One cannot obtain accurate stresses near a constraint nor also near a load unless you have the correct exact distribution of the load.
See St. Venant's principal.
In order to analyze the local stress you will either need to follow some type of best practice and/or include the 3D screw such that the constraint is far enough away from the area of interest, and the local compliance is appropriate. The fastener idealizations are really only effective at reducing simulation times and model complexity. You can do this with a simplified 3d model of the fastener. Also, the preload can be approximated with added loads or enforced displacements in some situations and I think this might be one of those situations.
On the best practice front, If I was using this types of slot fastening often I would do a separate study with some real parts and develop a standard for evaluating their performance by FEA with the necessary safety factors. Just changing to a slightly different fastener and changing the installation torque and mating hole parameters could make large differences in performance for screws.
There are likely experts at this exact situation, but anyway there are my thoughts, and hopefully they are at least a little useful.
