Literature search

The databases Medline, PsycInfo, Scopus, CINAHL, Sociological Abstracts, Social Services Abstracts and Applied Social Science Index & Abstracts (ASSIA) were searched for peer-reviewed research articles published between 1 March 2020 and 7 September 2022. We searched for research articles reporting gender-disaggregated results on mental health outcomes related to unpaid care work in the context of COVID-19 containment measures in Europe. An initial database search was conducted on 21 October 2021 and updated on 7 September 2022 using the same search strategy according to Bramer and Bain’s [41] guide for updating search strategies for systematic reviews. To this end, we re-ran the database searches and imported all results into EndNote 20 software (Thomson Reuters, USA). These were copied into a new EndNote file, and the original results were checked for completeness and duplicates removed. The reference lists of selected articles were hand-searched to identify additional relevant studies. Language limitations were not applied. The search results were exported and uploaded to EndNote and then imported into Covidence [42] for the original search and to Rayyan [43] for the updated search and screening. In addition, a search was conducted in Google Scholar in May 2022. It was not updated as most of the references retrieved were identified via other sources. This search identified relevant grey literature and was conducted in addition to a targeted hand-search on websites of international and European organisations and institutions, which was performed in April 2022 and updated in December 2022. The results of these searches were transferred into Excel spreadsheets for the screening. A detailed documentation of the grey literature search including a guide for a systematic literature search on Google Scholar created by the corresponding author is available in S1 Appendix.

Search strategy

The database search strategy consisted of a combination of key words and controlled vocabulary (i.e., MeSH-terms, subject headings, and thesaurus words) derived from the search themes (i) “unpaid care work”, (ii) “COVID-19 containment measures”, and (iii) “mental health outcomes”. Titles and abstracts were searched for themes (i) and (ii) which were combined with the Boolean operator AND; search terms of theme (iii) were initially defined to narrow down a potentially high number of hits, but were ultimately not used throughout the process. Further details on the search strategy and search terms can be found in the published study protocol [40]. An example of the search strategy used in the MEDLINE database (via Ovid) can be found in S2 Table. The search terms for Google Scholar were adapted to the database search strategy as described in S1 Appendix.

Study selection

After removal of duplicates, the remaining identified peer-reviewed research articles and grey literature were screened for eligibility using pre-defined inclusion and exclusion criteria (Table 1). In short, original research articles and grey literature were eligible if they analysed primary or secondary data from European countries collected after the onset of the COVID-19 pandemic in March 2020.They were required to report mental health outcomes and measure unpaid care work, which would enable a description of the differential impact of COVID-19 containment measures by gender. Records from the database and grey literature searches were screened by two independent researchers in two steps: first, titles and abstracts were screened (for the grey literature, the introduction or first two pages were screened where an abstract was missing), and second, the full-texts of included titles/articles were screened. Disagreements between two reviewers were discussed bilaterally and, if necessary, between all reviewers until a consensus was reached.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Inclusion and exclusion criteria.

https://doi.org/10.1371/journal.pone.0308381.t001

Data extraction

Data extraction was limited to the relevant data items shown in Table 2; additional information and results that were irrelevant to the research questions were disregarded. Data from the full texts were extracted by two researchers independently who compared their findings before discussing them within the research team. Further information on data extraction and charting can be found elsewhere [40]. As it is not a common practice to assess the quality of studies in scoping reviews, the quality of selected studies was not formally appraised. However, reported study limitations were extracted and summarised together with the limitations identified by the research team in S3 Table.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Data items extracted from included full texts.

https://doi.org/10.1371/journal.pone.0308381.t002

Data analysis

The extracted data was coded into quantitative categories, qualitative descriptions or authors’ statements and documented in Excel spreadsheets, which served as data charting forms. This formed the basis for the narrative and descriptive syntheses of relevant measures and characteristics. The results were described narratively and reported along the main categories of the research questions.

Patient and public involvement

Neither patients nor members of the public were involved in the development of this scoping review. However, we do recognise that participatory research methods, such as consulting the target population in the development of research questions and the discussions of results, would have led to a much more realistic representation of the lived realities of people with unpaid caregiving responsibilities.

