Inequalities in under-5 mortality. What about parental origin?
Are there differences in terms of under-5 mortality (U5M) depending on whether one is born
to an immigrant or a native parent?
This is the question that this paper addresses using the permanent demographic sample, the
largest French sociodemographic panel with a sample of 657,548 births from 1990 to 2019.
Parental origin is measured by the mother’s country of birth grouped into 5 regions: Sub-
Saharan Africa, North Africa, Asia, Europe, America-Oceania & Other compared to the french
natives. Two parametric longitudinal Weibull models were implemented. The first one is an
unadjusted model whose results show an excess U5M among children of immigrant mothers,
with the exception of those from Europe and Asia. As for the second model, it was implemented
by considering the weighting by the inverse of the probability of treatment. Indeed, the
differences in terms of socioeconomic characteristics between immigrants and natives (in favor
of the latter) make them incomparable. We therefore proceeded to rebalance them on the basis
of covariates relating to standard of living, socio-professional category and level of education,
which have been proven to be favorable to the natives. These covariates were used to construct
the propensity score, which was then used to assign a weight to each individual according to
the inverse probability of treatment method. At the end of this process, the results not only
confirm an excess U5M of immigrants (except for the case of Europe), including those of
mothers born in Asia, but a higher excess mortality for most children of immigrants. To ensure
our findings were robust to different methods based on propensity score, we conducted
sensitivity analysis by re-fitting the propensity score method using matching method based on
the same propensity score and compared it to Inverse Probability of Treatment Weighting
(IPTW) method. The findings from this sensitivity analysis confirmed the main result even
though it slightly overestimates the effect of the category “America-Oceania & other”.
In short, we can conclude that Mother origin plays an essential role in the construction of
inequalities especially in terms of under-5 mortality. These inequalities remain and are
reinforced after adjusting for socioeconomic characteristics; which can be indicative of