There has been no large study characterizing selection bias in allergy and evaluating school personnel’s ability to use an epinephrine auto-injector (EpiPen®). Our objective was to determine if the consent process introduces selection bias by comparing 2 methods of soliciting participation of school personnel in a study evaluating their ability to demonstrate the EpiPen®.
School personnel from randomly selected schools in Quebec were approached using a 1) partial or 2) full disclosure approach and were assessed on their ability to use the EpiPen® and identify anaphylaxis.
343 school personnel participated. In the full disclosure group, the participation rate was lower: 21.9% (95%CI, 19.0%-25.2%) versus 40.7% (95%CI, 36.1%-45.3%), but more participants achieved a perfect score: 26.3% (95%CI, 19.6%-33.9%) versus 15.8% (95%CI, 10.8%-21.8%), and identified 3 signs of anaphylaxis: 71.8% (95%CI, 64.0%-78.7%) versus 55.6% (95%CI, 48.2%-62.9%).
Selection bias is suspected as school personnel who were fully informed of the purpose of the assessment were less likely to participate; those who participated among the fully informed were more likely to earn perfect scores and identify anaphylaxis. As the process of consent can influence participation and bias outcomes, researchers and Ethics Boards need to consider conditions under which studies can proceed without full consent. Despite training, school personnel perform poorly when asked to demonstrate the EpiPen®.