The Use of Immunization Information Systems (IISs) in Vaccine Effectiveness Studies

Background:
Recently, many new vaccines have been added to the childhood immunization schedule. Post-licensure effectiveness studies of these vaccines have typically relied on data collected through active surveillance involving the enrollment of comparison groups. Obtaining immunization data for enrolled patients is time-consuming and often requires extensive follow-up. IISs could represent easily accessible sources of immunization data and be used as a source of control patients, thus eliminating the need to enroll a comparison group.

Objectives:
1. To validate immunization information obtained from an IIS against provider records;
2. To assess the utility of an IIS in evaluating vaccine effectiveness.

Methods:
A convenience sample of pediatric emergency room patients was enrolled in a study assessing the effectiveness of pentavalent rotavirus vaccine (RV5). When available, a copy of each child's immunization record was obtained from the parent/guardian. Permission was obtained to search the Houston-Harris County Immunization Registry (HHCIR) and contact the child's immunization provider(s). Vaccine effectiveness (VE) was calculated using 2 concurrently enrolled control groups (rotavirus-negative acute gastroenteritis (AGE) and acute respiratory infection (ARI) patients) and compared with VE using controls randomly selected from HHCIR.

Results:
45% (n=285) of enrolled patients had immunizations in HHCIR. Provider records were obtained for 93% (n=266) of these children. HHCIR data were the same as the provider record for 71% (n=189) of patients. The HHCIR record contained additional immunization data for 11% (n=30) and was less complete for 18% (n=47) of patients with both records.
VE for a full RV5 series was 89%, 85%, and 82% using rotavirus-negative AGE, ARI, and HHCIR-selected control patients, respectively.

Conclusions:
Although immunization data for only 45% of patients were available from HHCIR, VE calculated using IIS data was similar to estimates using the other control groups. IISs could represent a valuable source of data for VE evaluations.