Expected value of the additional state in evaluating the method of quantification and uncertainty of additional states in an analytical model of grade I hypertension
- Takeshi Uchikura^{1, 2}Email author,
- Makoto Kobayashi^{3},
- Masayuki Hashiguchi^{1} and
- Mayumi Mochizuki^{1}
DOI: 10.1186/s40780-014-0006-z
© Uchikura et al.; licensee BioMed Central. 2015
Received: 25 June 2014
Accepted: 30 October 2014
Published: 28 January 2015
Abstract
Background
In the construction of pharmacoeoconomic models, simplicity is desirable for transparency (people can see how the model is built), ease of analysis, validation (how well the model reproduces reality), and description. Few reports have described concrete methods for constructing simpler models. Therefore we focused on the value of additional states and uncertainty in disease models with multiple complications.
Objectives
The objective of this study was to examine the possibility of ranking additional states in disease models with multiple complications using a method for evaluating the quantification and uncertainty of additional states.
Methods
The expected value of additional states (EVAS) was formulated to calculate the value of additional states from the variation between analytic models using the net benefit method, and uncertainty was subtracted from the variation. We also verified the usefulness and availability of this method in grade I hypertension as a verification of the disease model. We assumed that stroke was recognized as an associated complication of hypertension in the basic model. In addition, stroke recurrence, coronary heart disease (CHD), and end-stage renal disease (ESRD) were assumed to represent other complications of hypertension. Ten thousand Monte Carlo simulations were performed, and the probability distribution was assumed to be the beta distribution in clinical parameters. The ranges of clinical parameters were ±6.25%, 12.5%, 25%, and 50% of the standard deviation from the mean value.
Results
The EVAS in complications of CHD showed the greatest uncertainty. In contrast, the EVAS of ESRD differed from stroke recurrence in the value ranking by uncertainty.
Conclusions
The EVAS has the potential to determine the ranking of additional states based on the quantitative value and uncertainty in disease models with multiple complications.
Keywords
Expected Value of the Additional State Pharmacoeconomics Hypertension Quantification UncertaintyBackground
Pharmacoeoconomic model analysis is conducted to combine various parameters (effect, side effects, quality of life [QOL], and cost data) using modeling techniques, e.g., the Markov model [1]. The construction of analytic models is a key issue affecting their results. The guidelines for modeling analysis refer to internal validity, e.g., the course of a disease or clinical process, and external validity, e.g., comparing model results with real-world results [2,3]. Thus, The Joint Report of the International Society for Pharmacoeconomics and Outcomes Research and the Society for Medical Decision Making (report of the ISPOR-SMDM Modeling Good Research Practices Task Force-2 [4]) mentions that the simplicity of models is desirable for ease of analysis, description, transparency (people can see how the model is built), and validation (how well the model reproduces reality) [5]. However, few reports have described concrete methods for constructing simpler models.
Generally, in the construction of analytical (disease) models, we assume that multiple complications may occur as the outcome of a specific disease. The complications in the analytical model are selected based on their importance (generally referred to as “a state”), but currently there is no evaluation method for that purpose. Furthermore, it is important to consider the influence of the uncertainty of the state itself, but that is not done at present. Therefore, if we could attempt to estimate the quantitative value of additional states in an analytical model and examine the possibility of ranking additional states in disease models with multiple complications, it would contribute to the establishment of useful analytical models. Therefore, in this study we attempted to estimate the quantitative value of additional states and rank those additional states, in addition to evaluating the robustness of ranking while considering uncertainty.
Objectives
The objective of this study was to examine the possibility of ranking additional states in disease models with multiple complications using a method for evaluating the quantification and uncertainty of additional states.
Methods
Method to evaluate quantification of the value of additional states and uncertainty of transition probabilities
where C_{A} = the cost of analytical model A, C_{B} = the cost of analytical model B, λ = 5,000,000 yen (the value of 1 QALY), E_{A} = the effectiveness (QALY) of analytical model A, and E_{B} = the effectiveness (QALY) of analytical model B.
Using Eq. (1), we can calculate each value of multiple complications (states) assumed in a specific disease based on costs and QALY and rank these complications (states) in the analytical model. Furthermore, we can estimate the robustness of the results based on the uncertainty of each state itself by estimating the EVAS from Eq. (2). Therefore, we can select a specific complication among multiple complications of a disease, because this method ranks multiple complications that consider the values of additional states and the robustness of the value of uncertainty of additional states.
