Cost-effectiveness of pharmacist-participated warfarin therapy management in Thailand

Thrombosis Research 132 (2013) 437–443

Contents lists available at ScienceDirect

Thrombosis Research journal homepage: www.elsevier.com/locate/thromres

Regular Article

Cost-effectiveness of pharmacist-participated warfarin therapy management in Thailand Surasak Saokaew a,b,c, Unchalee Permsuwan d,⁎, Nathorn Chaiyakunapruk c,e,f,g, Surakit Nathisuwan h, Apichard Sukonthasarn i, Napawan Jeanpeerapong j a

Clinical Epidemiology Program, Faculty of Medicine, Chiang Mai University, Chiang Mai, Thailand Department of Pharmaceutical Care, School of Pharmaceutical Sciences, University of Phayao, Phayao, Thailand Center of Pharmaceutical Outcomes Research (CPOR), Faculty of Pharmaceutical Sciences, Naresuan University, Phitsanulok, Thailand d Department of Pharmaceutical Care, Faculty of Pharmacy, Chiang Mai University, Chiang Mai, Thailand e Discipline of Pharmacy, Monash University Sunway Campus, Malaysia f School of Population Health, University of Queensland, Brisbane, Australia g School of Pharmacy, University of Wisconsin, Madison, USA h Faculty of Pharmacy, Mahidol University, Bangkok, Thailand i Division of Cardiology, Department of Medicine, Faculty of Medicine, Chiang Mai University, Chiang Mai, Thailand j Department of Pharmacy, Buddhachinaraj Regional Hospital, Phitsanulok, Thailand b c

a r t i c l e

i n f o

Article history: Received 21 February 2013 Received in revised form 26 June 2013 Accepted 27 August 2013 Available online 1 September 2013 Keywords: Anticoagulation clinic Bleeding Cost-effectiveness Pharmacist Thromboembolism Warfarin

a b s t r a c t Introduction: Although pharmacist-participated warfarin therapy management (PWTM) is well established, the economic evaluation of PWTM is still lacking particularly in Asia-Pacific region. The objective of this study was to estimate the cost-effectiveness of PWTM in Thailand using local data where available. Methods: A Markov model was used to compare lifetime costs and quality-adjusted life years (QALYs) accrued to patients receiving warfarin therapy through PWTM or usual care (UC). The model was populated with relevant information from both health care system and societal perspectives. Input data were obtained from literatures and database analyses. Incremental cost-effectiveness ratios (ICERs) were presented as year 2012 values. A base-case analysis was performed for patients at age 45 years old. Sensitivity analyses including one-way and probabilistic sensitivity analyses were constructed to determine the robustness of the findings. Results: From societal perspective, PWTM and UC results in 39.5 and 38.7 QALY, respectively. Thus, PWTM increase QALY by 0.79, and increase costs by 92,491 THB (3,083 USD) compared with UC (ICER 116,468 THB [3,882.3 USD] per QALY gained). While, from health care system perspective, PWTM also results in 0.79 QALY, and increase costs by 92,788 THB (3,093 USD) compared with UC (ICER 116,842 THB [3,894.7 USD] per QALY gained). Thus, PWTM was cost-effective compared with usual care, assuming willingness-to-pay (WTP) of 150,000 THB/QALY. Results were sensitive to the discount rate and cost of clinic set-up. Conclusion: Our finding suggests that PWTM is a cost-effective intervention. Policy-makers may consider our finding as part of information in their decision-making for implementing this strategy into healthcare benefit package. Further updates when additional data available are needed. © 2013 Elsevier Ltd. All rights reserved.

Introduction Warfarin is a common therapy to prevent thromboembolism (TE) in various conditions [1]. However, it might cause bleeding complications that sometimes are potentially fatal [2]. Thus, it requires individualized management to achieve effective and safe anticoagulation treatment. Pharmacist-participated warfarin therapy management (PWTM) is one of the anticoagulation monitoring service (AMS) models where a

⁎ Corresponding author at: Department of Pharmaceutical Care, Faculty of Pharmacy, Chiang Mai University, Chiang Mai 50200, Thailand. Tel.: +66 54944342; fax: +66 53222741. E-mail address: [email protected] (U. Permsuwan). 0049-3848/$ – see front matter © 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.thromres.2013.08.019

pharmacist plays an important role in a number of activities ranging from warfarin dosage adjustment, medication/drug interaction review to providing patient and/or health care provider education. PWTM has been accepted and implemented widely for managing warfarin therapy worldwide [1,3]. A recent meta-analysis evaluating the effect of PWTM on clinical outcomes confirmed that PWTM was associated with significant reduction of total bleeding and a trend of reduction major bleeding and thromboembolic events compared to usual care (UC) [3]. In Thailand, the concept and practice of PWTM was first introduced in 2002. Despite the initially slow uptake of such practice, PWTM has now gained national interest with increasing number of hospitals adopting this practice. A formal clinical outcomes evaluation of Thai PWTM was recently conducted in a pioneering tertiary care hospital [4]. The results showed that PWTM provided significantly better anticoagulation control