Search and study selection

The database search yielded 3,563 potentially relevant records and 271 additional records that were identified via hand-searching reference lists. After removing duplicates, 2,503 titles and abstracts were initially screened; 2,307 articles did not meet the inclusion criteria, leaving 170 full-text articles for review. A total of 14 research articles were selected for inclusion. The grey literature search yielded 1,704 references, of which 1,524 were identified via Google Scholar, 157 via websites and 23 via reference lists. Ultimately, 171 full texts were screened, of which 15 reports were included in this review. This scoping review reports on 29 studies that met the inclusion criteria. The study selection process is illustrated in Fig 1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. PRISMA flow chart of study selection process.

https://doi.org/10.1371/journal.pone.0308381.g001

Study characteristics

The key characteristics of the included studies are summarised in Table 3 and S3 Table. Settings included the UK (n = 10), Germany (n = 9), Italy (n = 2), Austria (n = 1), the Netherlands (n = 1), Slovenia (n = 1), Spain (n = 1), Türkiye (n = 1) and Europe (n = 3). Most of the articles and reports used quantitative research methods (n = 26), with only two qualitative studies and one mixed-methods study. Data was collected through surveys and interviews conducted exclusively online and/or by telephone. Seventeen studies used convenience samples and 12 studies used nationally representative datasets.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Study characteristics of included literature.

https://doi.org/10.1371/journal.pone.0308381.t003

The majority of included studies used a cross-sectional study design (n = 20), of which seven studies applied a quasi-pre-post design with retrospective questions to measure differences before and during the COVID-19 pandemic [22, 45, 46, 54, 55, 65, 67]. Longitudinal study designs were applied in seven studies, most of which used data from the Understanding Society: UK Household Longitudinal Study (UKHLS) [1, 49, 56, 62, 68]. An overview of the types of studies included in this scoping review is provided in Table 4.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Overview of type of studies.

https://doi.org/10.1371/journal.pone.0308381.t004

COVID-19 containment measures

As shown in Table 3, data was largely collected at the early stages of the COVID-19 pandemic, as early as at the onset of the public health crisis in March 2020 when the WHO declared the Coronavirus outbreak a public health emergency and stay-at-home orders started the first lockdown in most European countries. Most studies were conducted between April and July 2020, during the phase of the first lockdown in which containment measures including shutdowns of schools, workplaces, and non-essential shops had been in force for a few weeks (April-May 2020), and during phases of reopening of schools (April-June 2020) as well as gradual ease of lockdown restrictions (May-July 2020). The studies administered by the OECD (2021) [67] and Close the Gap & Engender (2021) [60] were conducted during the phase of the second lockdown (October-December 2020) when restrictions were re-imposed.

While most surveys and interviews were conducted during lockdowns in general, some studies focused on more specific COVID-19 containment measures and related personal outcomes such as working from home as a result of workplace closures [44, 47, 48, 61, 64], adaptation of work hours or work patterns with regards to childcare and home-schooling duties [1, 53, 64], and general work-family conflict during lockdowns [59, 61, 68]. In the longitudinal study reported by Zhou and Kan (2021), participants were asked whether they tested positive for COVID-19 during each study wave to control for contracting the coronavirus so that the period indicator could better represent the spread of the Coronavirus and COVID-19 related policy change at the macro level. The cross-sectional studies by Hipp & Bünning (2021) [53] and Illing et al. (2022) [65] included multiple study waves which allowed for outcome parameters to be compared over different phases of national lockdowns, as did the longitudinal studies by Hübgen et al. (2021) [64] and Zhou & Kan (2021) [56]. The latter study used different study waves that coincided with different stages of COVID-19 policy responses in the UK over a one-year timespan (April 2020-March 2021).

Measurements of unpaid care work

Table 3 depicts the definitions and operationalisation of unpaid care work underlying each study. Unpaid care work was mainly defined as housework (e.g., cleaning, preparing meals, day-to-day shopping) and personal care of household or family members within the nuclear family, consisting of mothers, fathers, (grand-)children and grandparents. Studies that focused on parents surveyed heterosexual (and predominantly coupled) parents. Only one study reported results for single parents [22] and two studies compared results of people with and without children in the household [58, 62]. The main measures were self-reported time spent on unpaid care work (n = 12) and the perception of equal sharing and primary caretaking between different-gender couples (n = 11). Seven studies provided subjective frequency data on the change in unpaid care work during lockdowns compared to the time before the COVID-19 outbreak. Examples of this were outlined in three qualitative studies (partly as part of a mixed-methods approach), in which changes in unpaid care work between women and men due to COVID-19 containment measures were examined in more detail [44, 47, 48].

