Disease clusters: should they be investigated, and, if so, when and how?

Individual cluster reports are subject to several difficulties in interpretation. Although they rarely lead to new aetiological insights, a public health response to delineate the size and extent of any excess risk may be warranted. Further investigation, where merited, should usually include an examination of data for different areas and/or different time periods. A statistical evaluation of disease clusters is often secondary to a detailed appreciation of issues such as the availability and quality of data, confounding and bias in the selection of areas for study.