Two factors are considered in determining whether the antipsychotic treatment is valid or invalid: whether the patient has an exclusionary diagnosis, and whether the patient is a threat. However, the threat state can occur at any time during the encounter, while the diagnosis must be pre-existing or start within 12 hours of the encounter start.

There doesn't seem to be any justification in the measure description for this difference. Why not just treat the diagnosis in the same way, so that any overlap with the encounter is enough to exclude the patient?