Article Text

Original research
Cost-effectiveness of behavioural counselling intervention compared with non-intervention for adult patients with metabolic syndrome to prevent cardiovascular diseases and type 2 diabetes in Japan: a microsimulation modelling study
  1. Yoko Akune1,
  2. Hisataka Anezaki2,
  3. Yoko M Nakao3,
  4. Rei Goto1,4
  1. 1Graduate School of Health Management, Keio University, Tokyo, Japan
  2. 2Graduate School of medicine, Kobe University, Kobe, Japan
  3. 3Graduate School of Medicine and Public Health, Kyoto University, Kyoto, Japan
  4. 4Graduate School of Business Administration, Keio University, Tokyo, Japan
  1. Correspondence to Dr Yoko Akune; yoko.akune{at}


Objectives Nationwide lifestyle intervention—specific health guidance (SHG) in Japan—employs counselling and education to change unhealthy behaviours that contribute to metabolic syndrome, especially obesity or abdominal obesity. We aimed to perform a model-based economic evaluation of SHG in a low participation rate setting.

Design A hypothetical population, comprised 50 000 Japanese aged 40 years who met the criteria of the SHG, used a microsimulation using the Markov model to evaluate SHG’s cost-effectiveness compared with non-SHG. This hypothetical population was simulated over a 35-year time horizon.

Setting SHG is conducted annually by all Japanese insurers.

Outcome measures Model parameters, such as costs and health outcomes (including quality-adjusted life-years, QALYs), were based on existing literature. Incremental cost-effectiveness ratios were estimated from the healthcare payer’s perspective. Deterministic and probabilistic sensitivity analyses (PSA) were conducted to evaluate the uncertainty around the model input parameters.

Results The simulation revealed that the total costs per person in the SHG group decreased by JPY53 014 (US$480) compared with that in the non-SHG group, and the QALYs increased by 0.044, wherein SHG was considered the dominant strategy despite the low participation rates. PSA indicated that the credibility intervals (2.5th–97.5th percentile) of the incremental costs and the incremental QALYs with the SHG group compared with the non-SHG group were −JPY687 376 to JPY85 197 (−US$6226 to US$772) and −0.009 to 0.350 QALYs, respectively. Each scenario analysis indicated that programmes for improving both blood pressure and blood glucose levels among other risk factors for metabolic syndrome are essential for improving cost-effectiveness.

Conclusions This study suggests that even small effects of counselling and education on behavioural modification may lead to the prevention of acute life-threatening events and chronic diseases, in addition to the reduction of medication resulting from metabolic syndrome, which results in cost savings.

  • Obesity

Data availability statement

No data are available.

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See:

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


  • A unique microsimulation model was developed for a cost-effectiveness analysis of specific health guidance to prevent the development of cardiovascular diseases, diabetes and chronic kidney diseases in Japan.

  • The model was used to evaluate the economic evaluation of the specific health guidance and the impacts of changes in each individual risk factor such as blood pressure on cost-effectiveness.

  • The impact of participation rates on cost-effectiveness was also evaluated.

  • Model input parameters were set mainly based on literature data.


Metabolic syndrome (MetS) is characterised by a combination of the following disorders: abdominal obesity, high blood pressure, dyslipidaemia and/or high blood glucose; these disorders increase the risk of developing cardiovascular diseases (CVDs) and type 2 diabetes mellitus (DM).1 2 MetS is also associated with an elevated risk of cancer and chronic kidney disease (CKD).3 4 Worldwide, several definitions of MetS prevail, with the two most popular coming from the National Cholesterol Education Programme Adult Treatment Panel III5 and the International Diabetes Federation.6 The Japanese MetS definition is similar to that of International Diabetes Federation but with differences in cutoffs for abdominal circumference, fasting glucose values and high-density lipoprotein cholesterol (HDLC) levels.7 Under the Japanese criteria, national MetS prevalence is 13.3%–18.9% for men and 1.5%–4.8% for women.8

