PURPOSE: Assess the validity of the Clinical Practice Research Datalink (CPRD) Aurum Mother-Baby Link (MBL). METHODS: We assessed the validity of CPRD Aurum MBL pairs using several strategies. Our study population was based on all available MBL pairs in a 100-practice sample of CPRD Aurum. We sent a questionnaire to the general practitioner of a sample of MBL patients asking if linkage was correct and whether delivery date was correct within 14 days. We also compared MBL delivery date to the matched child's birth date in CPRD Aurum. Finally, in the Hospital Episode Statistics (HES) eligible population, we compared MBL delivery date to birthing parent's HES delivery date. RESULTS: The study population included 124 962 MBL pairs: 124 962 livebirths, 91 218 birthing parents, and 123 562 deliveries. Based on 361 questionnaire responses standardized to the MBL population, 85% (95% CI 74%-96%) of pairs were confirmed as correctly matched, <
1% (0.04%-1.4%) were incorrectly matched, and 13% (10%-18%) could not be confirmed. For 71% (61%-82%) of the sample, delivery date was correct within 14 days. Among the entire population of 124 962 MBL pairs, 73% of MBL pairs had MBL delivery date within the baby's birth month recorded in CPRD Aurum. Among 79 834 patients with HES linkage, 94% of 107 080 MBL deliveries in CPRD Aurum were confirmed in HES. CONCLUSIONS: Our various assessments of CPRD Aurum MBL found the great majority of pairs were correctly matched and most delivery dates were correct within 14 days. Further refinement of delivery dates may be necessary for some studies.