The study of infectious intestinal disease in England: what risk factors for presentation to general practice tell us about potential for selection bias in case-control studies of reported cases of diarrhoea

Cases reported to national surveillance systems commonly form the sampling frame for population-based case-control studies of infectious intestinal diseases (IID). However, reported cases represent a minority of all cases in the population, differ systematically from unreported cases, and their use in case-control studies could introduce biases. Case reporting is a selective process involving multiple stages, the first of which requires presentation by a symptomatic person to the health services. Factors affecting a person's likelihood of presentation will thus affect the characteristics of reported cases. Data from a large study of IID in England were used to investigate factors influencing presentation to a general practitioner (GP) following an episode of IID. Multivariable logistic regression was performed, comparing GP presenters with non-presenters. Explanatory variables used were illness severity, recent foreign travel and socioeconomic indicators. Severe illness (OR = 12.54, 95% CI: 7.58-20.74), recent foreign travel (OR = 2.4, 95% CI: 1.39-4.14), leaving full-time education at an earlier age (OR = 2.06, 95% CI: 1.22-3.50) and housing categories representing lower socioeconomic status (SES) were all independently associated with GP presentation. Case reporting to national surveillance is shaped by complex biological and social factors, of which illness severity appears to be the most important. Results from case-control studies comparing cases of IID identified by surveillance with population controls are likely to generalize mainly to cases severe enough to be reported. Controlling for educational and SES (mostly housing) is required.