Background risk factors for MetS development include unhealthy lifestyle behaviours.9 Evidence supports the clinical benefit of behavioural modifications such as diet and exercise.10 Counselling and education help engage people in evidence-based behavioural modifications. In Japan, universal health check-ups for the general adult population aged 40–74 years have been conducted since 2008.11 The universal health check-ups are conducted by each insurer, and participants in the universal health check-ups who are diagnosed with MetS, including preliminary MetS, are eligible for specific health guidance (SHG). All insurers are required to conduct universal health check-ups and SHG annually under the Japanese government. SHG aims to prevent the development of CVD and DM and reduce high blood pressure, dyslipidaemia, and high blood glucose in the population through counselling and education aimed at changing unhealthy behaviours that contribute to obesity or abdominal obesity. The SHG consists of counselling by physicians, public health nurses or dietitians and is provided to eligible individuals after the universal health check-ups. Once the population has participated in SHG, they can participate in subsequent SHGs as long as they are eligible for SHG. Clinically, SHG was reported to improve several MetS risk factors12–14; however, the improvement effects were small, as with other behavioural counselling interventions.15 Additionally, the reported SHG participation rates were low, between 15% and 35%.16 Current data on its low health improvement effects and participation rates casts doubt on the efficient balance between economic inputs and outputs.17

Model-based economic evaluations help confirm the long-term effects of behavioural counselling interventions in terms of health outcomes and costs.18 19 To the best of our knowledge, the only cost-effectiveness analysis of SHG using model simulation was reported in the Health Labour Sciences Research Grant report.20 This report suggested that SHG was cost saving compared with no SHG in a 12-year simulation. However, it focused on CVD only and did not consider the reduction in the number of patients with conditions requiring medication (such as dyslipidaemia), the programme participation rate and the extent to which this affects SHG’s cost-effectiveness in Japan. We modelled SHG’s cost-effectiveness for 40-year-old Japanese adults with a 35-year time horizon from the healthcare payer perspective and further evaluated the cost-effectiveness impacts of (1) disease prevention of acute life-threatening events (myocardial infarction (MI) and stroke) and chronic diseases (DM and CKD); (2) reductions in the number of patients with medication owing to high blood pressure, dyslipidaemia and high blood glucose and (3) low intervention participation rates.


We performed a cost-effectiveness analysis based on literature data. The model and reporting follow the Consolidated Health Economic Evaluation Reporting Standards.21

Patient and public involvement

Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans for this study.

Specific health guidance

SHG helps the targeted individual assess their lifestyle habits and promote health behaviours to achieve their personal goals.14 22 There are two types of SHG—motivational and intensive health guidance. In the former, as a rule, one counselling session (initial counselling) is provided, and the latter consists of initial counselling and continuous counselling for more than 3 months. The time and number of counselling sessions per year are determined based on participants’ risk factors (online supplemental appendix 1 and table S1).

The effects of SHG on each risk factor in this model were based on the results of Nakao et al,12 who reported the effect of SHG on risk factors using 3-year nationwide observational data. The following literature-based model assumptions were considered: (1) The effect of motivational health guidance was assumed to be equivalent to that of intensive health guidance because the effects of the two were not reported separately. (2) The effects of SHG on continuous medication, prevention of smoking and reduced alcohol consumption were not considered. (3) The improvement effect on risk factors was considered both with and without SHG because information on their health status was provided to both groups. Additionally, the SHG effects were set to decay because the improvement achieved by SHG is known to be attenuated in the long term.23 Rates of participation, improvement for each risk factor and decay are provided in table 1 and online supplemental table S2.

Table 1

Rate of change in each risk factor with or without SHG

Microsimulation model

A microsimulation approach model was developed to estimate costs and health outcomes in the SHG group compared with the no-guidance group in a hypothetical population of 50 000 Japanese individuals aged 40 years who fulfilled the SHG criteria. The hypothetical population’s characteristics were set based on Nakao et al.12 The observational study included participants aged 40–74 years, with participants predominantly in their 40s–50s. In fact, 40% of the propensity score-matched population were in their 40s. We used the mean values of risk factors reported in the observational study, assuming that they were those of 40 years in our model. The proportion of men in the hypothetical population was also set at 78%, based on the baseline proportion of participants in the observational study. We followed each hypothetical participant until death or the age of 74 years, the oldest age for which SHG data were available.