438

S. Saokaew et al. / Thrombosis Research 132 (2013) 437–443

A No Thromboembolism No Bleeding

B Thromboembolism

F Death

D Major bleeding

those studies were exclusively derived from European and Western countries [5,6]. Direct application of such model to the developing countries would be of limited accuracy. This study, therefore, aimed to evaluate the incremental costs and health benefits of PWTM versus UC for long-term warfarin therapy in Thailand from both health care system and societal perspectives. Availability of such information would be very helpful in the decision making process of healthcare policy whether such service should be supported nationwide. Methods

C Sequelae

Overview

E Sequelae

Fig. 1. Schematic diagram of the Markov model with six health states.

compared with UC [4]. However, its impact on economic outcomes has not been evaluated. In the developing world where resources are scarce, evidence of cost-effectiveness analysis (CEA) is in great need. While several studies evaluating economic impact of PWTM are available, input parameters in

A Markov model was adapted from a previous study [7] to evaluate the cost and outcomes of PWTM compared with UC. We adapted only the clinical state model, and separated sequelae into 2 states to illustrate the difference of consequence of bleeding and TE events, thus six health states were specified (Fig. 1). This model assumed PWMT as an additional service to UC for warfarin management because PWTM is mostly implemented as a collaborative approach in Thailand [4]. The hypothetical cohort of 1,000 patients requiring warfarin therapy for all indications aged 45 years old [4] was entered into the model with 3-month cycle length. Patients received standard care from physician in the UC

Table 1 Input parameters, values and data sources used in the model. Parameters Probabilities (3-month) TE Bleeding Sequelae from TE Death with TE Sequelae from bleeding Death with bleeding Death with sequelae of TE Death with sequelae of bleeding Effectiveness of PWTM (risk ratios, RR) On TE On major bleeding Costs (THB, year of costing: 2012) Direct medical care costs Warfarin clinic setup and training TE (per episode) Bleeding (per episode) Sequelae (3-month) Pharmacist service cost (3-month) INR test (per test) Direct non-medical care costs Transportation (per visit) Additional food cost (per visit) Indirect costs (daily productivity loss by age) Age 15-29 Age 30-39 Age 40-59 Age 60-69 Age 70-79 Utilities (EQ-5D) Warfarin use (no event) TE Bleed Sequelae of TE Sequelae of bleeding

Base case

Range

Source(s)

0.0080 0.0112 0.2243 0.0719 0.0370 0.0561 0.0567 0.0446

0.0068 – 0.0092 0.0095 – 0.0129 0.1907 – 0.2579 0.0611 – 0.0827 0.0315 – 0.0426 0.0339 – 0.0765 0.0482 – 0.0652 0.0379 – 0.0513

[4] [4] [15] [16] [15] BCRH ⁎ BCRH ⁎ BCRH ⁎

0.79 0.64

0.33 –1.93 0.18 – 2.36

[3] [3]

68,000 – 92,000 41,101.9 – 46,808.0 32,726.6 – 73,940.6 2,273 – 17,204 408.0 – 522.0 -

Calculated † BCRH ⁎ BCRH ⁎

80,000 43,790.9 58,127.5 8,101.3 480.0 82.6 145.3 53.5

122.2 – 168.5 42.8 – 64.2

196 409 571 246 98

-

0.987 0.713 0.836 0.320 0.620

0.967 – 0.998 0.271 – 1 0.711 – 0.962 0.000 – 0.700 0.370 – 0.870

[27] [4,36] ‡ [25] § [25] ¶ [25] ¶ [39] ⁎⁎

[6,35] [29] [31] †† [32,33] [34]

TE, thromboembolism; THB, Thai Baht; INR, international normalized ratio; EQ-5D, European Quality of Life-5 Dimensions; ECH, extra-cranial haemorrhage; ICH, intracranial haemorrhage. ⁎ Calculated from Buddhachinaraj Regional Hospital (BCRH) database using ICD-10 for identification of patients. † Calculated from cost of pharmacists’ training and materials. ‡ Calculated by multiplying working time for warfarin clinic of pharmacists salary and other fringe benefits for 2 participating pharmacists, and divided by number of patients using our institution data [4]. § Mean INR test is every 3 months [4]. ¶ Used for patients’ cost and one relative for each visit [28]. ⁎⁎ Used for sensitivity analysis with incorporated productivity loss. †† Weighted from the proportion of bleeding type (ECH: ICH = 0.929: 0.071) [4].

