Background The COVID-19 pandemic has had detrimental and unequal repercussions on mental health. Working women, especially mothers, appeared to be disproportionately affected. However, to date studies have been limited and no research has explored whether and how workplace size moderates this relationship. This study aimed to estimate changes in working women’s mental health at the start of each COVID-19 lockdown and estimate the effect of motherhood and workplace size on mental health.

Methods We used UK Household Longitudinal Study (Understanding Society) data from women in paid employment, who participated in at least: one pre-COVID-19 Wave (9 or 10/11) and one COVID-19 lockdown wave (Lockdown 1: April 2020, Lockdown 2: November 2020, Lockdown 3: January 2021). Primary outcome was probable psychological distress, defined by ‘caseness’ (score≥4) in the General Health Questionnaire-12 (GHQ-12). In Model 1, exposure was motherhood (binary), interacting with a variable that split time in the pre-pandemic and lockdown periods. In Model 2, workplace size (Micro:1–24, Medium: 25–199, Large: More than 200 employees) was added as an exposure in Model 1 (3-way interaction) to investigate moderation effects. We fitted mixed-effects logistic regression models, adjusting for age, ethnicity, UK country of residence, cohabitation, educational qualifications, working hours, furlough, subjective financial difficulty and previous health condition.

Results Our final sample included 26,106 observations across 5,569 individuals. In the adjusted Model 1, pre-pandemic, odds of GHQ-12 caseness were lower for mothers compared to non-mothers (OR:0.89 95%CI:0.77,1.03). However post-pandemic, odds for mothers were higher than non-mothers, especially during lockdown 3 (Non-mothers: OR:1.92 95%CI:1.69,2.20; Mothers: OR:2.87 95%CI:2.36,3.49).

In Model 2, workplace size did not modify the relationship between motherhood and GHQ-12 caseness. Pre-pandemic, there was no difference in the odds of GHQ-12 caseness by workplace size; however, the differences observed in Lockdown 3 in Model 1 between non-mothers and mothers, are mainly attributed to differences in medium-sized and less for those working in micro or large enterprises (Non-mothers: OR:1.94 95%CI:1.53,2.48; Mothers: OR:3.59 95%CI:2.56,5.02).

Conclusion Motherhood did not have a differential effect on working women during lockdowns, apart from Lockdown 3 and specifically in medium-sized workplaces. Future analysis will attempt to investigate why these differences existed particularly in Lockdown 3. Nonetheless, policies should be designed to facilitate the working lives of mothers across all workplace sizes, but especially in medium-sized enterprises as these employees are likely to be affected more in times of extreme uncertainty.