Collections > Electronic Theses and Dissertations > Are We Studying Who We Think We're Studying? Role of Socioeconomic Status in the Validity of Estimates of Pneumococcal Conjugate Vaccine Effectiveness in the United States
pdf

Thirteen-valent pneumococcal conjugate vaccine (PCV13) was licensed for use in children in the United States in February 2010. Shortly thereafter the Centers for Disease Control and Prevention began a post-licensure vaccine effectiveness (VE) study in 13 surveillance sites around the US. Cases were identified through active surveillance and controls were matched to cases by age (+/- 14 days) and zip code, which was used as a proxy for socioeconomic status (SES). Due to issues locating and enrolled cases and controls in an era of increased cell phone usage, investigators were concerned that zip code may not provide adequate control for SES and that enrolled children may not be representative of eligible children, threatening both the internal and external validity of study results. We obtained data on SES for cases and controls from a parent interview, birth certificates, and via geocoding and linkage to the US Census Bureau’s American Community Survey. We used conditional logistic regression models to estimate the adjusted and unadjusted VE and to assess effect measure modification by SES of the estimated VE. Small differences were found between enrolled cases and enrolled controls; however, these differences did not meaningfully change our estimated VE and we therefore concluded that internal validity of our estimates was high. Similarly, small differences were found between enrolled and unenrolled cases, but we did not find effect measure modification, indicating that external validity of our estimates was high. In addition to providing reassurance that previously published VE estimates are valid, we show that, despite not being able to contact unenrolled children, an assessment of validity of observational data is feasible.