Use of near-real-time medical claims data to generate timely vaccine coverage estimates in the US: the dynamics of PCV13 vaccine uptake.

BACKGROUND: Vaccine coverage estimates lag by years in the US. Commercially available medical claims databases contain timely records of childhood vaccinations given in physician offices. We used such data to track the replacement of the 7-valent pneumococcal conjugate vaccine (PCV7) by PCV13- a new vaccine active against 6 additional serotypes- starting in March 2010. METHODS: We developed an age cohort model to compute vaccination coverage over time. We used age-stratified- national projections of monthly PCV7 and PCV13 doses administered to children <5 years based on physicians' office claims- January 2008-May 2012. We assumed doses were given on schedule- and tracked cumulative numbers of doses given to aging monthly cohorts to estimate the percentage of children fully PCV13-immunized. To account for children uninsured or in the Vaccines for Children program- estimates were projected using National Immunization Survey coverage data. RESULTS: PCV7 was phased out by June 2010. By March 2012- 82% of children 6-23 months were fully immunized with PCV13 and 42% of toddlers aged 15-59 months had received a catch-up PCV13 dose. For children aged 6-59 months- protective PCV13 coverage levels reached 33% and 56% by March 2011 and 2012- respectively- and were projected to reach 88% by March 2014. Our estimates for children aged 0-59 and 24-59 months are consistent with CDC's Immunization Information System sentinel sites data for 2011-2012. CONCLUSIONS: By using a simple analytic approach to compute vaccine coverage in aging cohorts from claims data- we show that PCV13 coverage rose rapidly as the PCV7 program was replaced. These estimates- validated against a CDC sentinel surveillance system in 8 states- should enable early documentation of the PCV13 impact on pneumococcal disease in the US. Moreover- they demonstrate the feasibility of tracking uptake patterns in near real-time even with simple summary counts of medical claims data.