Search tips
Search criteria 


Logo of bmjThis ArticleThe BMJ
BMJ. 2007 September 1; 335(7617): 450.
PMCID: PMC1962874

Remember the null hypothesis

Steve Ramplin, senior house officer, Royal Victoria Infirmary, Newcastle upon Tyne

My curiosity turned to excitement: I imagined my name in lights or, even more prestigious, the BMJ.

My eagerness concerned the case of a 65 year old man with alcohol dependency and comorbid depression. He had resumed drinking 12 months after having stopped because of an aortic dissection. The trigger for his relapse was his arrest for shoplifting.

His account of his crime was somewhat confused. On entering a large department store, he had bought some small items before noticing some expensive cameras. He explained that he had believed absolutely that the staff in the store were his friends and that, by demonstrating the ease with which goods could be stolen, he was being helpful. He placed a camera in his bag and proceeded to the exit. En route, he realised that being caught was a prerequisite for warning the staff about their vulnerability to theft. He became a prisoner of ambivalence, loitering just inside the store entrance unable to carry on.

Of course, the sharp eyed staff apprehended him the instant he stepped outside, when he suddenly grasped the bizarreness of his former belief that he was doing them a favour. He also then admitted stealing from the same store the previous week, when his theft of a clock radio had gone unnoticed. What was interesting was that he said he had received a corticosteroid injection into his left shoulder the day before each shoplifting incident.

We all know corticosteroids can cause psychosis. I wondered if intra-articular steroid injections had been similarly implicated, as the link seemed logical. An extensive literature search located but one case report.1 Before writing my case up, however, I needed more details. At this point things began to unravel: yes, my patient had received injections of methylprednisolone, but in April and May, not in June, when he committed the thefts. A later injection in August was not associated with criminal activity, and further inquiry revealed a previous caution for shoplifting in 2001.

I was reminded of the value of a key statistical assumption: there is no association until proved otherwise. My dreams of fame and fortune temporarily shattered, I fell to earth, brought down by the null hypothesis.


1. Robinson DE, Harrison-Hansley E, Spencer RF. Steroid psychosis after an intra-articular injection. Ann Rheum Dis 2000;59:927

Articles from The BMJ are provided here courtesy of BMJ Publishing Group