LAUSR

Background Selection of test-negative controls takes less time and costs less than traditional control selection for evaluating vaccine effectiveness (VE). Here, rotavirus VE was evaluated using hospital controls and compared with test-negative controls to determine whether using the latter can substitute for the former. Methods We recorded gastroenteritis in children from 2 months to 2 years of age at six medical facilities in Saga City between January 4th and May 31st, 2014. Stools from all identified acute gastroenteritis patients were tested for rotavirus using immunochromatography. Rotavirus gastroenteritis (RVGE) cases had test-positive stool, whereas test-negative controls had gastroenteritis but no rotavirus infection; hospital controls were outpatients visiting the same facility for indications other than gastroenteritis. Vaccination status was verified by inspecting maternal and child health records, and demographic data were obtained from a questionnaire completed by the patients’ guardians or from the medical records. Unconditional logistic regression analysis was used to adjust for possible confounding factors. Results Sixty-four RVGE cases, 260 test-negative controls, and 589 hospital controls were enrolled. The characteristics of the two control groups, including RV vaccination history, were similar. The RVGE cases were more likely to have used daycare services than children from either of the two control groups. The VE against RVGE estimated using hospital controls was 86.6% (95% confidence interval [CI], 55.9–96.0%), very similar to the VE using test-negative controls (84.9% [95% CI, 49.6–95.5%]). Conclusions The estimated VE using test-negative controls and hospital controls is similar. Therefore, test-negative controls are considered appropriate for establishing VE.