A systematic review of media parenting in the context of childhood obesity research

The American Academy of Pediatrics (AAP) recommends children spend no more than 1 to 2 h per day using screens [1], yet, it has been documented that children spend more time using screen media than attending school [2]. Screen media includes televisions (TV), video gaming systems, computers, cell- or smartphones, and other electronic devices such as tablets or laptops [3, 4]. According to a recent Kaiser Family Foundation report, children ages 8 to 10 years-old spend nearly 8 h using media each day, while teenagers spend more than 11 h [5]. While there is an evidence base on the harmful effects of TV viewing on children’s risk of obesity [6–8], we know much less about the effect of modern screen based devices such as smartphones and tablets [9]. Rapidly evolving forms of media demand creativity and adaptability on the part of researchers looking to capture device use among children and make recommendations to parents on how to intervene on their child’s use.

Parents shape child behaviors through modeling, guiding their children through reinforcement, and controlling their environments [10–14]. They play a significant role in monitoring child media consumption and helping them find alternative activities. Although studies have previously explored parenting practices associated with child exposure to specific types of media content, there has been limited research focused on the implications of media parenting practices for obesity prevention [3]. Jago, et al. recently published a systematic review [13] of 29 studies examining associations between media parenting and child screen viewing, with a particular focus on parenting measures. The authors reported inconsistent relationships between parenting and child screen use, and highlighted important shortcomings in parenting measures. This review did not focus on childhood obesity, nor did it provide a systematic and quantitative assessment of how child media use was operationalized.

As a result of the dramatic proliferation in child screen use, there is an increasing need for interventions to support parent modulation of child screen use. It is questionable, however, if the literature is at a stage of development to support the creation of such programs. Thus, at this time it is important to take stock of the available evidence. In particular, it is important to document the specific media devices and media parenting practices measured in studies, and the extent to which studies have examined the ecological context. This information is crucial to understand gaps in the literature that must be addressed. With this in mind, we conducted a systematic review to identify studies on childhood obesity and parenting, and draw out those studies with a particular focus on media parenting. By doing this, our analysis addresses two major aims: 1) to describe how media use and media-related parenting practices or skills are operationalized in studies; and 2) to explore whether studies included measures of ecological factors, which could be associated with media parenting.

For this study, a multi-stage process was used to: 1) identify articles on parenting in the context of childhood obesity; 2) screen for eligibility; and 3) compile a final sample of studies for full-text coding. Five databases (PubMed, EMBASE, Academic Search Premier, PsycINFO, and CINAHL) were searched to identify studies for inclusion. The final database search was conducted in December 2015 (see Fig. 1). We used search terms related to parents or parenting, and obesity or obesity-related risk behaviors (see Additional file 1 for an example of search terms used); included search terms related to media use, devices, and media parenting; and limited our search to parents/caregivers of children 0–18 years of age. We also used terms to exclude unrelated topics unrelated, such as family history or pregnancy, as well as non-research studies (e.g., letters, reviews, commentaries). A total of 5557 unique studies were identified after removing 4499 duplicates. An initial screening of study titles eliminated those meeting the following exclusion criteria: focused on unrelated topics (e.g. food safety), were not written in English, included animals, targeted a clinical population (e.g. children with spina bifida), or were not original research (e.g. commentaries, letters), resulting in a pool of 2080 studies. Two coders then applied eligibility criteria to study abstracts. Eligible studies included non-intervention, peer-reviewed research studies written in English that focused on parenting (e.g. measured parenting behaviors, styles, or cognitions), were relevant to childhood obesity (e.g. diet, sedentary behavior), included parents/caregivers as research participants, and were published between January 2009 and December 2015. This time parameter was chosen to capture the most recent literature available and for the feasibility of coding, given the high volume of studies identified in this seven-year time period alone. Intervention studies were excluded from the sample because they require a separate coding scheme, which was beyond the scope of the present study. Consistent with systematic review guidelines [15] only one article per study was included in the analysis. When multiple articles from the same study were identified, we included the first published article. When multiple articles from the same study were first published in the same month, we randomly selected one article to include in the sample.

