Cost-effectiveness analysis for HbA1c test intervals to screen patients with type 2 diabetes based on risk stratification

The number of patients with type 2 diabetes (T2DM) is expected to be close to 600 million worldwide by 2030. The HbA1c test to detect new cases of T2DM in a healthy population is one of the most effective ways to prevent and start early treatment for type 2 diabetes; however, the best frequency for HbA1c testing remains unclear. Although multiple studies have tried to investigate the optimal frequency for HbA1c testing, current guidelines about test frequencies to screen type 2 diabetes patients in a healthy population still rely on expert opinion.

Kahn et al. [1] suggested that 30- to 45-year-old people are required to have 3- to 5-year interval screening on the basis of cost effectiveness analysis. However, this study does not consider the possibility of an incidence rate for type 2 diabetes based on risk factors. Multiple studies have reported that subjects with obesity are more likely to have type 2 diabetes than people with a normal BMI [2‐4]. In addition to body mass index level, age also matters in the development of type 2 diabetes. The prevalence of type 2 diabetes increases as people live longer, and the complication rate among older adults with type 2 diabetes is significantly higher for both acute and chronic microvascular and cardiovascular diseases than that among younger adults with type 2 diabetes. People with a higher risk of onset of type 2 diabetes seem to require screening at shorter intervals than those with a lower risk [5, 6]. Previously, we reported and recommended that subjects with obesity aged 30–44 should be screened every 2 years, while those with a normal body mass index in the same age group would not need screening for the next 10 years because the paces of HbA1c progression are different based on age and BMI [6]. These results considered only HbA1c test characteristics and did not include economic impact in determining the optimal interval for HbA1c screening. People whose screening results were false positive may receive unnecessary treatment, and people whose screening results were false negative may incur higher treatment costs. When optimal screening intervals are introduced based on patient risk stratification, it would be possible to eliminate unnecessary tests as well as to minimize the chance of failing to detect affected patients.

Our study aimed to determine the optimal HbA1c test interval strategy to detect new type 2 diabetes mellitus (T2DM) cases in a healthy population stratified by age and body mass index (BMI), considering HbA1c test characteristics and cost effectiveness.

We built a state transition model of screening results and type 2 diabetes disease progression to simulate lifetime diabetes-related health care costs and QALYs. Our target population comprised individuals who had no history of T2DM or cardiovascular events. We stratified the population into age categories of 30–44, 45–59 and 60–74 years old. For each age category, we also stratified patients into BMI status as follows: underweight, BMI < 18.5; normal weight, BMI 18.6–24.9; overweight, BMI 25–29.9; and obesity, BMI ≥30 [7, 8]. In total, 12 stratifications were used to create the state transition models. We used the same tree structure for all 12 models with different parameter values. The tree structures in the state transition model comprise three main branches (Fig. 1), “screening results by HbA1c test”, “no screening year” and “T2DM progression”. Screening result branches consist of parameters: incidence of T2DM and sensitivity and specificity of HbA1c testing to determine how many individuals from the population go to the T2DM progression branch. In the no screening year branch, people are classified as the condition that was determined in the screening results branch. For example, if a person was categorized as false negative, they had a higher complication rate while they were in the no screening year because they technically missed a chance for early detection and treatment. In the T2DM progression branch, people die or may experience complications based on each relative risk. A first-order Monte Carlo simulation (microsimulation) of a hypothetical cohort of 50,000 people was performed to estimate the lifetime expected costs and expected QALY. The cycle length of the model was set to 1 year. A willingness-to-pay (WTP) threshold of 50,000 USD per QALY gained was used as the acceptable level for ICER, and an annual discount rate of 2% was used to calculate both costs and benefits, following the current guidelines [9]. The incremental cost-effectiveness ratio (ICER) was estimated as an indicator of cost effectiveness of the test interval using the formula: ICER = (Cost _{interval_a} – Cost _{interval_b})/(QALYd _{interval_a} – QALY _{interval_b}). The TreeAge Pro 2016 (TreeAge Software, Williamstown, MA, USA) was used for model construction and analyses.

