PharmacoEconomics (2015) 33:521–531 DOI 10.1007/s40273-015-0269-8

ORIGINAL RESEARCH ARTICLE

Cost-Utility Analysis of Intravenous Immunoglobulin for the Treatment of Steroid-Refractory Dermatomyositis in Thailand Naruemon Bamrungsawad • Nathorn Chaiyakunapruk Nilawan Upakdee • Chayanin Pratoomsoot • Rosarin Sruamsiri • Piyameth Dilokthornsakul

•

Published online: 14 March 2015  Springer International Publishing Switzerland 2015

Abstract Introduction Intravenous immunoglobulin (IVIG) has been shown to be effective in treating steroid-refractory dermatomyositis (DM). There remains no evidence of its cost-effectiveness in Thailand. Objective Our objective was to estimate the cost utility of IVIG as a second-line therapy in steroid-refractory DM in Thailand. Methods A Markov model was developed to estimate the relevant costs and health benefits for IVIG plus Electronic supplementary material The online version of this article (doi:10.1007/s40273-015-0269-8) contains supplementary material, which is available to authorized users. N. Bamrungsawad
N. Chaiyakunapruk
N. Upakdee
P. Dilokthornsakul Department of Pharmacy Practice, Faculty of Pharmaceutical Sciences, Naresuan University, Phitsanulok, Thailand N. Chaiyakunapruk
R. Sruamsiri
P. Dilokthornsakul Department of Pharmacy Practice, Faculty of Pharmaceutical Sciences, Center of Pharmaceutical Outcomes Research, Naresuan University, Phitsanulok, Thailand N. Chaiyakunapruk (&) School of Pharmacy, Monash University Malaysia, Jalan Lagoon Selatan, 46150 Bandar Sunway, Selangor, Malaysia e-mail: [email protected] N. Chaiyakunapruk School of Population Health, University of Queensland, Brisbane, Australia N. Chaiyakunapruk School of Pharmacy, University of Wisconsin-Madison, Madison, WI, USA C. Pratoomsoot Faculty of Public Health, Naresuan University, Phitsanulok, Thailand

corticosteroids in comparison with immunosuppressant plus corticosteroids in steroid-refractory DM from a societal perspective over a patient’s lifetime. The effectiveness and utility parameters were obtained from clinical literature, meta-analyses, medical record reviews, and patient interviews, whereas cost data were obtained from an electronic hospital database and patient interviews. Costs are presented in $US, year 2012 values. All future costs and outcomes were discounted at a rate of 3 % per annum. One-way and probabilistic sensitivity analyses were also performed. Results Over a lifetime horizon, the model estimated treatment under IVIG plus corticosteroids to be cost saving compared with immunosuppressant plus corticosteroids, where the saving of costs and incremental quality-adjusted life-years (QALYs) were $US4738.92 and 1.96 QALYs, respectively. Sensitivity analyses revealed that probability of response of immunosuppressant plus corticosteroids was the most influential parameter on incremental QALYs and costs. At a societal willingness-to-pay threshold in Thailand of $US5148 per QALY gained, the probability of IVIG being cost effective was 97.6 %. Conclusions The use of IVIG plus corticosteroids is cost saving compared with treatment with immunosuppressant plus corticosteroids in Thai patients with steroid-refractory DM. Policy makers should consider using our findings in their decision-making process for adding IVIG to corticosteroids as the second-line therapy for steroid-refractory DM patients.

522

Key Points for Decision Makers Intravenous immunoglobulin (IVIG) plus corticosteroids could be perceived as a cost-saving second-line treatment for steroid-refractory patients with dermatomyositis (DM) in Thailand. Policy makers should consider using our findings in their decision-making process for adding IVIG to corticosteroids as the second-line therapy for steroidrefractory DM patients.

N. Bamrungsawad et al.

In 2011, it was proposed that IVIG be added with additional indications (of which aggressive DM was one) to the Thailand National List of Essential Medicines (NLEM). The main issues are the high cost of IVIG and the chronic nature of DM, requiring patients to use it intermittently throughout their lifetimes. Thus, it has significant implications for the national budget. To date, no evidence has been reported for the cost effectiveness of these immunotherapeutic agents in steroid-refractory DM. The objective of this study was to estimate the cost utility of IVIG as a second-line therapy in steroid-refractory DM patients in Thailand.

2 Methods 1 Background

2.1 Overall Description of Cost-Utility Analysis

Dermatomyositis (DM) is an idiopathic inflammatory disorder characterized by inflammation of the skeletal muscles, with proximal muscle weakness and a distinctive skin rash [1]. It is a rare disease, with a low incidence and prevalence. The incidence estimates range from 0.14 to 1.4 cases per 100,000 person-years [2–7], and prevalence ranges from 5.9 to 21.42 per 100,000 population [2, 4, 6]. DM affects all ages, but females are affected twice as often as males [5]. The onset of the disease in adults is most common in those aged 50–70 years [8]. The life-threatening nature of DM as a result of muscle weakness and visceral involvement is associated with increased morbidity and mortality rates [9, 10]. Progressive muscle weakness may cause permanent disability. Approximately 15–20 % of cases of DM are associated with malignancy [2, 11]. The goal of DM therapy is to improve muscle strength and function in activities of daily living and to ameliorate muscular manifestations [12]. Corticosteroids are the preferred first-line treatment [10, 13, 14]. The goal of this therapy is to maintain steroid-responsive patients in a state of remission at the lowest dose with the lowest number of adverse effects (AEs). To achieve this, a steroid-sparing immunosuppressant, such as azathioprine or methotrexate, may be added [15, 16]. However, a number of studies have shown that about 30–50 % of patients using combinations of prednisone, immunosuppressants, and plasma exchange became resistant to the aforementioned therapies and remained disabled [17]. In the disease state where immunosuppressive therapies are inadequate to produce a response or improve muscle strength, intravenous immunoglobulin (IVIG) may be recommended as a second-line drug [13, 15, 16, 18].