Gender-specific differences in unpaid care work were mainly found for women in almost all studies. Ashencaen Crabtree et al. (2021) [44] and Balenzano et al. (2020) [45] found evidence of an unequal division of care work for women only. Although fathers were relatively more involvement in unpaid care work, mothers remained the primary caregivers and homemakers [45, 48, 55, 58, 64, 65]. While Chung et al. (2020) [59] found a tendency towards a more balanced distribution within couples as a result of the COVID-19 pandemic, Cannito & Scavarda (2020) [48] summarise that the division of unpaid care work did not change during lockdowns, especially for less egalitarian couples. Zoch et al. (2020) [70] emphasise that conforming to traditional gender roles seemed to increase the likelihood of an unequal care work division in the form of lower involvement of fathers and exclusive involvement of mothers in childcare.

Parenthood was identified as an important influencing factor in that gender differences in unpaid care work were greater for parents than for non-parents [52] or that parenthood was a more important predictor of reduced paid work during the pandemic than being a woman [53]. Being a single parent was associated with a greater childcare burden than being a coupled parent, especially being a single mother with children under 12 years of age [22].

Illing et al. (2022) [65] reported a higher childcare burden for parents of school-age children compared with parents of younger children. Zoch et al. (2020) [70] found that the share of fathers involved in childcare was greater for parents with younger children (under 14 years), while four other studies reported a greater gender care gap for parents of children aged 12 or younger at the expense of mothers [22, 64, 65, 67]. Parents of younger children also had greater difficulty reconciling paid work and unpaid work [44, 47].

The educational level of parents appeared to increase the likelihood of a more egalitarian division of additional care work. Before the COVID-19 pandemic, academic parents reported sharing childcare more equally than parents with lower education levels. During the pandemic, academic mothers reported being solely responsible for additional childcare tasks compared to academic fathers. Regardless of this, childcare during the pandemic was more evenly distributed among academic parents than among parents with lower levels of education [64]. Additionally, parents with a higher level of formal education reported having more access to formal emergency care in the first months of the pandemic and were more likely to rely on relatives for childcare [70].

In terms of parents’ employment status, being unemployed, working part-time and short-time were associated with taking on the majority of care work [53, 58, 64]. Although his was true for both women and men, working arrangements (i.e., working part-time vs. full-time) remained highly gendered [47, 53], and both working and non-working women spent more time doing housework than men [22]. According to a survey by the OECD (2021) [67], the gender care gap was largest in constellations where the mother was not working and the father was working. At the social policy level, the gender care gap in OECD countries was greater in countries where there were prolonged school closures and less was spent on family support (e.g., Türkiye, Poland and Ireland) and vice versa (e.g., Norway and Denmark).

Working from home was an important lockdown-induced change in working arrangements, and women were more likely to be working from home than men [61, 64]. As opposed to working on site, it was associated with an increase in unpaid care work for women and men [53]. In cases where fathers worked remotely, mothers were significantly more likely to become exclusive caregivers. In contrast, mothers’ status as essential workers, working longer hours or working on-site did not affect the likelihood of greater paternal or shared unpaid care work [70]. Hübgen et al. (2021) [64] saw a greater chance for equal housework division when both parents work from home.

Only one study reported findings for unpaid care work by ethnicity and showed that ethnic minority or racialised people performed more unpaid care work during compared to before the lockdown than white people [69]. According to this study, the gender gap was greater among white people; white men were the subgroup that performed the least unpaid amount of care work, while ethnic minority and racialised women performed the most.

While the average number of hours per week spent on informal care for sick, disabled or elderly persons increased by nearly a fifth during the lockdown, as European study shows [63] women caregivers spent more time on informal care. Lockdown-related restrictions on health and social services and other COVID-19-related difficulties led to an intensification of informal caregiving tasks, including emotional support, remote communication, practical help, care coordination and support, and help with transportation. In all these tasks, women caregivers reported a higher intensity than men caregivers [63]. The gender gap in informal caregiving was confirmed by two other included studies [62, 69]. One study also found that ethnic minority or racialised people, especially women, spent more time caring for adults, while white people were less likely to do so, with white men spending the least amount of time on informal care [69].

