A statistical approach to an outbreak of endophthalmitis following cataract surgery at a hospital in the West of Scotland

The number of cases of endophthalmitis following cataract surgery caused considerable concern in a West of Scotland hospital throughout 1998 and early 1999. A multi-disciplinary team including infection control nurses, doctors, public health officials,<$>epidemiologists and statisticians was set up to investigate the situation. This paper examines the statistical issues surrounding the investigation. A method based on the Poisson distribution showed that the number of cases was significantly higher than expected. Fisher's Exact Test and Logistic Regression were then applied to the data from two related case control studies. These analyses showed that a higher risk of endophthalmitis was associated with being female, having a vitrectomy or having a previous history of respiratory disease. Finally, a method was devised to enable staff to recognize more quickly when the number of cases of endophthalmitis was becoming higher than expected. The method should find application in other clinical situations where the probability of rare events is known.