I"m a doctor. I saw a case of this about 10 years ago when I was a medical student. The patient was a veteran in his late 20"s. A few years prior, he had completed his enlistment and enrolled in college. He was healthy and had not had any mental health problems. He smoked marijuana for the first time. Later that day, the police were called after he had stabbed his mother at home. He was floridly psychotic and spent almost a year on a psych ward at the VA. While he was there he almost killed a nurse, so every time you pull up his chart a big flashing warning is triggered. Remarkably, he eventually recovered completely, went back home, re-enrolled in college, and got off his antipsychotic meds. A year or two later (this is when I saw him), his college friends somehow talked him into trying marijuana one more time. So he smoked for a second time, and within hours was again psychotic and attacked his friends, who called the police. He was admitted to the psych ward again. I rounded on him for 6 weeks every morning, flanked by four armed federal police officers who we had to call every time we unlocked the door to his room. He was heavily sedated however so there was not much risk. Toward the end of my psychiatry rotation, he had improved to the point where he could talk a little and denied violent thoughts, but he was probably pretty far from going home again. I"ve asked some psychiatrist friends about this phenomenon and all of them say that they"ve seen it at least once. I would guess that the determination of causation would be based on observations of the killer"s mental state prior to the episode and his overall medical history. I"d still be skeptical that they could prove a link between the urine test and what he did.