Identifying and oversampling Hispanics by the Passel-Word surname list for enrollment in a web-based nutritional intervention

Abstract

Objective: To describe the enrollment rates and characteristics of Hispanics and non-Hispanics from Kaiser Permanente Colorado invited to participate in a web-based intervention promoting increased fruit and vegetable consumption.

Design: Hispanics were identified by the Passel-Word Spanish surname list. Characteristics associated with the likelihood of enrollment overall and by ethnicity were examined by logistic regression.

Results: A total of 174 (6.1%) probable Hispanics and 340 probable non-Hispanics (11.8%) enrolled. Hispanics were 48% less likely to enroll than non-Hispanics, females were almost four times as likely to enroll as males, and those living in a census tract associated with higher income levels were 41% more likely to enroll than other income groups. Among Hispanics, females were 87% more likely to enroll than males and those living in a census tract associated with higher income levels were 62% more likely to enroll than other income groups. Among non-Hispanics, the odds for enrolling increased 14% for each decade increase of age, females were 43% more likely to enroll than males and those living in a census tract associated with higher income levels were 68% more likely to enroll than those in other income groups.

Conclusion: Identifying Hispanics through surname for oversampling can be successful in terms of sampling yield and accuracy. However, our results suggest that Hispanics are less likely to enroll in a web-based nutritional intervention. Additional research is needed to identify methods of attracting more Hispanic subjects to these kinds of interventions.