2009 H1N1 and seasonal influenza immunization among pregnant women: a comparison of different sources of immunization information.

Validity of prenatal immunization data from different sources has not been assessed. We evaluated prenatal 2009 H1N1 and seasonal influenza (FLU) data obtained from state immunization information systems (IIS), medical record abstraction (MRA), and participant recall using medical care logs (NCS-MCL). 2009 H1N1 and FLU data were obtained from IIS and MRA for 325 pregnant women participating in the National Children's Study at three locations (SD/MN, NC, WI). Women recalled immunizations at first pregnancy visit and at 16-17 and 36 weeks' gestation (NCS-MCL). The proportion of women with vaccine information obtainable from each data source was determined, and proportions immunized as determined using different data sources were compared. IIS data were available for 82%, MRA for 97%, and NCS-MCL for 93% of women. No mention of either vaccine occurred in 29% (range 4-48%) of IIS, 40% of MRA (25-59%), and 59% (43-82%) in NCS-MCL. Best agreement between sources was 2009 H1N1 vaccine in MRA versus IIS [kappa (95% CI) of 0.44 (0.32-0.55)], with poorest agreement for FLU in IIS versus NCS-MCL [0.11 (-0.03 to 0.25)]. IIS was the most sensitive method for identifying women receiving 2009 H1N1 vaccine (92%); MRA was most sensitive for FLU vaccine (81%). IIS provided the most complete and sensitive data for 2009 H1N1 immunizations and MRA the most complete and sensitive data for FLU; IIS data were available for a smaller percent of population than MRA. NCS-MCL was the least sensitive method for identifying vaccinated women.