S. Saokaew et al. / Thrombosis Research 132 (2013) 437–443

group, while additional service was provided by pharmacists in the PWTM group. We performed the analysis from both societal and health care system perspectives using lifetime horizon. All costs and outcomes were discounted at a rate of 3% [8] and adjusted to 2012 values in line with recommendations of Thai Health Technology Assessment (HTA) guideline [9]. The analyses were performed using Microsoft Excel® (Microsoft Corp., Redmond, WA). Likelihood of Events The probabilities of clinical events used in the Markov model are shown in Table 1. To reflect Thai population, the probabilities used in this study were derived from Thai literature or Thai electronic hospital database if possible. Buddhachinaraj Regional Hospital (BCRH) is a 1,000-bed tertiary hospital located in northern Thailand. We use BCRH database due to its richness and comprehensiveness of data. In addition, this database has been accepted and widely used in literature [10–13]. The probabilities of developing TE and bleeding in the UC group were derived from previous study [4]. The probabilities of TE and major bleeding in patients who received warfarin in the UC group were estimated using exponential parametric survival analysis. Then, we converted 1-year risk to 3-month risk using the standard equation [14]. With the total 154.9 person-year follow-up, the probability of developing TE and major bleeding was 0.80% and 1.12% per 3-month, respectively. Ischemic stroke was chosen to represent TE since all TE patients diagnosed in our previous study were ischemic strokes [4]. Patients with ischemic stroke had a 22.43% chance of having long-term sequelae [15] and a 7.19% chance of dying within 3 months of the event [16]. The probability of death from major bleeding within 3 months (5.61%) was derived from BCRH database. Major bleeding was categorized as either intracranial hemorrhage (ICH) or extracranial hemorrhage (ECH). ECHs were assumed to be gastrointestinal (GI) bleeding, since almost anticoagulation-related ECH is located in GI tract [17] and patients with GI bleeding recovered within a month [7] and have no deficits [18]. Thus, in our model, we used the probability from ICH to represent the probability of transitioning from bleeding to sequelae. The ICH patients were assigned a 3.70% chance with sequelae [15]. Minor bleeding events were not included as no treatment attention was sought. We assumed that the stroke and bleeding survivors with no sequelae returned to the no-events state. The probability of TE or bleeding in the PWTM group was determined by multiplying probability of TE or bleeding in the UC group by the summary risk ratio (RR) of TE or bleeding which was based on the previous meta-analysis of randomized controlled trials (RCTs) [3]. This study showed that PWTM was associated with a 21% reduction in thromboembolism (RR, 0.79; 95%CI, 0.33 - 1.93), and a 36% reduction in major bleeding (RR, 0.64; 95%CI, 0.18 - 2.36) compared with UC. The probabilities of death from sequelae were derived from BCRH database with a death rate of 5.67% and 4.46% per 3-month for sequelae of TE and bleeding, respectively. The transition probability from no-events to death was based on the age-specific mortality rate (ASMR) for Thai population [19]. A study in mixed patients population (e.g. ischemic stroke, cerebral hemorrhage, atrial fibrillation, valvular defect) showed that patients receiving warfarin therapy was associated with improved survival [20]. However, this study has shown that patients receiving warfarin therapy remain at higher risk of death compared to general population, thus ASMR of Thai population was not applicable. We calculated transition probability for no-events to death by multiplying AMSR of Thai population by 1.3 to represent ASMR for Thai patients receiving warfarin therapy [20]. Costs The costs included (i) direct medical care costs e.g. costs of TE event, cost of bleeding event, cost of pharmacist service, cost of sequelae,

439

(ii) direct non-medical care costs e.g. cost of transportation, additional food cost (Table 1). Indirect cost was excluded to avoid doublecounting. Because quality-adjusted life year (QALY) has already captured health-related quality of life of patients with morbid conditions, measuring the productivity loss associated with morbidity would result in double counting [21]. All costs were estimated from Thai literature or Thai database. The costs were then converted to 2012 value using the Medical Care consumer price index (CPI) [22]. From the societal perspective, all costs were considered, while only direct medical care costs were considered for the health care system’s perspective. For inter-country comparisons, money values in Thai baht (THB) can be converted into United States dollar (USD) using exchange rate of 1 USD ≈ 30 THB [23]. Cost of TE management was obtained from BCRH database using ICD-10 to identify the patients and calculated using cost-to-charge ratio method. The average 3-month cost for TE management was 48,884 THB (1,629.5 USD), and cost of major bleeding was 63,548 THB (2,118.3 USD) (Table 1). Cost of warfarin clinic setup was calculated by summing the cost of pharmacists’ training and material used. This setup cost was included into the model only the first year of calculation. Cost of pharmacist services was calculated by multiplying working time in warfarin clinic and salary with other fringe benefits (Table 1). Since this model tracked the patient life-long, the pharmacist salary was assumed to increase every year by 5% [24]. The cost of international normalized ratio (INR) testing was derived from the standard cost lists for HTA in Thailand [25]. We did not include warfarin drug costs because the costs would likely be similar between PWTM and UC groups. In addition, including warfarin cost would have a relatively small effect on total cost due to its very low cost in Thailand (~1-3 THB or 0.03-0.1 USD per tablet) [26]. We did not include cost of physician since it was assumed to be similar between both groups. Cost of sequelae management included cost of rehabilitation and cost of care. Cost was derived from the existing study conducted in 320 stroke survivors from 9 centers in Thailand [27]. The cost of sequelae management was 8,101 THB (270 USD) per 3 months (Table 1). Direct non-medical costs included cost of transportation, and additional food cost. These costs were based on the standard cost lists for Thai HTA [25]. The costs were counted for patient and one relative since a previous study showed that a stroke survivor on average came to a hospital with one caretaker [28].

