The demographic bias of Email as a survey method in a pediatric emergency population

Email has been considered as a communication medium between patients and clinicians in pediatric emergency departments, but the demographic bias involved in using email has not been fully explored. We developed a paper based survey to explore access, willingness to participate and the demographic bias of email within our parent population. Methods: To 1733 possible subjects, 1200 surveys were distributed with a return of 1018, a survey response rate of 85%, and a population response rate of 59%. Results: Subjects from families with incomes less than $60,000 per year had lower access rates (OR = 0.40, 95% CI [0.25, 0.62]), as did those with lower education (OR = 0.37, [0.17, 0.81]). Employment outside of the home was associated with increased email access rates (OR = 1.79, 95% CI [1.19, 2.70]). Visible minority status was associated with an increased willingness to participate (OR = 1.84, 95% CI [1.10, 3.06]) as was low education (OR = 2.12, 95% CI [1.04, 4.32]). The population of theoretical responders to an email based quality assurance process would have been significantly different from the base population of adults accompanying children to our emergency department as a result of these biases. Conclusion: We have demonstrated a degree of demographic bias in email access rates, negatively affecting those individuals with lower income, less employment, and lower education. Email based surveys directed at parents in pediatric emergency departments should include questions on income, employment and education in order to permit those who analyze the data to correct for these variables. More research is needed to confirm these findings.