OBJECTIVES: In 2009 Yoo and Grosse (Y/G) compared screening/no screening for CAH in a CEA, including a PSA using triangular distributions with minimum (MIN), mode and maximum (MAX) values for parameters.  These distributions have been criticized for use in PSAs (Briggs et al).  This research  reproduces the Y/G analysis and then conducts a PSA with more appropriate distributional assumptions to evaluate potential bias from triangular distributions. METHODS: We limited our reanalysis to a change in distributions to focus on the question of distributional sensitivity. For parameters in Y/G with symmetric triangular distributions (MIN and MAX values equidistant from their mean/mode), we used a normal distribution with the same mean and chose a Standard Error based on a 95% confidence interval defined by the MIN and MAX values of the triangular.  For non-symmetric distributions, we fit a beta distribution to the triangular, keeping the means equal and ensuring that the tails of the distributions extended slightly beyond the MIN and MAX of the triangular. RESULTS: We reproduced the Y/G deterministic ICER ($292,841 vs. Y/G reported $292,000).  We also reproduced the triangular distribution-based PSA ICER from Y/G (256,947 vs Y/G $255,700).  In our revised PSA, our mean ICER was higher ($273,187), but the CEACs were nearly superimposable.  For low Willingness to Pay (WTP) values, the probability of being cost-effective - P(CE) - was very similar between the two analyses. There was some divergence at higher WTPs. For a WTP of $270,000, the P(CE) for the no screen option was .482 with the TRI distributions and .535 with our revised distributions. CONCLUSIONS: In this case, despite the criticism of triangular distributions generally, CEA results were not appreciably different qualitatively or quantitatively with the change to more accepted distributional assumptions.  For an intervention with an ICER closer to the CE threshold, the importance may be greater.