Adjusting for misclassification error in follow-up analysis

Hello!

I have used OpenMx to undertake LGMM, and have identified the classes / number of trajectories, as well as the posterior probabilities of class membership (which I have used to make modal class assignment of my sample). Following this, I am attempting to look at the relationship between class membership and external variables (predictors and distal outcomes) via logistic / linear regression.

The recommended approach appears to be the 3 step method (either using the ML OR the Bolck–Croon–Hagenaars [which appears most fitting as it allows for both continuous and categorical outcomes] approach); these bias-corrected methods account for the misclassification error introduced in the classifying of individuals using modal assignment. While software such as Mplus have settings for implementing this, I have been struggling to find resources surrounding how to do this in R with an OpenMx model. I have attempted to work out the misclassification probabilities, and transform these into a form that can be used in logistic regression (and can share the code if that would be helpful), but am unsure if I have done so correctly. 

Any assistance or tips would be greatly appreciated! Thanks!!