Cost-effectiveness analysis of hospital treatment volume and survival outcomes in endometrial cancer in Japan
Abstract
Objective
Hospital treatment volume affects survival in patients with endometrial cancer; notably, initial treatment at high-volume centers improves survival outcomes. Our study assessed the effect of hospital treatment volume on cost-effectiveness and survival outcomes in patients with endometrial cancer in Japan.
Methods
A decision-analytic model was evaluated using the following variables and their impact on cost-effectiveness: 1) hospital treatment volume (low-, intermediate-, and high-volume centers) and 2) postoperative recurrent risk factors based on pathological findings (high- and intermediate-risk or low-risk). Data were obtained from the Japan Society of Obstetrics and Gynecology database, systematic literature searches, and the Japanese Diagnosis Procedure Combination database. Quality-adjusted life years (QALY) was used as a measure of effectiveness. The model was built from a public healthcare perspective and the impact of uncertainty was assessed using sensitivity analyses.
Results
A base-case analysis showed that the incremental cost-effectiveness ratio at high-volume centers was below a willingness-to-pay (WTP) threshold of ¥5,000,000 with a maximum of ¥3,777,830/4.28 QALY for the high- and intermediate-risk group, and ¥2,316,695/4.57 QALY for the low-risk group. Treatment at the high-volume centers showed better efficiency and cost-effectiveness in both strategies compared to intermediate- or low-volume centers. Sensitivity analyses showed that the model outcome was robust to changes in input values. With the WTP threshold, treatment at high-volume centers remained cost-effective in at least 73.6% and 78.2% of iterations for high- and intermediate-risk, and low-risk groups, respectively.
Conclusion
Treatment at high-volume centers is the most cost-effective strategy for guiding treatment centralization in patients with endometrial cancer.
Synopsis
The study examined the relationship between hospital treatment volume, cost-effectiveness, and survival outcomes in endometrial cancer patients in Japan. Treatment at high-volume centers was potentially the most cost-effective strategy. Treatment at intermediate- and low-volume centers falls within the acceptable range by the willingness-to-pay threshold.
INTRODUCTION
Uterine cancer is the most common cancer of the female reproductive organs worldwide. In 2020, 417,367 new cases and 97,370 deaths of endometrial cancer were projected to occur globally [1]. In Japan, endometrial cancer is the most common gynecologic malignancy, with approximately 17,880 women estimated to be newly diagnosed in 2019, and its incidence continues to increase every year [2]. Minimally invasive surgery for endometrial cancer has been covered by Japanese public health insurance since 2014, and the number of procedures has been increasing [3]. Depending on the risk factors associated with radiotherapy and chemotherapy, these treatments have also been covered by the public health insurance [4].
The public medical insurance program in Japan includes all citizens and covers standard treatments for endometrial cancer. However, healthcare costs in Japan have risen drastically over the past few decades and are expected to continue increasing. The increasing incidence of endometrial cancer, availability of new surgical techniques such as minimally invasive surgery, and new drug therapies such as molecular-targeted therapies and immune checkpoint inhibitors, are considered to increase healthcare expenditure [5]. Therefore, the cost-effectiveness of endometrial cancer treatment is important.
Previous studies showed that initial treatment at high-volume centers was associated with improved survival in patients with endometrial cancer in Japan [6, 7]. However, there remains a lack of high-quality evidence on the cost-effectiveness of the treatment of endometrial cancer in Japan and on the relationship between hospital treatment volume, survival outcomes, and healthcare costs. Therefore, the primary objective of this decision-analytic modeling study was to examine the impact of hospital treatment volume on cost-effectiveness and survival outcomes in patients with endometrial cancer stratified by recurrence risk.
MATERIALS AND METHODS
1. Study design and patient selection
The study population included patients with invasive endometrial cancer in Japan who underwent initial surgery in 2016. Patients who did not undergo surgical treatment or pathological evaluations were excluded.
2. Data sources
The Japan Society of Obstetrics and Gynecology (JSOG) database and the Diagnosis Procedure Combination/Per-Diem Payment System (DPC/PDPS) databases were used. Both databases are publicly available and were used with prior approval. The JSOG database is a national hospital-based gynecological cancer registry that records comprehensive information on the initial treatment of endometrial cancer, including the tumor characteristics, treatment types, and survival outcomes [8]. The DPC/PDPS is a bundled payment program database for inpatients supported and managed by the Japanese Ministry of Health, Labor, and Welfare that records information on the costs of endometrial cancer treatment [9]. This study was approved by the hosting Institutional Review Board; the JSOG Ethics Committee, which has the authority to approve any study conducted using the JSOG tumor registry data; and the National Cancer Center Ethics Committee, which has the authority to provide DPC/PDPS data.