Utility We conducted a literature search for the utility of each health state for the QALY estimation. The utility value for TE was derived from the quality of life in Thai stroke patients measured by SF-36 [29]. The values from SF-36 were converted into EQ-5D using the equation of Ara et al. [30]. The equation was derived from an ordinal least regression models from 6,350 patient level data of 12 clinical trials. Through this method, an equation was constructed to allow conversion from average values of eight dimension scores of SF-36 to a utility score based on EQ-5D. Based on such equation, the utility of TE in our study was calculated as 0.713 (Table 1). The utility values for bleeding events were derived for ECH and ICH, and then weighted by proportion of such events. Since we did not find any literature about utility of bleeding in Thailand, we adopted the utilities of ECH and ICH from a previous study conducted in Western population [31] which used the utility 0.84 and 0.79, respectively. After taking into account of the proportion of ECH and ICH in Thailand (ECH per ICH; 92.9%:7.1%) [4], the utility value was 0.836 (Table 1). For the utility of sequelae of TE, we adopted the utility by Post et al. which was 0.32 [32]. In addition, to reflect Thai population, the utility of patients after stroke (EQ-5D ~0) in Thai setting was used for sensitivity analysis [33]. For the utility of sequelae of bleeding, we also adopted the utility from previous study by Christensen et al. which was 0.62 [34]. For

440

S. Saokaew et al. / Thrombosis Research 132 (2013) 437–443

patients taking warfarin without complications, the utility was estimated at 0.987 [35].

Table 3 Sensitivity analysis according to included or not included cost of productivity loss into the model by perspectives, discounting, and initial age group.⁎

Base-Case Analyses Using the societal perspective, we calculated the expected costs and outcomes in warfarin users aged 45 years old. This age was chosen to represent Thai population based on the average age of warfarin users in a recent report [4]. In this base-case analysis, we assumed that the pharmacists’ services are 4 hours per day and 1 day per week as generally practiced in Thailand. In addition, we assumed pharmacists’ service cost was 20,000 THB (666.7 USD) per month (salary 12,000 THB [400 USD] plus other fringe benefits 8,000 THB [266.7 USD] per month) [36]. Based on previous study, the average number of patients per month is 25 patients and the average frequency of patient visit is 4 times per year [4]. The results were presented as an incremental cost per quality-adjusted life years (QALYs) gained (incremental costeffectiveness ratio; ICER) for PWTM versus UC. According to WHO recommendation [37], strategy was considered ‘very cost-effective’ if ICER less than 1 times the per-capita gross domestic product (GDP) [150,000 THB or 5,000 USD in 2011] [38], and ‘cost-effective’ if between 1 and 3 times the per-capita GDP.

THB, Thai Baht; QALY, quality-adjusted life-year. ⁎ Values were represented as incremental cost-effective ratio (ICER) of THB per QALY with discounted at 3% per year, and shaded frame indicated base-case analysis.

was provided to illustrate the relationship between the values of ceiling ratio (willingness to pay for a unit of outcome, i.e. QALY gained) and the probability of favoring each strategy [42].

Sensitivity Analysis Results To determine the robustness of the estimates from base-case analysis, we also performed CEA and incorporated the cost of productivity loss into the model [21] by multiplied daily productivity loss cost with the number of day loss for illness. The average daily income was varied by age based on socioeconomic survey [39]. We used 1 day loss for regular follow-up, while 12 and 8 days loss (estimated length of hospital stay from BCRH database) for TE and bleeding event, respectively. Since life expectancy depends on patient’s age, we also vary patient age group at initial treatment to figure out the ICER for each age group. In addition, a series of one-way sensitivity analyses were performed to investigate the effects of altering parameters within the plausible ranges including epidemiologic data, effectiveness of PWTM, costs, and utilities (Table 1). Discount rate on cost and outcome of 0% - 6% [8] was also included in one-way sensitivity analysis. Furthermore, probabilistic sensitivity analysis (PSA) was undertaken to address uncertainty in the assumptions underlying the model by allowing all of input parameters values to vary simultaneously over their respective feasible ranges within the model. This analysis requires thousand iterations. When specific ranges or confidence intervals were not available, it was assumed that the range varies by ±15% to present the uncertainty in the absence of confidence intervals for each estimate. All input parameters were assigned a probability distribution to reflect the feasible range of values the each parameter could attain. The rationale for distributional assumption selection for each variable has been given detail elsewhere [40,41]. The Monte Carlo simulation then drawn one value at a time for each input variable and calculates the expected cost and effectiveness of strategy. Cost-effectiveness acceptability curve

Base-Case Analysis In base-case analysis from societal perspective, PWTM and UC results in a discounted gain of 39.5 and 38.7 QALY, respectively. Thus, PWTM increase QALY by 0.79 compared with UC. PWTM costs 92,491 THB (3,083 USD) more and has an incremental cost-effectiveness ratio of 116,468 THB (3,882.3 USD) per QALY gained. From health care system perspective, PWTM results in 0.79 QALY, and increase costs by 92,788 THB (3,093 USD) compared with UC (ICER 116,842 THB [3,894.7 USD] per QALY gained) (Table 2).

