Bayesian Optimization - Minimising False Negatives

I am trying to tune a Random Forest Classifier Model (Using TreeBagger). I have used the bayesopt function to optimize hyperparameters such as the number of categories and splits. This has led to improvements in my model, however it has primarily improved precision at the cost of recall.

For this particular dataset (classifying heart disease), false negatives are more costly than false positives. Is there any way to set parameters for the bayesopt so that it favours hyperparameters that punish false negatives? If not, is there any alternative way to do this through Matlab?

Thanks