3. Study definition
According to the International Federation of Gynecology and Obstetrics 2009 staging system, low-risk endometrial cancer is defined as grade 1 and 2 stage I disease with <50% myometrial invasion, whereas intermediate-risk endometrial cancer is defined as grade 1–2 stage I disease with >50% myometrial invasion. Patients with postoperative grade 3 disease or more than stage II disease were considered to be at high risk. The average annual hospital treatment volume was calculated as the number of women with endometrial cancer treated at a given institution divided by the number of years that the institution had participated in the JSOG tumor registry database. Institutions were categorized into high- (≥50 cases/year), intermediate- (20–49), and low-volume centers (≤19) based on a prior study [6].
4. Model structure
A decision-analytic model was designed to evaluate the cost-effectiveness of the three assessment strategies grouped according to hospital treatment volume (low-, intermediate-, and high-volume centers). The decision tree for each strategy is shown in Fig. 1. The model was stratified according to the use of adjuvant therapy based on postoperative risk factors (low, intermediate, or high risk). Additionally, a treatment model based on recurrence type was examined (Fig. S1). Intermediate- or high-risk treatment with initial surgical management can be followed by postoperative adjuvant therapy (chemotherapy or radiotherapy) based on JSOG guidelines [10]; for low-risk, neither adjuvant therapy was used. Most patients with recurrence in Japan are treated with systemic chemotherapy [6]. Patients with local recurrence are treated with radiotherapy, followed by systemic chemotherapy for re-recurrence.
Fig. 1
Decision tree representing the treatment pathway for EMCA.
EMCA, endometrial cancer.
A decision tree representing the clinical pathway corresponding to each diagnostic strategy and a Markov model simulated a five-year follow-up in terms of survival, health-related quality of life (QOL), and healthcare costs (Fig. 1). The possible health states in the Markov model were post-initial treatment (complete remission [CR]), recurrence, and death (Fig. 2). Patients moved between different health states in the Markov part of the model according to a set of transition probabilities (Table 1). Post-treatment health state starts with CR. Patients may experience disease recurrence, and its occurrence and consequences of which are incorporated into the post-treatment health state. Death can be either cancer-related or due to other causes. Background mortality was based on 2016 Japanese statistics on age-related deaths among women in the general population (Table S1). Patients with endometrial cancer may have comorbidities, and their life expectancy is time-adjusted depending on their risk [11]. After each one-year cycle, patients may move between health states. For each year that patients remain in particular health state, they are assigned corresponding outcomes in terms of QOL and associated costs [12, 13, 14]. The model was run for a total of 5 years, corresponding to five cycles in the model and to the average time of observation at each center in the JSOG database. Additionally, since the median age of patients diagnosed with endometrial cancer in the JSOG database is 60 years, the model age started at 60 years [6].
Fig. 2
Markov model representing health status after the initial treatment.
CR, complete remission.
Table 1
Input parameters: clinical parameters, cost estimates, and utility values
5. Parameters of probabilities, cost, and utility in the model
An overview of the input values is provided in Table 1. Among patients with endometrial cancer in the intermediate- to high-risk groups, the average probability of disease recurrence was 15%, based on previous trials [15, 16]. The corresponding recurrence rate per year in the intermediate- or high-risk groups in the three types of centers was calculated based on JSGO data and previous studies [15, 16, 17]. Similarly, the probability of disease recurrence in the low-risk group was calculated based on previous studies. The annual probability of progression to death after recurrence was calculated to be 20% based on JSGO data. The probability of remission due to other causes of death was estimated using the relative risk and the Japanese National Statistics [11, 17].
A summary of the costs and detailed explanations are provided in Table 1. Costs were based on the diagnosis treatment combination prices from the DPC/PDPS database and reference prices from the Japan Pharmacotherapy Compass [18, 19]. All costs are reported at 2016 prices and included initial surgical management following adjuvant chemotherapy and radiotherapy, annual follow-up examinations during outpatient hospital visits (including office visits), laboratory checks, and imaging tests in line with the JSOG recommendations, and recurrence treatment with systematic chemotherapy. The annual discount rate was set at 2% according to Japanese guidelines [20, 21].