Measurements of mental health

Definitions of mental health, wellbeing, quality of life, life satisfaction, and related concepts in health sciences vary widely in the included literature. These concepts and corresponding outcome measures are often intertwined (e.g., quality of life is a subjective aspect of mental health and wellbeing is an aspect of quality of life). Some measures relate to more objective levels of mental health, while others relate to subjective aspects of wellbeing. The most commonly used mental health measure in the included studies was satisfaction with various aspects of life (n = 11), such as satisfaction with work life and family life [48, 53, 58, 61] and general life satisfaction [54, 64, 65, 69], which was measured using predefined scales. The perceptions and frequency of (parenting) stress [45, 54], anxiety [60], emotional exhaustion [46], and more general depressive symptoms [48, 54, 57, 63, 64, 66, 69] were also frequently used. Five studies investigated subjective wellbeing [22, 47, 52, 59, 67]; the Eurofound (2020) [22] study used the 5-item World Health Organization Well-Being Index (WHO-5). A further five studies, all using UKHLS data, used psychological distress as their primary mental health outcome, measured with the 12-item General Health Questionnaire (GHQ) [1, 49, 56, 62, 68]. Czymara et al. (2021) [51] addressed the concept of mental load (i.e., concerns, worries, personal experiences) as a theoretically relevant dimension of gender inequality. They refer to cognitive labour, i.e. the work of anticipating needs and identifying options for meeting those needs, to illustrate gender inequalities in experiences and concerns during the lockdown. An overview of the definition of mental health adopted in the included studies is provided in in Table 3.

Most studies (n = 21) showed that it was mainly women’s mental health that deteriorated during the pandemic. Analyses of UKHLS panel data suggested a connection between the amount of time spent on unpaid care work and gender differences in mental health: Respondents who already had childcare responsibilities before the pandemic showed lower levels of wellbeing during the pandemic [62]. The deterioration in wellbeing was significant for parents with substantial childcare responsibilities (i.e. over 15 hours a week), which applied to 70% of the surveyed mothers. An increase in weekly housework hours for women also appeared to be associated with a deterioration in their wellbeing. The wellbeing levels of men, who took on a much smaller share of housework already, appeared to be unaffected. The increase in informal care for people outside the household appeared to affect the mental health of women and men carers equally, despite the gender gap in informal caregiving (51% women vs. 45% men).

Tani et al. (2021) [68] came to similar conclusions: an intensive amount of time spent on childcare and home-schooling of at least 20 hours per week was associated with a deterioration in mental health. Working mothers and parents whose pre-pandemic income was above the median were particularly affected. The authors conclude that working mothers (vs. working fathers) and working parents of lower income households (vs. working parents of higher income households) had a poorer mental health and financial situation during the pandemic. Financial insecurity was identified as a predictor of poorer mental health for mothers and fathers in two other studies [49, 62]. A privileged living and working environment (i.e., more living space, remote working opportunity, family support) also seemed to make it easier for Austrian parents to reconcile working from home and home-schooling during the lockdown [46].

Generally speaking, persons with childcare responsibilities seemed to be more severely affected by lockdown-related mental health deterioration than non-parents [46, 53, 59, 62]. In contrast, results from the Mannheim Corona Study [58] suggested that women and men without children in their household were particularly dissatisfied with their family life as a result of the contact restrictions in Germany.

The age or schooling stage of children seemed to make a difference in how parents were coping with lockdown-induced additional caregiving responsibilities [47, 55, 58, 65]. Parents of younger children more frequently reported difficulties in reconciling work and family life [47]. For example, German mothers with children over the age of 6 were particularly dissatisfied with their work and family life [58]. The life satisfaction of German fathers, on the other hand, did not appear to be influenced by the age of their children [65]. Yerkes et al. (2020) [55] found significant gender differences in the quality of life among Dutch parents: Mothers reported a greater loss in their leisure time compared to father (57% vs. 36%); the presence of children of primary school age was a significant explanatory factor. A deterioration in work-life balance and a significant increase in disagreements about childcare division during the lockdown was observed for parents regardless of gender. Having primary school children appeared to be a more relevant explanatory factor.

Caregivers’ working arrangements seemed to have had an influence on parental wellbeing. For example, mothers who worked from home reported to be more satisfied with their work and family life than parents and non-parents who had to work on-site [58, 61]. Moreover, unpaid caregivers in essential work were more likely to report worse mental health outcomes than non-essential workers [55, 57].

They way in which parents perceived changes to their work and family life caused by the lockdown appeared to be gendered. The qualitative accounts of Italian parents reported by Cannito & Scavarda (2020) [48] showed that fathers believed that family interfered with work, leading to reduced productivity, feelings of guilt toward the workplace and higher stress levels. For the interviewed mothers, on the other hand, work interfered with family life, preventing them from responding to their children’s needs and from mentally detaching themselves from their work, accompanied by feelings of guilt toward their children. Similarly, British mothers were more likely than fathers to report that their work prevented them from spending time with their family (48% vs. 40%) and that their family prevented them from spending time with their work (49% vs. 32%) [59]. Czymara et al. (2021) [51] summarise that these differences in concerns as one part of cognitive labour reflected the social situation of women and men, namely one in which the traditional division of paid and unpaid work follows the male breadwinner model, even among the relatively well-educated.

