Objective Assessing characteristics of non-responders in epidemiological studies is often complicated by lacking information. In nested studies however such information is often available. An example is our study on the impact of vaccinations on atopic disorders involving optional blood donation, where characteristics of non-responders to blood donation were available. We assessed these characteristics and their potential for having biased our previous results.
Methods In a study on vaccinations and atopic disorders we assessed characteristics of non-response to blood donation, an optional part of study participation. Possible bias due to this non-response was evaluated.
Results Having refused consent to consult vaccination registration data, not having answered the question on income, lower school class, lower birth order, infrequent consumption of vegetables, and not having a history of pertussis were significant determinants of non-response to blood donation. Weighted analysis and sensitivity analysis yielded results similar to those in the original study.
Conclusions In this study of characteristics of non-responders to blood donation we found that refusal to donate blood was related to reluctance to disclose private information in general and to sensitivity on the subject of vaccinations in particular. Additionally, parents of younger children with less older siblings and without a history of pertussis, probably having less experience with health care for their children, were more likely to be a non-responder. Selective participation to blood donation may have affected our assessment of the reliability of the reported vaccination status, but leaves our conclusions from the original study unaffected.