Bias in Pertussis Incidence Data and Its Implications for Public Health Epidemiology

Abstract

A fter a decline of more than 99% from 1934 to 1976, pertussis incidence in the United States has increased dramatically over the last 3 decades. Although the initial decrease was attributed to the introduction of vaccination programs, the muchpublicized recent resurgence has occurred despite continued high-vaccination uptake estimates.4 Unfortunately, an incomplete understanding both of the underlying factors causing this resurgence in incidence and of the key drivers of pertussis transmission in general confounds creation of effective vaccination strategies and public health responses. A powerful approach to dissecting and elucidating the relative contributions of these drivers is the development of disease transmission models and their confrontation—via statistical inference methodology—with highly resolved incidence data. Despite the existence of highly resolved data collected through the Supplementary Pertussis Surveillance System (SPSS), and generous stated policy guidelines regarding data sharing/release, these data are not publicly accessible for researchers. Here, we discuss the lack of resolution and important biases in the incidence reports that are publicly available and argue in favor of more transparency in data sharing policies and broader dissemination of incidence data such as those collated through the SPSS. In addition to mortality and morbidity, the public health consequences of pertussis include an economic burden from lost productivity, as well as financial expenditures in areas such as associated medical costs, public awareness campaigns, and increased vaccinations. The most significant increase in pertussis in the United States (and other developed countries with vaccine programs) has occurred among adolescents and adults, with smaller increases observed for infants and older children. Nonetheless, infants make up the plurality of reported cases and nearly all of the fatalities,