Autism spectrum disorders and fetal hypoxia in a population-based cohort: Accounting for missing exposures via Estimation-Maximization algorithm

Table 3 Association of Fetal Hypoxia at Birth¹ and Diagnosis of Autism Spectrum Disorders²: Results of Logistic Regression Fitted with EM Algorithm for Handling Unknown Exposures That are Modeled From Maternal and Birth Characteristics Under MAR Assumption (r = 1) Stratified on Child's Gender

All gestational ages at birth
	Unadjusted			Adjusted
Child's Gender	Both	Male	Female	Both³	Male⁴	Female⁴
Odds ratio estimate	1.15	1.14	0.93	1.11	1.13	0.94
95% confidence interval	0.98, 1.33	0.97, 1.34	0.62, 1.40	0.95, 1.29	0.96, 1.33	0.62, 1.41
P-value	0.08	0.11	0.74	0.19	0.14	0.76
Full-term births only
	Unadjusted			Adjusted
Child's Gender	Both	Male	Female	Both³	Male⁴	Female⁴
Odds ratio estimate	1.11	1.10	0.92	1.07	1.09	0.93
95% confidence interval	0.94, 1.31	0.93, 1.31	0.60, 1.42	0.91, 1.34	0.91, 1.29	0.60, 1.56
P-value	0.21	0.23	0.71	0.42	0.35	0.74

1: at least one of the three tests for fetal hypoxia is positive, i.e. fetal scalp pH < 7.25, or umbilical artery pH < 7.20, or venous pH < 7.28
2: singletons born in Alberta 1998-2004 and followed up till March 31, 2008
3: adjusted for gender, birth year and socio-economic status
4: adjusted for birth year and socio-economic status

ISSN: 1471-2288