Background Despite being considered high risk, little is known about the perinatal health of refugees in developed countries. Our objectives were to examine whether: (1) the healthy migrant effect applies to infants born to refugee women with respect to severe neonatal morbidity (SNM); (2) refugee status was a risk factor for SNM among immigrants; (3) refugee sponsorship status was a risk factor for SNM by comparing asylum-seekers to sponsored refugees; and (4) refugees were at greater risk of specific SNM subtypes. Methods Immigration records (1985-2010) linked to Ontario hospital data (2002-2010) were used to examine SNM. We calculated adjusted risk ratios (ARR) with 95 % confidence intervals (95 % CI) for SNM and unadjusted risk ratios with 99 % CI for SNM subtypes using log-binomial regression. Results There were borderline differences in SNM among refugees (N = 29,755) compared to both non-immigrants (N = 860,314) (ARR = 0.94, 95 % CI 0.89, 0.99) and other immigrants (N = 230,847) (ARR = 1.10, 95 % CI 1.04, 1.18) with a larger difference comparing other immigrants to non-immigrants (ARR = 0.83, 95 % CI 0.81, 0.85). Asylum-seekers did not differ from sponsored refugees (ARR = 1.07, 95 % CI 0.90, 1.27). Though rare, several SNM subtypes were significant with large effect sizes. Conclusion With respect to SNM risk, the healthy migrant effect clearly applies to non-refugee immigrants, but is weaker for refugees and may not apply. Among immigrants, refugee status was a weak risk factor for SNM and may not be clinically important. Sponsorship status was not associated with greater risk of SNM. Further investigation of several SNM subtypes is warranted.
Keywords: Administrative data; Healthy migrant effect; Refugees; Severe neonatal morbidity.