A cost-utility analysis was performed to estimate the costs and health outcomes of IVIG plus corticosteroids versus immunosuppressant plus corticosteroids over a lifetime in steroid-refractory DM patients. A Markov model was built based on a hypothetical cohort of patients with steroid-refractory DM (defined as no objective benefit or no remission within 3 months of starting steroid therapy [13, 14]) at the age of 40 years based on the average age of patients in clinical studies [19, 20]. The treatment sequence used in the model was based on literature review [13–16, 18, 21] and was confirmed by two neurologists as being applicable in Thailand. This study was undertaken using a societal perspective as recommended by Thailand’s Health Technology Assessment (HTA) guideline [22]. A cost-utility analysis was performed to arrive at the incremental cost per quality-adjusted life-year (QALY) gained for IVIG plus corticosteroids versus immunosuppressant plus corticosteroids. The study protocol was approved by the Ethics Committees of Naresuan University and Ramathibodi Hospital, Mahidol University. In order to define the scope of the study, key stakeholders, including representatives from the Royal College (neurologists) and the pharmaceutical industry as well as policy makers, payers, patients, and health economists, were invited to participate in two meetings. The first meeting focused on defining the scope of the study and model approval. The model was developed to mimic disease progression in DM and clinical practice. It was judged and revised through consultations with two neurologists with expertise in DM and a health economist. The second meeting was hold to reach agreement on model input values and the sources of all parameters.

CUA of IVIG in Steroid-Refractory DM Patients

523

Fig. 1 Structure of IVIG plus corticosteroid treatment arm. DM dermatomyositis, IVIG intravenous immunoglobulin

2.2 Economic Model Figure 1 shows the model structure for IVIG plus corticosteroids. The model consisted of four health states, including ‘symptom/no response/relapse’ (S), ‘remission’ (RM), ‘disability’ (DB), and ‘death’ (D). The health state ‘S’ was categorized into three sub-health states based on response to initial treatment as (1) initial treatment (S1), (2) no response/relapse (S2) (including no response to the treatment or relapse after remission), and (3) no response after additional IVIG or immunosuppressant (S3). Response was assessed with a modified Medical Research Council (MRC) scale to measure muscle strength. It was performed on 18 proximal muscle groups and graded on a scale of 0–5, with a healthy state represented by a score of 90 [19, 20, 23]. Remission was defined as spontaneous improvement of DM symptoms, and an increase of total MRC scores of five or more [19, 20]. Disability was defined as someone who is unable to perform several routine activities of daily living due to physical problem(s) (expert opinion). Relapse was defined as recurrence of disease activity (as determined by clinical or biochemical parameters) after remission in the absence of an alternative etiology [24, 25]. For IVIG plus corticosteroids treatment, all patients entered the model at the initial IVIG plus corticosteroids health state (S1). Each state represented a 12-week cycle based on the duration of assessment and follow-up period [19, 20]. After the end of initial treatment, patients who responded to treatment moved to the remission state (RM), and were assumed to switch to corticosteroids alone.

Patients not responding to initial treatment were assumed to receive additional IVIG plus corticosteroids (S2) for another cycle. After the second cycle of IVIG, patients who responded to the additional IVIG plus corticosteroids moved to the remission state (RM), and were assumed to receive corticosteroids alone for the subsequent 12-week cycles until they relapsed or died (D). Patients who relapsed after corticosteroids alone were assumed to receive additional IVIG plus corticosteroids (S2) for a cycle. Patients not responding to additional IVIG treatment (S2) were assumed to switch to immunosuppressant plus corticosteroids (S3) until they were disabled (DB) or dead (D). The model structure for immunosuppressant plus corticosteroids treatment was similar to that of IVIG plus corticosteroids, except that treatments with IVIG were replaced with an immunosuppressant in states S1 and S2, and an immunosuppressant was added to corticosteroids in the RM state. The economic model was based on the following assumptions: • •

•

The population entering the models was not disabled. The long-term use of IVIG was not associated with a declining effect of the drug; the dose of IVIG was thus assumed not to increase in order to give the same therapeutic effect. All patients would have a chance of receiving IVIG of at least twice, and when an individual did not respond to IVIG in state S2, the individual would not have a chance to receive the next IVIG. This was based on the evidence that additional IVIG is not likely to be

