Methods: Resource utilization and outcome data was collected prospectively on patients presenting to St. Paul's hospital for surgical treatment for HPT, 2002-2005. The primary measure of effectiveness was the rate of complications (hypocalcemia, paresthesias) post-surgery. Net benefits were compared between the three treatment options using a range of cost-effectiveness threshold (lambda) values. Non-parametric bootstrapping was applied to evaluate uncertainty around estimates of costs and effectiveness, from which cost-effectiveness acceptability curves (CEACs) were derived for each treatment strategy within a range of lambda values.
Results: Patient-level data on a total of 94 patients (BNE=50, UNE=19, MIP=25) provided estimates of mean costs between treatment arms (BNE=$4843; SE=($944), UNE=$4881 ($519), MIP=$5954 ($842)) as well as estimates of rates of complications (BNE=0.10, UNE=0.16, MIP=0.04). Subjects from the 3 treatment groups did not differ significantly with regard to age, sex, mean pre- and post-operative calcium levels, or hospital length of stay. The gold standard BNE strategy displayed 1st-order stochastic dominance over the UNE strategy, and 2nd-order stochastic dominance over the MIP strategy at lambda=$15,000. The CEAC showed that up to lambda=$5,000 per complication avoided, the BNE strategy was the most cost-effective with no uncertainty. It was only at lambda=$30,000 per complication avoided that a (marginally) higher proportion of iterations of net benefits of the MIP strategy were higher.
Conclusion: Our results suggest that in the experience of HPT surgery at St. Paul's Hospital, newer, costlier strategies of treatment of HPT may be less cost-effective than the gold standard Bilateral Neck Exploration.