The model includes the following four diseases: MI, stroke, DM and CKD. Furthermore, we considered three DM complications: CKD, diabetic foot ulcer and diabetic retinopathy. The risks of the four diseases were calculated using risk prediction tools validated for Japanese individuals. MI and stroke risks were determined using the Suita CVD risk model based on the Suita study.24 This prediction tool is widely known in Japan, including its use in Japan Atherosclerosis Society guidelines for preventing atherosclerotic cardiovasuclar diseases.25 The type 2 DM risk model was based on the Japan Epidemiology Collaboration on Occupational Health Study.26 CKD risk equation was developed based on observational data from annual universal health check-ups in a specific prefecture in Japan.27 For stroke, a risk prediction tool based on the Japan Public Health Center-based Prospective Study (JPHC study)28 was also available in addition to the Suita study. We used the Suita risk prediction tool because the age range of the epidemiological data used to develop the tool was 30–79 years, which was wider than that of the JPHC study (40–69 years). Other than the above risk prediction tools for MI, DM and CKD, none of the risk prediction tools disclosed the coefficients for each risk factor in the risk prediction equation for the Japanese population. Each equation was composed of various risk factors, such as the body mass index (BMI), weight circumference, blood pressure (systolic blood pressure (SBP) and diastolic blood pressure (DBP)), lipid profile (HDLC, low-density lipoprotein cholesterol (LDLC) and non-HDLC), fasting blood glucose (FBG), haemoglobin A1c (HbA1c) and estimated glomerular filtration rate (eGFR) (online supplemental appendix S2, table S3). The risk factors such as total cholesterol, eGFR or frequency of drinking, which were not reported in the observational study,12 were set based on open data from surveys of the general Japanese population (online supplemental appendix S3, table S4). In addition, non-HDLC and LDLC were not reported in the observational study,12 and thus these risk factors were calculated using HDLC, TG and total cholesterol (online supplemental appendix S3). The microsimulation was performed using TreeAge Pro 2020, with an annual cycle. The model structure is presented in online supplemental figure S1 and the comparator is described in online supplemental appendix S4.

Medications and risk factors after age 41

Participants in the initial hypothetical population (ie, age 40) were not medicated for high blood pressure, dyslipidaemia and high blood glucose because the target population fulfilled the SHG criteria, which does not target the population under medication. As the population may receive medication after initial guidance, we considered opportunistic screening for these three disorders per the medication criteria in Japan (online supplemental table S5). The rates of change in each risk factor under medication and the drop-out rates of medication were set based on the literature (online supplemental table S5).

After age 41, the risk factors were modified with age in this model. Considering that there are no data on the population eligible for SHG at age 40 followed until age 74, the rates of change in each risk factor with age (online supplemental table S6) were calculated using the mean values of risk factors by age for the population that receives universal health check-ups. The risk factor for each age was calculated by multiplying the risk factor at age 40 by the rate.

Health state transition probabilities

The risks of MI, stroke, DM and CKD were calculated using each risk prediction tool (online supplemental appendix S2), whereas the other transition probabilities between health states in the model were set based on literature reporting Japanese data (online supplemental tables S7–S8). The disease-wise prognosis is illustrated in online supplemental figures S2–S7. Among the four diseases included in the model, MI and stroke were set to be exclusive, that is, an individual with MI would not have a stroke and vice versa. In contrast, DM (including complications), CKD, high blood pressure and dyslipidaemia with high prevalence were not set to be exclusive. For example, patients with DM could have comorbidities such as MI or stroke, CKD, high blood pressure and dyslipidaemia in this model.

Health benefits

The SHG outcomes were measured using quality-adjusted life-years (QALYs), which were aggregated time spent in each state weighted by the health utility of that state. The health utility values were obtained from the literature29–31 reported for the Japanese population with the disease(s) covered by this model (online supplemental table S9). The associated disutility was considered when MI and stroke occurred.32 For example, in a first-ever stroke, only disutility was applied in the year of the event. Utility scores associated with the modified Rankin Scale (mRSs) were applied in subsequent years. For recurrent stroke, disutility was subtracted from the utility score associated with the mRS determined for the first-ever stroke. For an individual with several diseases, the lowest utility score was applied.33 34

In addition to QALYs, we estimated other health benefits of SHG: prevention rate of disease development and a change in the proportion of patients receiving medication.