Table 1 provides an overview of the utility for each health state. The effectiveness of the strategies, expressed in quality-adjusted life years (QALY), was measured as a combination of survival and QOL. QOL was calculated as a single index utility ranging from 0 to 1, representing death and perfect health, respectively [22]. Utility data were derived from previous studies, and the applicability of these data to our model was discussed with experts [12, 13]. Patients without any side effects or recurrent disease were assumed to be in perfect health and were assigned a utility value of 1. The disutility value was subtracted for each health state to determine the proportion of patients with side effects or local recurrence.
6. Statistical analysis
Costs and effects were compared between strategies, and the incremental cost-effectiveness ratio (ICER) was calculated when applicable. The ICER represents the extra cost required to gain one extra QALY from one strategy to another. A strategy is deemed cost-effective if the ICER is lower than the willingness-to-pay (WTP) for the QALY. A WTP threshold of ¥5,000,000/QALY was used, which is the lowest reference value for WTP per QALY [23]. Comparisons were made in terms of cost, effect (in QALYs), and ICERs. Several scenarios were analyzed, including (i) patients with a high-to-intermediate recurrence risk, (ii) patients with a low recurrence risk, (iii) patients with recurrence treated with systemic chemotherapy, and (iv) patients with local recurrence treated with radiotherapy.
Model development and analysis were performed in Tree Age Pro 2022 (Tree Age Software, Williamstown, MA, USA). To assess model parameter uncertainty, extensive one-way and probabilistic sensitivity analyses were performed. In the one-way sensitivity analysis, each variable or parameter was individually varied to assess its effect on the results. Probabilities and utility scores were varied according to their 95% confidence intervals or range, and costs were varied by ±20. In the probabilistic sensitivity analysis, all input variables were varied simultaneously. As suggested in the literature, probabilities were modeled as beta distributions, relative risks/odds ratios/hazard ratios as log-normal distributions, costs as gamma or lognormal distributions, and utilities (which are bounded by 0 and 1, just like probabilities) as beta or, less commonly, log normal. For the probabilistic sensitivity analysis, we obtained 1,000 estimates of the incremental costs and effects by sampling from the distributions of each variable. A cost-effectiveness acceptability curve was then plotted to show the probability of all patients being treated, with each center being cost-effective at a WTP threshold of ¥5,000,000 (rate of 34,652 USD as of July 1st, 2023). The model validity was verified using the CHEERS checklist [24]. The internal model validation was undertaken through tests of descriptive, technical, and face validities.
RESULTS
1. Base case analysis
Among patients with endometrial cancer, regardless of adjuvant therapy use (Table 2), initial treatment at high-volume centers was the most effective strategy (high-to intermediate-risk group: 4.28 expected QALYs, least cost: ¥3,777,830; low-risk group: 4.57 QALYs, least cost: ¥2,316,695), followed by treatment at intermediate-volume centers with 4.25 QALYs and ¥3,836,413 (high to intermediate-risk group) and 4.54 QALYs and ¥2,360,745 (low-risk group) at the end of the model duration. The least effective strategy was at the low-volume center, with a total of 4.21 QALYs and ¥3,892,959 (high to intermediate risk group) and 4.49 QALYs and ¥2,444,621 (low-risk group) after 5 years.
Table 2
Incremental analysis
The ICERs for both the intermediate- and low-volume center strategies stratified by recurrent risk were considered less cost-effective strategies since treatment at these centers cost more and were less effective than that in high-volume centers; this was visually identified using a cost-effectiveness graph (Fig. S1). Intermediate-volume centers had an incremental cost and effect of ¥58,582 and -0.04 for the high-to-intermediate-risk group, and ¥44,040 and -0.02 for the low-risk group, respectively. Low-volume centers had an incremental cost and effect of ¥115,129 and -0.07 for the high-to-intermediate-risk group and ¥127,296 and -0.08 for the low-risk group, respectively (Table 2).
