Gestational age reporting and preterm delivery.

This study examines recent trends in the reporting completeness and quality of gestational age estimates derived from the date of the last normal menses (DLNM) as reported in South Carolina vital records from 1974 to 1985. Noteworthy improvements in the completeness of reporting emerged during this period with a decline from 31.1 percent missing information in 1974 to 6.6 percent missing in 1985. Completeness of reporting and strategies for imputing values for missing data were analyzed for their impact on the calculation of the percentage of preterm live births. The results indicate that the underreporting of gestational age can lead to marked underestimation of the preterm percentage in a population and to misinterpretation of trends in these percentages. Based on the results of this analysis, it is recommended that preterm percentages be based on cases with DLNM gestational age values between 20 and 50 weeks. Since cases with missing or implausible gestational age data have a greater risk of a poor pregnancy outcome, these findings emphasize the importance of identifying both the completeness of data reporting and the use of imputation and deletion strategies when employing population-based DLNM data to calculate gestational age related indicators.