Sensitivity Analyses A sensitivity analysis showed that PWTM was unlikely to be costeffective at a willingness-to-pay (WTP) 450,000 THB (15,000 USD; 3 GDP) per QALY if the initial patient age was 65 year old even included or not included cost of productivity loss into the model, used societal or health system perspectives, or discounted or non discounted (Table 3). A series of one-way sensitivity analysis showed that the most influential parameter was discount rate. When the discount rate was varied from 3% to 0% and 6%, the ICER of PWTM compared with usual care was shifted to 85,603 THB (2,853 USD) and 154,612 THB (5,154 USD) per QALY, respectively (Fig. 2). In the probabilistic sensitivity analysis, at a 150,000 THB (5,000 USD) per QALY threshold, 49.6% of the simulations were cost effective and at

Table 2 Results of base-case analysis (n = 1,000).⁎ Strategy

Health system perspective Usual Care PWTM Societal perspective Usual Care PWTM

Cost (THB)

Effectiveness (QALYs)

Incremental Cost (ΔTHB)

Incremental Effectiveness (ΔQALYs)

ICER (ΔTHB/ΔQALYs)

56,016,778.00 148,804,433.73

38,710.65 39,504.78

92,787,655.73

794.13

116,841.85

73,580,805.06 166,071,599.84

38,710.65 39,504.78

92,490,794.78

794.13

116,468.03

THB, Thai Baht; QALY, quality-adjusted life-year; PWTM, pharmacist-participated warfarin therapy management; ICER, incremental cost-effectiveness ratio. ⁎ Discounted at 3% per year.

S. Saokaew et al. / Thrombosis Research 132 (2013) 437–443

441

Discount rate (0% - 6%) Cost of warfarin clinic setup and training (68,000 - 92,000 THB) Probability of death with bleeding (3-month) (3.39% - 7.65%) Utility of sequelae of TE (0 - 0.7) Probability of TE (3-month) (0.68% - 0.92%) Probability of bleeding (3-month) (0.95% - 1.29%) Probability of sequelae from TE (3-month) (19.07% - 25.79%) Cost of bleeding (per episode) (32,727 - 73,941 THB) Pharmacist service cost (3-month) (408 - 522 THB) Utility of sequelae of bleeding (0.370 - 0.870) Utility of TE (0.271 - 1) Cost of sequelae of TE (3-month) (2,273 - 17,204 THB) Cost of sequelae of bleeding (3-month) (2,273 - 17,204 THB) Utility of warfarin use (no event) (0.967 - 0.998) Utility of bleeding (0.711 - 0.962) Probability of death with TE (3-month) (6.11% - 8.27%) Probability of sequelae from bleeding (3-month) (3.15% - 4.26%) Cost of TE (per episode) (41,102 - 46,808 THB) Probability of death with sequelae of bleeding (3-month) (3.79% - 5.73%) Probability of death with sequelae of TE (3-month) (4.82% - 6.52%) Cost of INR test (per test) (70 - 95 THB) Cost of transportation (per visit) (122 - 169 THB) Cost of additional food (per visit) (43 - 64 THB)

Using the low parameter value Using the high parameter value

85,000

105,000

125,000

145,000

165,000

Incremental cost per quality-adjusted life year gained (THB/QALY gained) Fig. 2. Tornado diagram showing a series of one-way sensitivity analyses comparing PWTM and usual care. The horizon bars represent the range of the incremental cost-effectiveness ratio (ICER) for one-way sensitivity over the range of parameters in parenthesis. The wider the horizon bar, the more uncertainty that parameter introduces. The vertical line represents the base-case ICER.

Discussion To the best of our knowledge, this is the first formal costeffectiveness study evaluating PWTM in comparison with UC in Thailand. This is also the only study reflecting the Asia-Pacific context since clinical probabilities in previous studies were derived mostly from European and North American context [5]. We developed a model to help inform policy decisions and populated it with the latest available data. This cost-effectiveness study revealed that PWTM is very cost-effective with ICER of 116,468 THB (3,882.3 USD) per QALY.

Incremental cost (x1000 THB)

250

200

150

100

50

Our findings are consistent with results in previous analyses of the cost-effectiveness of AMS [5,6]. You et al. [5] evaluated costeffectiveness in two models of management for patients on chronic warfarin therapy using a Markov model during 10-year time horizon. They found that anticoagulation clinic was cost-effective compared to routine medical care from the perspective of public health organisation. In addition, Sullivan et al. [6] estimated the lifetime societal costs and health benefits of warfarin therapy to prevent stroke in atrial fibrillation 80

Probability that PWTM is cost-effective (%)

450,000 THB (15,000 USD) per QALY level, 65.5% of the simulations were cost effective from societal perspective (Figs. 3, 4).