From the healthcare payer perspective, we included SHG costs and direct medical costs per person but did not consider productivity loss and time costs to receive SHG and medical treatment. The cost per participant for motivational and intensive health guidance was set at JPY8000 (US$72) and JPY25 000 (US$226), respectively, based on the literature,20 although the effects of both were assumed to be the same. SHG is performed at medical providers, for example, medical facilities, contracted by insurers. SHG costs per participant represent the outsourcing costs under those contracts. The SHG costs of the literature,20 referenced in this study, were based on the national average of SHG outsourcing costs which were provided by the Ministry of Health, Labour and Welfare. The details of the outsourcing cost were not available. In addition, SHG costs may be reduced depending on the contracts when participants do not complete all counselling programmes in the SHG. We did not consider the changes in SHG costs due to SHG discontinuation in this study because details of the contracts and discontinuation rates were not known. We set direct medical costs based on the literature (online supplemental appendix S5, table S10). The direct medical costs in the literature35–38 were used without conversion to the reference year (2018) because the costs in the literature were the mean values of aggregated costs over the study period, and costs for each year were not available. In addition, the rates of reimbursement revision which is used to convert the costs varied from −1.36% to 1.55% (the maximum range of costs reported in the literature (online supplemental table S11),39 with small fluctuations. All costs were reported in JPY and 2018 US dollars (US$1=JPY110.4).40

Cost-effectiveness, sensitivity and scenario analyses

To evaluate SHG’s cost-effectiveness compared with no guidance, we estimated the ratio of incremental cost to incremental QALY gained (ie, incremental cost-effectiveness ratio, ICER). Costs and QALYs were discounted at 2% based on Japan’s economic evaluation guidelines.41 We considered the intervention to be cost-effective at a willingness-to-pay (WTP) threshold of JPY5 million/QALY (US$45 280/QALY).42 Furthermore, we evaluated the uncertainty pertaining to model input parameters using deterministic sensitivity analysis (DSA) and PSA (see online supplemental appendix S6). Moreover, the SHG participation rate varied from 10% to 100%, irrespective of sex and age, when assessing the participation rate’s impact on cost-effectiveness.

Four scenario analyses (scenarios A–D) were performed. First, we used the improvement effects on risk factors by SHG in the Fukuma et al study (BMI-only scenario, scenario A),13 which reported that only BMI was improved by SHG (online supplemental table S12). The estimation of the improvement effects on risk factors by SHG in the Fukuma et al study was based on SHG data for some insurers, whereas the study by Nakao et al12 used for the base-case analysis was based on SHG data for all insurers in Japan. These two studies were retrospective. The former was adjusted for confounding by quasi-experiments with regression discontinuity designs, whereas the latter was adjusted for confounding by propensity score matching with observed variables. There may be a bias due to unobserved variables in the effects of SHG. Second, the improvement rate for each SHG group risk factor was assumed to be equivalent to that of the without SHG group to evaluate the contributory effect of each risk factor’s improvement on the cost-effectiveness of SHG (scenario B). Third, to estimate the impact of sex differences on the cost-effectiveness analysis, we simulated men-only and women-only conditions in scenario C, respectively. Finally, to assess the impact of the starting age of the model simulation on the cost-effectiveness analysis, we conducted a scenario analysis (scenario D) with the simulation starting at age 50. Scenario D simulated 25 years from age 50. The values of the risk factors at the starting age in scenario D were assumed to be the same as in the base-case analysis setting.


The mortality rates of MI and stroke and the number of patients with DM and/or CKD were estimated for model validation. We used epidemiological data from the Japanese general population for validation because the data that continuously followed participants in SHG for a long term were not available. We did not observe major deviations, although the model results were not consistent with the epidemiological data (online supplemental appendix S7, figures S8–S9).


Base-case analysis

A hypothetical population that met the SHG criteria was simulated from 40 to 74 years of age. The cumulative total costs and QALYs gained per person with SHG were JPY3 013 990 (US$27 301) and 22.434 QALYs, respectively, whereas those without SHG were JPY3 067 004 (US$27 781) and 22.390 QALYs, respectively. The cost saving by SHG was JPY53 014 (US$480), and the incremental QALY gained was 0.044, indicating that SHG was a dominant strategy (table 2). The cumulative costs for each disease and each medication are shown in online supplemental table S13.

Table 2

Results of base-case analysis

Cumulative total costs comprised the SHG cost and direct medical costs, such as disease-related and medication costs. The cumulative SHG cost was JPY87 401 (US$792) per person. The disease-related and medication cost savings by SHG were JPY40 991 (US$371) and JPY99 424 (US$901) per person, respectively (table 2). The trajectory of incremental costs was positive up to age 65 but became negative after age 66 (online supplemental figure S10a), although the uncertainty around the incremental costs was large according to the PSA (see the Sensitivity analyses section). The trajectory of SHG costs showed that the increases slowed down after age 65, while the discrepancy between the disease-related and medication costs with SHG and without SHG increased after about age 63 (online supplemental figure S10b).