Ethnic minority and racialised people appear to be disproportionately affected by COVID-19 containment measures, not only in terms of a greater caregiving burden, as previously mentioned, but also in terms greater deterioration in mental health. In Scotland, mothers were more likely to report higher anxiety levels than fathers; ethnic minority and racialised parents, especially mothers, reported the highest anxiety levels [60]. In line with these findings, another study from the UK shows although there were no statistical differences in coping levels with social isolation at the intersection of gender and ethnicity/racialisation, scores for life satisfaction and happiness were lowest for women of colour and highest for white men. Consistently, anxiety levels were highest for women of colour and lowest for all men. Overall, white men were most likely to have high levels of life satisfaction and happiness and, along with men of colour, the least likely to have high anxiety scores. All women were equally least likely to have high life satisfaction and happiness scores and high anxiety scores [69]. One study reported anxiety levels by disability status and found that disabled mothers were most likely to report high anxiety levels (58%) compared to non-disabled mothers (32%) and non-disabled fathers (24%) [60]. However, the authors did not provide data on unpaid care work by disability status. A summary of the findings on gender differences in unpaid care work and mental health can be found in Table 5.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 5. Gender and intersectional differences in unpaid care work and mental health.

https://doi.org/10.1371/journal.pone.0308381.t005

Concepts of gender and intersectionality

To account for the degree of their inclusion, we extracted data on the conceptualisation of gender, gender (in-)equality and mention of intersectionality (see S4 Table). The gender or biological sex of respondents was measured as a binary variable in all studies, with the exception of two studies in which it was measured as a categorical variable (female, male, diverse) but reported as a binary variable due to the small number of diverse individuals in the samples [51, 54]. Some studies referred to theories and theoretical concepts to support their arguments and hypotheses, including neoliberalism and the ideal worker notion [44], role theory [47], different fatherhood models, the concepts of femininity and masculinity [48], and ‘doing gender’ [51, 53]. Other authors discussed gender inequalities in the context of socio-cultural narratives and geographical contexts, highlighting traditionalist or conservative societal norms that promoted gender gaps in employment and unpaid care work in Italy [45, 48], Germany [53], the Netherlands [55], the UK [56], and Türkiye [66]. Two studies collected data on attitudes towards gender roles and found that more traditional gender role attitudes were associated with lower levels of fathers’ childcare involvement [70]. Hübgen et al. [64] found that women in Germany appeared to have more progressive views towards gender roles than men; the same was true for university graduates compared to non-academics.

The theoretical concept of intersectionality was explicitly mentioned in two reports [60, 69], which emphasised the importance of including gender, racialisation, disability and other lived realities in the analyses of COVID-19 related mental health and social disadvantage of racialised and disabled women. Nine other publications implicitly integrated intersectional approaches into their gender analyses by including other dimensions of social inequality such as socio-economic status [45, 52], working arrangements and employment situation [53, 58, 61, 70], parenthood [53], education [54, 64, 70], and financial situation [68].

Summary of results

The aim of this scoping review was to explore the impact of European policy measures to contain the spread of COVID-19 on the division of unpaid care work and related changes in mental health between gender groups, with a focus on intersectional dimensions of potential gender inequalities. As expected, the collection of evidence published up to two years after the outbreak of the pandemic largely confirms that women bore the greatest burden of the COVID-19 crisis, as predicted by empirical work before [7–10] and during the pandemic [22]. The time spent on childcare, housework and informal caregiving was already gendered before the pandemic, with women taking on these traditional caregiving responsibilities more often than men [2, 3]. Although men’s unpaid care work appears to have increased during lockdown phases, especially among different-gender coupled parents, it did not reach women’s degree of involvement. What is more, women caregiver’s mental health was the most affected during the pandemic when compared to men caregivers, largely due to greater involvement of women in unpaid care work. Although a trend toward a more egalitarian division of unpaid care work can be observed in European countries, the gender care gap remains.