70 60

-10

0

10

20

30

40

49.6%

50 40 30 20

3 GDP

1 GDP

10 0

0 -20

65.5%

0

100

200

300

400

500

600

700

800

900 1000 1100

Willingess to pay (x1000 THB)

Incremental effectiveness (QALY gained) Fig. 3. Cost-effectiveness scatter plot. Each point represents the incremental cost (year 2012 values) and QALYs between PWTM and usual care from the Monte Carlo simulation. PWTM, pharmacist-participated warfarin therapy management; QALY, quality-adjusted life year; THB, Thai Baht.

Fig. 4. Cost-effectiveness acceptability curve. The curves provide the probability of PWTM being the cost-effective at any willingness to pay value for an additional QALY. The curve was generated from the Monte Carlo simulation. PWTM, pharmacist-participated warfarin therapy management; QALY, quality-adjusted life year; GDP, gross domestic product; THB, Thai Baht.

442

S. Saokaew et al. / Thrombosis Research 132 (2013) 437–443

and at high risk of stroke patients in the U.S. using Markov decision model with 30-days cycle length and 10-year time horizon. They found that anticoagulation management service improved effectiveness and reduce costs resulted in 91% of Monte Carlo simulation was dominant. Considering PWTM as an additional service to usual care in this analysis, the results could have been more pronounced if PWTM is considered as a substitute to the usual care. This is due to the fact that, exclusion of the cost of physician in the PWTM arm would result in an even lower cost of PWTM than what was used in our model. Key strengths of our study should be highlighted. First, the input parameters for effectiveness of PWTM were derived from the best recent meta-analysis of RCT comparing PWTM and other health care providers on TE and bleeding outcomes [3]. In addition, we used input parameters derived from local Thai data or Asia-Pacific information as much as possible to reflect the Asian population and can be implemented to healthcare benefit package in this region. We also did the intensive sensitivity analysis to determine the robustness of the results. Several potential limitations should be acknowledged. First, the most important issue is limited availability of data. Although data used to construct the model were based on Thai database and literature in Asia-Pacific countries whenever possible, some data were derived from Western countries particularly the probability of sequelae, and some utilities. However, these data were not much influential to the finding of the study as illustrated by one-way sensitive analysis. The utility of sequelae derived from the systematic review [32] was the only one parameter that influenced the ICER in the model. However, the ICER still falls within the acceptable threshold and it is important to note that borrowing utility value from the literature is not an uncommon practice [43]. Second, cost of TE and bleeding, and the probability of death due to bleeding were calculated from BCRH database. This might not be generalisable to other settings or other levels of healthcare setting because the use of only one source of data may not represent the whole population of warfarin users in Thailand. Another important issue is that the case-fatality rate of bleeding was based on what being reported within hospitalization as death. Given an anecdotal report indicating that patients prefer to die at home, it is possible that a number of patients may not be documented as death in hospital. So, the actual case-fatality rate may be higher than those used in this model. Fig. 2, however, illustrates that the higher probabilities of deaths are, the lower ICERs would be realized. This means that PWTM might be more cost-effective if probabilities of deaths are higher. Third, the ability of our results to be extrapolated to other countries may be uncertain since the incidence of TE and bleeding, age-specific mortality rate, and cost of care and service differ between countries, and all have an impact on cost-effectiveness results. However, it is likely that countries especially those countries in Asia-Pacific region with similar characteristics to those retained in our analysis might show similar cost-effectiveness results. Despite limitations, our findings revealed that PWTM illustrated the clinical and economic benefits for the management of patients receiving warfarin therapy. Please note that our study used percapita GDP concept as a willingness to pay (WTP) for consideration of strategy to be cost effective. If another concept such as 1.2 times gross national income (GNI) per capita (160,000 THB or 5,333 USD) [44] was used, the conclusion of economic values of strategy might be different. Fortunately, even we used GDP or GNI concept as a WTP, PWTM remained lower than the threshold that to be considered cost effective. At present, economic evaluation study has been widely accepted and used to inform policy decisions for health benefits packages in Thailand (e.g. the implementation of smoking cessation program at the community pharmacy) [45–47]. Thus, this present study is the valuable information for policy-makers to be considered as part of information in their decision-making process for implementing this strategy into health care system and reallocate health care resources toward greater promotion of PWTM.