The two coders achieved 90 % agreement in applying the eligibility criteria, and all discrepancies were discussed and resolved. A total of 1170 studies did not meet the eligibility criteria or could not be located, resulting in a sample of 667 studies representative of recent literature on parenting and childhood obesity. A separate analysis was conducted on the total pool of eligible studies and is reported elsewhere [16]. Studies were eligible for the present analysis if they were coded as measuring media parenting practices (e.g., rules around TV/computer/video game or other media use) and/or the home media environment (e.g., parents’ own media use, the presence of a media device the bedroom), resulting in a final sample of 103 studies.

Forty three studies (42 %) measured a single ecological factor; 17 % measured multiple factors. Family-level factors include parent cognitions (i.e. knowledge, beliefs, attitudes, intentions) and intrafamilial dynamics and were measured in about a third of studies (29 and 35 %); however, contextual factors at other levels were not often measured. Since child media use, exposure, and parenting is centered at home, we wanted to explore measurement of additional elements within the intrafamilial domain which could affect parenting practices and skills [25], such as parents’ ability to set rules around screen time. Four percent or less of studies measured any of these factors, including parental self-efficacy (4 %), mental health (4 %), social support (2 %), parent stress (2 %), social norms (1 %), parental control (1 %), food insecurity (1 %), acculturation (1 %), and child care challenges (1 %). No studies measures housing instability, family conflict, or parent self-esteem.

Given the ever-increasing ubiquity of media and screens in children’s lives [5, 17] and the documented relationship between screen use and childhood obesity [26–28], the paucity of studies measuring media use and media parenting within the literature of parenting and childhood obesity is concerning. Results from this analysis can provide explicit guidance on future work needed in this area.

The vast majority of studies on media parenting originated from the United States, Europe/United Kingdom, and Australia. Although child obesity rates are increasing in low/middle income regions [29], there is a lack of representation of those areas in our sample, with only 7 % of studies originating from Asia and Mexico/Central America, and no studies from Africa or South America. While it is possible that there are still disparate levels of screen use in low/middle versus high-income countries, which would temper the association between screen behaviors and childhood obesity, we cannot know this without measurement.

Our analysis examined parent populations targeted for participation, and found more than half of studies included both female and male parent participants, which presents an opportunity to engage both mothers and fathers around media parenting as part of efforts to reduce child obesity. However, greater work is clearly needed in this area, as we found that sample size distributions between male and female participants were skewed to include many more mothers/female caregivers than fathers/male caregivers. These disparities are addressed in a forthcoming manuscript. Our study also revealed other shortcomings in specific parent group inclusion. Although parenting around obesogenic behaviors remains important in adolescence [28, 30], parents of teenagers were also largely absent from the studies we reviewed.

While 35 % studies included minority parents, the lack of inclusion of Asian and Indigenous parents, particularly in U.S.-based studies, is concerning given that Asians are the fastest growing minority group in the U.S., followed by Native Hawaiian/Pacific Islanders [31]. It has been suggested that parent attitudes and beliefs about media use vary by culture [32], and that youth media use varies by race/ethnicity [33–36], so it is important to include parents across multiple racial/ethnic groups to document those nuances. It is concerning that a limited number of studies recruited low income/SES parents, given the pervasiveness of childhood obesity and the fact that barriers experienced by parents in these groups have been documented [37, 38]. Specific recommendations for recruiting under-resourced groups have been previously noted [39, 40]; we would add that our analysis highlights the continued need for protocols that specifically target low-income parents in order to document persisting disparities, identify the circumstances within which people are parenting, and to strengthen strategies to target underserved populations.

