OBJECTIVES: Advanced sensitivity methods including value of information were developed to quantify overall decision uncertainty and to assess the cost-effectiveness of additional research that would reduce that uncertainty.  Our objective was to compare the information gained by utilizing three alternative sensitivity methods with increasing complexity: a simple one-way sensitivity analysis; probabilistic analysis of covariance (ANCOVA); and expected value of partial perfect information (EVPPI) of input parameters. METHODS: We replicated and expanded a published HIV/AIDS cost-effectiveness Markov model (zidovudine vs. zidovudine plus lamivudine in the UK) using TreeAge®.  Health states included three HIV/AIDS states and death.  Our outcome of interest was the incremental net monetary benefit (INMB) assuming a willingness-to-pay of £20,000/QALY.  We generated one-way and probabilistic sensitivity analyses of the INMB using published input parameter uncertainties.  One-way sensitivity analysis identified the 10 most influential parameters. A total of 10,000 Monte Carlo draws were used to estimate the ANCOVA results from the same ten parameters.  EVPPI for each of the same ten parameters was estimated specifying 1000 inner and 1000 outer Monte Carlo draws.  We ranked the parameters based on their influence on variation for each sensitivity method and compared them using Spearman’s rank correlation.   RESULTS: Mean INMB was £9694 in favor of combination therapy.  The two most influential inputs were the same across all methods, contributed 78% of variation in outcome (ANCOVA), and were the only inputs with non-zero EVPPI values.  The rank order for the top ten inputs from all methods was similar (correlation=0.99 for one-way vs. ANCOVA, 0.70 for one-way vs. EVPPI and 0.70 for ANCOVA vs. EVPPI, all p-values <0.05). CONCLUSIONS: The correlation was significant between one-way and more advanced sensitivity analyses.  Although each method provides unique information, the additional resources needed to generate advanced analyses should be weighed, especially when the outcome decision uncertainty and therefore value of information is low.