The authors constructed a one-year decision tree that modelled the likelihood of virologic suppression, ART retention, injectable drug use, loss to follow-up, and death, with ART, based on the two methadone maintenance delivery options. Patients who became resistant to first-line ART were allowed to progress to second-line ART. Third-line ART was not modelled as it was not commonly available in Vietnam. The authors stated that a Vietnamese health services perspective was taken.
The key effectiveness data were the odds ratios for having virologic suppression. The data for the baseline risk of virologic suppression with ART were from a Swiss study. Those for the two other interventions and for active injectable drug users, were from several non-Vietnamese, non-randomised studies. Other inputs were ART retention, mortality, drug use rates, and the rates of loss to follow-up or ART, for each strategy. For ART alone, these data were from the Vietnamese Ministry of Health. The same retention rate was used for all three options. The other clinical data were from Europe, USA, or Thailand. The model assumed that there was no difference in ongoing drug use between the same-site methadone and the separate-site methadone options.
Monetary benefit and utility valuations:
The utility values were estimated for each treatment, and for continued use of injectable drugs or for drug abstinence. These utilities were derived from some published Vietnamese studies.
Measure of benefit:
The primary measure of benefit was quality-adjusted life-years (QALYs), which were discounted at a 5% annual rate.
The costs included the drugs, personnel, management, laboratory tests, and site operation and maintenance. Site operation and maintenance costs were assumed to be identical for same-site and for separate-site therapy. The costs were from published Vietnamese sources and they were discounted at a 5% annual rate. The costs were converted from Vietnamese dong to US $ and inflated to 2009 prices.
Analysis of uncertainty:
The authors conducted a probabilistic sensitivity analysis, using Monte Carlo simulation. The results were presented as cost-effectiveness planes; in a cost-effectiveness acceptability curve; and in a table. Confidence ellipses and threshold lines were presented for a willingness-to-pay for a QALY that was derived from World Health Organization guidelines for cost-effectiveness in developing countries.