Implications for registry-based vaccine effectiveness studies from an evaluation of an immunization registry: A cross-sectional study

Population-based electronic immunization registries create the possibility of using registry data to conduct vaccine effectiveness studies which could have methodological advantages over traditional observational studies. For study validity, the base population would have to be clearly defined and the immunization status of members of the population accurately recorded in the registry. We evaluated a city-wide immunization registry, focusing on its potential as a tool to study pertussis vaccine effectiveness, especially in adolescents.

We conducted two evaluations – one in sites that were active registry participants and one in sites that had implemented an electronic medical record with plans for future direct data transfer to the registry – of the ability to match patients' medical records to registry records and the accuracy of immunization records in the registry. For each site, records from current pediatric patients were chosen randomly. Data regarding pertussis-related immunizations, clinic usage, and demographic and identifying information were recorded; for 11–17-year-old subjects, information on MMR, hepatitis B, and varicella immunizations was also collected. Records were then matched, when possible, to registry records. For records with a registry match, immunization data were compared.

Among 350 subjects from sites that were current registry users, 307 (87.7%) matched a registry record. Discrepancies in pertussis-related data were common for up-to-date status (22.6%), number of immunizations (34.7%), dates (10.2%), and formulation (34.4%). Among 442 subjects from sites that planned direct electronic transfer of immunization data to the registry, 393 (88.9%) would have matched a registry record; discrepancies occurred frequently in number of immunizations (11.9%), formulation (29.1%), manufacturer (94.4%), and lot number (95.1%.) Inability to match and immunization discrepancies were both more common in subjects who were older at their first visit to the provider site. For 11–17-year-old subjects, discrepancies were also common for MMR, hepatitis B, and varicella vaccination data.

Provider records frequently could not be matched to registry records or had discrepancies in key immunization data. These issues were more common for older children and were present even with electronic data transfer. These results highlight general challenges that may face investigators wishing to use registry-based immunization data for vaccine effectiveness studies, especially in adolescents.