OBJECTIVES: To test whether the health of the mothers is affected by the presence of a disabled or chronically-ill child. METHODS: Using the National Longitudinal Survey of Children and Youth (1994-2000), a nationally representative household survey in Canada, ordered Probit analysis was used to test the effect of the presence of a disabled child on self-reported health, controlling for other factors. Firstly, we modeled the health effect of a disabled child in the house currently. Secondly, utilizing the longitudinal aspect of the data, we modeled the effect of the presence of a disabled child: a) currently only; b) previous years only; and c) both currently and previously. Married and lone mothers are analyzed separately for children aged 6-15 years of age in 2000 for a sample size of 5217 and 1177 for married and lone mothers, respectively. Bootstrap standard errors were calculated to account for multi-staged, non-random survey design. To cope with causal inference issues we: 1) control for previous health status, and 2) test results using propensity score matching (PSM). RESULTS: Evaluating at the sample means, model 1 indicates a shift down in the health distribution for mothers with disabled children. There is a lower probability of excellent health of 10[8.8-11.2] percentage points for married mothers with a disabled child compared to households with no disabled children. Lone mothers show a difference of 9[7-11]percentage points although health for lone mothers is lower in general compared to married mothers. Model 2 shows the strongest association for those households with a disabled child both currently and in the past. PSM analysis confirms our results.CONCLUSION: The presence of a disabled or chronically-ill child has a statistically significant effect on the health of mothers. Consequences for maternal health are larger in the long-run, if the child's condition persists.