Mind the gap: bridge methods to allocate multiple-race mothers in trend analyses of birth certificate data

To examine the effects of proposed methods of redistributing multiple-race mothers to single-race categories when computing trend data from birth certificates.

Low birthweight and multiple (twin and higher-order) birth rates for California were calculated for non-Hispanic mothers from birth certificate data for 2000. Births to the 1.9% of mothers identified as multiple-race were reassigned to single-race groups according to 12 "bridging" methods. Bridge methods utilized population-based whole allocation, fractional allocation, and other methods, primarily depending on first race listed.

For large race groups, there was little difference in low birthweight and multiple birth rates regardless of the bridge method employed. For smaller groups such as Native Hawaiians and other Pacific Islanders and American Indians/Alaska Natives, there was substantial variation by bridge method in observed rates.

Tracking trends in birth outcomes across the change in data collection will challenge public health researchers. This paper outlines advantages and disadvantages of various bridge methods.