524

• •

•

•

•

•

•

N. Bamrungsawad et al.

effective if there is no improvement by the end of the second infusion [13, 26]. All patients would have a chance to receive life-long IVIG if they responded to IVIG after all relapses. After IVIG use for induction of remission, if the patients responded to IVIG, the treatment would be switched to corticosteroids alone for maintenance therapy until they relapsed. This was based on draft DM guidelines for IVIG use, reflecting the real practice in Thailand. Furthermore, patients who were steroid resistant may respond again to steroids after a few infusions of IVIG [27]. Immunosuppressants used in these models included azathioprine and methotrexate. The proportion of immunosuppressant use between azathioprine and methotrexate was 50:50 based on real-world practice in Thailand (expert opinion). Patients were assumed to be disability free during the RM state, which was based on expert opinion where patients in the RM state would respond well to treatment, resulting in spontaneous symptom improvement. Patients who received immunosuppressant plus corticosteroids after not responding to additional IVIG (S3) would not have a chance to return to the RM state. This was because treatment with immunosuppressant plus corticosteroids during the S3 sub-health state to control symptoms and slow the progression to disability was inadequate to substantially increase muscle strength or return to complete remission [14, 28]. Based on a lack of evidence that continued therapy with low-dose corticosteroids or an immunosuppressant can prevent further disease progression in an unresponsive patient [13], it was assumed that disabled patients did not respond to any immunotherapeutic agents and would not have a chance to receive any immunotherapeutic agents. The AEs were assumed to have no effects on costs, mortality, or quality of life (QOL) because IVIGrelated AEs were generally minimal and tolerable [10, 28, 29].

2.3 Clinical and Utility Data Clinical and utility data were derived from literature reviews, medical records, patient interviews, and expert opinion. The process for eliciting expert opinion was included in the aforementioned stakeholder meetings. We used a consensus elicitation method to obtain expert opinion by asking two neurologists to provide model input values and sources of model parameters.

2.3.1 Response Rates The relative response rate for IVIG compared with placebo for steroid-refractory DM was taken from a meta-analysis of two randomized controlled trials (RCTs) [19, 20] in which patients were randomized to receive IVIG or placebo over a 12-week period. The pooled relative response to IVIG was 2.09 (95 % confidence interval [CI] 0.97–4.50) with heterogeneity (I2 = 53.3 %) (Table 1) (Appendix Supplement 1). In terms of the response rate for immunosuppressant plus corticosteroids, no previous studies have been conducted to compare immunosuppressant alone or in combination with corticosteroids versus placebo or IVIG. We assumed that the response rate for immunosuppressant plus corticosteroids was similar to the pooled placebo response rate of 42.3 %, which was calculated from Dalakas et al. [19] and Miyasaka et al. [20]. The pooled placebo response rate was not statistically significantly different from the pooled response rate of immunosuppressants (cyclosporine and methotrexate), which was calculated from Vencovsky´ et al. [30] (Appendix Supplement 1). 2.3.2 Relapse Rates Only the study by Ponyi et al. [31] reported a relapse rate and was able to be translated into time to relapse. The author plotted cumulative relapse-free proportions of 35 adult DM patients against the follow-up time using the Kaplan–Meier method. The 86-month relapse rate in this curve was 52 %, which was equivalent to a 12-week relapse rate of 2.53 %. The constant relapse rate of 2.53 % for IVIG plus corticosteroids responder health state (from state RM to S2) was assumed to be equivalent to that for the immunosuppressant plus corticosteroids responder health state in each cycle (Table 1). 2.3.3 Disability In order to reflect the Thai context, we derived disability data from those of DM patients from a medical school hospital where we reviewed the medical records from January 2009 to December 2011. Given the rare condition, we included patients with DM, polymyositis (PM), or chronic inflammatory demyelinating polyneuropathy (CIDP) to obtain time to disability because progressions of disability in these three conditions were similar according to disease characteristics and expert opinion. We used the MRC scale as a proxy of disability outcome evaluation because patient data related to the definition of disability could not be extracted from their medical records. A

CUA of IVIG in Steroid-Refractory DM Patients

525

Table 1 Model parameters in base-case analysis Model parameters

Value (95 % CI)

Distribution

Data source

Efficacy RR response to IVIG

2.091 (0.970–4.500)

Log-normal

Meta-analysis of two studies [19, 20]

Prob. of response to immunosuppressant ? steroids

0.423 (0.130–0.716)

Beta

Assumption from a pooled analysis of two studies [19, 20]

Prob. of relapse in pts responding to IVIG ? steroids

0.025 (0.011–0.036)

Beta

Ponyi et al. [31]

Prob. of relapse in pts responding to immunosuppressant ? steroids

0.025 (0.011–0.036)

Beta

Ponyi et al. [31]

0.098 (–)

Beta

Medical record reviews

Beta

Sigurgeirsson et al. [11]

Beta

MOPH [32]

Disability Prob. of becoming disabled in pts non-responding to therapy Mortality Standardized mortality ratios of DM

2.10 (1.70–2.50)

Age-specific mortality in Thai population Utility Pre-disabled

0.677 (0.575–0.779)

Beta

Interview

Disabled

0.105 (–)

–

Interview

Costs ($US, year 2012 values) Direct medical costs Drug cost (per 12 weeks) IVIG 5 % w/v (5 g/100 mL/vial)a

Gamma

NHSO

Prednisolone 5 mga

3525.74 (2996.88–4054.60) 10.70 (9.10–12.31)

Gamma

DMSIC [39]

Azathioprine or methotrexate (average)a

54.47 (46.30–62.64)

Gamma

DMSIC [39]

OPD service cost (per year) Pre-disabled

1672.92 (1421.98–1923.85)

Gamma

Hospital database

Disabled

1326.17 (1127.24–1525.09)

Gamma

Hospital database

423.95 (360.35–487.54)

Gamma

Hospital database

420.21 (357.18–483.24)