The cumulative number of patients with MI, stroke, DM and CKD with SHG compared with those without SHG was reduced by 1.2%, 2.0%, 1.4% and 0.1%, respectively. The number of patients receiving medication for high blood pressure, dyslipidaemia or high blood glucose was also reduced by 2.0%, −0.4% and 1.3%, respectively.

Sensitivity analyses

A one-way deterministic analysis was conducted to evaluate the input parameters’ uncertainty. The base-case analysis result was most sensitive to variability in the probability of a recurrent ischaemic stroke. Among the top 25 parameters that had the greatest impact on ICER, 8 were related to CVD; the incremental QALYs were positive for all parameters; in 7 parameters only, the incremental costs were positive, and thus, these parameters differed from the base-case analysis result. However, the ICERs of these seven parameters were below the WTP threshold (online supplemental table S14). In PSA, the credibility intervals (2.5th–97.5th percentile) of the incremental costs and incremental QALYs with SHG compared with no SHG were −JPY687 376 to JPY 85 197 (−US$6226 to USD 772) and −0.009 to 0.350, respectively. The probability of SHG providing cost savings compared with without SHG was estimated at 78.8%. The probability of ICER being positive and below the WTP threshold (JPY5 million/QALY, US$45 280/QALY) with SHG was 12.8% compared with without SHG (figure 1).

Figure 1

Results of probabilistic sensitivity analysis. (A) Incremental cost-effectiveness scatter plot for 1000 iterations. The solid line represents the willingness-to-pay threshold of JPY5 million/QALY. (B) Cost-effectiveness acceptability curve. JPY, Japanese yen; QALY, quality-adjusted life-year.

We calculated the ICER by varying the participation rate from 10% to 100% (online supplemental tables S15). The incremental QALY gains were positive for all participation rates, and the incremental costs were positive for participation rates of 50% or more. The relationship between participation rates and each cost, such as SHG, disease related and medication, is shown in online supplemental figure S11. The SHG costs increased linearly with increasing participation rates, but the disease-related and medication costs with SHG were not reduced enough to compensate for the increased SHG costs. However, the ICERs for the participation rates from 50% to 100% were JPY326 388/QALY (US$2956/QALY) to JPY1 755 111/QALY (US$15 898/QALY), and this improvement in participation rate was cost-effective based on the WTP threshold.

Scenario analyses

In scenario A, when only the change in BMI was considered as the effect of SHG, the ICER was estimated as JPY6 170 929/QALY (US$55 896/QALY; table 3), and SHG with scenario A effects was not cost-effective based on the WTP threshold. In scenario B, we set the improvement in each SHG group risk factor to that of the non-SHG group to assess the effect on the cost-effectiveness analysis (table 3). Assuming no improvement in SBP and DBP specific to the SHG group (ie, the same improvement rate in both strategies), the ICER was estimated as JPY352 197/QALY (US$3190/QALY); in FBG and HbA1c, the ICER was estimated as JPY129 128/QALY (US$1170/QALY), whereas no improvement for other risk factors changed the dominant result in the base-case analysis. Although SHGs without improvement in SBP and DBP, and in FBG and HbA1c were not cost saving, those SHGs were cost-effective based on the WTP threshold.

Table 3

Results of scenario analyses

In scenario C, the impact of sex differences on SHG resulted in cost savings for the men-only condition (JPY74 294 saving, US$673 saving) and for the women-only condition (JPY13 583 saving, US$123 saving) (online supplemental table S16). In scenario D, the ICER was JPY2 374 015/QALY (US$21 504/QALY) for the starting age of 50 (online supplemental table S16).


Our cost-effectiveness analysis revealed that the SHG in Japan may provide cost savings compared with no SHG via the reduction in the development of diseases and in the number of patients on medication(s), though its improvement on risk factors was small. The DSA revealed that the variation in all parameters was cost saving or cost-effective, and the cost-saving probability exceeded 70% in the PSA. Variations in parameters related to ischaemic stroke had a large impact, although the ICER was less than JPY5 million/QALY (US$45 280/QALY).

