Children are spending an increasing amount of their time engaged in activities involving screen time, which is defined as time spent watching television (TV), DVD or video, or playing computer or video games. The Canadian Paediatric Society recommends that pre-school children watch ≤1 h of TV per day(1) and the American Academy of Pediatrics recommends that children over the age of 2 years watch ≤2 h of TV per day(Reference Strasburger2). TV viewing has been associated with important health outcomes, including delayed language development, aggressive behaviour, cigarette smoking in children(Reference Christakis, Gilkerson and Richards3–Reference Gidwani, Sobol and DeJonh5) and overweight and obesity(Reference Lumeng, Rahnama and Appugliese6–Reference Crespo, Smit and Troiano13). Some studies have shown associations between screen time and BMI in this young age group(Reference Dennison, Erb and Jenkins8, Reference Rey-Lopez, Vicente-Rodriguez and Biosca14, Reference Marshall, Biddle and Gorely15). For example, in very-low-income pre-school children, the presence of a TV in the bedroom was associated with increased BMI(Reference Dennison, Erb and Jenkins8).
There is compelling evidence that obesity prevention in the primary-care paediatrician's office should be focused on pre-school children(Reference Perrin, Finkle and Benjamin16–Reference Mikkila, Rasanen and Raitakari18). Understanding factors associated with screen time in young children may provide opportunities to develop targeted interventions to reduce screen time and prevent overweight and obesity in young children. Studies to date have not focused on modifiable parental factors such as parental rules about screen time(Reference Dubois, Farmer and Girard19, Reference Kourlaba, Kondake and Liarigkovinos20). A recent systematic review of correlates of screen viewing in young children identified variables including demographic (e.g. parents' educational level) and sociocultural (e.g maternal depression) factors, as well as family structure (e.g single parent), and concluded that research is needed to test modifiable variables (including parental correlates) of screen time(Reference Hoyos Cillero and Jago21). The objective of the present study was to identify child and parental factors associated with increased screen time in healthy 3-year-old children that may be targeted using interventions implemented in the primary-care practice setting to reduce screen time.
Methods
Children were recruited at their 3-year-old well-child visit to a large primary-care paediatric group practice in Toronto, Canada, which is part of TARGet Kids! (Toronto Applied Research Group), a paediatric primary-care research network of child health practitioners, researchers and policy makers. Exclusion criteria included non-English-speaking parents who could not complete the questionnaire, and children with significant hearing or visual impairment or developmental delay that may affect screen time use and behaviours. After consent and enrolment, a research assistant reviewed a parent self-administered questionnaire on screen time including: child and family demographics (parents’ age, educational level, employment, mother's country of birth, child's age, attending daycare), screen ownership, location, common viewing times and parental screen-viewing factors (rules about screen time, parents' screen time duration, meals with the TV switched on). Items in the questionnaire were derived from the Canadian Community Health Survey(22) and from other published research on screen time in children(Reference Lumeng, Rahnama and Appugliese6, Reference Christakis, Ebel and Rivara7, Reference Gortmaker, Must and Sobol9). Outcome measures used were the sum of each reported child weekday and weekend day screen time. Screen time was defined as the time when the child was in a room with the TV or video or DVD switched on, when the child was using the computer and when the parent was using the computer for non-work-related activities, or playing video games. Data are presented as means with sd and frequencies with 95 % CI as appropriate. Linear regression models (using maximum likelihood estimates) were used to test the univariate associations between screen time and selected potential associated factors. Selected variables were included in a multivariable model if the univariable P value was <0·1, using backward variable selection. All regression models were mathematically transformed using the Poisson distribution to account for the skewed distribution of screen time. Missing values were imputed using mean values. Estimated effect and 95 % CI were reported. All statistical analyses were performed using the SAS statistical software package version 9·1 (SAS Institute Inc., Cary, NC, USA). Ethical approval was obtained from the Research Ethics Board, Hospital for Sick Children, Toronto, Canada.
Results
A total of 157 children were enrolled in the present study, with a recruitment rate of 91 % (sixteen did not consent; eleven did not meet the eligibility criteria). In all, 50 % of children were female and 34 % attended pre-school or daycare; 62 % of mothers were born in Canada, 47 % reported having completed a university degree and 81 % were employed part time or full time. The families had a median of two TV sets, two DVD or video machines and one computer in their homes. A total of 10 % of the children had a TV in their bedroom and 44 % were able to view a TV from their main dining area; 59 % of children ate at least one meal in a room with the TV switched on; 81 % of parents reported having household rules about watching TV (Table 1). The mean reported weekday screen time was 104 min, with 69 min watching TV, 28 min watching a DVD or video and 6 min on the computer. There was no significant difference in weekday and weekend day screen time (Table 1).
