(228) Characteristics of Linked and Unlinked Mother-Infant in The Healthcore Integrated Research Database

Publication Number: 559

Background: For pregnancy safety, population-based studies, linking maternal and infant records in administrative health claims is an indispensable tool for studying pregnancy and infant outcomes. Little is known about how mothers with and without linked infants may differ, informing analyses and interpretation. Not all infant outcomes can be assessed on maternal records, as not all are evident at birth. Thus, there is a need to assess some outcomes using infants’ health claims. October 2018, Mother‐Infant Linkage (MIL) table was added to the Sentinel Common Data Model (SCDM); HealthCore implemented the use of the MIL table into the HealthCore Integrated Research DatabaseSM (HIRD) in March 2020

Objectives: To examine the characteristics, including demographics, for women with linked infants and women without linked infants in the HIRD

Methods: This is a secondary analysis of a large existing cohort of all live births in the HIRD, March 2020 and March 2021, ages 18-49 years at delivery with a minimum of 6-months of continuous enrollment. For the MIL in the HIRD, deterministic matching on subscriber ID was applied to link mothers with delivery claims to their infant(s). A mother was classified as linked to her infant if infant’s date of birth was within three days of the delivery date. We examined the distributions of maternal demographic characteristics stratified by infant-linkage status. For continuous measures, mean and standard deviation (SD) were calculated, and proportions for categorical measures

Results: As of March 2021, the HIRD has 1.9 million mothers and 3.4 million infants, resulting in 1.4 million mother-infant linkages (73.4%). For this analysis, 101,580 women with live birth claims were identified in the HIRD between March 2020 and March 2021. Of these, 64,400 (63.3%) had linked infants and 37,180 (36.6%) had no linked infant. Linked mothers were older (mean=32.1 years (SD:4.3)) than those who were unlinked (mean=28.8 years (SD=5.8 Linked mothers had a similar pattern of continuous enrollment (mean=6.8 years (SD: 4.2)) as unlinked mothers (mean=7.0 years (SD:4.4)). Both linked and unlinked mothers had similar US geographic distribution, 41% linked vs. 39% unlinked mothers were living in the South

Conclusions: Mother-infant linkage is essential for drug safety database studies in pregnant women and studying infant health. While most mothers could be linked to their infants, current MIL algorithms do not capture all linked infants in the HIRD. Understanding different characteristics between linked and unlinked mothers could help facilitate improvements in future MIL linkage, increasing the precision and generalizability of pregnancy studies.