Conclusion On the basis of currently available data, this cost-effectiveness analysis suggests that PWTM is very cost-effective strategy in Thai population compared with usual care for the management of patients receiving long-term warfarin therapy. Policy-makers may consider our finding as part of information in their decision-making process. Conflicts of Interest Statement No potential conflicts of interest exist. Acknowledgements We would like to thank the Office of the Higher Education Commission, Ministry of Education, Thailand, for supporting by grant fund under the program Strategic Scholarships for Frontier Research Network for the Ph.D. Program Thai Doctoral degree for this research. We also thanks to Buddhachinaraj Regional Hospital for giving permission and providing database to be analyzed for input parameters and Miss Ubonwan Sapoo for suggestions on PWTM program. References [1] Ansell J, Hirsh J, Hylek E, Jacobson A, Crowther M, Palareti G. Pharmacology and management of the vitamin K antagonists: American College of Chest Physicians Evidence-Based Clinical Practice Guidelines (8th Edition). Chest 2008;133:160S–98S. [2] Pirmohamed M. Warfarin: almost 60 years old and still causing problems. Br J Clin Pharmacol 2006;62(5):509–11. [3] Saokaew S, Permsuwan U, Chaiyakunapruk N, Nathisuwan S, Sukonthasarn A. Effectiveness of pharmacist-participated warfarin therapy management: a systematic review and meta-analysis. J Thromb Haemost 2010;8:2418–27. [4] Saokaew S, Sapoo U, Nathisuwan S, Chaiyakunapruk N, Permsuwan U. Anticoagulation control of pharmacist-managed collaborative care versus usual care in Thailand. Int J Clin Pharmacol 2012;34(1):105–12. [5] You JHS, Chan FWH, M. Wong RS, Cheng G. Cost-effectiveness of two models of management for patients on chronic warfarin therapy – a Markov model analysis. Thromb Haemost 2003;90:1106–11. [6] Sullivan PW, Arant TW, Ellis SL, Ulrich H. The cost effectiveness of anticoagulation management services for patients with atrial fibrillation and at high risk of stroke in the US. Pharmacoeconomics 2006;24(10):1021–33. [7] Meckley LM, Gudgeon JM, Anderson JL, Williams MS, Veenstra DL. A policy model to evaluate the benefits, risks and costs of warfarin pharmacogenomic testing. Pharmacoeconomics 2010;28(1):61–74. [8] Permsuwan U, Guntawongwan K, Buddhawongsa P. Handling time in economic evaluation studies. J Med Assoc Thai 2008;91(Suppl. 2):S53–8. [9] Teerawattananon Y, Chaikledkaew U. Thai health technology assessment guideline development. J Med Assoc Thai 2008;91(Suppl. 2):S11–5. [10] Jenghua K, Jedsadayanmata A. Rate and predictors of early readmission among Thai patients with heart failure. J Med Assoc Thai Jul 2011;94(7):782–8. [11] Chirakup S, Chaiyakunapruk N, Chaikledkeaw U, Pongcharoensuk P, Ongphiphadhanakul B, Roze S, et al. Cost-effectiveness analysis of thiazolidinediones in uncontrolled type 2 diabetic patients receiving sulfonylureas and metformin in Thailand. Value Health Mar 2008;11(Suppl. 1):S43–51. [12] Chaiyakunapruk N, Thanarungroj A, Cheewasithirungrueng N, Srisupha-olarn W, Nimpitakpong P, Dilokthornsakul P, et al. Estimation of financial burden due to oversupply of medications for chronic diseases. Asia Pac J Public Health May 2012;24(3):487–94. [13] Chaiyakunapruk N, Asuphol O, Dhippayom T, Poowaruttanawiwit P, Jeanpeerapong N. Statins utilisation pattern: a retrospective evaluation in a tertiary care hospital in Thailand. Int J Pharm Pract Apr 2011;19(2):129–35. [14] Briggs A, Sculpher M. An introduction to Markov modelling for economic evaluation. Pharmacoeconomics 1998;13(4):397–409. [15] Connock M, Stevens C, Fry-Smith A, Jowett S, Fitzmaurice D, Moore D, et al. Clinical effectiveness and cost-effectiveness of different models of managing long-term oral anticoagulation therapy: a systematic review and economic modelling. Health Technol Assess 2007;11(No. 38). [16] Limwattananon S. Report 3: Acute myocardial infarction, ischemic and hemorrhagic strokes in universal coverage 2008-2010. Bangkok: Health Insurance System Research Office; 2011. [17] Fang MC, Go AS, Chang Y, Hylek EM, Henault LE, Jensvold NG, et al. Death and disability from warfarin-associated intracranial and extracranial hemorrhages. Am J Med Aug 2007;120(8):700–5. [18] Patrick AR, Avorn J, Choudhry NK. Cost-effectiveness of genotype-guided warfarin dosing for patients with atrial fibrillation. Circ Cardiovasc Qual Outcomes Sep 2009;2(5):429–36. [19] World Health Organization. Global Health Observatory Data Repository. [cited 2012 June 17]; Available from: http://apps.who.int/ghodata/?vid=61640; 2011.