Table 3 shows each QALY, incremental QALY, cost, incremental cost and ICER for 12 stratified groups based on age and BMI level. The HbA1c test interval to detect T2DM in a healthy population varies by age and BMI. Three-year intervals were the most cost effective in obesity at all ages —30-44: $15,034/QALY, 45–59: $11,849/QALY, 60–74: $8685/QALY—compared with the other two interval strategies. The three-year interval was also the most cost effective in the 60–74-year-old groups—underweight: $11,377/QALY, normal: $18,123/QALY, overweight: $12,537/QALY—and overweight in the 45–59-year-old group; dominant. In other groups, the screening interval for detecting T2DM was found to be longer than 3 years, as previously reported. Annual screenings dominated in many groups with low BMI and in younger age groups. Suggested screening strategies for all groups are shown in the Table 4. According to PSA, the results were consistent with the basic analysis (Additional file 1). Based on the probability distribution of ICER, the QALY does not show much difference among any groups.

According to our cost-effectiveness analysis results, T2DM screening programs for healthy populations should consider risk stratification for T2DM. The optimal interval seems to vary from 3 to 10 years, and some groups do not require the 3-year screening interval that current guidelines suggest. There were no groups that warrant annual screening, which is required by Japanese law.

As we reported previously, HbA1c progression in the high-risk group was remarkable, while HbA1c stayed at the same level or plateaued in the low-risk group. HbA1c progression in the lower risk group in the short interval is often affected by noise, which is originally possessed by HbA1c [6]. When deciding the appropriate screening testing frequencies, we should consider how well a screening program can distinguish true type 2 diabetes patients from true nondiabetic patients with the lowest financial burden to public health. To achieve this, health policies need to simultaneously integrate all influences according to patients’ risk factors, HbA1c test characteristics and financial impact.

Annual intervals dominated for all age groups of underweight and normal BMI individuals and the 30–44-year-old overweight group. In other groups, annual intervals did not dominated; however, annual intervals were not warranted based on the ICER.

A 3–5 year monitoring interval suggested by Kahn et al. [1] while providing good evidence for cost effectiveness, does not apply to those over 45 years old. Chen et al. [21] concluded that a 5-year interval for all age groups would be the most cost-effective strategy. Interestingly, Hoerger et al. [22] concluded that the most cost-effective strategy is targeted screening of 55- to 75-year-old patients with hypertension. Moreover, Brateanu et al. [5] suggested that the optimal interval for type 2 diabetes screening should be decided by patients’ risk score regarding the cause of type 2 diabetes; they concluded that patients in the highest risk group could be rescreened after 8 months, while those in the intermediate and lowest risk categories could be rescreened between 3 and 5 years. Our study also supports that risk stratification should be considered when deciding the optimal interval for type 2 diabetes. To the best of our knowledge, this study is the first to investigate the best interval for T2DM screening, considering both test characteristics and cost effectiveness. We have previously suggested optimal HbA1c test screening results based on risk stratification of patients and test characteristics of the HbA1c test. This time, we found that the results could change when we integrated economic evaluation as well as patient risk stratification and HbA1c test characteristics. Patient risk stratification, HbA1c test characteristics and economic impact on public health not only should be used as the basis of health policy but also should be integrated and evaluated to maximize the effectiveness of mass screening.

As in many other studies, this study also includes uncertainties in our models. All patients will receive treatment once the screening test results are positive. We estimated that patients with T2DM will live if appropriate treatment is provided based on a previous report. In the real world, there should be a certain number of patients who do not visit clinics for treatment even after they receive positive results by screening tests. We estimated a lower complication rate if patients received treatment (true positive) and a higher complication and mortality rate if patients did not receive treatment (false negative). It was impossible to obtain published data for patients who did not receive treatment; thus, we had to retrieve data from the report comparing new treatment regimens versus conventional treatment regimens [17]. The intention of our study is not to identify the best risk stratification strategy. We stratified patients by age and current BMI level; however, there is room for consideration that BMI level in childhood should be applied instead of current BMI level, which has been reported to be more important for predicting the cause of new type 2 diabetes in adulthood [23]. We also did not consider due to data inaccessibility data for family history and patient medical history such as presence of prediabetes, hyperlipidemia and hypertension, which could be the candidate factors for robust stratification.

Annual screening to detect T2DM was not cost effective and should not be used for any population. The three-year screening interval was optimal for all elderly populations, the obesity at all ages and the overweight 45–59-year-old group. Among low BMI and younger age groups, the optimal HbA1c test interval can be longer than 3 years.