Validation of the new Swedish vaccination register - Accuracy and completeness of register data.

OBJECTIVE: The aims of this study are to validate infant vaccination data in the Swedish Vaccination Register (SVR) to the Swedish administrative coverage reports, and to assess differences in register-based vaccination coverage estimates between providers using different data reporting methods. METHODS: The study population included all infants born in Sweden with a Swedish Personal Identity Number during 2014 and 2015 (n = 230,220). Data on all National Immunisation Programme vaccinations administered before 24 months of age were collected from the SVR and from administrative coverage reports. Information regarding data registration methods in the SVR were collected from national and regional authorities. Coverage from health care providers using single registration methods, where vaccination data were transferred automatically from the electronic health care record to the SVR, was compared to that from providers using double registration methods where data had to be added into the SVR in a separate process. RESULTS: For 98,4% of the study population at least one vaccination was recorded in the SVR. The coverage of 3-dose DTP-containing (87,1%) and 1 dose MMR (91,1%) in the register did not reach administrative data coverage (97,4% for 3-dose DTP-containing and 97,0% for MMR). Single registration procedures yielded significantly higher coverage than double registration procedures (92,24% vs 87,10%, p < 0,0001). A regional switch from double to single registration increased coverage from 80,0 to 95,2%. CONCLUSIONS: The SVR is a valuable data source for vaccination coverage monitoring. For research purposes, the SVR provides valuable data, since every health care provider is obliged to register all vaccine doses given within the national immunisation program. The SVR shows a high completeness validated by comparison to a very well-functioning administrative data system. Single-registration procedures give more complete data and should be supported by health systems while creating health care registers.