OBJECTIVE: To present a method for evaluating uncertainty due to unknown parameter correlations in stochastic decision models. METHODS:  The use of probabilistic sensitivity analysis (PSA) has grown significantly in health economic decision modeling. When parameter correlations are known various methods exist to evaluate uncertainty in PSAs to account for correlations (eg, Cholesky decomposition). However, in cohort analyses, using literature-based data, parameter correlations are seldom known and it is therefore typically assumed that parameters are uncorrelated and independent. We present a method and worked example to explore uncertainty due to parameter correlation in the absence of known correlations. For a decision model with n parameters a n x n diagonal correlation matrix defining all parameter correlations is developed. With the SIMTOOLS add-in, the CORAND function and correlation matrix are used to generate correlated random numbers for the model simulation. For the base case all parameters in the matrix are assumed to be independent and the correlations are set to zero. Systematic analyses can then be conducted in which model parameters, individually or in groups, are correlated to explore the potential impact of parameter correlation on the model outcomes. We report the results of a sample analysis and show that parameter correlations can have a significant impact on model uncertainty. While the median ICERs did not change significantly, the 95% confidence intervals ranged widely as the shape of the ICER-scatterplots changed. Parameters with little impact in deterministic sensitivity analyses were observed to contribute to significant uncertainty in the correlation analysis. CONCLUSIONS: In stochastic decision models where parameter correlations are unknown, it is possible to evaluate uncertainty due to potential parameter correlations in Excel-based decision models. Unknown parameter correlations may be a significant source of uncertainty. Future research is needed to validate this method in comparison to methods for evaluating known parameter correlations.