Analytical approach:
The analysis was based on a Markov model, with a five-year horizon. The authors stated that the perspective of society was adopted.
Effectiveness data:
The data on the effectiveness of acupuncture were identified by a systematic review of the literature, in standard and Korean databases, from 1999 to 2009. The inclusion and exclusion criteria were partly reported and 16 studies were included (10 clinical trials, five systematic reviews, and one health technology assessment). A meta-analysis was undertaken to pool the evidence from multiple sources. The data on the natural history of disease came from other studies that included a prospective cohort study. Korean life tables were used for mortality.
Monetary benefit and utility valuations:
The utility values were from two sources; a subgroup analysis of 2007 Korean National Health and Nutrition Survey (KNHNS) data, and a published pragmatic (clinical practice) clinical trial conducted in Germany.
Measure of benefit:
Quality-adjusted life-years (QALYs) were the summary benefit measure and they were discounted at an annual rate of 5%.
Cost data:
The economic analysis included the direct medical costs of treatments (visits, diagnosis, and drugs), as well as non-medical costs and indirect costs, such as travel expenses and waiting time. The costs of medical treatment were from the South Korean National Health Insurance Reimbursement for standard medical procedures. The resource use data were from published pragmatic clinical trials. Other economic data were from the 2006 KNHNS database. The unit costs and quantities of resources used for the direct medical costs were reported for most items. Average wages were used to assess the indirect costs. Costs were in Korean won (KRW) and in US dollars ($). The price year was 2009 and a 5% annual discount rate was applied.
Analysis of uncertainty:
Deterministic one-way and probabilistic analyses were carried out. The indirect non-medical costs were excluded in the base case and included in the sensitivity analysis. The expected value of perfect information (EVPI) was calculated, using a probabilistic approach that assumed 57,400 cases per year, based on 2007 KNHNS data.