Disease model used for examining the ranking of the values of additional states
Transition probabilities, QOL, and cost data on states in the disease model
Clinical parameters
Mean | |
---|---|
Annual incidence of stroke in grade I hypertension (1000 people/year) | |
Men [10] | 4.9 |
Annual incidence of CHD in grade I hypertension (1000 people/year) | |
Men [10] | 1.17 |
Annual incidence of ESRD in grade I hypertension (1000 people/year) | |
Men [11] | 0.21 |
Recurrence rate of stroke [12]* | |
First year | 12.9% |
Second to fifth year | 8.2% |
Mortality after stroke [13]*† | |
First year | 20.7% |
Second year | 6.7% |
Third year | 5.8% |
Fourth year | 5.9% |
Fifth year | 5.9% |
In hospital | 7% |
One year (1000 people/year) | 20.4 |
Mortality after ESRD [16] | |
One year | 9.7% |
The degree of disability after stroke was scored using the modified Rankin Scale (mRS) [18,19], and allocations of patients were derived from the proportion of stroke type and mRS score at hospital discharge [20,21]. The degree of disability was assumed to remain constant over time.
The cost of stroke treatment in the acute phase was obtained from the average number of hospitalized days based on the highest number of diagnostic procedure combinations in patients with each type of stroke [25]. The cost of recovery-phase rehabilitation facility care was calculated based on an average of 88 hospitalized days [26]. The annual cost of care was estimated from the amount of maximum payment for the categories of allowances for nursing care per month [27], which correspond to mRS scores of 2–5, and it was calculated until death. In addition, we calculated work lost until the age of 65 years if the patients died at age 64 years or younger.
Categorization of assumed costs related to medical treatment and others after stroke, CHD, and ESRD
Severity | Acute-phase hospital care cost | Recovery-phase care cost | Ambulant treatment cost | Nursing care cost | Work lost* | ||
---|---|---|---|---|---|---|---|
After stroke | First year | mRS0-1 | ● | ● | ● | ||
mRS2-5 | ● | ● | ● | ● | ● | ||
Second year or later | mRS0-1 | ● | ● (for clinic visits) | ||||
mRS2-5 | ● | ● | ● | ||||
After CHD | First year | - | ● | ● | ● (in hospital) | ||
Second year or later | - | ● | ● (for clinic visits) | ||||
After ESRD | - | - | ● (hemodialysis or conservative management) | ● (for clinic visit) |
The data on medical costs were obtained from official prices in Japan [25,30,31]. The costs of work loss reflected the average wages of each age cohort for regular employees [32].
Discounting
This study assumed that the discount rate was 3% and 0% (no discounts) for reference values.
External validity
The external validity of analytical models compared the overall mortality from grade I hypertension patents in the JPHC Study [10] with the overall mortality of each analytical model.
Statistical analyses
Ten thousand Monte Carlo simulations were performed for each evaluation. The probability distribution for transition probabilities was assumed to be the beta distribution [33]. The ranges of transition probabilities were ±6.25%, 12.5%, 25%, and 50% of the standard deviation (SD) from the mean. The simulation software TreeAge2013 (TreeAge Software, Inc. USA) was used.
Results
Evaluation of external validity
The overall mortality rate per year from the JPHC Study [10] was 0.0084, in the basic model (stroke alone) it was 0.0086, in the stroke-recurrence model it was 0.0087, in the stroke-CHD model it was 0.0087, and in the stroke-ESRD model it was 0.0086.
Evaluation of the value of additional states
Expected value of additional state (EVAS) of each complication
Value of additional state* | Uncertainty** | EVAS^{†} | |
(±6.25% SD) | |||
(±12.5% SD) | |||
(±25% SD) | |||
(±50% SD) | |||
Recurrence of stroke | 116 (140) | 7 (8) | 109 (132) |
13 (16) | 102 (124) | ||
26 (32) | 89 (108) | ||
52 (60) | 64 (80) | ||
CHD | 295 (356) | 15 (18) | 280 (338) |
30 (36) | 265 (320) | ||
60 (72) | 235 (284) | ||
117 (142) | 178 (214) | ||
ESRD: Dialysis treatment | 137 (167) | 7 (8) | 130 (159) |
14 (16) | 123 (151) | ||
27 (33) | 120 (134) | ||
53 (64) | 84 (103) | ||
Conservative treatment | 53 (65) | 3 (4) | 50 (61) |
6 (7) | 47 (58) | ||
12 (15) | 41 (50) | ||
23 (28) | 30 (37) |
Evaluation of the EVAS
The EVAS when uncertainty was considered tended to become smaller if the variable of the transition probabilities of the state grew larger. However, the ranking of additional states did not differ markedly when uncertainty was not taken into consideration. Therefore, the EVAS of CHD was the greatest of the additional states examined, even when the SD was ±50%, and the EVAS of ESRD with conservative management was the least in each case. In contrast, the EVAS of ESRD with dialysis was greater than that for stroke recurrence when uncertainty was not considered and reached 84 when the SD was ±50%. This estimate was less than the 89 for the EVAS of stroke recurrence when the SD was ±25% (Table 4).
