The case-cohort design in outbreak investigations.

The use of the case-cohort design for outbreak investigations has been limited. Here we discuss its strengths and limitations based on real and fictitious examples. The case-cohort is a case–control study where controls are sampled from the initial population at risk, and may thus include both cases and non-cases. An advantage of the design, compared to traditional case-control studies, is that risk ratios can easily be obtained directly from the cross-product of exposed and unexposed cases and controls (rare disease assumption is not required). We illustrate this in the context of point source gastrointestinal outbreaks and in field studies on vaccine effectiveness. The design is also useful to investigate multiple outcomes with a unique sample of controls or to test hypotheses when different case-definitions (from the most sensitive to the most specific) are used for a particular outcome. Strengths and limitations are presented, and discussed in the context of outbreak investigations.