal glycemic control on a day-to-day basis.8^,9

Despite widespread use, there is controversy regarding the benefits of SMBG, especially in patients with type 2 diabetes mellitus not using insulin.10^–13 In addition, the optimum frequency of testing has not been defined in any population.7^,14 Clinical and economic evidence is needed to inform the optimal prescribing and use of SMBG. As the prevalence of diabetes in Canada and rates of self-monitoring increase, the costs associated with SMBG are also rising.8 In 2005/2006, the Nova Scotia Seniors’ Pharmacare Program spent $4 million on blood glucose test strips, approximately 60% of which was spent on beneficiaries who were not using insulin agents.15 In Saskatchewan, of the $6.5 million spent on diabetic testing supplies in 2001 (most of it on blood glucose test strips), approximately half was for people who were not using insulin agents.16 Evidence relating to the optimal prescribing and use of SMBG may assist policy decision-makers, consumers, and health care providers make informed decisions for patients with diabetes.

Objective

The objective of this assessment is to assess the cost-effectiveness of SMBG in patients with type 2 diabetes.

Methods

Before conducting this economic analysis, a systematic review17 was conducted to identify primary studies that compared SMBG with no SMBG, or different frequencies of SMBG, in patients with diabetes. Results from this systematic review are presented in detail elsewhere.17 In general, this systematic review elicited few studies that explored the effect of SMBG in patients with insulin-treated diabetes. However, the evidence was more robust for patients with non–insulin-treated type 2 diabetes.

The current analysis reports cost-effectiveness results for patients with:

• non–insulin-treated type 2 diabetes

• insulin-treated type 2 diabetes.

An incremental cost-utility analysis of SMBG, compared with not performing SMBG, was conducted using the United Kingdom Prospective Diabetes Study or UKPDS Outcomes Model.18 The UKPDS Outcomes Model18 is a computer simulation model used to forecast long-term health outcomes and cost consequences in patients with type 2 diabetes. The risk of developing seven diabetes-related complications (i.e., fatal or non-fatal myocardial infarction, other ischemic heart disease, stroke, heart failure, amputation, renal failure, and blindness) is estimated based on data from 3,642 patients with type 2 diabetes who were enrolled in the UKPDS. Each equation18 estimates the absolute risk of developing a complication, based on patient characteristics (e.g., age and sex, glycosylated hemoglobin [A1C], systolic blood pressure, cholesterol, body mass index, smoking history, history of diabetes-related complications). Model projections have been validated against published clinical and epidemiological studies.19

This economic evaluation took the perspective of a Canadian publicly funded ministry of health.20 Patients with type 2 diabetes who were not using insulin therapy were assumed to use 1.29 test strips per day based on a recent utilization study. Unit costs for blood glucose test strips were obtained from the Ontario Drug Benefit Plan Formulary/Comparative Drug Index, July 28, 2008.21 For the reference case, a unit cost of $0.73 per test strip was used, and a $7.00 dispensing fee was applied for every 100 test strips claimed.22 The cost of glucose meters and lancet devices were not incorporated, because they are often provided for free23 or covered under other diabetes funding programs.24 Sensitivity analyses were performed to test robustness of results to changes in the unit cost of blood glucose test strips. Resource utilization and costs associated with managing diabetes-related complications were obtained from the Ontario Ministry of Health and Long-Term Care.25 The average annual cost for patients without diabetes-related complications who were not performing SMBG was $1,507.25 In the reference case, both costs and quality-adjusted life-years (QALYs) were discounted at a rate of 5%, as recommended by CADTH guidelines.20 Sensitivity analyses were performed for discount rates of 0% and 3%.20

The primary outcome measure in this analysis was the QALY,26^,27 which is an outcome measure that simultaneously captures quantity and quality of life (i.e., mortality and morbidity).26^,27 Utility scores for health states in the model were obtained from a US catalogue of EuroQol-5 Dimension (EQ-5D) scores for chronic conditions.28^,29 Patients with non–insulin-dependent type 2 diabetes without a history of diabetes-related complications were assumed to have an EQ-5D score of 0.753.28^,29 Disutilities for diabetes-related complications, during the year of the event, were based on EQ-5D scores for relevant International Classification of Diseases, 9th edition codes.28 For subsequent years, disutilities were based on quality priority conditions estimates, which correspond to individuals who have ever been diagnosed with a particular disease.28^,29

Results

Based on the observed difference in A1C of −0.25% across seven RCTs comparing SMBG with no SMBG in patients not using insulin, the economic model projected that patients using SMBG would have a slightly lower cumulative incidence of diabetes-related complications over a 40-year period. Absolute risk reductions for diabetes-related complications, which represent differences in cumulative incidence rates between cohorts, ranged from 0.08% to 0.40%. Consequently, the number of patients who would need to be treated with SMBG, relative to no SMBG, to avert one diabetes-related complication over a 40-year period ranged from 228 to 1,299, depending on the outcome.

