An ecological momentary assessment study of physical activity behaviors among mothers of toddlers from low-income households

Physical activity is a central component of a healthy lifestyle for reducing poor health outcomes [1]. Only about half of US adults report that they meet the nationally recommended levels of physical activity (50.2% in 2017) [2, 3]. Disparities in physical activity level by socio-demographic factors exist, such that lower physical activity is associated with low-income status, Black race or Hispanic ethnicity, and female gender [4‐11]. These associations suggest that women from low-income communities may experience barriers to meeting the recommended physical activity levels [12].

Based on the Ecological Model of Active Living, an individual’s physical activity behavior is dependent on their social context and physical environment, including caregiving activities, physical location (outside vs. inside), work patterns, electronic usage, and social support [13‐18]. Frequent television viewing and lack of time have been associated with low physical activity in children, whereas social support has been associated with increased physical activity in mothers of young children [17, 19]. These social and community-level factors, in combination, may be related to physical activity behaviors among mothers of young children living in low-income communities [15].

Ecological Momentary Assessment (EMA) is a recurrent, real-time assessment of an individual’s behavior that minimizes recall bias by addressing the context in which these behaviors are occurring [20]. A recent EMA study found higher toddler physical activity when outside (versus inside; with stronger associations observed for older children versus younger), other children were nearby (versus no other children present), and engaging with the mother (versus no engagement) [21]. This manuscript extends the toddler study by examining the association between the home environment and physical activity among mothers of toddlers using the same constructs. To expand the knowledge-base on mother and toddler physical activity, we tested the hypotheses that mothers are more active when: (1) outside versus inside, (2) TV off versus on, (3) other adults present versus absent, 4) children present versus absent, and (5) engaging with their toddler versus not engaging. Further exploratory analysis examined whether recruitment site (urban versus semi-urban) and Body Mass Index (BMI) moderated relations between environment factors and physical activity.

Chi-square tests for categorical variables and t-tests for continuous variables were used to assess differences in baseline characteristics and response rate between individuals included and excluded from the analysis. Response rate by prompt time (e.g. morning versus evening, weekday versus weekend) was also assessed.

Linear mixed-effects regression models with random intercepts were used to assess the real-time relation between activity counts collected using accelerometry and environmental factors based on EMA responses, accounting for the clustering of responses within each mother. Separate models assessed activity counts in relation to environmental factors (TV on/off, presence of adults, presence of children, engagement with child, physical location) based on EMA responses, respectively. Recruitment site, number of children, time of day, age, and maternal BMI were included in the models as covariates [27]. Race/ethnicity was highly correlated with recruitment site and therefore was excluded as a covariate to avoid collinearity (non-Hispanic White/other compared to urban/semi-urban r = 0.63; p < 0.001).

We explored the moderating effects of recruitment site and maternal BMI on the relations between each home environment factor and activity counts, by including the interaction terms in the models. The moderating effect of engaging (i.e. talking, singing, playing) with the child was also explored using similar models with interaction terms for between-group and within-group relations, separately. Stratified analyses by recruitment site and by maternal weight followed if significant interactions existed (Table 4). SPSS (IBM SPSS Statistics 22) and SAS 9.4 statistical software (SAS Institute Inc, Cary, NC, USA) were used for the analyses.

Using EMA in conjunction with accelerometry in real time, five factors within the Ecological Model of Active Living were tested, yielding within-group (for individual participants) and between-group (average of all participants) findings. Study findings align with the guiding model by identifying home environment factors associated with maternal physical activity such as screen time, the presence of others (i.e., adults and children), physical location (i.e., outside vs. inside) as well as moderating factors such as maternal weight status and recruitment site (i.e., urbanicity).

The within-group finding that individual participants were less active when the television was on (versus off) extends similar findings that television usage is associated with low physical activity, as reported on recall surveys [15, 28]. In contrast, a study using the same dataset found no relation for toddler physical activity between television on versus off [21]. We did not collect data on the type of television programming and whether it was geared towards mothers or toddlers. Future research should explore the type of programming and the intended audience (mother vs. toddler) and how this may affect their physical activity levels.

Findings related to physical activity in the context of other adults varied. Across participants, mothers who reported being alone more frequently had more physical activity counts over 30 min compared to mothers who reported being with other adults. However, among individual mothers, rates of physical activity were not associated with the presence of other adults, possibly due to competing contexts of why other adults were present. For example, mothers may be with other adults at a sedentary job or with other adults who are providing social support to be physically active, as shown in other studies [18]. Future studies may prioritize understanding the context of the social setting to parse out associations with physical activity.