2. Deterministic sensitivity analysis
Tornado diagrams show the impact of ICER uncertainty from multiple inputs on the model (Fig. S2). The dominance of treatment in high-volume centers was associated with the following factors: high-to intermediate-risk group, recurrence probability within 0.165, and initial treatment cost of ¥2,564,359; low-risk group: initial treatment cost of ¥1,628,567. Deterministic variations in transition probabilities, costs, utility values, and the proportion of recurrent disease in the treatment center types over a range of values assumed possible showed that these variations were associated with the results. The model was sensitive to variations in the probability of survival outcomes in high-volume centers and the costs of initial treatment associated with high- and intermediate-volume centers. However, variations in other parameters were mostly below the WTP threshold of ¥5,000,000.
3. Probabilistic sensitivity analysis
Probabilistic sensitivity analysis was performed to examine the effectiveness of the model, in which initial treatment at high-volume centers was more cost-effective than that at intermediate- or low-volume centers when other parameters were varied. The incremental costs and effects of 1,000 iterations for both treatments at high- versus low-volume centers or high- versus intermediate-volume center strategies stratified by risk factors are shown in Fig. 3. The cost-effectiveness acceptability curve showed that initial treatment at high-volume centers, maintained cost-effectiveness for 73.6% and 78.2% of iterations in the high- to intermediate-risk and low-risk groups, respectively, with a WTP of ¥5,000,000/QALY. With this threshold, intermediate-volume centers were cost-effective in 16.3% and 14.7% of iterations in the high-to intermediate-risk and low-risk groups, followed by low-volume centers, which were cost-effective in 10.0% and 7.1% of iterations in the high- to intermediate-risk and low-risk groups, respectively (Fig. 4).
Fig. 3
Incremental cost-effectiveness scatterplot.
QALY, quality-adjusted life years; WTP, willingness-to-pay.
Fig. 4
Cost-effectiveness acceptability curves (probabilistic sensitivity analyses).
WTP, willingness-to-pay.
4. Validation analysis
The ICERs for both the intermediate- and low-volume center strategies stratified by recurrence risk were considered to be inferior since they were less cost-effective than the high-volume center strategies; however, there was less than a 2% difference in the costs among the three volumes. Intermediate-volume centers had an incremental cost and effect of ¥37,316 and −0.04, respectively, for the high-to-intermediate risk group, and ¥20,300 and −0.03, respectively, for the low-risk group. Low-volume centers had an incremental cost and effect of ¥73,715 and −0.07, respectively, for the high-to-intermediate risk group, and ¥54,787 and −0.09, respectively, for the low-risk group (Table S2).
The superiority of treatment in high-volume centers was also associated with survival outcomes. All other parameters were below the WTP threshold. Additionally, probabilistic sensitivity analysis showed that treatment at high-volume centers was cost-effective in 70.2% and 63.3% of iterations in the high to intermediate-risk and low-risk groups, respectively, regarding the WTP/QALY ratio. Using this threshold, intermediate-volume centers were cost-effective in 17.5% and 27.7% of iterations in the high-to intermediate-risk and low-risk groups, respectively, followed by low-volume centers, which showed cost-effectiveness in 12.3% and 9.0% of iterations in the high-to intermediate-risk and low-risk groups, respectively (Fig. S3). The cost-effectiveness acceptability in this scenario was slightly lower than those of the other strategies.
DISCUSSION
Our economic evaluation suggests that initial treatment at high-volume centers could be considered a cost-effective strategy for patients with endometrial cancer in Japan compared to initial treatment at low-volume centers. Conversely, treatment effectiveness and cost were nearly equivalent between high- and intermediate-volume centers. A deterministic sensitivity analysis showed that the model was sensitive to variations in the probability of developing recurrence and initial treatment costs at high-volume centers. A WTP threshold of ¥5,000,000 barely exceeded the variations in probability, costs, and utility values. Probabilistic sensitivity analysis demonstrated that initial treatment at high-volume centers remains a cost-effective strategy across iterations with the WTP threshold.
There has been a recent paradigm shift from “evidence-based medicine” to “value-based medicine” regarding cancer treatment [25]. The integrated use of multiple approaches, including social and economic evidence, to scientific evidence to provide individualized treatment for patients with cancer, suggest a cost-beneficial treatment, with the selection of the most appropriate care plan. It is increasingly essential to consider patients’ cancer treatment decisions and provide sustainable medical treatment from a public healthcare perspective. However, high-quality evidence regarding cost-effective strategies for endometrial cancer treatment is lacking in Japan.