Gamma

Hospital database

IPD service cost (per year) Pre-disabled Disabled Direct non-medical costs Transportation cost (per visit)

4.62 (3.89–5.36)

Gamma

Riewpaiboon [40]

Meal cost (per visit)

1.70 (1.36–2.04)

Gamma

Riewpaiboon [40]

Accompanying person real income loss (per visit)

3.10 (0.85–5.35)

Gamma

Riewpaiboon [40]

Facilities at home (per year) Pre-disabled Disabled

48.26 (41.02–55.50)

Gamma

Interview

214.50 (182.33–246.68)

Gamma

Interview

3096.86 (2632.33–3561.39)

Gamma

Interview

Nursing home (per year) Disabled

CI confidence interval, DM dermatomyositis, DMSIC Drugs and Medical Supplies Information Center, IPD inpatient, IVIG intravenous immunoglobulin, MOPH Ministry of Public Health, NHSO The National Health Security Office, OPD outpatient, prob. probability, pts patients, RR relative risk a

IVIG 100 g/course (2 g/kg/course, 50 kg), prednisolone 60 mg/day, azathioprine 100 mg/day, methotrexate 20 mg/week

disabled patient was defined as an MRC scale for one muscle group of B3. As a result, among 99 patients, 28/46 (60.87 %), 19/24 (79.17 %), and 24/29 (82.76 %) of DM, PM, and CIDP patients were disabled, respectively (International Classification of Diseases, version 10 [ICD-10] codes were M331 for DM, M332 for PM, and G618 and

G619 for CIDP). No significant difference between times to disability was observed for patient groups using the logrank test. The time to disability for each disease was assumed to be similar and was pooled together to calculate the probability of becoming disabled in patients not responding to therapy. Exponential survival analysis was

526

fitted to these data to allow extrapolation. The probability of becoming disabled in patients not responding to therapy in all health states was 0.098 (Table 1). 2.3.4 Mortality The risk of death was estimated by multiplying the relative risk by the Thai age-specific mortality rate (ASMR) [32]. A number of studies have shown increased mortality in patients with DM, but reports have not been consistent. A wide range of standardized mortality ratios (SMRs) or relative risks have been reported from 2.10 to 7.63 [11, 33, 34]. In base-case analysis, we used SMR of 2.10, which was derived from data from a large population-based cohort including 392 patients with DM in Sweden, from 1963 through 1983 [11]. The SMRs from other studies were not used because of a mixed population of DM and PM patients [34], and a high mortality rate (SMR 7.68) [33], which was considered to be an outlier when compared with previous studies [11, 34]. 2.3.5 Utilities No previous studies have reported DM-specific utility data or any health-related quality of life (HRQoL) suitable for conversion into utility values. Consequently, utility values were derived from the Thai EuroQol-5D (EQ-5D) questionnaire [35] in patients at a medical school hospital from April to May 2012. Given the rare nature of DM, which affects a small proportion of the population, we included patients with DM and PM to obtain utility data. Based on disease characteristics and expert opinion, we assumed that QOL for DM and PM would be similar in terms of progression of disability and severity. The utility data were obtained from pre-disabled and disabled patients. No significant difference was found between utility scores of patient groups using a non-parametric Mann–Whitney test. The utility scores for each disease in pre-disabled patients were assumed to be similar and were pooled together to calculate the utility scores. The utility scores of pre-disabled and disabled patients were 0.677 and 0.105, respectively (Table 1) (Appendix Supplement 2). 2.4 Cost Data This study was undertaken from the societal perspective; therefore, cost data included both direct medical and direct non-medical costs. We assumed that lost or impaired ability to work or engage in leisure activities due to morbidity would be captured in the disutility of QALYs; therefore, indirect costs were not included so as to avoid double counting [36]. In our model, we calculated health resource use and healthcare costs from different sources,

N. Bamrungsawad et al.

including a hospital database, local standard costs, and patient interviews. Costs were divided into two main groups that consisted of costs in pre-disabled and disabled patients to reflect the difference in costs between the two groups of patients. All costs were converted to $US, year 2012 values ($US1 = Thai baht [THB] 31.08), using the consumer price index (CPI) [37]. Direct medical costs included drug costs and outpatient department (OPD) and inpatient department (IPD) visit costs. OPD and IPD costs were obtained from the outpatient and inpatient department database of a medical school hospital from 2009 to 2011. These data included all records of resource use for OPD and IPD medical visits. Of 69 patients, 28/45 (62.22 %) and 19/24 (79.17 %) of DM and PM patients were disabled, respectively. The frequencies of OPD medical visits were 19.25 and 15.26 visits per year, and that of IPD medical visits were 0.33 and 0.32 visits per year in pre-disabled and disabled patients, respectively. OPD service costs were $US1672.92 and $US1326.17 per year, whereas IPD service costs were $US423.95 and $US420.21 per year in pre-disabled and disabled patients, respectively (Table 1). No significant difference was observed in medical costs between the patient groups. Annual direct medical costs of pre-disabled and disabled patients with DM and PM were calculated from actual charges from the medical school hospital database. A cost-to-charge ratio of 0.73, derived from a Ministry of Public Health Network of Unit Cost report in 2011 [38], was applied to convert hospital charges to costs. The drug cost was based on dose, frequency, and cost per treatment. The IVIG dose was based on clinical guidelines for DM treatment [13–16, 18, 21]. The cost for IVIG was provided by the National Health Security Office (NHSO) in 2011 values (5 % w/v 100 mL/vial = $US176.29). Based on a 2 g/kg/course [14, 28] and a patient weight of 50 kg, the total cost per IVIG administration was calculated as $US3525.74 per 12-week cycle. The cost for corticosteroids was obtained from the reference price database of the Drugs and Medical Supplies Information Center (DMSIC) in 2010 values (prednisolone $US0.01/5 mg tablet) [39]. Based on a prednisolone dosage of 60 mg/day [13, 14], the total cost was calculated as $US10.70 per 12-week cycle. The costs for immunosuppressants were also obtained from the DMSIC in 2010 values (azathioprine $US0.55 per 50 mg tablet, methotrexate $US0.18 per 2.5 mg tablet) [39]. Based on a 50:50 proportion of immunosuppressant use and dosages of azathioprine 100 mg/day and methotrexate 20 mg/week [13, 14], the total cost for immunosuppressants was calculated as $US54.47 per 12-week cycle. Direct non-medical costs, including those for transportation, meals, and accompanying person real income loss, were obtained from standard cost lists from the HTA

