OBJECTIVES: After the introduction of the 7-valent pneumococcal conjugate vaccine (Prevnar®) in the US, the incidence of invasive pneumococcal disease (IPD) caused by the 7 vaccine serotypes declined dramatically in vaccinated children as well as unvaccinated adults (via herd effect benefit).  In 2008, a transmission dynamic equation model was developed to capture direct and indirect vaccination effects. This model accurately predicted the total incidence of IPD (caused by all serotypes) after Prevnar® introduction.  This original model was refined in the present analysis to predict the dynamics of IPD caused by specific serotypes. METHODS: The model simulates the acquisition of asymptomatic carriage of pneumococci and the development of fatal and non-fatal IPD among vaccinated and unvaccinated individuals aged <2, 2–4, 5–17, 18-49, 50-64, and >=65 years old. Categories of pneumococcal serotypes include PCV7-type (4, 6B, 9V, 14, 18C, 19F, and 23F) and non-PCV7-type (all others).  The model was calibrated by approximating serotype-specific US IPD surveillance data from the years 1998-2006. RESULTS: The previous model structure fit the disease incidence caused by the PCV7 serotypes quite well, but was inadequate to predict increases in disease caused by the non-PCV7 serotypes due to serotype replacement. Additionally, the surveillance data showed limited increase in IPD caused by non-PCV7 serotypes in the vaccinated <2-year-old group.  A subsequent recalibration and reformulation resulted in a revised model able to replicate closely the observed IPD incidence stratified by pneumococcal serotypes. CONCLUSIONS: The revised model validates the accuracy of the original model to replicate the incidence of IPD caused by PCV7 serotypes and may be used to predict the future incidence of disease given the increases in IPD caused by non-PCV7 serotypes.