Evaluation of a volunteer sample in nasopharyngeal colonization surveys for Streptococcus pneumoniae in rural Alaska

Objective. To compare characteristics of persons in rural northern communities who participated in a study on antimicrobial use and drug-resistant Streptococcus pneumoniae (SP) to those who did not participate. Study Design. The original study (1998-2000) was a community-based, controlled intervention trial designed to determine the penicillin susceptibility of nasopharyngeal SP isolates in relation to community-wide use of antibiotics. The study continued after 2000, in a subset of the original communities, to prospectively evaluate the impact of the heptavalent pneumococcal conjugate vaccine on the carriage of SP. The results presented here are an analysis of the first five years of data. Methods. We conducted annual surveys (1998-2002) for nasopharyngeal colonization of SP using a volunteer sample of residents in rural communities. Medical chart reviews for health clinic visitation and antibiotic use were completed for all village residents. Results. Participants were younger (22.8 vs. 28.4 years), had more health clinic utilization (3.3 vs. 2.4 visits) and received more antibiotics (1.0 vs. 0.6 courses) than non-participants. Differences between participants and non-participants were similar across all years of the study. Conclusions. Our study provides further empirical evidence that selection bias should be considered when designing studies. However, a volunteer sample provided internal consistency for comparison of our main study outcomes across time. (Int J Circumpolar Health 2005; 64(1):16-25) Keywords: Alaska, community intervention, non-participation, non-randomized sample, participation, selection bias