Discussion
In an attempt to construct a simpler model, this study ranked the importance of multiple complications in hypertension to estimate the value of additional states in a quantitative analytical model including transition probabilities, QOL, and cost data. In addition, we evaluated the robustness of the ranking of these values when considering the uncertainty of transition probabilities for the EVAS. To the best of our knowledge, this was the first experiment on quantification and ranking of the value of states along with uncertainty.
The ranking of the value of additional states was in the order CHD, ESRD with dialysis, stroke recurrence, and ESRD with conservative management in the analytical model of 55-year-old men with grade I hypertension. The differences in the assumed treatment in ESRD influenced the ranking of the value of the additional state, because the cost of dialysis treatment is 100, but that of conservative treatment is 16 (the value of QALY conversion). Thus, the ranking of ESRD with dialysis and stroke recurrence will change with an assumption of the uncertainty of transition probabilities of the EVAS.
In the development of analytical models, the difference in internal validity and results of analytical models must be weighed [34,35], particularly in analytical models constructed by academic groups in competitions like the Mount Hood Challenge Meeting [36]. Few reports referred to concrete methods for constructing simpler models. If our method enables the ranking of multiple complications in a specific model based on analytical values, more important and robust additional states could be identified. This would contribute to the development of a simpler model for pharmacoeconomic evaluation.
The limitations of this study were that: 1) only grade I hypertension was examined; 2) the setting of the threshold (λ) in this method could be refined; and 3) the variation in all transition parameters was assumed to have the same range in a state. Although only grade I hypertension was considered, Eq. (2) shows that there would be no difference in the ranking of the value of the state based on the EVAS in other diseases. Therefore, this method has sufficient potential to be widely applied. The threshold (λ) was set at 5 million yen for 1 QALY because 1 QALY in Japan was reported to be between approximately 5 and 6 million yen [37]. However, care must be taken because the results (the value of states) will change based on the setting of the threshold (λ). Finally, the ranking of states (complications) may be different in some results from the EVAS based on variations in transition probabilities. The variation of all transition parameters was assumed to be in the same range in a state in this study. However, the actual uncertainty of transition probabilities can differ. It will be necessary to consider this in an actual analytical model in future.
Conclusions
This is the first report on devising a method for the evaluation and ranking of the quantitative value of additional states with uncertainty. It allows cumulative evaluation of transition probabilities, QOL, and cost data of a state using the EVAS. Generally, when we construct a pharmacoeconomic model, we consider states based on transition probabilities (e.g., incidence or mortality rate), but QOL and cost data are also components of a state. This evaluation method is useful because it quantifies the value of a state. In addition, the ranking of additional states in a disease model with multiple complications can be determined to identify which have less impact on a pharmacoeconomic model. The EVAS can therefore take into consideration the value of additional states with the uncertainty of transition probabilities. In the next phase of this continuing study, we plan to demonstrate the relation between the EVAS and pharmacoeconomic evaluation, i.e., the analytic results of interventions, and/or to examine the results in other chronic diseases that have multiple complications, e.g., chronic hepatitis C, etc., because we think that it is necessary to verify that this method will help to construct a simpler analytical model.
Declarations
Acknowledgments
Yuma Higashino contributed to data collection.
Authors’ Affiliations
References
- Stahl JE: Modelling methods for pharmacoeconomics and health technology assessment: An overview and guide. Pharmacoeconomics 2008, 26:131–148.PubMedView ArticleGoogle Scholar
- National Institute for Health and Clinical Excellence: The Guidelines Manual 2012. http://www.nice.org.uk/article/PMG6/chapter/1%20Introduction.
