Table 1. Main Differences in the Design and Analytic Approach Between the US and Nordic Analyses.
Study Characteristic | US Study | Nordic Study |
---|---|---|
Cohort | Live births | Live singleton births |
Exclusions | Known teratogens | Gestational wk ≤22 or >44 |
Birth weight <300 or >7000 g | ||
Exposure window | First trimester: LMP to LMP + 90 d | First trimester: LMP to LMP + 97 d |
Reference group | No exposure from 3 mo before LMP to LMP + 90 d | No exposure from LMP to LMP + 97 d |
Outcome assessment | First 3 mo of life | First year of life, except Norway (first 3 mo of life) |
Covariate assessment | ICD-9 | ICD-9 and ICD-10 |
3 mo before LMP to LMP + 90 d | Maternal nonpsychiatric: 1 y before LMP to delivery | |
Healthcare utilization: 3 mo before LMP to LMP | Maternal psychiatric: 5 y before LMP to delivery | |
Concomitant drug use: 90 d before LMP to LMP + 97 d | ||
Confounding adjustment | Fine stratification on the propensity score | Multivariate regression |
Abbreviations: ICD-9, International Classification of Diseases, Ninth Revision; ICD-10, International Classification of Diseases, Tenth Revision; LMP, last menstrual period.