^{*}School of Pharmacy, Sungkyunkwan University, Suwon, Korea.
^{†}College of Pharmacy, Kyungpook National University, Daegu, Korea.
^{‡}Department of Internal Medicine, Seoul National University Bundang Hospital, Seoul National University College of Medicine, Seongnam, Korea.
This study was performed to investigate the cost effectiveness of
A decision Markov model was used to estimate the effectiveness and economic impact of an
The total costs per patient were US $2,454 for the
The
This study was performed to investigate the cost effectiveness of
A decision Markov model was used to estimate the effectiveness and economic impact of an
The total costs per patient were US $2,454 for the
The
The prevalence of
A recent systematic review reported that the annual incidence of peptic ulcer disease (PUD) in Western countries is 0.10% to 0.19% for physician-diagnosed PUD cases and 0.03% to 0.17% based on hospitalization.^{3} Over time, the incidence or prevalence of PUD has slightly decreased, which may have been caused by decreased rates of
Eradication of
There are other reasons for the preference of
Based on this background, the aim of the present study was to evaluate the cost-effectiveness of
A combined decision tree and Markov process model were used to estimate the effectiveness and economic impact of a
For the base model, patients entered the model at the age of 40 years old. The 3-month long cycle reflected disease processes such as treatment or exacerbation. Patients in this model died based on the death transition probability due to ulcer complication or natural causes depending on age-stratified life expectancy. It was assumed that all of these individuals died before the age of 100 years.
There were five possible health statuses into which a patient could transit: well, ulcer, ulcer complication (bleeding), death due to bleeding, and death due to other cause.
The definitions of ulcer, ulcer complication, death due to bleeding, death due to other cause were as follows: Well is no gastric intestinal symptoms in addition to no endoscopic lesion required treatments; Ulcer is defined endoscopical ulcer of which diameter at least 5 mm which needs treatments; Ulcer complication is defined as perforated and bleeding peptic ulcers; Death due to bleeding is mortality due to ulcer complication; Death due to other cause means natural death due to other causes than ulcer complication.
The probability of disease transition for each health status was obtained from the literature.^{10-12} For the present study, values for all causes of mortality rate were obtained through the Complete Life Table published in 2009 by the South Korean Statistical Office,^{13} and values for death rate due to complication were derived from published data sources.^{11} Transition into the next health status depended on the current health status. This model was established after several meetings with clinicians and methodologists to reflect real-world situations. The basic model is described in Fig. 1. The conceptual framework for this Markov process model was similar to that of a previous study.^{5}
We used South Korean-specific
Since South Korean data for the transition probability for PUD development and complications when comparing
Utility weights were applied to five health statuses (i.e., well, ulcer, complication, death due to complication, and death due to other cause) in this model to reflect quality adjusted life year (QALY). Since utility weight worsens with aging, participants designated with "well" and "ulcer" status had the utility weights depending on their age. Specific utility weights for applicable ages were obtained using the 2007 to 2009 Korean National Health and Nutritional Examination Survey Data, which was a national survey representative of the South Korean population.^{15} For example, the utility weights for the "well" and "ulcer" status at 40 years of age were 0.931 and 0.889, respectively, using the same database.^{15} The utility weights of complication during the initial 3 months of "bleeding" status was 0.460 based on a previous report.^{16} Death was valued as zero regardless of the cause (Table 1).
The screening/eradication strategy was associated with screening and eradication costs; the no-screening strategy was not. Costs for the screening strategy were determined by multiplying each health examinations by unit costs. Screening costs included costs for physician visits, endoscopic examination, and biopsy for the
Patients with a "well" status had no associated costs because they did not have any clinical symptoms. Costs for treating ulcers with bleeding or without bleeding were drawn from a previous report,^{19} In this South Korean report, the annual cost for treating ulcers without bleeding was around US $930 (benign gastric ulcers, US $959.6; duodenal ulcer, US $901.4). When bleeding occurred due to PUD then the cost increased 2.6 times compared to ulcers without bleeding (benign gastric ulcers, US $2,553.1; duodenal ulcers, US $2,316.4).^{19} These costs were divided by four before being entered into the model to account for the periods for each cycle. Costs for death were entered as zero (Table 1). All costs are adjusted for 2011 currency values.
In the base model, the cost-effectiveness of the
To evaluate the impact of the uncertainty of input data, probabilistic sensitivity analysis with 10,000 simulations was carried out. Costs and disease transition probabilities were assumed to have a gamma distribution while utility weights were assumed to have a triangular distribution. From the applicable distribution for cost, transition probability, and utility weight, one random value was drawn, and then incremental cost and QALY were calculated. If the process is repeated 10,000 times, 10,000 plot values are drawn in the plane of the incremental cost and QALY. Using this scatter plot, we could determine whether the study results were robust or not. A discount rate of 3% was applied to both the effectiveness and cost. TreeAge 2009 (TreeAge Software Inc., Williamstown, MA, USA) and Microsoft Excel 2007 (Microsoft Corp., Redmond, CA, USA) were used for these analyses.
Total costs per patient for
Results of different cohort group analysis for participants aged 30, 50, and 60 years old were similar to those of base analysis. The
The probabilistic sensitivity analysis with 10,000 simulations showed that incremental QALYs (i.e., QALYs for the
In the present study, a
In a randomized study performed in China with a sample size of 100 patients, showed that,
Second, the
This study had several limitations. First, we used meta-analysis results to determine the rates of ulcers and ulcer-related complications among NSAIDs or aspirin users depending on
In conclusion, the present study found that the
NSAIDs, nonsteroidal anti-inflammatory drugs.
Model Input Parameters
Data are presented as percentage or transition probabilities.
^{*}SD was calculated using the mean value of ±25%.
^{*}SD was calculated using the mean value of ±25%.
Base Case Analysis and Sensitivity Analysis According to Cohort Starting Age and NSAIDs Naïve-Patients
NSAIDs, nonsteroidal anti-inflammatory drugs; QALY, quality-adjusted life years; ICER, incremental cost-effectiveness ratio.
^{*}SD was calculated using the mean value of ±25%.
Model Input Parameters
Data are presented as percentage or transition probabilities.
^{*}SD was calculated using the mean value of ±25%.
^{*}SD was calculated using the mean value of ±25%.
Base Case Analysis and Sensitivity Analysis According to Cohort Starting Age and NSAIDs Naïve-Patients
NSAIDs, nonsteroidal anti-inflammatory drugs; QALY, quality-adjusted life years; ICER, incremental cost-effectiveness ratio.
^{*}SD was calculated using the mean value of ±25%.
NSAIDs, nonsteroidal anti-inflammatory drugs.