- Fukuda T: Guidelines for analytical methods in health economics evaluation research. In Health and Labor Sciences Research Grant: Research on the Health Benefit Plan Applying Health Economics Evaluation. Tokyo; 2013.Google Scholar
- Roberts M, Russell LB, Paltiel AD, Chambers M, McEwan P, Krahn M: Conceptualizing a model: A report of the ISPOR-SMDM Modeling Good Research Practices Task Force-2. Med Decis Making 2012, 32:678–689.PubMedView ArticleGoogle Scholar
- Eddy DM, Hollingworth W, Caro JJ, Tsevat J, McDonald KM, Wong JB, for the ISPOR-SMDM Modeling Good Research Practices Task Force: Model transparency and validation: a report of the ISPOR-SMDM Modeling Good Research Practices Task Force-7. Med Decis Making 2012, 32:733–743.PubMedView ArticleGoogle Scholar
- Claxton K, Neumann PJ, Araki S, Weinstein MC: Bayesian value-of-information analysis. An application to a policy model of Alzheimer's disease. Int J Technol Assess Health Care 2001, 17:38–55.PubMedView ArticleGoogle Scholar
- Naversnik K, Rojnik K: Handling input correlations in pharmacoeconomic models. Value Health 2012, 15:540–549.PubMedView ArticleGoogle Scholar
- Uchikura T, Yokoi N, Hashiguchi M, Hashiguchi M, Mochizuki M: Pharmacoeconomic evaluation of hypothetical over-the-counter angiotensin-converting enzyme (ACE) inhibitors for the prevention of stroke in patients with grade I hypertension. Yakugaku Zasshi 2011, 131:571–580.PubMedView ArticleGoogle Scholar
- Japanese Society of Hypertension: Guidelines for the Management of Hypertension 2009. Tokyo: Japanese Society of Hypertension; 2009 [In Japanese].Google Scholar
- Ikeda A, Iso H, Yamagishi K, Inoue M, Tsugane S: Blood pressure and the risk of stroke, cardiovascular disease, and all-cause mortality among Japanese: The JPHC study. Am J Hypertens 2009, 22:273–280.PubMedView ArticleGoogle Scholar
- Klag MJ, Whelton PK, Randall BL, Neaton JD, Brancati FL, Ford CE, Shulman NB, Stamler J: Blood pressure and end-stage renal disease in men. N Engl J Med 1996, 334:13–18.PubMedView ArticleGoogle Scholar
- Hata J, Tanizaki Y, Kiyohara Y, Kato I, Kubo M, Tanaka K, Okubo K, Nakamura H, Oishi Y, Ibayashi S, Iida M: Ten year recurrence after first ever stroke in a Japanese community: The Hisayama study. J Neurol Neurosurg Psychiatry 2005, 76:368–372.PubMed CentralPubMedView ArticleGoogle Scholar
- Investigative Commission on Measures of Stroke: Interim Report. Tokyo: Japanese Ministry of Health and Welfare, Health Service Bureau, Lifestyle-Related Disease Measure Office; 1999 [In Japanese].Google Scholar
- Watanabe J, Iwabuchi K, Koseki Y, Fukuchi M, Shinozaki T, Miura M, Komaru T, Kagaya Y, Shirato K, Kitaoka S, Ishide N, Takishima T: Declining trend in the in-hospital case-fatality rate from acute myocardial infarction in Miyagi prefecture from 1980 to 1999. Jpn Circ J 2001, 65:941–946.PubMedView ArticleGoogle Scholar
- Japanese Coronary Artery Disease (JCAD) Study Investigators: Current status of the background of patients with coronary artery disease in Japan. Circ J 2006, 70:1256–1262.View ArticleGoogle Scholar
- Japanese Society for Dialysis Therapy, Committee of Renal Data Registry: An overview of regular dialysis treatment in Japan. http://docs.jsdt.or.jp/overview/.
- Japanese Ministry of Health, Labor and Welfare, Minister’s Secretariat Statistics and Information Department: 2007 Abridged Life Table. Tokyo: Ministry of Health, Labor and Welfare; 2008. http://www.mhlw.go.jp/toukei/saikin/hw/life/life07/index.html]. [In Japanese].Google Scholar
- Van Swieten JC, Koudstaal PJ, Visser MC, Schouten HJ, van Gijn J: Interobserver agreement for the assessment of handicap in stroke patients. Stroke 1988, 19:604–607.PubMedView ArticleGoogle Scholar
- Shinohara Y, Minematsu K, Amano T, Ohashi Y: Reliablity of modified Rankin scale—introduction of a guidance scheme and a questionnaire written in Japanese. Jpn J Stroke 2007, 29:6–13 [In Japanese].View ArticleGoogle Scholar
- Kobayashi S (Ed): Japanese Stroke Data Bank 2009. Tokyo: Nakayama Shoten; 2009 [In Japanese].Google Scholar
- Kobayashi S (Ed): Japanese Stroke Data Bank 2005. Tokyo: Nakayama Shoten; 2005 [In Japanese].Google Scholar
- Noto S, Yanagi H, Tomura S: Measuring utilities for various functional outcomes after stroke, comparison of rating scale and time trade-off methods. Nippon Koshu Eisei Zasshi 2002, 49:1205–1216 [In Japanese].PubMedGoogle Scholar
- Takura T, Kawanishi H, Minakuchi J, Nagake Y, Takahashi S: Cost-effectiveness analysis of on-line hemodiafiltration in Japan. Blood Purif 2013, 35:85–89.PubMedView ArticleGoogle Scholar
- Lacey EA, Walters SJ: Continuing inequality: gender and social class influences on self perceived health after a heart attack. J Epidemiol Community Health 2003, 57:622–627.PubMed CentralPubMedView ArticleGoogle Scholar
- 2012 DPC electronic score list. [online]. [http://www.mhlw.go.jp/topics/2012/03/tp0305-02.html]. [In Japanese].