CUA of IVIG in Steroid-Refractory DM Patients

by Riewpaiboon [40] and multiplied by total number of visits for each patient, as obtained from the hospital database. These were $US4.62, $US1.70, and $US3.10 per visit, respectively. Other direct non-medical costs, including accessories, home renovation, and nursing home, were obtained through questionnaires completed by DM and PM patients and their relatives. Among ten patients, 5/7 (71.43 %), 1/3 (33.33 %) of DM and PM patients were predisabled, respectively. The total costs of facilities per year were $US48.26 and $US214.50 in pre-disabled and disabled patients, respectively, and the total costs of nursing homes for disabled patients was $US3096.86 per year. All cost parameters are presented in Table 1.

527

using tornado diagrams (Figs. 2, 3). Probabilistic sensitivity analysis (PSA) was conducted to simultaneously examine the effects of all parameter uncertainty using a Monte Carlo simulation, performed using Microsoft Excel 2003 [42] and run for 1000 simulations. The results of the PSA are presented as cost-effectiveness acceptability curves (Figs. 4, 5). The expected net monetary benefit (NMB) was calculated for WTP threshold of the NLEM 2013 in Thailand to show the probability that IVIG is cost effective for monetary values that a decision maker might be willing to pay. We also performed threshold analysis by varying the response rate of immunosuppressants to determine the change in ICER.

2.5 Analyses 3 Results 2.5.1 Base-Case Analysis 3.1 Base-Case Results The Markov model was simulated to calculate the expected lifetime costs and outcomes for each intervention in steroid-refractory DM patients. The results were presented as incremental cost per QALY gained for IVIG plus corticosteroids versus immunosuppressant plus corticosteroids in second-line therapy to determine an incremental costeffectiveness ratio (ICER). Both costs and effects were discounted at a rate of 3 % per annum. The interpretation of cost effectiveness of the findings was based on the willingness-to-pay (WTP) threshold set by the Subcommittee for Development of the NLEM in Thailand in 2013 of 1.2 times the gross national income (GNI) per capita per QALY gained (or $US5148 or THB160,000 per QALY gained) [41]. 2.5.2 Sensitivity Analysis One-way sensitivity analyses were performed to identify the effects of altering parameters within plausible ranges, including all clinical effects, costs, utilities, and discount rates. The one-way sensitivity analysis results are presented Fig. 2 A tornado diagram showing the change in the incremental QALYs attributable to the change of each individual parameter. DM dermatomyositis, IVIG intravenous immunoglobulin, Prob. probability, QALYs quality-adjusted life-year

Under base-case analysis over a patient’s lifetime, the total costs for the IVIG plus corticosteroids and immunosuppressant plus corticosteroids were $US45,386 and $US50,124, respectively, while the estimated QALYs were 6.297 and 4.333, respectively. The model estimated the saving of costs and incremental QALYs for IVIG plus corticosteroids compared with immunosuppressant plus corticosteroids to be $US4739 and 1.964 QALYs, respectively. Treatment with IVIG plus corticosteroids was both more effective and less costly; therefore, it is considered dominant (Table 2). 3.2 Sensitivity Analyses Figures 2 and 3 show the results of the one-way sensitivity analyses. The most influential parameter on incremental QALY and cost was the probability of response to immunosuppressant plus corticosteroids. When the probability of response to immunosuppressant plus corticosteroids was varied from 0.423 to 0.130 and 0.716,

528

N. Bamrungsawad et al.

Fig. 3 A tornado diagram showing the change in the incremental cost attributable to the change of each individual parameter. DM dermatomyositis, IPD inpatient, IVIG intravenous immunoglobulin, OPD outpatient, Prob. probability

Fig. 4 Cost-effectiveness plane of IVIG plus corticosteroids in base-case scenario. IVIG intravenous immunoglobulin, QALY quality-adjusted life-year

the incremental QALY shifted to 3.290 and 0.685 QALY, and the incremental cost shifted to -$US11,305 and $US1600, respectively. Figure 4 is a graphical representation (scatter plot) of the incremental costs and QALYs for IVIG plus corticosteroids versus immunosuppressant plus corticosteroids. The majority of incremental cost and QALY pairs fell to the lower-right quadrant. This shows that IVIG plus corticosteroids is cost effective (dominant). Figure 5 presents the results of the PSA in the form of cost-effectiveness acceptability curves. This shows the probability that an intervention is cost effective at different decision thresholds. At a societal WTP threshold in Thailand of 1.2 GNI per capita per QALY gained (or $US5148

or THB160,000 per QALY gained), the probability that IVIG plus corticosteroids is cost effective is 97.6 %. In addition, the threshold analysis showed that the use of IVIG plus corticosteroids was changed from dominant to not cost saving and not cost effective when the response rate to the immunosuppressant was increased by approximately 52 and 80 %, respectively.