Contrary to expectations, individual mothers had lower activity counts when in the presence of one or more children compared to when they were alone. Previous studies have shown that maternal and child physical activity are associated, including a similar study on toddlers in this same population [21]. Toddlers often engage in “secure base” behaviors, in which they play in the vicinity of their caregivers, running back and forth to check-in [29], which does not require much activity for mothers to supervise their child [30]. Although when individual mothers were engaging with their child, they were more active compared to when they were together, but not engaging with their child. In the toddler study, the toddlers were also more active when engaging with the mother, supporting the connection between mother and toddler physical activity [21]. These findings suggest that mothers are most active without their toddler nearby. Future studies could examine whether mother physical activity drives toddler physical activity or vice versa, since mothers in low-income communities are exposed to barriers associated with low physical activity [12].

The finding that individual mothers and mothers on average (marginally) were more active when outside (versus inside) is consistent with the toddler study, where within individual toddlers, physical activity counts were higher when outside compared to inside [21]. Outside location could be associated with exercise or active transportation if there are yards, sidewalks, or public transit near the home that provide opportunities to be physically active [15]. In this study, the availability of green space and perhaps the accessibility and infrastructure of surrounding areas, including parks and walking paths, may encourage physical activity and active transport when outdoors [15, 31].

Recruitment site significantly moderated the within-group relation between physical location and physical activity among participants. Mothers recruited from the urban site had higher activity counts when outside compared to inside, which was greater than the activity counts of mothers recruited from the semi-urban site. Prior studies have found that active transport is associated with physical activity, which may be more prevalent in the urban sample compared to the semi-urban sample [15]. Social environmental factors that may restrict physical activity, such as crime, poverty, and maintenance of infrastructure, may impact both the urban and semi-urban mothers [32]. Decreased access to exercise facilities may also support these findings because mothers may be getting their activity mainly outdoors from transport, instead of exercising in a facility or indoors, as they are a relatively sedentary group [33]. Previous literature suggests that individuals of low socioeconomic status often get their physical activity from physically demanding occupations, which may apply in this sample [34].

This study found a difference in physical activity by weight status among mothers when outside versus inside such that mothers with overweight/obesity were less active when outside compared to mothers who had normal weight status. In general, adult healthy weight women are more active compared to women with overweight/obesity [27, 35]. These findings suggest that targeting active transport and more time outside of the home among mothers with overweight/obesity and their toddler age children may be an area for future intervention.

There are limitations that should be considered. First, the direction of the relations between home environmental factors and physical activity cannot be determined in this repeated cross-sectional investigation due to the contemporary relations (15 min of physical activity before/after prompt). Second, the mean response rate is low for EMA responses among mothers, with a wide range of response numbers per participant. However, this rate is an estimate using the maximum possible prompts available for each person as the denominator. We are unable to know if the non-responses were due to true participant non-response or due to device-related reasons, (e.g., non-functional battery) or if the device was returned before all 53 prompts reached the participant. These factors can be taken into consideration when calculating the response rate for future studies. Despite the limitations, there was not a moderating effect of participant response rate. Furthermore, the EMA questions were designed to be brief and did not require the participant to report the details of the context, only whether they were or were not in a particular setting. For example, the question asked whether the TV was on, not whether the mother or toddler were watching. Additionally, questions asked about the presence of other adults or children and whether mothers were engaged with the child, but they did not ask for information around the specifics of the activities. Given the limited number of EMA studies conducted with low-income women to understand the context of their physical activity, it is difficult to compare the current results with other studies. Therefore, future research is needed to compare how context relates to physical activity in other populations and settings. By combining other environmental measures, such as Geographic Information Systems (GIS) with EMA and accelerometry, we could gain a greater understanding of the home and community environmental context, including outdoor space, which was not examined in the current study.

Guided by the Ecological Model of Active Living and through a novel approach using EMA paired with accelerometry, this study identified home physical and social environment factors that were associated with maternal physical activity. To increase maternal physical activity, findings call for a reduction in screen time, increasing active mother-toddler engagement in play, and promoting active transport methods. Additionally, findings suggest that future interventions target mothers with overweight/obesity and from semi-urban areas as they may be particularly vulnerable to low physical activity. Understanding factors associated with physical activity during this critical period in a woman’s life could help reduce disparities surrounding inactivity.