Intersections of parenthood, paid work, and gender

Parenthood has proven to be one of the biggest influencing factors on caregivers’ mental health during the lockdown. Parents, most notably mothers and working parents, showed a greater overall deterioration in mental health than non-parents [52]. Parents of school-age children seemed to be particularly affected by school closures and home-schooling efforts [44, 47, 65]. Balancing work and family life during the pandemic appeared to be a particular challenge for working parents. This was strongly linked to the amount of time spent on childcare and home-schooling [49, 62, 68] as well as financial insecurity [49, 68]. Lockdown-induced work and care arrangements were highly gendered [47, 53] and both working and non-working women spent more time on unpaid care work than men [22, 67]. Working mothers had a higher overall burden from the combination of paid work, childcare and education, housework and commuting [65] and showed poorer mental health outcomes than working fathers [49, 55, 68].

Intersections of working arrangements and gender

Working constellations affected how caregiving was distributed between partners: the gender care gap appeared to be largest in the most traditional work-care arrangements, where women were not employed and men were employed [67]. However, the results are inconclusive: Analyses from the initial phase of the lockdown in Germany (March and April 2020) showed that in constellations in which only one partner was employed or working onsite, this partner performed less unpaid care work than their non-employed or remote-working partner, regardless of gender [53]. A few months into the German lockdown, in May and June 2020, fathers’ remote work contributed significantly to exclusive maternal caregiving, while mothers’ working constellations did not promote paternal or shared caregiving [70]. Although working from home increased caregiving for both gender groups [53], women were more likely to work remotely than on-site compared to men [61, 64]. Parents working in essential sectors (e.g., healthcare, social services, day care, education, supermarkets) were more likely to report poorer mental wellbeing during the COVID-19 pandemic than non-essential workers [55, 57], especially if their children were attending primary school [55]. At the same time, women were more likely to work in essential sectors during the pandemic and took on the majority of unpaid care work. While the COVID-19 pandemic brought challenges, it also provided working parents with opportunities to experience positive changes in their daily routines and family dynamics. Remote work allowed some parents more flexible working arrangements and a better work-family balance, especially for mothers [48, 58, 61]. This required the possibility of working remotely and prerequisites that contributed to a privileged and less stressful working-from-home environment, such as a spacious apartment with a separate room [47]. Moreover, remote-working parents had more opportunities to spend time with their children, which appeared to be beneficial for mothers’ wellbeing [44].

Intersections of financial security and gender

Financial insecurity has proven to be an important predictor of poorer mental health for both mothers and fathers [49, 62], with working mothers and working parents of poorer households experiencing higher levels of mental and financial distress [68]. Working women suffered less earnings reduction than working men during the COVID-19 pandemic, largely due to the higher proportion of women working in essential fields such as health and social care. Nevertheless, women’s wellbeing and mental health were more sensitive to COVID-19 containment measures compared to men, with greater gender differences during the more restrictive phases of lockdowns [56].

Intersections of racialisation/ethnicity, disability, and gender

The results of this scoping review highlight that, among European women, population groups that are often understudied, such as women who are single parents, disabled or of colour, have the greatest increase in unpaid care work and the greatest deterioration in wellbeing [22, 60, 69]. The lack of diversity in study populations is a short-coming of the studies included in this scoping review. All included studies that reported results at the intersection of gender and racialisation, ethnicity or migration status used the umbrella terms ‘BAME’ (black, Asian and minority ethnic) and ‘BME’ (black and minority ethnic). This is highly critical as it emphasises some groups to the exclusion of others and portrays people of colour as a homogenous group, leaving no room for differences between and within different migrant and racialised groups. Similarly, disabled unpaid caregivers had much higher levels of anxiety than able-bodied parents [60], however all but one study failed to consider a disability perspective in their analyses altogether. In terms of intersecting social dimensions and lived realities, researchers need to consider more diverse study populations and more innovative ways to reach them, e.g., through qualitative interviews and participatory approaches.

Intersections of societal norms and gender

The division of unpaid care work between gender groups, especially in different-gender coupled households, is not only the result of negotiation processes between gender groups (especially between romantic partners), but it is also strongly influenced by the underlying societal norms that shape the roles of women and men in society and the expectations placed on them in terms of paid and unpaid labour [4, 5]. As the uptake and division of unpaid care work between women and men is highly gendered and normalised, measures to contain COVID-19 would naturally affect gender groups differently, as women and men are likely to adjust their behaviour to what is expected of them based on gender-normative ideas [71].