Our results are consistent with those of several studies that have evaluated the cost-effectiveness of behavioural counselling interventions for obesity. For example, Cecchini et al evaluated the cost-effectiveness of several behavioural counselling interventions, such as physician counselling and school-based interventions, using a microsimulation model to simulate the development of stroke, ischaemic heart disease and cancers.18 They found that a small change in risk factors such as BMI, blood pressure, cholesterol and glycaemia by physician counselling prevented the development of chronic diseases; the intervention proved to be a cost-effective strategy in Brazil, England, Mexico and Russia. Bates et al estimated the maximum justifiable cost-per-person—defined as costs that remain cost-effective (£20 000–£30 000 for the National Institute for Health and Care Excellence)—for weight loss maintenance interventions using an individual patient-level model to simulate the development and prognosis of CVD, DM, cancer, osteoarthritis and depression.19 The intervention was considered cost-effective when the maximum justifiable cost-per-person was £104.64 and £137.78 for the high BMI population. However, despite including medication costs, these studies did not address whether each cost, such as intervention, disease related and medication was greater on the ICER. In this study, a cost-effectiveness analysis was conducted to determine each cost. The results demonstrated that cost reductions owing to reduced medication use were greater than cost reductions from preventing life-threatening diseases. As high blood pressure, dyslipidaemia and high blood glucose usually precede the development of life-threatening conditions, the large cost savings in medication costs might suggest cost savings and prevention of life-threatening diseases after age 75, although longer-term evaluations are required. As presented in online supplemental figure S10a, the cost-saving effect of SHG did not emerge in the short term but did so after age 66 in the base-case analysis. This could be attributed to three factors. First, SHG participants aged 65–74 are not eligible for expensive intensive health guidance (online supplemental table S1), and thus, SHG costs slowly increased after age 65 (online supplemental figure S10b). Second, the older population is more susceptible to diseases such as MI, stroke, DM and CKD, as well as high blood pressure, dyslipidaemia and high blood glucose. This might result in the large discrepancy between the disease-related and medication costs with and without SHG in older adults (online supplemental figure S10b). Third, the proportions of people aged 40–59 receiving medication were lower than those of aged 60–74 in this study. This implied that the prevention effects of medications with SHG compared with those without SHG were unlikely to be reflected in the incremental costs for those aged 40–59.

Our scenario analyses indicated that SHG was not cost-effective for a BMI-only improvement effect. However, the improvement in blood pressure and blood glucose contributed to the cost-saving programme. Blood pressure risk factors, such as SBP, are included in all risk prediction tools considered, and the improvement in SBP exhibited a greater impact on medication costs as well as disease-related costs. Improved blood glucose impacted medication costs. In the base-case analysis, the medication-related cost with SHG decreased by approximately JPY100 000 (US$906) compared with that without SHG, whereas it was reduced to approximately JPY40 000 (US$362) without improvement in blood pressure by health guidance. Similarly, the assumption of no effect on glycaemia decreased the incremental cost related to medication by approximately JPY50 000 (US$453). Thus, behavioural counselling interventions aimed at improving BMI alone are inadequate, and highly effective programmes are needed for blood pressure and blood glucose. Note that SHG which provided no improvement on blood pressure or blood glucose was considered cost-effective based on the WTP threshold compared with no SHG. In scenario C, the impact of sex differences in both men-only and women-only conditions resulted in cost savings, but the savings were smaller for women. The calculated SHG costs (online supplemental table S16) indicated low SHG participation rates for women. This implies that efforts should be made to increase the participation rates of women who are eligible for SHG. In fact, the participation rates for women in their 40s and 50s were slightly lower than those for men (online supplemental table S2). When the participation rates for women in their 40s and 50s were the same as those for men, the cost savings for women were JPY27 039 (US$245) (data not shown). However, since our model did not account for sex differences in all parameters. For example, the utility scores did not take sex differences into account at all. Further studies based on sex-specific data are required to clarify the impacts of sex differences on the cost-effectiveness analysis. In scenario D, the starting age of the simulation had a considerable impact on the cost-effectiveness analysis. For the starting age of 50, there was only a small difference in the disease-related and medication costs with and without SHG compared with that of the based-case analysis. This might be attributed to the risk factors that change with age. Most risk factors increased with age, except for total cholesterol and eGFR. This means that most risk factors at age 50 in the base-case analysis would always be higher than the risk factors at age 50 in scenario D, and the population in the base-case analysis was higher risk than that in scenario D. The results of scenario D indicated that the cost savings in the base-case analysis might be overestimated, but the SHG was still cost-effective based on that scenario.

