Annual and sub-annual rhythms in human conception rates. I. Effective correction and use of public record LMP dates.

Most studies of rhythms in human conception rates have used monthly total births back-dated 9 months. Such data are of little use for studying any pattern less than a few months in length. Even at that level, they are poor material for studying the possibility of altered conception rhythms in anomalous births. The high frequency of premature delivery in such births makes a simple 9-month offset simplistic and misleading. Dates of last normal menses (LMP) provide appropriate detail, but have generally been dismissed as unreliable. In a large public birth record dataset, we have identified a clear artifact of poor LMP recall, highly correlated with passage of time between conception and onset of prenatal care. The majority of the variation in daily LMP counts due to this recall artifact can be corrected statistically to yield population data suitable to more detailed analyses.