For patients with type 2 diabetes who are not using insulin, use of SMBG generated an additional 0.02385 QALY and increased costs by $2,711, resulting in an incremental cost-utility ratio (ICUR) of $113,643 per QALY gained. According to the sensitivity analyses, ICURs ranged from $189,376 to $47,512 per QALY when A1C treatment effects were varied between −0.15% (lower limit of 95% confidence interval from meta-analysis of seven RCTs) and −0.57% (estimate from an observational study30). ICURs ranged from $6,322 to $152,095 per QALY when mean testing frequency of SMBG was varied between one and 12 strips per week. When the cost of blood glucose test strips ($0.71 per strip) was decreased by 25%, 50%, and 75%, ICURs decreased to $86,129, $58,615, and $31,101 per QALY gained, respectively. When the cost of the least expensive test strip in the ODBP was applied, the ICUR decreased to $63,892 per QALY. In contrast, ICURs increased to $123,143 per QALY when an alternative price per strip from other publicly funded drug plans in Canada was used in the analysis.

There was limited clinical evidence on which to model the cost-effectiveness of SMBG in patients with type 2 diabetes using insulin. Therefore, cost-effectiveness estimates were calculated for several SMBG frequencies and over a plausible range of A1C differences between SMBG users and non-users. When an A1C difference of 1% was assumed, SMBG frequencies of up to 14 per week appeared to be cost-effective (e.g., ICUR for 14 tests per week was $33,412 per QALY gained). For a testing frequency of 21 per week to have an ICUR of less than $50,000 per QALY gained, the A1C difference compared with no SMBG would have to be between 1% and 1.5%.

Limitations

The results of this analysis are limited by the quality of the available clinical evidence, particularly in patients with insulin-treated type 2 diabetes. Consequently, results should be interpreted with caution.

Another limitation is that the model used A1C, a surrogate end point, to project the occurrence of long-term diabetes-related complications. Although the validity of surrogate outcomes continues to be debated in the literature,31^–34 A1C is routinely used in clinical practice as an indicator of treatment success35 and is therefore the best estimate of efficacy for these analyses in the absence of data on long-term complications of diabetes.

A number of morbidities and intermediate states are not included in the UKPDS Outcomes Model.36 However, this limitation should not be overstated, since very small absolute risk differences have been reported for these morbidities in a previous economic analysis of SMBG.37 In the future, post-monitoring of data from the UKPDS Outcomes Model36 will provide additional data for these intermediate states, at which time a reassessment of the economic evaluation of SMBG in patients with type 2 diabetes may be warranted.

There is a lack of clinical data demonstrating that SMBG decreases the incidence of hypoglycemia and, in particular, severe episodes. Moreover, the UKPDS Outcomes Model36 cannot accommodate hypoglycemia. Therefore, this analysis does not incorporate benefits that may be incurred because of decreased hypoglycemia risk. Hypoglycemic episodes, however, are rare in patients with type 2 diabetes not using insulin;38 consequently, this limitation should not significantly alter cost-effectiveness estimates for this population. For patients with insulin-treated type 2 diabetes, the incidence of severe hypoglycemia is greater than in non–insulin-treated patients, although it is still less frequent than in patients with type 1 diabetes.39^–41

Finally, we assumed in the reference case analysis that SMBG was not associated with a decrement in health-related quality of life (HRQoL), despite evidence from the DiGEM trial suggesting the contrary.42^,43 Thus, current estimates may be biased in favour of SMBG. Future investigators should include methodologically rigorous HRQoL assessments in their study protocols. If additional data become available and clearly demonstrate that SMBG decreases HRQoL, then a reassessment of the economic evaluation of SMBG may be warranted.

Conclusion

The strength of CADTH’s economic conclusions is limited by the quality of the available clinical evidence. Overall, the quality of evidence for patients with insulin-treated diabetes (either type 1 or type 2 diabetes) was poor. For patients with non–insulin-treated diabetes, the clinical evidence was more robust. Within the limitations of modelling and the available data, we conclude that:

• Routine use of SMBG (one or more test strips per day) in patients with non–insulin-treated type 2 diabetes is associated with an incremental cost of $113,643 per QALY gained, relative to no SMBG.

• A reduction in the price of blood glucose test strips would improve the cost-effectiveness of SMBG. For patients with insulin-treated type 2 diabetes, SMBG testing frequencies beyond 21 test strips per week require unrealistically large A1C estimates of effect to achieve favourable incremental cost per QALY estimates.

Because of conflicting evidence, the CADTH model did not incorporate HRQoL data from clinical trials. Potential benefits associated with a reduction in hypoglycemia were also not incorporated in the model because of lack of adequate evidence. Therefore, more well-designed RCTs are needed to explore the impact of SMBG on HRQoL and incidence of hypoglycemia among patients with either insulin- or non–insulin-treated type 2 diabetes.

[Adapted from Canadian Agency for Drugs and Technologies in Health. Cost-Effectiveness of Blood Glucose Test Strips in the Management of Adult Patients with Diabetes Mellitus. (Optimal therapy report; vol.3 no.3). Ottawa: The Agency; 2008.]

References