Low participation rates have remained an issue in SHG implementation. A high participation rate is necessary to obtain sufficient health benefits from SHG. Notably, a high participation rate increases the cost of guidance per person. Our study indicated that 15%–35% participation rates (base-case setting) resulted in cost savings, whereas SHG was cost-effective even if all participants who fulfilled the SHG criteria participated. The relationship between the participation rates and each cost (online supplemental figure S11) indicated that, in the higher participation rates, the disease-related and medication costs were not reduced with SHG enough to compensate for the increased SHG costs. This implied the importance of improving other aspects of SHGs, such as SHG cost setting and the development of programmes that increase their effectiveness, along with increasing participation rates.

This study has several limitations. First, it simulated only up to the age of 74 years—a short time horizon for the lifetime analysis. No data are available after 75 years of age, as SHG targets those aged 40–74 years. Although it is unclear whether the effects of SHG (risk factor improvement) will persist in older adults not subject to health guidance, the prevention of disease development in the 40–74 age group is likely to contribute to higher healthy life expectancy after 75 years of age by delaying disease development and severity. Therefore, the analysis of effectiveness up to 74 years of age may have underestimated the effect of SHG. Second, the modelled diseases were limited to MI, stroke, DM and CKD. For example, the relationship between MetS, including obesity and cancer, is relatively well established and has been modelled in the reported studies.18 19 43 44 With the exception of MI, stroke, DM and CKD, which were examined in this study, other MetS-related diseases, such as cancer, need to be considered in further studies. Third, the relationship between risk factors (eg, BMI and lipid profile) was not considered. This can lead to underestimations of medication effectiveness. Fourth, universal health check-ups that were combined with SHG were not modelled to evaluate the individual cost-effectiveness of SHG in this study. The participants in SHG are selected through universal health check-ups, and thus, our model does not consider the cost of screening to identify high-risk groups who meet the SHG criteria, leading to underestimation. To comprehensively evaluate screening and SHG, constructing a model that includes universal health check-ups is necessary. Fifth, the analysed population was a high-risk group for the development of diseases and was characterised by a predominantly male population. This makes it difficult to conduct an evaluation and validate the model using data from the general population. Moreover, universal health check-ups and SHG are being reviewed for more efficient and effective implementation and should continue to be evaluated for cost-effectiveness. Sixth, our cost-effectiveness analysis was based on literature data and various assumptions. The costs data reported in the literature were not converted to the reference year (2018). The conversion would result in slightly different costs. The PSA results indicated that the uncertainty in model input parameters had a significant impact on the cost-effectiveness analysis. Furthermore, online supplemental table S15 shows that the increase in QALYs was not uniform as participation rates increased. This could be attributed to stochastic uncertainty (called Monte Carlo errors). Further studies are needed to reduce uncertainty in the simulation estimates. Finally, there are limitations to simply generalising our results to other countries with different MetS prevalence and healthcare systems because this study was optimised for SHG in Japan.


This study indicated that the SHG, universal behavioural counselling intervention for MetS (and preliminary MetS) in middle-aged Japanese (40–74 years) may be a cost-saving strategy; the model simulation reflected the possibility of achieving disease prevention. The medication reduction by SHG also significantly affected cost-effectiveness. Although differences in the prevalence of MetS and healthcare systems in Japan and elsewhere preclude simple generalisations, this study’s findings and model structure can contribute to developing and implementing further behavioural counselling interventions.

Data availability statement

No data are available.

Ethics statements

Patient consent for publication


Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.


  • Contributors Concept and design: YA, HA, YMN and RG. Acquisition, analysis or interpretation of data: YA, HA and RG. Drafting of the manuscript: YA. Critical revision of the manuscript for important intellectual content: YA, HA, YMN and RG. Statistical analysis: YA and HA. Obtained funding: RG. Administrative, technical or material support: YA and RG. Supervision: YA, HA, YMN and RG.

  • Funding This study was supported by the Ministry of Health, Labour and Welfare (MHLW) Comprehensive Research on Lifestyle Related Diseases including Cardiovascular Diseases and Diabetes Mellitus Program Grant Number JPMH22FA1006. We acknowledge the support by National Institute of Public Health for the article publishing charge.

  • Competing interests YMN reports a study grant from Bayer, outside the submitted work. No other disclosures are applicable.

  • Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.