S. Saokaew et al. / Thrombosis Research 132 (2013) 437–443 [20] Friberg L, Hammar N, Pettersson H, Rosenqvist M. Increased mortality in paroxysmal atrial fibrillation: report from the Stockholm Cohort-Study of Atrial Fibrillation (SCAF). Eur Heart J Oct 2007;28(19):2346–53. [21] Riewpaiboon A. Measurement of costs. J Med Assoc Thai 2008;91(Suppl. 2):S28–37. [22] Bureau of Trade and Economic indices, Ministry of Commerce Thailand. Report of consumer price index of Thailand. [cited 2012 Dec 11]; Available from: www. price.moc.go.th/content1.aspx?cid=1. [23] Bank of Thailand. Foreign Exchange Rates. [cited 2013 Jan 14]; Available from: http://www.bot.or.th/english/statistics/financialmarkets/exchangerate/_layouts/ Application/ExchangeRate/ExchangeRateAgo.aspx; 2013. [24] Ministry of Finance. [cited 2012 Jun 10]; Available from: http://www.mof.go.th/ home/salary.html; 2012. [25] Riewpaiboon A. Standard cost lists for health technology assessment. [cited 2012 Dec 9]; Available from: http://www.hitap.net/costingmenu/; 2010. [26] Drugs, Medical Supplies Information Center (DMSIC), Ministry of Public Health. Reference drug price. [cited 2011 Sep 9, 2011]; Available from: http://dmsic.moph. go.th/price/price1_1.php?method=drug; 2011. [27] Archongka Y, Manimmanakorn N, Kuptniratsaikul V, Solunda S, Yee-heng P. Unit cost of stroke rehabilitation. J Med Assoc Thai Aug 2008;91(8):1257–62. [28] Singhpoo K, Tiamkao S, Ariyanuchitkul S, Sangpongsanon S, Kamsa-ard S, Lekbunyasin O, et al. The expenditures of stroke outpatients at Srinagarind hospital. Srinagarin Med J 2009;24(1):54–9. [29] Singhpoo K, Tiamkao S, Kuchaisit C, Ariyanuchitkul S, Sangpongsanon S, Kamsa-ard S, et al. The quality of life of stroke outpatients at Srinagarind Hospital. J Med Assoc Thai Dec 2009;92(12):1602–9. [30] Ara R, Brazier J. Deriving an algorithm to convert the eight mean SF-36 dimension scores into a mean EQ-5D preference-based score from published studies (where patient level data are not available). Value Health Dec 2008;11(7):1131–43. [31] Eckman MH, Rosand J, Greenberg SM, Gage BF. Cost-effectiveness of using pharmacogenetic information in warfarin dosing for patients with nonvalvular atrial fibrillation. Ann Intern Med 2009;150:73–83. [32] Post PN, Stiggelbout AM, Wakker PP. The utility of health states after stroke: a systematic review of the literature. Stroke 2001;32:1425–9. [33] Chaiyawat P, Kulkantrakorn K. Effectiveness of home rehabilitation program for ischemic stroke upon disability and quality of life: A randomized controlled trial. Clin Neurol Neurosurg Feb 7 2012;114(7):866–70.

443

[34] Christensen MC, Mayer S, Ferran JM. Quality of life after intracerebral hemorrhage: results of the Factor Seven for Acute Hemorrhagic Stroke (FAST) trial. Stroke May 2009;40(5):1677–82. [35] Gage BF, Cardinalli AB, Owens DK. The effect of stroke and stroke prophylaxis with aspirin or warfarin on quality of life. Arch Intern Med Sep 9 1996;156(16):1829–36. [36] The Civil Service Commission NR 1008.1/V31. [cited 2012 May 11]; Available from: http://www.ocsc.go.th/ocsc/th/uploads/circular/ccl0000537.pdf; 2010. [37] WHO Commission on Macroeconomic and Health. Macroeconomics and health: investing in health for economic development. Geneva: World Health Organization; 2001. [38] The World Bank. GDP per capita (current US$). [cited 2012 Dec 29]; Available from: http://data.worldbank.org/indicator/NY.GDP.PCAP.CD; 2011. [39] National Statistical Office. Socio-Economic Survey project; 2009 [Bangkok]. [40] Limwattananon S. Handling uncertainty of the economic evaluation result: sensitivity analysis. J Med Assoc Thai Jun 2008;91(Suppl. 2):S59–65. [41] Briggs AH. Handling uncertainty in cost-effectiveness models. Pharmacoeconomics May 2000;17(5):479–500. [42] Zethraeus N, Johannesson M, Jonsson B, Lothgren M, Tambour M. Advantages of using the net-benefit approach for analysing uncertainty in economic evaluation studies. Pharmacoeconomics 2003;21(1):39–48. [43] Luangasanatip N, Chaiyakunapruk N, Upakdee N, Wong P. Iron-chelating therapies in a transfusion-dependent thalassaemia population in Thailand: a cost-effectiveness study. Clin Drug Investig 2011;31(7):493–505. [44] The International Bank for Reconstruction and Development/The World Bank. Economy Profile: Thailand. [cited 2012 Dec 9]; Available from: http://espanol. doingbusiness.org//media/fpdkm/doing%20business/documents/profiles/country/ THA.pdf; 2012. [45] Teerawattananon Y, Tantivess S, Yothasamut J, Kingkaew P, Chaisiri K. Historical development of health technology assessment in Thailand. Int J Technol Assess Health Care Jul 2009;25(Suppl. 1):241–52. [46] Mohara A, Youngkong S, Velasco RP, Werayingyong P, Pachanee K, Prakongsai P, et al. Using health technology assessment for informing coverage decisions in Thailand. J Comp Eff Res 2012;1(2):137–46. [47] Thavorn K, Chaiyakunapruk N. A cost-effectiveness analysis of a community pharmacist-based smoking cessation programme in Thailand. Tob Control Jun 2008;17(3):177–82.