Our study suggests that it is important to have a system that designates specialized hospitals to patients with endometrial cancer and centralizes treatment. However, several issues remain unresolved. In Japan, the number of comprehensive cancer centers and healthcare workers differs according to the geographic location [26]. Most cancer treatment specialists belong to comprehensive centers and academic hospitals, and most high-volume centers are in eastern Japan, particularly in urban areas [6]. Moreover, Japan is surrounded by water, and the mountainous terrain creates barriers to transportation. The lack of appropriate infrastructure and social accessibility may impair patients’ access to healthcare systems.
The maldistribution of healthcare workers and access to healthcare in Japan creates a need for telemedicine. Robot-assisted surgery for endometrial cancer has been covered under insurance since 2018 in Japan [3]. however, the surgical team and patients need to be on-site for this surgery, and the conditions for telesurgery are only being verified. In the future, remote surgery for endometrial cancer may be adopted to help resolve the resource maldistribution in Japan.
The cost disparity in endometrial cancer treatment tends to increase with the utilization of high-cost medical interventions such as robotic surgery or new drugs. Therefore, the benefits of initiating treatment for endometrial cancer at high-volume centers are significant. Our 2016 study model emphasizes the importance of initial treatment, as these high-cost medical interventions were not covered by public medical insurance programs.
However, considering that approximately 20% of patients receive care at high-volume centers [7], according to a previous study, if all Japanese women currently receiving care at intermediate- and low-volume centers were suddenly redirected to high-volume centers, it would undoubtedly have direct implications for the healthcare sector. The cost and effectiveness difference between high- and intermediate-volume centers was minimal, and the cost-effectiveness of treatment at intermediate-volume centers remained within the acceptable range defined by the WTP threshold. Our results highlight the significance of centralizing cancer treatment and bottom-up support that provides specialized technical knowledge and skills to intermediate- and low-volume centers.
Moreover, the benefits of a value-based healthcare system can extend to patients' decision-making, healthcare providers, and the society as a whole. In Japan with declining birth rates and aging populations, where healthcare expenditures have recently increased significantly [27], value-based care has the potential to substantially reduce overall healthcare costs and increase patient satisfaction. The widespread adoption of value-based healthcare may lead to transformative changes in how physicians and hospitals deliver care.
The strength of this study is that it is the first study to evaluate a cost-effective strategy for endometrial cancer treatment in Japan. The study included real-world data relating to patient, tumor, and treatment demographics including treatment costs using data from the JSOG and DPC/PDPS databases. Our findings provide valuable insight into designing a treatment system for gynecological malignant tumors in Japan, taking into account the nation's future population trends.
This study has some limitations. First, the model was built from a public healthcare perspective regarding the appropriateness of institutional disparities in endometrial cancer treatment and not from the perspective of medical costs for patients. Therefore, the model did not include non-medical costs such as access to a hospital (transportation expenses), labor losses (loss of working hours), and unpredictable costs. Second, confounding factors may have affected the results. For instance, a lack of information regarding patient comorbidities may lead to an underestimated treatment cost. Third, hospital-specific conditions were not considered. High-volume hospitals may have better capacity (number of beds, number of gynecologic oncologists and staff with high-quality clinical experience, and availability of operating rooms, intensive care unit, facilities for investigational modalities, and so on) than low- and intermediate-volume hospitals, which may result in higher treatment costs. The high-volume centers in this study had slightly higher costs for initial cancer treatment. Moreover, although the number of treatment cases per facility was specified, details regarding the number of gynecologic oncologists affiliated with each facility were not provided. The number of gynecologic oncologists at each institution had social mobility of human resources and varied from year to year and could not be used in the present analysis.
In conclusion, our findings suggest that initiating treatment at high-volume centers could be a cost-effective strategy for patients with endometrial cancer in Japan. However, considering patients' accessibility, treatment at intermediate- and low-volume centers also falls within the acceptable range defined by the WTP threshold of ¥5,000,000/QALY. To address treatment disparities in Japan, it is crucial to establish centralized treatment in high-volume centers and provide bottom-up support for intermediate- and low-volume centers for cancer treatment.
SUPPLEMENTARY MATERIALS
2016 life table (women) in JapanTable S1
Incremental analysisTable S2
Cost-effectiveness analysis.Fig. S1
Deterministic sensitivity analysis.Fig. S2
Cost-effectiveness acceptability curves (probabilistic sensitivity analyses).Fig. S3
Funding:This work was supported by Bristol Myers Squibb.
