Ecological studies and debate on rotavirus vaccine and intussusception

Ecological analyses have two major shortcomings. The first arises from the fact that they do not link the exposure in the individual to the outcome of that individual; thus, it could be that the children who are not exposed are the ones who get the disease-this is the so-called "ecological fallacy". The second difficulty is the lack of control of confounding. Thus, if the reasons for choosing to have a vaccine are related to a cause of intussusception, then an ecological study will show a relation that is not a result of the vaccine itself. This effect might go in either direction: if it is wealthy white children who receive rotavirus vaccine (as at least one study suggests) and these children have a low risk of intussusception compared with the population as a whole, then the risk of intussusception after vaccination will be underestimated. In Pennsylvania 62% of children received rotavirus vaccine, whereas this proportion in Missouri was only 39%. If those children not receiving vaccine are more likely to be poor and black and at increased risk of intussusception, then a weak relation between vaccine coverage and rate of admission to hospital may well be seen, and indeed this is what [Simonsen] and colleagues found.