Our analysis shows that measures of child media use in studies focused on childhood obesity are largely centered on television viewing, which does not reflect the varied modes of exposure children experience. This could potentially be explained by low popularity of “newer’ devices (e.g. tablets) among children, particularly in the earlier publication years included in our analysis. However, reports about the changing landscape of child device use [41], and rapid increases in such use [42] indicate that child use of these devices was already underway in the earlier years of our search parameters. Additionally, while several studies measured the presence of such devices in children’s bedrooms (i.e. laptops, smartphones, and tablets), only one study measured use of laptops, and no studies measured parent or child use of tablets. We encourage researchers to focus on capturing the full landscape of media devices used by children in future research. Many studies measuring media use utilize questions that collapse device categorization (e.g. defining computers as “desktops/laptops/tablets”), creating unspecified results which obscures device-specific media use by children. As some devices lend themselves to potentially increasing child exposure (e.g. portable devices such as handheld gaming devices versus stationary devices such as TVs), accurately documenting child media exposure by device-type is crucial to understanding their exposure across time and settings, as well as how parents are or are not setting bounds to that exposure.

Relatedly, our analysis shows that although data on media parenting is being collected, measures generally focus on the existence of rules regarding time limits and do not account for settings outside of the home, where children could be using media, such as in the car on the way to school or during afterschool programs. Additionally, greater consideration should be given to measuring the ways media is used as a mode of parenting, such as the use of media for managing child behavior, incidence of media viewing as a family activity, or screen use during meals. More work is needed in order to better understand the ways in which children are using media and how parents are allowing or responding to that use.

Another surprising omission from this literature is measurement of parent media literacy, or other relevant skills such as digital literacy. Measures of media literacy in the literature typically focus on assessments among children or parent acquisition of media literacy skills [43, 44]. Consistent definitions and validated measures of this are currently lacking, but are needed in order to understand and quantify existing parent knowledge about media. Parents cannot provide proper supervision of media they are not familiar with.

Observational studies on parenting and childhood obesity, particularly as precursors to intervention efforts, are meant to have a thorough understanding of parent experience, which should include accounting for the contexts within which parenting occurs. Family ecology and social and emotional contexts shape and drive families’ day-to-day experiences and behaviors. Our analysis shows that more attention could be paid to neighborhood/community factors, including social capital, neighborhood safety, and deprivation. Child-specific factors, such as temperament and self-regulation were also absent from the literature. Just as with family-specific factors, these domains may also impact parenting practices around media use, and their measurement can inform behavior change strategies or associations with risk of child overweight.

This analysis is not without limitations. We restricted our review to English-language and non-intervention studies, which may have limited our final sample of studies available for analysis. We refined our initial search terms to collect studies that included parents as research participants and those with a focus on parenting related to childhood obesity (i.e. parenting practices related to diet, physical activity, and sedentary behaviors). In so doing, we will not have captured studies that explore parenting solely related to media content, and may have missed, for example, studies that measure parent regulation of television food advertisements and commercials. By placing these limits on our search terms, we avoided having many irrelevant studies in our initial sample of studies, and we are confident that our sample accurately reflects the literature on parenting of media specifically as it relates to childhood obesity. An area of future research could be an analysis of measurements of the content or platforms of media used by children, such as active gaming or social media. We did not assess risk of bias across studies or strength of evidence for study outcomes because these were not relevant to our research aims. There are also many strengths of this analysis. The studies reviewed were drawn from a broad range of publication disciplines, included both U.S.-based and non-U.S. studies, broadly examined media parenting in the context of childhood obesity, and examined the presence of environmental factors that may impact screen viewing and parenting practices thereof.

More detailed measurements of media use that reflects current technology trends and diverse contexts of use are needed to better understand media use and parent regulation of child media exposure. Future observational studies should consider recruitment protocols which target more diverse samples of parents, and consider including measures of the ecological context within which families live to more fully understand factors that may impact media parenting and, subsequently, child risk for overweight.