Identifying potential biases and confounders that may affect data quality is an important consideration when evaluating surveillance systems. Having the benefit of predictable temporal trends is a key requirement to improve upon the specificity of detecting outbreaks. Identification of factors that impact on the reliability of the temporal trends observed in the data may provide for the ability to improve the capability to identify aberrations in those trends. During a retrospective study of a dataset of microbiology orders from the veterinary teaching hospital at The Ohio State University for 2003 we noticed regular intervals when increases in the number of culture orders were not accompanied by proportional increases in the number of isolates. These instances appeared to occur at intervals that coincided with the clinical rotation of senior veterinary students within the hospital.
Objective
This paper reports on a potential confounder discovered during an investigation of microbiology orders in a veterinary teaching hospital as a possible data source for outbreak detection.