4 Discussions The use of IVIG plus corticosteroids as a second-line therapy was both more effective and less costly than

CUA of IVIG in Steroid-Refractory DM Patients

529

Fig. 5 Cost-effectiveness acceptability of IVIG plus corticosteroids versus immunosuppressant plus corticosteroids in the base-case scenario. IVIG intravenous immunoglobulin, QALY qualityadjusted life-year

Table 2 Results from base-case analysis Intervention

Cost

Effectiveness (QALYs)

Incremental costa

Incremental effectiveness (QALYs)

ICER ($US/ QALY)

IV immunoglobulin plus corticosteroids

45,385.60

6.297

-4738.92

1.964

IVIG dominant

Immunosuppressant plus corticosteroids

50,124.52

4.333

ICER incremental cost-effectiveness ratio, IV intravenous, QALY, quality-adjusted life-year a

$US, year 2012 values

immunosuppressant plus corticosteroids for steroid-refractory DM patients. It could be perceived as a cost-saving treatment for steroid-refractory DM patients throughout the patient’s lifetime in Thailand, given the WTP value. Our study is the first economic study evaluating the cost effectiveness of adding IVIG to corticosteroids in steroidrefractory DM patients. A cost-effectiveness study conducted in DM patients was previously published; however, it was conducted in juvenile DM, compared intermittent high-dose intravenous corticosteroids with oral corticosteroids, and was not specific for steroid-refractory DM patients [43]. We believe that the cost-saving findings result mainly from the effect of IVIG, which is twofold higher than that of an immunosuppressant. We recognize that the assumption that the response rate for the immunosuppressant is equivalent to that of the placebo is a major assumption in this study. Therefore, we conducted a threshold analysis and found that the response rate needed to increase approximately 80 % to reach the WTP threshold. We thus ensured that our assumption did not influence our findings. Our findings showed that the results were robust to changes in utility value of disability as demonstrated by

one-way sensitivity analysis. The low utility score for the disabled state when compared with the pre-disabled state may result from difficulties in self-care and activities of daily living experienced by disabled patients who might assess their level of severity to be high as a result of limited public facilities (i.e. transportation, toilet) in Thailand. In addition, 55 % of patients with myositis, including DM, had moderate (requires some help, but able to walk unassisted) to severe (requires constant nursing care and attention, bedridden, incontinent) disability [44]. Therefore, disabled DM patients could have severe problems with self-care and usual activities, resulting in a low utility value. This emphasizes the need for policy makers to give steroid-refractory DM patients access to IVIG. The strength of our study is the high validity of our findings for several reasons. First, key stakeholders such as representatives of the Royal College, policy makers, payers, patients, and researchers were invited to participate in two meetings to share their ideas and concerns in relation to the study. Second, where possible, we endeavored to use local data to reflect the Thai context in the analysis. We obtained utility data and direct non-medical costs from patient interviews, and time to disability from medical

530

record reviews at a medical school hospital. These ensure our results are more reliable for the Thai context. Finally, our study was conducted in accordance with Thailand’s HTA guideline [22], and the reported results were also assessed by peer reviewers who were members of the Health Economic Working Group under the Subcommittee for Development of the NLEM. A number of assumptions and limitations in our study should be discussed. First, DM is a rare condition affecting a small number of the population. We therefore included patients with DM, PM, and CIDP to obtain some input parameters. To our knowledge and from the expert opinion received, the disability progressions in the three conditions are similar; therefore, we obtained time to disability for those diseases. In addition, input parameters related to disability and severity, including utility data and direct non-medical costs from DM and PM patients were pooled because their level of severity was similar. However, further studies conducted specifically in DM patients are needed to ensure the credibility of the research results. Second, efficacy data used in the model were obtained from RCTs with short-term follow-up [19, 20]. This would only show the immediate impact of treatment on outcomes. Future clinical studies may require longer periods of follow-up to measure the long-term effects. However, when the upper bound of the most influential parameter (the probability of response to treatment in immunosuppressant plus corticosteroids assumed from the pooled placebo response rate of IVIG studies) [19, 20] was used in the model, IVIG plus corticosteroids was still cost effective at a WTP threshold in Thailand. Third, the costs of treatment for AEs were not considered in the study. In particular, corticosteroids and immunosuppressants may result in more adverse drug events or side effects. However, we believe that the use of IVIG plus corticosteroids would still be cost effective, even when parameters related to AEs were incorporated. The reason is that available data indicate that IVIG has a good safety profile [10, 28, 29], particularly in comparison with immunosuppressants or corticosteroids. Finally, the direct medical costs of the interventions were obtained from a medical school hospital database. It may not reflect actual clinical practice for all institutional practice sites in Thailand. However, only a small number of hospitals have specialists in rheumatology or dermatology with experience in diagnosing and treating DM. It seems reasonable to assume that it represents hospitals providing DM patient treatment services. Our findings suggest that IVIG plus corticosteroids is cost saving for steroid-refractory DM patients as a secondline therapy in base-case analysis. Policy makers should consider using our results to support decision making. However, cost-effectiveness analysis is not the only tool in

N. Bamrungsawad et al.

the decision-making process. A budget-impact analysis should be conducted to predict future expenditure.

