Herring WL, Mladsi D, Mauskopf J. Patient-level simulation modeling for economic evaluations: opportunities and challenges in a practical setting. Poster presented at the 2013 ISPOR 18th Annual International Meeting; May 1, 2013. New Orleans, LA. [abstract] Value Health. 2013 May; 16(3):A28-9.

OBJECTIVES: In the spectrum of methodologies used to conduct economic evaluations, patient-level simulation modeling is noteworthy for its flexibility in reproducing patient experiences that closely mirror reality. This flexibility is especially important when modeling diseases with continuous or multidimensional health states or with non-Markovian dependence on disease history, but often comes at the expense of advanced data and software requirements and reduced computational efficiency. Methods for exploiting the opportunities while mitigating the challenges associated with using simulation modeling are needed.

METHODS: Based on our experience developing patient level simulation models in Microsoft Excel for a complex, progressive disease, we identified steps that can be taken during the development and presentation of a simulation model to capitalize on the advantages inherent in the flexibility and transparency of this approach while mitigating some of the the associated difficulties.

RESULTS: Three methods were identified for mitigating the challenges of simulation modeling. First, the implementation of the model can be simplified by minimizing the dependence on random number draws wherever possible. For example, a single, cumulative probability of treatment discontinuation can replace a series of separate, time-dependent discontinuation probabilities. Second, the transparency and efficiency of the computations can be improved by anticipating all the random draws required to determine a patient’s experiences and organizing the calculations so that a sufficient batch of random numbers can be generated at the beginning of the patient’s sojourn through the model. Lastly, and perhaps most importantly, the face validity of the model can be ensured through the visual representation of sample patient experiences that highlight the ability of the model to more accurately represent reality.

CONCLUSIONS: Our modeling experiences have demonstrated that meaningful steps can be taken to capitalize on the flexibility of patient-level simulation modeling while maintaining critical aspects of transparency and efficiency.

Share on: