Dietary habits, physical activity, and sedentary behaviour of children of employed mothers: A systematic review

Highlights • Dietary pattern is poorer among children of employed mothers.• Children of employed mothers are more physically active.• Children of employed mothers experience greater prevalence of sedentary activity.


Introduction
Two important worldwide trends can be identified in recent years: increasing prevalence of childhood overweight/obesity and increasing participation of women in the paid labour force. Childhood obesity is an emerging salient public health challenge of the 21st century (WHO, 2020). Childhood obesity is risky as it has strong associations with likelihood of adult obesity, which has led to the increasing risk of morbidity, including non-communicable diseases (NDCs) such as cardiovascular disease, type 2 diabetes mellitus, some cancers, poor skeletal health, and some aspects of mental health Das and Horton, 2012;Lee et al., 2012;Wolin et al., 2010).
According to World Development Indicator (World Bank, 2020), worldwide female employment rate increased substantially in the last century. Employment creates a double burden for women as they often take the family responsibilities of unpaid household tasks and childcare due to traditional division of labour. Balancing with daily multiple roles and responsibilities, employment may impact upon the wellbeing of children if, as hypothesized, employed mothers spend less time on household activities centred on children, such as children's diet and physical activity (Bianchi, 2000;Cawley and Liu, 2012). However, employment may contribute to greater economic opportunities and resources, which may also enhance health and wellbeing (Waddell and Burton, 2006).
Literature from the USA (Datar et al., 2014;Anderson et al., 2003), U.K. (Hawkins et al., 2008), Canada (Chia, 2008) and Germany (Baten and Böhm, 2010) have demonstrated that children of employed mothers demonstrate a trend towards being overweight due to changes in food intake patterns (e.g., homemade food vs meals from outside, more processed food and 'junk' food), reduced physical activity and increased sedentary behaviour. The latter is defined as sitting or lying with low energy expenditure during waking hours . Poorer health behaviours among children (e.g., unhealthy dietary patterns, physical inactivity, sedentary behaviours) serve as gateways towards poorer health trajectories and increased health comorbidities in adulthood, including being overweight and obese (Mu et al., 2017). Childhood adiposity as well as physical inactivity and sedentary behaviour among children; represent key modifiable risk factors, to enhance both current and future health outcomes (Raynor et al., 2012).
There has been a visible shift in women's employment in low to middle-income countries (LMIC) over the past two decades (Dodzin and Vamvakidis, 2004;Lopez-Arana et al., 2013). The increase in women's participation in the labour force parallels the increasing prevalence of overweight among children (BMI Z-score > 2), which is believed to occur as a result of the country's nutrition transitions (referred to as characteristic changes in food and physical activity patterns that occur as a result of macro-level changes in economic development, globalization and urbanization) (Lopez-Arana et al., 2013). While the literature confirms that children of working mothers in developed countries demonstrate a trend for being overweight (Datar et al., 2014;Anderson et al., 2003;Baten and Böhm, 2010), research related to dietary patterns, physical activity and sedentary behaviour of children in LMICs are scarce to confirm any definite relationship. Thus, the relationships between maternal employment and children's dietary patterns, physical activity and sedentary behaviour are largely unexplored in LMIC.
No systematic review has investigated how dietary patterns, physical activity and sedentary behaviours (in combination) among children are related to maternal employment. Some previous research that has been identified investigated the association of maternal employment with child obesity and discussed solely one or two of these behaviours as influencing factors towards overweight and obesity (Duch et al., 2013;Mech et al., 2016;Hoyos Cillero and Jago, 2010;Shrewsbury and Wardle, 2008). However, not only single behaviours, but the combination of multiple risk behaviours ultimately determines the risk of being overweight or obese. It is already identified that consumption of energy dense foods, low levels of physical activity and high levels of recreational screen use (e.g., TV watching and computer use) are key behavioural determinants of overweight and obesity in children and adolescents (Barnett et al., 2018;Rennie et al., 2005). Dietary and physical activity habits are developed at early stages of life (Savage et al., 2007) and have been tracked into adulthood (Kelder et al., 1994;Tammelin et al., 2014), suggesting the importance of increasing our understanding of the roots and development of these behaviours in children. The relationships among maternal employment and these three key behavioural variables remains largely unknown. Since approximately 40% of the global workforce are women (World Bank, 2020), a comprehensive understanding of association of maternal employment with child dietary patterns, physical activity and sedentary behaviour needs more focus considering its short-and long-term impacts on health and wellbeing trajectories over the life course.
This systematic review aims to identify the association between the dietary habits, physical activity and sedentary behaviour of children, with the employment status of mothers. Our main research question is, therefore, 'does the employment status of mothers with children aged 6-18 years affect children's dietary habits, physical activity and sedentary behaviour?'

Methods
The research protocol of this study is registered in PROSPERO, an international prospective register of systematic reviews (registration number: CRD42020145438).

Search strategy
The review followed the protocol of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (Moher et al., 2009). Literature searches were conducted using the bibliographic databases of Scopus, PubMed, Science Direct, JSTOR, Google Scholar and ProQuest. For the primary search, no restriction was imposed on publication type, or study design, however, only English language papers were considered. Initial search was applied to title, abstract and key words. For working mothers and physical activity, the following search terms were used: (physical activity OR physical inactivity OR exercise OR sedentary behaviour) AND (maternal work OR working mother OR maternal employ* OR working women) AND ('child' OR 'adolescent' OR 'youth' OR 'juvenile'). For diet, 'dietary pattern OR dietary intake OR food intake OR dietary habits OR feeding behaviour' were used instead of 'physical activity' or 'sedentary behaviour'. Additional literature and document searches were undertaken via backward searching through the key words identified from the literature review and the secondary literature search of the reference lists of all full text articles selected in the primary search. The search strategies are presented in Supplementary Material.

Inclusion criteria
Studies were incorporated in the present review if they (i) reported on maternal employment status, and dietary patterns (DP), physical activity (PA), or sedentary behaviour (SB) of children aged 6 to 18 years; (ii) were written in English; (iii) published as a peer reviewed journal article, conference paper, or thesis at masters or doctoral level. All research methods, designs, as well as measurement instruments were included. Studies were not considered for inclusion in the systematic review if: (i) the population of the study was only adults, (ii) had obesity or overweight as the focal point of research outcomes, (iii) published as a literature review, (iv) did not provide information about the age of the study population, (v) full text was not available and (iii) was not written in English.

Study selection and data extraction
All studies identified by database search and additional searches were screened for eligibility based on title, abstract and full text by two independent reviewers (SA and LB). Any disagreements were resolved by discussion with the other reviewers/authors.
The data for each included study were extracted by the first author using a standardized extraction form and verified by other authors. Data were extracted on: (1) characteristic of publication [title of the article, author(s), year, country/data source, study design], (2) sample characteristics [sample size, age/age group, employment status/working hours], (3) primary and secondary outcomes as well as measurement methods used for dietary patterns, physical activity and sedentary behaviour. A narrative synthesis of results of included studies was provided. The number of studies included in the systematic literature review were too diverse in outcome as well as measurement to pool data to conduct a meta-analysis.

Coding associations with dietary outcome, physical activity and sedentary behaviour
Studies with significant associations identified between maternal employment and domains of dietary patterns, physical activity and sedentary behaviour were not discussed unless three or more studies were available (for each category) (Sallis et al., 2000). Conceptually similar domains were combined if there were not enough studies to examine the domains individually. For example, 'snack food including fast food & junk food' domain combined fast food, junk food and processed food. Domains relating to physical activity in included studies were too diverse to report on separately, thus conceptually similar domains were aggregated as moderate to vigorous intensity physical activity (MVPA).
Studies with significant associations between maternal employment and variables (dietary patterns, physical activity and sedentary behaviour) were included in the 'Related to maternal employment' column of Table 2; and associations were classified and coded as: positive association (+), negative association (− ). Studies reporting no significant associations were entered in the "Unrelated to maternal employment" column. The coding process was completed following the rules used by Sallis et al. (2000). Studies with low risk of bias scores are presented in bold numbers in the Table 2. Included studies typically used univariate tests for assessing the statistical significance of associations. However, even if multivariate tests were conducted, univariate tests were reported for consistency across studies to ensure meaningful comparisons of key findings.

Summary codes
Numbers in the second and the fourth columns of Table 2 refer to the study numbers in Table 1. Studies that examined multiple domains of dietary patterns, physical activity or sedentary behaviour, multiple associations with maternal employment were recorded. The column 'number of samples' includes the number of samples that have been studied for each identified domain. The 'Summary' column contains a code to summarize the state of the domain for that variable. After assessing all the studies, calculating the percentages of findings supporting the overall association, each domain was classified as no association (0%-33 % of studies supporting the association), indeterminate/ inconsistent (34%-59 % of studies supporting the association) and positive or negative association (60%-100 % of studies supporting the association) and coded as '0 ′ , '?' and '+/− ' respectively. These rules for classifying variables strength of evidence and direction of association are in accordance with Sallis et al. (2000).

Risk of bias
Risk of bias of the included studies was assessed using a modified version of Cochrane Collaboration tool adopted for observational studies following Higgins et al. (2011), and the JBI Critical Appraisal Checklist was used for the qualitative study. The adapted tool of Cochrane Collaboration has been used in prior studies (Poitras et al., 2016;Prince et al., 2017;Castro et al., 2018). The tool for observational studies focused on six potential sources of bias: selection bias (sampling method), performance bias (measurement of maternal employment), detection bias (measurement of DP, PA, SB), attrition bias (completeness of outcome data), selective reporting bias (selective outcome reporting), and other bias (control for confounding). Each type of bias was marked as "high", "low", or "unclear" according to pre-specified criteria. The comprehensive explanation of these criteria is provided in the supplementary document. One reviewer [SA] assessed the risk of bias score while the other reviewers verified these by assessing randomly selected 2 studies each and discussed any conflicting results (initially 83% consistency was attained between reviewers). Further disagreements were resolved through team discussion. The overall risk of bias score was determined by summing the total number of criteria marked as 'low risk of bias', 'high risk of bias' and 'unclear risk of bias' according to the preestablished criteria. The JBI Critical Appraisal Checklist was used to assess the quality of the qualitative study based on study methodology. The corresponding score (out of 10 with 10 the highest) and the JBI Level of Evidence of Meaning (range from 1 to 5 with 5 the lowest) was applied. (Detail documents are available in supplementary files).

Results
The search of bibliographic databases yielded 14,306 potentially relevant citations, with a further 88 identified through the secondary backward reference searching. Full text papers were reviewed for 108 studies, of which 68 were excluded. A further two papers were identified from reviewing reference lists of included papers, providing a total of 42 papers for the review, as shown in Fig. 1. The papers were published between 1984 and 2020, with 95.2% published in 2000s (see Fig. 2). Most studies were peer-reviewed journal articles (85.7%), with others being theses (9.5%), conference papers (2.3%), and working papers (2.3%). Results showed an association of maternal employment with all three variables of dietary patterns, physical activity and sedentary behaviour in 9.5% papers, while 59.5% of studies focused on any one of the three variables. The remaining studies (30.9%) included a combination of two variables (dietary pattern-physical activity; dietary pattern-sedentary behaviour or physical activity-sedentary behaviour). This review identified 10 domains for dietary patterns, two for physical activity and four for sedentary behaviour.

Maternal employment and dietary patterns
Among the 42 studies included in the review 26 assessed dietary patterns, with 11 assessing dietary patterns using standard dietary pattern questionnaires, three with 24 h dietary recall, two studies used 3-day food diaries, four used food frequency questionnaires (FFQ), two used a healthy eating index (HEI), two used a youth and adolescent food frequency questionnaire (YAQ) and Global school based student health survey (GSHS) questionnaire, and 2-day food diary each were used by one study. Ten studies used self-reporting or parent reporting measures to assess dietary patterns. The review identified ten domains of dietary patterns, and eight were studied three or more times. Snack's food including fast food & junk food was the most assessed domain of dietary pattern used in 9 studies followed by family meals assessed in 8 studies. Four studies (Bauer et al., 2012;Datar et al., 2014;Gaina et al., 2009;Meyer, 2016) among nine reported an increase in snack food consumption when the mother was in full-time employment, three studies (Brown et al., 2010;Sweeting and West, 2005;Taylor et al., 2012) reported no association, and two studies (Adbi et al., 2017;Pearson et al., 2009) reported a negative association with employment status of the mother. Family meal (eating together and meals with family members) had a negative relationship with maternal employment in seven (Anderson, 2012;Bauer et al., 2012;Chang and Lee, 2012;Chang, 2012;Gwozdz et al., 2013;Nadia, 2012;Neumark-Sztainer et al., 2003) out of eight papers, with one (Nie and Sousa-Poza, 2014) showing no association. Maternal employment was negatively associated with children's fruit and vegetables consumption in five studies (Bauer et al., 2012;Datar et al., 2014;Meyer, 2016;Nadia, 2012;Sethi et al., 2014) and one study (Taylor et al., 2012) reported no association. Maternal employment was negatively associated with children's juice, soda, and soft/ sports drinks consumption in one sample (Nadia, 2012), positively in one sample (Datar et al., 2014) and no association in another study (Taylor et al., 2012). Healthy eating habits was negatively related to maternal employment in three studies (Bauer et al., 2012;Nadia, 2012;Sweeting and West, 2005) and positively related in one study (Honajee et al., 2012). Five studies (Chowhan and Stewart, 2014;Fitzsimons and Pongiglione, 2019;Gaina et al., 2009;Nadia, 2012;Shuhaimi and Muniandy, 2012) reported a negative relationship with maternal employment for eating meals regularly.
Dietary quality showed a positive relationship with maternal employment in one study (Touliatos et al., 1984), a negative relationship in two studies (Ben-Shalom, 2010;Li et al., 2012) and was unrelated in another study (Taylor et al., 2012). Maternal employment was Table 1 Characteristics of included studies, along with the results of the study quality assessment for each study (n = 42).

Study
Type of country

Parents interview on Tv watching hours/ week
In married couple family's food-intake quality decrease with maternal employment, but this association is weaker for single mother families (first study) and children are more likely to get rapid exercise when their mothers work more hours per week (second study). Gwozdz et al., 2013 (34) Cross sectional study 5-9 years (n = 7000)

High risk of bias
Multiple regression and  negatively associated with children's energy/calorie, protein, and fat intake in two samples (Gwozdz et al., 2013;Shuhaimi and Muniandy, 2012), and unrelated in one sample (Nie and Sousa-Poza, 2014). The effect of intensity of maternal employment on children's DP was reported in 8 studies (Bauer et al., 2012;Datar et al., 2014;Fitzsimons and Pongiglione, 2019;Gaina et al., 2009;Li et al., 2012;Meyer, 2016;Nadia, 2012;Sweeting and West, 2005). Four studies (Bauer et al., 2012;Datar et al., 2014;Gaina et al., 2009;Meyer, 2016) reported a positive association of full-time maternal employment with consumption of soda and fast food whereas another four studies found three domains (eating meals regularly, fruits and vegetables consumption, dietary quality) to be negatively associated with full-time maternal employment (Fitzsimons and Pongiglione, 2019;Bauer et al., 2012;Gaina et al., 2009;Meyer, 2016). Dietary patterns and maternal part-time employment were reported in three studies (Fitzsimons and Pongiglione, 2019;Gaina et al., 2009;Sweeting and West, 2005) -positively associated with two domains (snacking and eating dinner regularly) in one study (Gaina et al., 2009) and negatively in two domains (eating breakfast regularly and healthy eating) in two studies (Fitzsimons and Pongiglione, 2019;Sweeting and West, 2005).
In brief, 28 samples reported negative, 9 samples positive, and 9 samples reported no association between children's dietary patterns and maternal employment. Overall, the associations between maternal employment and children's dietary patterns showed more adverse than favourable directions.
Briefly, 11 samples reported positive association whereas 5 samples observed negative association with maternal employment. Overall, working mothers were more likely to have active children.

Differences in outcome between countries differing in income status
The World Bank's income classification of countries was used in this review. Based on this classification, 32 of 42 reviewed studies were from HIC, 7 from UMIC, and 3 from LMIC. Table 3 summarizes the associations between maternal employment and dietary patterns, physical activity and sedentary behaviour in high-, upper-middle and lower-middle income countries. The association of dietary patterns, physical activity and sedentary behaviour showed an indeterminate relationship with maternal employment in LMIC and UMIC. Twenty-six studies examined dietary patterns in HICs with the majority from the USA (n = 16). The association between "healthy dietary patterns of children in HIC" were indeterminately related to maternal employment. Physical activity and sedentary behaviour were found to have positive associations with maternal employment in HIC. Eight out of thirteen studies reported positive associations between children's physical activity and maternal employment. Twenty-one studies reported sedentary behaviour, with 14 positively related to maternal employment.

Risk of bias assessment
Across all studies, 57.1% (n = 24) had high risk of bias score. The remaining 40.5% studies (n = 17) were classified as low risk of bias score, while 2.4% (n = 1) had unclear risk of bias. Concerning each criterion of risk of bias for observational studies, 56.1% of studies (n = 23) used probability sampling methods and were hence identified as low risk of bias. The majority of the studies (70.8%) had high risk of performance bias as most cases of 'maternal employment' were measured using non-validated tools. Similar to performance bias, over half of detection bias (51.2%) had high risk for using non-validated measurements for dietary patterns, physical activity and sedentary behaviour. Over one third of studies (41.5%; n = 17) did not manage < 20% of missing data and were marked as high risk for attrition bias. One third of studies (34.1%; n = 14) did not provide any reasons for missing data and were identified as unclear risk of attrition bias. Reporting bias is low for two-thirds of the studies (70.7%; n = 29) and 17.1% (n = 7) had high risk of reporting bias. Nearly two-thirds of studies (63.1%; n = 26) reported about the statistical methods to control for potential confounding factors, and hence were coded as low risk of bias; and 22% (n = 9) did not provide sufficient information regarding confounding factors. Detailed risk of bias results is available in the supplementary material.
Effect size is measured for individual studies. Effect sizes ranged from − 0.08. to 3.8. Majority of studies fail to produce medium to large effect size.

Discussion
The aim of this systematic review was to determine whether maternal employment was associated with children's health behaviours, specifically dietary patterns, physical activity, and sedentary behaviour. A wide range of domains for these behaviours among children within HIC and LMIC were identified. These domains were assessed using various tools and the association with maternal employment was varied. The review shows that the number of studies on these lifestyle variables among children related to maternal employment has expanded in the last two decades, with two studies between 1980s and 1990s, and all remaining studies published in the 2000s.
Results showed that maternal employment was inversely associated with children's family meals (Anderson, 2012;Bauer et al., 2012;Chang and Lee, 2012;Chang, 2012;Gwozdz et al., 2013;Nadia, 2012;Neumark-Sztainer et al., 2003), fruits and vegetables consumption (Bauer et al., 2012;Datar et al., 2014;Meyer, 2016;Nadia, 2012;Sethi et al., 2014), healthy eating habits (Bauer et al., 2012;Gwozdz et al., 2013;Sweeting and West, 2005), eating meals regularly (Chowhan and Stewart, 2014;Fitzsimons and Pongiglione, 2019;Gaina et al., 2009;Nadia, 2012;Shuhaimi and Muniandy, 2012), and energy/calorie, protein, and fat intake (Gwozdz et al., 2013;Shuhaimi and Muniandy, 2012). In dual income families it is expected that maternal employment may allow families to spend more on healthy foods (Lowery et al., 2019), however, employment is likely to create a time constraint for meal related behaviours (Devine et al., 2003;Jabs et al., 2007). Studies show that employed mothers spent significantly less time in meal preparation (Beshara et al., 2010;Cutler et al., 2003) and consume more meals prepared away from home (Kant and Graubard, 2004). Their children are less likely to eat fruits and vegetables (Hawkins et al., 2009) and more likely to skip meals and have unfixed snacking time (Watanabe et al., 2011). In addition to this, employed mothers have less time to supervise their children, which creates opportunities for poorer diets (Datar et al., 2014). Hence, time pressures for employed mothers may have a greater detrimental effect on children's food intake patterns as reflected in fewer family meals, less consumption of fruits and vegetables, less healthy eating habits, and greater irregularity or 'skipping' of meals (breakfast and dinner). Juice, water, soda, soft/sports drinks were found to have no association with maternal employment (Datar et al., 2014;Nadia, 2012;Taylor et al., 2012). In this review an indeterminate association with maternal employment was found in children's consumption of snack food, including fast food and junk food (Adbi et al., 2017;Brown et al., 2010;Datar et al., 2014;Gaina et al., 2009;Meyer, 2016;Sweeting and West, 2005;Taylor et al., 2012), and dietary quality (Ben-Shalom, 2010;Liet al., 2012;Taylor et al., 2012;Touliatos et al., 1984). Overall, further investigations are needed to determine more definite conclusions for those domains (e.g., milk and milk products, eating out at restaurant) that had indeterminate outcome because of the small number of studies available. MVPA was positively associated with maternal employment (Anderson, 2012;Aniza and Fairuz, 2009;Ben-Shalom, 2010;Chia, 2008;Cho, 2017;Datar et al., 2014;Gwozdz et al., 2013;Koca et al., 2017;Meyer, 2016;Morrissey et al., 2011) whereas sports participation showed a negative association (Lopoo, 2007;Wijtzes et al., 2014) in the current review. Overall, it is indicated that children of working mothers were sufficiently physically active. This finding may seem somewhat contradictory. Working mothers bear a double or triple burden of responsibilities at home and at work (Bond and Sales, 2001), and thus may lack sufficient remaining time and energy to more fully supervise and actively engage with their children (Cawley and Liu, 2012;Fertig et al., 2009). However, monetary support from employed mothers may lead a better quality of life. There is evidence that higher socioeconomic status (SES) of families provides more opportunities for their children to do more activities, some of which are physical activities (Park and Kim, 2008;Stalsberg and Pedersen, 2010), and they could financially support the enrollment of their children into organized physical activity such as active sports clubs (Datar et al., 2014;Kantomaa et al., 2007).
In terms of TV watching in children as a sedentary behaviour, most studies reported positive associations with maternal employment (Chia, 2008;Chowhan and Stewart, 2014;Datar et al., 2014;Fitzsimons and Pongiglione, 2019; Nadia, 2012; Richards and Duckett, 1994; Sweeting  and West, 2005;Ziol-Guest et al., 2013). TV viewing has become a preferred leisure time activity of children during unsupervised time (Datar et al., 2014) and the reason may be the unavailability of outdoor facilities or due to safety reasons. Existing literature support positive associations between maternal employment and TV viewing of children (Fertig et al., 2009). Screen time (TV, DVD, video, or movie, playing video/computer games) indicated a positive relationship with maternal employment (Cho, 2017;Ham et al., 2013;Martin et al., 2018;Meyer, 2016). Children and adolescent's increasing exposure to screen-based activities are evident in other reviews (Thomas et al., 2019). Results show that children of employed mothers were physically active, but at the same time, children spent more time on sedentary pursuits. While this may appear contradictory, it has been shown that physical activity can be independent of how much time children spend in sedentary behaviours over the day (Marshall et al., 2004;Pearson et al., 2014;Sallis et al., 2000). For example, a child can indulge in high levels of MVPA but also in sedentary screen time. Within a 24-h day, time can be displaced to lighter forms of physical activity or sleep.

