The study aimed (i) to examine the poverty trajectories of families with and without a child with NDD and (ii) to investigate socioeconomic factors associated with the trajectories. We tracked families from the year prior to child birth until the child reached 25 years. Based on resource investments required by families with a NDD child, we hypothesized that families of children with a NDD would follow more unfavorable poverty trajectories.
Methods: We used data from the US Panel Study of Income Dynamics and the Child Development Supplements (n=3317). We classified children with a NDD if parents reported a medical diagnosis of epilepsy, speech impairment, retardation, developmental delay, autism, or learning disability. Family poverty was measured from household income and national poverty thresholds. We calculated the predicted probability of household poverty for each year from a pooled logistic regression model. We used latent class growth modelling to identify distinct poverty trajectories and regressed family socioeconomic characteristics and child NDD status on the probability of belonging to each trajectory using a multinomial logistic regression.
Results: A fifth of children in the sample had a NDD. Across all years, families of children with NDD had a 14.9% probability of living in poverty compared to 10.7% among other families (difference of 4.1%, 95% confidence interval (CI) 1.6%, 6.6%). The difference in experiencing poverty extended back to the year prior to child birth. We uncovered five distinct trajectories of poverty over time: constant non-poverty (52.6%), fast progression into poverty (7.5%), slow progression into poverty (10.4%), slow exit out of poverty (16.7%), and constant poverty (12.7%). Compared to other families, families of children with NDD had a 45.3% greater probability (95% CI 18.7, 71.8) of remaining in constant poverty than constant non-poverty. After adjusting for socioeconomic factors in the year prior to birth the probability reduced to 30% (95% CI -8.5%, 68.8%). No significant associations were found for other trajectories.
Conclusions: The findings suggest that poverty is more prevalent and persistent among families of children with NDD. Importantly, we report evidence that part of this difference is accounted for by family socioeconomic characteristics that were present before the child was born. This finding suggests that the magnitude of difference in poverty risk between families of children with NDD compared to those without generated from cross-sectional studies are biased. The results support policies that aim to reduce economic hardship among families of children with NDD as well as those that address poverty in the context of NDD prevention.