5 Conclusions The use of IVIG plus corticosteroids was both more effective and less costly than immunosuppressant plus corticosteroids for steroid-refractory DM patients. It could be perceived as a cost-saving treatment for steroid-refractory DM patients throughout the patient’s lifetime. Policy makers should consider using our findings when making decisions about adding IVIG to corticosteroids as a secondline therapy for steroid-refractory DM patients. Acknowledgments This work was funded by the Subcommittee for Development of the NLEM, Thailand. The authors thank the experts and stakeholders for their helpful comments and suggestions. Finally, we wish to thank the patients for their participation, and Ramathibodi Hospital, Bangkok, Thailand and the National Health Security Office (NHSO) for database management. Conflict of interest The authors declare that they have no conflicts of interest. Author contributions NB participated in the design of the study, developed the cost-effectiveness model, conducted stakeholder meetings, searched input parameters, and conducted the cost-effectiveness analyses. NC participated in the design of the study, developed the cost-effectiveness models, identified data sources, conducted stakeholder meetings, and conducted the cost-effectiveness analyses. NU and CP participated in the design of the study and identified data sources. RS and PD participated in the design of the study, searched input parameters, and conducted the cost-effectiveness analyses. Each author also contributed to interpreting data and results, drafting the manuscript, critically reviewing the manuscript for important intellectual content, and has read and approved the final version.

References 1. Callen JP, Wortmann RL. Dermatomyositis. Clin Dermatol. 2006;24(5):363–73. 2. Bendewald MJ, Wetter DA, Li X, Davis MD. Incidence of dermatomyositis and clinically amyopathic dermatomyositis: a population-based study in Olmsted County, Minnesota. Arch Dermatol. 2010;146(1):26–30. 3. Prieto S, Grau JM. The geoepidemiology of autoimmune muscle disease. Autoimmun Rev. 2010;9(5):A330–4. 4. Reeder MJ, Wetter DA, Li X, Davis MDP. Incidence of dermatomyositis and clinically amyopathic dermatomyositis: a population-based study in Olmsted County, Minnesota. Arch Dermatol. 2010;146(1):26–30. 5. Cox S, Limaye V, Hill C, Blumbergs P, Roberts-Thomson P. Idiopathic inflammatory myopathies: diagnostic criteria, classification and epidemiological features. Int J Rheum Dis. 2010;13(2):117–24. 6. Furst DE, Amato AA, Iorga S¸ R, Gajria K, Fernandes AW. Epidemiology of adult idiopathic inflammatory myopathies in a U.S. managed care plan. Muscle Nerve. 2012;45(5):676–83.

CUA of IVIG in Steroid-Refractory DM Patients 7. Limaye V, Blumbergs P, Scott G, Hakendorf P, Roberts-Thomson PJ. The epidemiology of dermatomyositis in South Australia. APLAR J Rheum. 2007;10(2):94–100. 8. Yamasaki Y, Yamada H, Ohkubo M, et al. Longterm survival and associated risk factors in patients with adult-onset idiopathic inflammatory myopathies and amyopathic dermatomyositis: experience in a single institute in Japan. J Rheumatol. 2011;38(8):1636–43. 9. Dalakas MC, Hohlfeld R. Polymyositis and dermatomyositis. Lancet. 2003;362(9388):971–82. 10. Marie I, Mouthon L. Therapy of polymyositis and dermatomyositis. Autoimmun Rev. 2011;11(1):6–13. 11. Sigurgeirsson B, Lindelof B, Edhag O, Allander E. Risk of cancer in patients with dermatomyositis or polymyositis. A populationbased study. N Engl J Med. 1992;326(6):363–7. 12. Dalakas MC. Polymyositis, dermatomyositis and inclusion body myositis. In: Kasper DL, Braunwald E, Fauci AS, et al., editors. Harrison’s principles of internal medicine. 17th ed. New York: McGraw-Hill; 2008. p. 2696–703. 13. Dalakas MC. Immunotherapy of myositis: issues, concerns and future prospects. Nat Rev Rheumatol. 2010;6(3):129–37. 14. Dalakas MC. Inflammatory myopathies: management of steroid resistance. Curr Opin Neurol. 2011;24(5):457–62. 15. Dalakas MC. Immunotherapy of inflammatory myopathies: practical approach and future prospects. Curr Treat Options Neurol. 2011;13(3):311–23. 16. Dalakas MC. Therapeutic advances and future prospects in immune-mediated inflammatory myopathies. Ther Adv Neurol Disord. 2008;1(3):157–66. 17. Cherin P. Recognition and management of myositis. Drugs. 1997;54(1):39–49. 18. Dalakas MC. Therapeutic targets in patients with inflammatory myopathies: present approaches and a look to the future. Neuromuscul Disord. 2006;16(4):223–36. 19. Dalakas MC, Illa I, Dambrosia JM, et al. A controlled trial of high-dose intravenous immune globulin infusions as treatment for dermatomyositis. N Engl J Med. 1993;329(27):1993–2000. 20. Miyasaka N, Hara M, Koike T, Saito E, Yamada M, Tanaka Y. Effects of intravenous immunoglobulin therapy in Japanese patients with polymyositis and dermatomyositis resistant to corticosteroids: a randomized double-blind placebo-controlled trial. Mod Rheumatol. 2012;22(3):382–93. 21. Iorizzo LJ 3rd, Jorizzo JL. The treatment and prognosis of dermatomyositis: an updated review. J Am Acad Dermatol. 2008;59(1):99–112. 22. Thai Working Group on Health Technology Assessment Guidelines in Thailand. Health Technology Assessment Guideline. J Med Assoc Thai. 2008;91(suppl 2):S1–88. 23. Rider LG, Werth VP, Huber AM, et al. Measures for adult and juvenile dermatomyositis, polymyositis, and inclusion body myositis. Arthritis Care Res (Hoboken). 2011;63(Suppl 11):S118–57. 24. Phillips BA, Zilko P, Garlepp MJ, Mastaglia FL. Frequency of relapses in patients with polymyositis and dermatomyositis. Muscle Nerve. 1998;21(12):1668–72. 25. Dourmishev LA, Dourmishev AL. Dermatomyositis: advances in recognition, understanding and management. 1st ed. Berlin: Springer; 2009. 26. Dalakas MC. Intravenous immunoglobulin in autoimmune neuromuscular diseases. JAMA. 2004;291(19):2367–75.

