TABLE 4

Unadjusted ORs and aORs From Logistic Regression Models for Each of the Criteria for the Mediation Analysis: Massachusetts Singleton Births, July 1, 2004, to December 31, 2008

Criteria for mediationUnadjusted ORa (95% CI)PaORb (95% CI)P
ART versus Fertile (N = 312 790)
 1. Relationship between ART and EI enrollment1.17 (1.10 ̶ 1.25)<.0011.34 (1.26 ̶ 1.44)<.001
 2. Relationship between ART and preterm1.70 (1.57 ̶ 1.85)<.0011.45 (1.32 ̶ 1.59)<.001
 3. Relationship between preterm and EI enrollment3.21 (3.11 ̶ 3.31)<.0012.60 (2.51 ̶ 2.68)<.001
Subfertile versus Fertile (N = 311 858)
 1. Relationship between Subfertile and EI enrollment1.07 (1.00 ̶ 1.15).061.21 (1.13 ̶ 1.31)<.001
 2. Relationship between Subfertile and preterm1.28 (1.16 ̶ 1.42)<.0011.18 (1.06 ̶ 1.31).003
 3. Relationship between preterm and EI enrollment3.18 (3.08 ̶ 3.28)<.0012.57 (2.48 ̶ 2.66)<.001
  • a For the first 2 assumptions, the unadjusted OR is the crude OR of the association. For the third assumption, the unadjusted OR is the association of EI enrollment (outcome) and preterm (mediator), controlling only for fertility status (exposure).

  • b Models adjusted for maternal demographic characteristics (maternal age, race, education, marital status, nativity), insurance, smoking, prenatal care, parity, gender, labor and delivery complications, and abnormal conditions of the newborn.