The analysis was conducted on an intention-to-treat basis: all costs and outcomes were ascribed to the initial drug used, regardless of downstream events. The economic outcome was the expected cost per patient treated for each regimen (i.e., once started on the antibiotic). Clinical outcomes calculated were the expected rates of success (i.e., the sum of all rates across all branches of the tree that ended in success) and expected rates of amputation (calculated in a similar manner). In the case of dominance, the expected cost per success was calculated for each drug. In the case of incremental cost and incremental benefit, the incremental cost-effectiveness ratio (ICER) was calculated.
One-way and two-way sensitivity analyses were performed to test the robustness of the decision-tree model, with variation in clinical success rates, costs of antibiotics, and length of hospital stay. Probabilistic sensitivity analyses were also performed using Monte Carlo simulations over 10 000 iterations across presumed distributions of variables (mean ± 10%). We used log-normal distributions for all antibiotic costs and beta distributions for clinical success rates.
Several assumptions were made in this model. First, the model examined the cost-effectiveness of oral antibiotic regimens for mild foot infections, which generally excluded patients with osteomyelitis, gangrene, or extensive deep tissue infections. However, moderate to severe infections, which might have included osteomyelitis, gangrene, and deep tissue infections, were considered if the 10-day course of oral antibiotic failed. Low resistance rates were assumed, and the most common organisms involved with mild infections were assumed to be aerobic gram-positive microorganisms, specifically staphylococci and streptococci. Another assumption was that once oral antibiotic therapy had failed, the patients were admitted to hospital for the full 14 days before the outcome of clinical success, amputation, or death occurred. The possibility of parenteral antibiotics or step-down therapy to oral antibiotics (to complete therapy on an outpatient basis) was considered but not included in this model, because once oral therapy had failed and the patients were admitted to hospital, the infections were considered moderate to severe and potentially limb-threatening, which could lead to an over- estimation of costs. To test whether shorter hospital stays would change the decision of the model, a sensitivity analysis was performed. The last assumption of the model was that adverse effects were low. This assumption was based on the studies used to construct the model, which reported low incidences of adverse effects and mild to moderate adverse effects that resolved spontaneously without treatment; in addition, the rate of discontinuation of therapy because of antibiotic-associated adverse effects was low. However, this may represent a limitation, since these studies had small sample sizes and were underpowered, rather than reflecting current incidences of antibiotic-associated adverse events.
canadian cialis online