Conflict of Interest:Honorarium, Bristol Myers Squibb (H.M) and none for others.
Author Contributions:
Conceptualization: H.M, K.M, M.M.
Data curation: M.H., M.K., H.T., A.D., E.T., O.A., K.H., N.S., M.M., Y.N., Y.W., M.M.
Formal analysis: H.M.
Funding acquisition: M.M.
Investigation: M.H., M.K., H.T., A.D., E.T., O.A., K.H., N.S., M.M., Y.N., Y.W., M.M.
Methodology: H.M., K.M.
Project administration: H.M., M.M.
Resources: M.H., M.K., H.T., A.D., E.T., O.A., K.H., N.S., M.M., Y.N., Y.W., M.M.
Software: H.M.
Supervision: M.H., M.K., H.T., A.D., E.T., O.A., K.H., N.S., M.M., Y.N., Y.W., M.M.
Validation: H.M., K.M.
Visualization: H.M.
Writing - original draft: H.M.
Writing - review & editing: M.H., M.K., H.T., A.D., E.T., O.A., K.H., N.S., M.M., Y.N., Y.W., M.M.
References
-
Cancer Information Service. Cancer statistics in Japan [Internet]. Tokyo: National Cancer Center; 2023 [cited 2023 May 5]. [Accessed 2023 May 5].Available from: https://ganjoho.jp/reg_
stat/statistics/stat/cancer/1_ all.htm.
-
-
Nagase S, Ohta T, Takahashi F, Yamagami W, Yaegashi N. Board Members of the 2020 Committee on Gynecologic Oncology of the Japan Society of Obstetrics and Gynecology. Annual report of the Committee on Gynecologic Oncology, the Japan Society of Obstetrics and Gynecology: annual patient report for 2018 and annual treatment report for 2013. J Obstet Gynaecol Res 2022;48:541–552.
-
-
Inoue S, Xu H, Maswana JC, Kobayashi M. Forecasting of future medical care expenditure in japan using a system dynamics model. Inquiry 2022;59:469580221091397
-
-
Saito T, Takahashi F, Katabuchi H. 2016 Committee on Gynecologic Oncology of the Japan Society of Obstetrics and Gynecology. Annual report of the Committee on Gynecologic Oncology, Japan Society of Obstetrics and Gynecology: patient annual report for 2014 and treatment annual report for 2009. J Obstet Gynaecol Res 2017;43:1667–1677.
-
-
Ministry of Health, Labour and Welfare of Japan. Report on the results of the 2016 “Survey of Discharged Patients” pertaining to the impact assessment of the introduction of DPC [Internet]. Tokyo: Ministry of Health, Labour and Welfare of Japan; 2023 [cited 2023 May 10].Available from: https://www.mhlw.go.jp/stf/shingi2/0000196043.html.
-
-
Yamaguchi H, Iwasaki K, Shoji A, Kokubo K, Igarashi A. Health-related quality of life (QoL) in Japanese patients with cancer; a large-scale questionnaire survey using EQ-5D-5L. Ther Res 2020;41:949–955.
-
-
Japanese Government Statistics. Statistics of Japan [Internet]. Tokyo: Japanese Government Statistics; 2023 [cited 2023 May 10].Available from: https://www.e-
stat.go.jp/en.
-
-
Walker JL, Piedmonte MR, Spirtos NM, Eisenkop SM, Schlaerth JB, Mannel RS, et al. Recurrence and survival after random assignment to laparoscopy versus laparotomy for comprehensive surgical staging of uterine cancer: Gynecologic Oncology Group LAP2 Study. J Clin Oncol 2012;30:695–700.
-
-
Yasunaga H. Real world data in Japan: chapter II the diagnosis procedure combination database. Ann Clin Epidemiol 2019;1:76–79.
-
-
Center for Outcomes Research and Economic Evaluation for Health, National Institute of Public Health. Guideline for preparing cost-effectiveness evaluation to the central social insurance medical council. (C2H) NIoPH [Internet]. Wako: Center for Outcomes Research and Economic Evaluation for Health, National Institute of Public Health; 2022 [cited 2023 May 10].Available from: https://c2h.niph.go.jp/tools/guideline/guideline_
en.pdf.
-
-
Drummond MF, Claxton K, Stoddart GL, Torrance GW. In: Methods for the economic evaluation of health care programmes. 4th ed. Oxford: Oxford Medical Publications; 2015.
-