Most of the included studies lacked theoretical approaches to analyse the underlying mechanisms of gender inequality. These include most prominently ‘doing gender’ [71], bargaining position [72] and time availability, as highlighted by Steinmetz et al. (2022) [73], which are all necessary to understand gender inequalities during COVID-19 (see S4 Table). For example, changes in time availability affect the bargaining position in couple households [73] in that less time for commuting and school trips means more time for childcare and housework for parents, especially for fathers. Adherence to gender-normative beliefs, according to which women and men adapt their behaviour to societal expectations, would naturally lead to a greater involvement of mothers in unpaid care work and less or unchanged involvement by fathers [70]. Contrary to initial debates, we find no evidence of a re-traditionalisation of gender roles, whereby women, especially mothers, would shoulder the additional unpaid care work induced by the various containment measures [74]. However, we cannot conclude that the COVID-19 pandemic promoted a more egalitarian division of unpaid care work either, although we do see a trend towards greater involvement of men and fathers in personal care and housework [59].

This is particularly evident in more conservative welfare states with more traditional gender roles, such as in Italy and Türkiye, where women took on additional care work, especially childcare, when external care services were not or were no longer available, as has been the case during the COVID-19 pandemic [67]. This conservatism in European countries is reflected in the gender care gap, with a more traditional division of unpaid care work especially in Italy [45, 48], Germany [50, 51, 58, 70], Austria [47], and the Netherlands [55]. In some liberal welfare states such as the UK, however, there is a trend towards a more egalitarian division of paid and unpaid labour between women and men [59]. This is confirmed by studies in other liberal welfare states, such as the United States [75], Canada [76] and Australia [77].

The changing demands on unpaid care work also translate into changes to working arrangements [70]. In terms of working conditions, more working women than working men reduced their working hours or even quit their jobs to care for their children during the pandemic, although the male-breadwinner and female-caretaker models do not seem to be holding [53]. The reduction in working hours suggests that the ‘one-and-a-half-earner’ model [78] appeared to prevail in European countries and was being reinforced during the pandemic. This favoured a more traditional division of unpaid care work, even if this is not true for all people and population groups studied.

Scope and limitations

Despite rigorous research practices underpinning this scoping review, there are some limitations to our study. Firstly, while our goal was to include a wide range of studies to capture different perspectives, we acknowledge the potential for selection bias, as studies from high-income Western European countries predominate in the final selection and other parts such as Eastern Europe were excluded. The gender-specific division of paid and unpaid labour is embedded in the specific socio-economic and socio-political structures of a society and fundamentally determined by these circumstances. This also applies to mental health, which is influenced by interactions between individual and socio-structural conditions. The relationship between gender, unpaid care work and mental health, on which this review is based, is situated in a specific socio-economic and socio-political context, namely in high-income Western European countries with predominantly liberal and conservative welfare states. Therefore, the generalisability of our results to other societal and structural contexts is limited. This limitation reflects the availability of research on this topic. Since gender inequalities were found even in these contexts, we assume that these differences are more pronounced in other economies and more conservative welfare states. Nonetheless, the literature searches were conducted in English-language sources using English search terms, which may have led to an under-representation of non-English literature that could have contributed to a more diverse set of findings.

Secondly, although we tried to include a range of study designs, the predominance of quantitative studies in the final selection could introduce a bias in favour of certain research methods. Future studies could employ more diverse study designs, especially qualitative ones, to mitigate this limitation. Most studies were based on online surveys, which require access to digital devices and a certain level of digital literacy to participate. While online surveys were essential for data collection during the COVID-19 pandemic, there is a risk of over-representation of people with better access to digital devices and higher levels of digital literacy. This could affect the generalisability of the results to populations with limited internet access and digital literacy.

Thirdly, a limitation of scoping reviews is the neglect of quality assessment of included studies. The studies included in this review differ in terms of study populations, cultural contexts of countries, pandemic responses and the type of data used. In the early stages of the COVID-19 pandemic, ad hoc data of varying quality was collected, ranging from highly selective convenience samples to probability-based longitudinal data (see Table 4). Most studies lack pre-pandemic information, while some use retrospective measures with a potential recall bias [79]. Only a few studies investigated the impact of specific COVID-19 containment measures or included them as measurable instruments in their analyses. These include longitudinal studies with multiple study waves that correlate with different lockdown phases. The most recent survey in this scoping review refers to spring 2021, meaning that our study covers the short-term impacts of COVID-19 containment measures on unpaid care work and mental health. To investigate longer-term impacts and causal effects, more recent studies and further longitudinal studies with elaborate methodology are needed.

