Response to Letter Regarding Article, “Characteristics of Children Hospitalized With Infective Endocarditis”
Drs Bachhuber and Lott are correct that we did not use weights in our article in Circulation.1 We did not seek to estimate the national prevalence of endocarditis. Instead, our goal was to “assess the characteristics of children with infective endocarditis.”
In the Kids’ Inpatient Database, sampling strata consist of uncomplicated births, complicated births, and other pediatric admissions. The sampling fractions are 10%, 80%, and 80%, respectively. No admission with a diagnosis of endocarditis was classified as an uncomplicated birth. Therefore, we should have an 80% sample of endocarditis cases from each hospital in each of the states contributing to the Kids’ Inpatient Database (36 states in 2003, 27 states in 2000).
In summary, we do not present national estimates, nor did we state in our article that numbers of cases reported represented the US population. We are grateful to Drs Bachhuber and Lott for clarifying this point for readers who may have made the assumption that we were doing so.