STATISTIK-FORUM.de

[Franz]

Lässt sich die ausgeschnittene Passage aus der beigefügten Publikation (pharmakoökonomische Analyse) auch so formulieren, dass Sie ein Nicht-Fachmann versteht? Es muss doch möglich sein, die Vorgehensweise (bzw. deren Begründung) umgangssprachlich so anschaulich darzustellen, dass man sie versteht, oder?

“To address shortcomings in performing only univariate sensitivity analysis, we also performed probabilistic sensitivity analysis (PSA) taking the form of a Monte Carlo simulation (MCS) in which key model parameters were simultaneously varied by replacing parameter estimates with an appropriate distribution [2]. For example, parameters with respect to the proportion of once daily, twice daily injections are bounded on the interval 0 – 100, so it would be inappropriate to apply a distribution that gave a non-negligible value outside of that range. In this case, a beta distribution was specified to reflect the normal distribution and restriction to values between zero and 100 (%). With respect to insulin consumption, variation in resource use values (final dose per kg) was not reported in the trial itself. A range of values was therefore approximated assuming a 10% variation of the mean final dose. For MCS, a gamma distribution was fitted by estimating the standard deviation. For unit costs a 10% variation of base case prices were assumed to reflect the minimum and maximum negligible values.“