Fourthly, as noted in some studies, the populations under study consisted of highly educated and more financially stable individuals, which means our results are biased towards middle- or higher-class populations. At the same time, racialised individuals, migrants and ethnic minorities, as well as working-class individuals and those in precarious or irregulated working arrangements are heavily underrepresented, further limiting the generalisability of our results.

What is more, conducting a scoping review on a phenomenon such as the COVID-19 pandemic is subject to limitations and challenges given how rapidly it is changing. This scoping review mainly included cross-sectional studies, which are not able to depict changes in outcome parameters over time. Therefore, most studies were not able to compare different pandemic periods and related unpaid care work and mental health outcomes. The longer-term impacts and effects on the consequences of the gendered division of unpaid care work on the mental health of caregiving women, men and gender-diverse persons remain unclear. We expect further studies on this topic in the upcoming years. So far, we have been able to identify basic patterns of how COVID-19-related containment measures affect unpaid care work and mental health differently. We were also able to identify potentially at-risk-groups for increased caregiving burden and worsening mental health.

Finally, given the wide range of measures of unpaid care work and mental health included in this scoping review, the results must be interpreted with caution. There are a variety of measures of unpaid care work, with few studies using measures of time-use and others relying on subjective relational and retrospective measures, which are more prone to bias. Additionally, self-reported measures of unpaid care work may underestimate women’s share of caregiving within couple relationships, while men tend to overestimate their actual share of childcare and housework [80]. There is an even greater variety of measures in mental health, ranging from screening instruments for psychological distress and overall mental health to measures of wellbeing and quality of life. Arguably, measures of wellbeing, satisfaction and quality of life are not measures of mental health, but rather mediators that help explain how different phases of COVID-19 lockdowns affect mental health, or moderators that influence the magnitude or direction of the impact on mental health [81]. Although not necessarily equivalent to direct measures of mental health disorders, measures of wellbeing and quality of life provide valuable insights into individuals’ daily lives and challenges [82]. Our aim was to include as many studies as possible using a broad definition of mental health. Future studies should use multiple types of caregiving and mental health measures to obtain a more nuanced picture.

Gaps and future directions

Unpaid care work remains largely invisible, not only in traditional families, but also in academia. In many studies on the impact of COVID-19 containment measures on mental health, unpaid care work was not measured, and when included, many authors did not report their results by gender. This is reflected in the number of full-text articles (n = 43) we had to exclude (see Fig 1). Surprisingly, informal caregiving to elderly, sick and disabled people was the least gender-disaggregated form of unpaid care work, although informal caregivers make up 10–25% of the European population and are predominately women [83].

None of the included studies considered non-heterosexual individuals, same-sex couples and non-heteronormative constellations of families and unpaid care work, which was often due to the far too small number of non-heterosexual or non-cis-gendered individuals in the various samples. Investigating the distribution of unpaid care work in queer couples and differences in mental health between partners, for example, could shed light on negotiations in partnerships and households where heteronormative beliefs may have less influence. Evidence from before the COVID-19 pandemic suggests that same-sex couples have a more egalitarian distribution of housework compared to opposite-sex couples, with gender-affirming behaviour being less important and the availability of time and relative economic resources being more taken into account [84, 85].

Future research and policy must broaden their focus from the ‘typical’ unpaid care work and its ‘traditional’ distribution between women and men, mothers and fathers, to the cognitive and emotional dimensions of gender inequality. Unpaid care work consists not only of practical activities, but often also of cognitive and psychological processes, such as organising, planning and managing family work, worrying and anticipating needs. These types of cognitive and emotional labour have far more stressful impacts on caregivers and may lead to greater gender disparities in mental health [51, 86], but are largely overlooked in the literature on gender inequalities. Our findings suggest that cognitive and emotional labour was gendered and resulted in higher mental load for women, particularly mothers, as evidenced by higher parental stress levels [59, 68], worries related to children’s wellbeing [48] and greater dissatisfaction with unpaid care work division [45, 50] compared to men and fathers.

Another factor that should be taken account in future studies is the contextualisation of results with regard to other policies (e.g., family, equality, and labour-market policies) and the socio-political characteristics of individual European countries. Some of the studied countries reacted very differently to the pandemic, which does not make it easy to compare different settings. Future studies need to address unpaid care work, gender equality and diversity, as well as policies to promote them. Measures to combat the pandemic also varied greatly and need to be analysed in terms of gender inequality and other disadvantages, as well as policies that address them.