METHODS: The data included Minnesota birth records from the 2014 birth cohort and HBV surveillance records from the hepatitis disease surveillance system. Variables related to hepatitis B in the two datasets were compared to identify discrepancies for infants born to known HBV-positive women. To identify potentially missed births and evaluate data validity, infection status was investigated in all women marked as HBV-positive in birth records who had not been reported to the hepatitis surveillance program. LinkPlus record linking software was used to identify known HBV-positive women who gave birth in 2014 in order to identify missed births.
RESULTS: Of the 439 infants identified as being born to HBV-positive women in 2014, 230 (52%) had accurate data for all four variables related to hepatitis B. The most consistent variable was the birth dose of the hepatitis B vaccine, which was accurately documented in 419 (95%) of records. Only 333 (76%) records had accurate documentation of HBIG administration. For maternal characteristics, the single-select variable for hepatitis B infection status was more likely to be accurate than the variable with HBV status included on a list of maternal infections, 339 (77%) and 313 (71%) respectively.
There were 200 women documented as HBV-positive on birth records with no evidence of current infection in surveillance records. Following investigation, none were found to be currently HBV-positive.
By investigating matched cases and confirming diagnoses with prenatal clinics, we identified 12 infants born to HBV-positive women who were not previously reported to PHBPP.
CONCLUSIONS: Birth records data can be a valuable tool in identifying infants born to HBV-positive women. However, inconsistencies in the data persist and more work needs to be done to ensure hepatitis B data on birth records is accurate before it can be used to ensure appropriate PEP of at-risk infants and identification of HBV-positive women.