Strength and limitations of the study
Applying a comprehensive and systematic approach, this review included a detailed summary and critical narrative synthesis of 42 published papers. Additional strengths of the review are the inclusion of all study designs as well as all type of measures of dietary patterns, physical activity and sedentary behaviour. Furthermore, this review included multiple domains of dietary patterns, physical activity and sedentary behaviour.
This systematic review has some limitations. Although we tried to identify as many studies as possible, we may have inadvertently missed some eligible studies due to limited search strategy and beyond published English language studies. Furthermore, a majority of the studies were cross-sectional, thus conclusions regarding causality of association are not possible. Although device-based measures of physical activity and sedentary behaviour are more reliable, only a small number of studies used them. Most studies used self-report or maternal-report data, hence contributing to the possibility of reporting bias. Risk of bias among included studies were relatively high since many studies did not report on the reliability and validity of measures used to assess maternal employment and dietary patterns, physical activity and sedentary behaviour. The diverse nature of measures and outcomes prevented the use of meta-analysis. Conceptually similar domains were combined which may also narrow down the depth of analysis and generalizability of these findings. Though, multivariate tests are more accurate, use of univariate test for assessing statistical significance is a limitation of this study. Finally, most of the studies are from HIC, hence findings may not be similar in the context of LMIC.

Conclusion
To our knowledge, this is the first systematic review that summarises the evidence for links between dietary patterns, physical activity, and sedentary behaviour of children with the employment status of their mothers. Findings suggest that maternal employment was associated with poor dietary patterns but more physical activity and more time on sedentary activities. The latter were particularly for TV viewing and other screen-based activities of children. These findings provide an indication of how maternal employment may increase the risk of childhood obesity. We also identified a lack of validated measurers of dietary patterns and few studies using device-based assessment of physical activity and sedentary behaviour. The findings of this systematic review have important implications in the context of growing participation of women in the labour force. Considering that the employment and economic activity of women will continue to increase in the future, interventions should support employed mothers with an aim to promote healthier children's dietary patterns and decrease sedentary time. Little can be deduced from the inadequately studied domains (for example-milk and milk products, eating out at restaurant, mothers playing with their children, sitting activities (reading-writing, playing musical instruments)), hence future studies need to focus on these domains of dietary patterns, physical activity and sedentary behaviour. Domains that have inconsistent results also require further testing. In addition to this, future research needs to pay attention to UMIC and LMIC because research related to maternal employment and children's lifestyle variables are scarce in those countries. Moreover, the use of device-based measures of physical activity and sedentary behaviour are needed in order to produce accurate estimates of total time spent in physical activity and sedentary behaviour.

Financial support
This research did not receive any financial support from any organization.

Availability of data and materials
Data and other material will be provided as supplementary document.