- Koga M, Uehara T, Nagatsuka K, Yasui N, Hasegawa Y, Naritomi H, Okada Y, Minematsu K: Current role of convalescent rehabilitation units in community-based referral systems for stroke patients in Japan. Jpn J Stroke 2008, 30:735–743 [In Japanese].View ArticleGoogle Scholar
- Ministry of Health, Labor and Welfare: Public notice of no. 33 of 2000 (No. 258 revision). [In Japanese].Google Scholar
- Fujimura Y: Evaluation of the diagnosis procedure combination payment system: from the viewpoint of patients, medical and co-medical workers, and managements. Jpn J Political Economy Health Health Care 2010, 29:29–40 [In Japanese].Google Scholar
- Ikeda S, Kobayashi M: Prediction of long-term prognosis for diabetes patients. Development of risk simulation software for cost-effectiveness analysis. Bunshi Touyoubyougaku no Shinpo; 2003:191–194 [In Japanese].Google Scholar
- Medical fees scoreboards (Ikashinryohosyu Tensuhyo). 2012th edition. Tokyo: Shakai Hoken Kenkyujyo; 2012 [In Japanese].Google Scholar
- National Health Insurance Price Dictionary (Hokenyaku Jiten). 2012th edition. Tokyo: Jiho, Inc; 2012 [In Japanese].Google Scholar
- Ministry of Health, Labor and Welfare, Minister’s Secretariat Statistics and Information Department: Basic Survey on Wage Structure, 2009. Tokyo: Ministry of Health, Labor and Welfare; 2008 [http://www.e-stat.go.jp/SG1/estat/NewList.do?tid=000001011429]
- Briggs AH, Weinstein MC, Fenwick EA, Karnon J, Sculpher MJ, Paltiel AD: ISPOR-SMDM Modeling Good Research Practices Task Force: Model parameter estimation and uncertainty: a report of the ISPOR-SMDM Modeling Good Research Practices Task Force—6. Med Decis Making 2012, 32:722–732.PubMedView ArticleGoogle Scholar
- Doan QV, Chiou CF, Dubois RW: Review of eight pharmacoeconomic studies of the value of biologic DMARDs (adalimumab, etanercept, and infliximab) in the management of rheumatoid arthritis. J Manag Care Pharm 2006, 12:555–569.PubMedGoogle Scholar
- Ramos GF, Kuiper S, Dompeling E, van Asselt AD, de Grauw WJ, Knottnerus JA, van Schayck OC, Schermer TR, Severens JL: Structuring and validating a cost-effectiveness model of primary asthma prevention amongst children. BMC Med Res Methodol 2011, 11:150–163.PubMed CentralPubMedView ArticleGoogle Scholar
- Palmer AJ, Clarke P, Gray A, Leal J, Lloyd A, Grant D, Palmer J, Foos V, Lamotte M, Hermann W, Barhak J, Willis M, Coleman R, Zhang P, McEwan P, Betz Brown J, Gerdtham U, Huang E, Briggs A, Carlsson KS, Valentine W, for the Mount Hood 5 Modeling Group: Computer modeling of diabetes and its complications: a report on the Fifth Mount Hood Challenge meeting. Value Health 2013, 16:670–685.PubMedView ArticleGoogle Scholar
- Shiroiwa T, Sung YK, Fukuda T, Lang HC, Bae SC, Tsutani K: International survey on willingness-to-pay (WTP) for one additional QALY gained: what is the threshold of cost effectiveness? Health Econ 2010, 19:422–437.PubMedView ArticleGoogle Scholar
Copyright
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.