Early motherhood and mental health in midlife: A study of British and American cohorts

Abstract

Objectives: Examine the relationship between early age at first birth and mental health among women in their fifties.

Methods: Analysis of data on women from a British 1946 birth cohort study and the U.S. Health and Retirement Study birth cohort of 1931–1941.

Results: In both samples a first birth before 21 years, compared to a later first birth, is associated with poorer mental health. The association between early first birth and poorer mental health persists in the British study even after controlling for early socioeconomic status, midlife socioeconomic status and midlife health. In the U.S. sample, the association becomes non-significant after controlling for educational attainment.

Conclusions: Early age at first birth is associated with poorer mental health among women in their fifties in both studies, though the pattern of associations differs.

Keywords:

• birth timing
• mental health
• aging

Acknowledgements

This analysis was funded by the UK Economic and Social Research Council award RES-000-0394.

Notes

Notes

1. The regression-based linear transformation is: (predicted standard 20-item CES-D) = 1.6212 × (11-item CES-D score) +0.7415.

2. The choice of cutpoints for the early birth measure was based on inspection of a dummy variable regression for each country, using individual ages at first birth. The data suggested that age 20 or younger would be most appropriate for the British sample and age 19 or younger would be appropriate for the U.S. sample. We adopted one definition for both countries to produce a more parallel analysis. In preliminary work, we estimated alternative models for the United States using the age 19 and younger cutpoint. The coefficients for early birth were greater in Equations (1) and (2), but the pattern of results was similar to those reported in Table 2, with early birth becoming non-significant when education is added to the equation. The one difference is that, in Equation (6), neither of the age at first birth contrasts is significant when the age 19 cutpoint is used.

3. The outcome measures, the CES-D and the GHQ-28, are both skewed to the right. The large number in both samples means that the sampling distributions of regression coefficients are likely to approximate the t-distribution. In preliminary work, we estimated the models allowing a gamma error which yielded results generally parallel those reported. However, the large number of zero values in the British data prevented convergence of model 6 (probably due to negative predicted values during the maximum likelihood iterative estimation).