531 27. Dalakas MC. The role of high-dose immune globulin intravenous in the treatment of dermatomyositis. Int Immunopharmacol. 2006;6(4):550–6. 28. Wang DX, Shu XM, Tian XL, et al. Intravenous immunoglobulin therapy in adult patients with polymyositis/dermatomyositis: a systematic literature review. Clin Rheumatol. 2012;31(5):801–6. 29. Donofrio PD, Berger A, Brannagan TH 3rd, et al. Consensus statement: the use of intravenous immunoglobulin in the treatment of neuromuscular conditions report of the AANEM ad hoc committee. Muscle Nerve. 2009;40(5):890–900. 30. Vencovsky´ J, Jarosova´ K, Macha´cek S, Study´nkova´ J, Kafkova´ J, Bartu˚nkova´ J. Cyclosporine A versus methotrexate in the treatment of polymyositis and dermatomyositis. Scand J Rheumatol. 2000;29(2):95–102. 31. Ponyi A, Constantin T, Balogh Z, et al. Disease course, frequency of relapses and survival of 73 patients with juvenile or adult dermatomyositis. Clin Exp Rheumatol. 2005;23(1):50–6. 32. Health Information Group. Public Health Statistics A.D.2010. Bangkok: Bureau of Policy and Strategy, Ministry of Public Health (Thailand). Available from: http://bps.ops.moph.go.th/Health information/statistic53/statistic53.pdf. Accessed 4 Mar 2015. 33. Kuo CF, See LC, Yu KH, et al. Incidence, cancer risk and mortality of dermatomyositis and polymyositis in Taiwan: a nationwide population study. Br J Dermatol. 2011;165(6):1273–9. 34. Airio A, Kautiainen H, Hakala M. Prognosis and mortality of polymyositis and dermatomyositis patients. Clin Rheumatol. 2006;25(2):234–9. 35. Tongsiri S. The Thai population based preference scores for EQ5D health states: Ministry of Public Health; 2009. 36. Brouwer W, Rutten F, Koopmanschap M. Costing in Economic Evaluations. In: Drummond M, McGuire A, editors. Economic evaluations in health care: merging theory with practice. New York: Oxford University Press; 2001. 37. Bank of Thailand. EC_EI_027: Thailand’s Macro Economic Indicators. Available from: http://www2.bot.or.th/statistics/Report Page.aspx?reportID=409&language=eng. Accessed 5 July 2014. 38. Tienjaruwattana W, Sakulpanich T, Pongpattarachai D, Chiangchaisakulthai K, editors. Unit cost manual for hospital. 1st ed. Bangkok: MUCC, Ministry of Public Health Network of Unit Cost; 2011. 39. Drug Medical Supply and Information Center (DMSIC). Reference price database of the Drugs and Medical Supplies Information Center (DMSIC). Bangkok: Ministry of Public Health (Thailand). Available from: http://dmsic.moph.go.th/dmsic/index. php?p=1&type=3&s=3&id=drug_normal. Accessed 4 Mar 2015. 40. Riewpaiboon A. Standard cost lists for health economic evaluation in Thailand. J Med Assoc Thai. 2014;97(supp l5):S127–34. 41. Thavorncharoensap M, Leelahavarong M, Doungthipsirkul S, Sompitak S, Teerawattananon Y. Assessing a societal value for a ceiling threshold in Thailand. Nonthaburi, Thailand: Health Intervention and Technology Assessment Program (HITAP), Ministry of Public Health; 2013. 42. Briggs A, Sculpher M, Claxton K. Decision modelling for health economic evaluation. Oxford: Oxford University Press; 2006. 43. Klein-Gitelman MS, Waters T, Pachman LM. The economic impact of intermittent high-dose intravenous versus oral corticosteroid treatment of juvenile dermatomyositis. Arthritis Care Res. 2000;13(6):360–8. 44. van de Vlekkert J, Hoogendijk JE, de Visser M. Long-term followup of 62 patients with myositis. J Neurol. 2014;261(5):992–8.