Table 1 Screen-time factors by weekday and weekend day: duration, viewing-time periods, meals with the TV switched on and parental factors
TV, television; N/A, not applicable.
In univariable analyses (see Table 2), factors associated with increased child screen time included higher number of TV sets in the home, TV in the child's bedroom, TV viewing at mealtime, TV viewing in the morning and increased parental screen time. Family rules about screen time reduced screen time. For weekend days, the only additional factor that was associated with increased child screen time was a TV in view from the dining area. In multivariable analysis, controlling for maternal education and age, an increase in total minutes per day of child ‘weekday’ screen time was associated with eating lunch (96 (95 % CI 30, 192) min/d) and dinner in front of the screen (42 (95 % CI 12, 90) min/d) and with the mother being employed (36 (95 % CI 6, 72) min/d). An increase in total minutes per day of child ‘weekend’ screen time was associated with eating lunch in front of the screen (78 (95 % CI 36, 132) min/d) and an increase of 1 h of parental screen time (12 (95 % CI 6, 18) min/d). Family rules decreased child weekend screen time by 30 (95 % CI 6, 54) min/d.
Table 2 Univariable linear regression analysis: risk factors associated with screen time
TV, television.
*P < 0·05.
Discussion
Our study showed that eating meals in front of the screen, mother's employment and parental screen time are associated with increased screen time, and that the existence of a parental rule about screen time is associated with decreased screen time, in pre-school children. An association between parental screen time rules and screen time has not been identified in this age group before. Strengths of the study include a thorough investigation of child and parental factors related to screen time in a healthy population of 3-year-old children. Limitations of the study include an absence of a validated measure of screen time. We used a developmentally appropriate measure of screen time exposure (being awake in a room with the TV switched on)(Reference Lumeng, Rahnama and Appugliese6), compared with other studies that ask about time ‘watching TV’. Postulated mechanisms associated with screen time and obesity (e.g exposure to food advertising(Reference Kotz and Story23)) occur during TV exposure, not only when children are engaged in ‘watching TV’. Parental report of child screen time provides accurate estimates of TV-viewing times, compared with videotaped observation (r = 0·70), and other methods (including diaries) were not used because of burden on families. The estimates of screen time use in our study were similar to those of other national and international studies(Reference Wake, Hesketh and Waters4, Reference Lumeng, Rahnama and Appugliese6, Reference Gable, Chang and Krull24). Maternal mental health was not measured in the present study, and has been reported to affect child screen time(Reference Burdette, Whitaker and Kahn25). Daycare type or direct data from daycare were not collected. The families in the present study had similar sociodemographics compared with the population of Toronto(26).
Our study confirmed results from Dubois et al.(Reference Dubois, Farmer and Girard19) showing an association between eating meals in front of the screen and total screen time, in an even younger population of 3-year-old children. Mechanisms to explain the associations between TV viewing during mealtime and increased BMI include alterations in satiety signals and effects of food advertising(Reference Kotz and Story23, Reference Matheson, Killen and Wang27, Reference Bellissimo, Pencharz and Thomas28). We also identified associations between parental and child screen time, as shown in a population of pre-school children in Greece(Reference Kourlaba, Kondake and Liarigkovinos20). Parents who are exposed to increased screen time themselves may not be as concerned about the amount of screen time that their children are exposed to. A qualitative study in parents of young children showed that parents were concerned with the content but not the amount of screen time(Reference He, Irwin and Sangster Bouck29). Dennison et al.(Reference Dennison, Erb and Jenkins8) showed that 40 % of very-low-income pre-school children in New York State had a TV in their bedroom, and this was associated with increased BMI. A total of 10 % of children in our study reported having a TV in their bedroom, and this factor was not significantly associated with total screen time, confirming results from the recent systematic review(Reference Hoyos Cillero and Jago21). Differences in these results may be related to the different socio-economic status of the patient samples(Reference Wake, Hesketh and Waters4, Reference Lumeng, Rahnama and Appugliese6, Reference Dennison, Erb and Jenkins8).
Parental rules regarding mealtime TV viewing were found to mediate the relationship between maternal education and screen time in a large sample of pre-school children in Australia(Reference Hesketh, Ball and Crawford30). In our study, we found an association between total screen time and a general family rule related to screen time. Other parenting strategies, as they relate to screen time, have not been explored in the literature.
The present study identified several modifiable parental factors that may be tested in interventional trials in this population, including not consuming meals in front of the screen, modifying parental screen time and developing family rules regarding screen time. Longitudinal data investigating patterns of viewing over time and changes in parental factors would enhance understanding and may identify optimal timing for interventions. Developing and testing interventions that focus on parenting strategies as they relate to screen time may be important subsequent steps in this age group.
Acknowledgements
The present study was supported in part by a Paediatric Consultants Research Grant, Hospital for Sick Children, Toronto. This funding organization was not involved in any of the following: design and conduct of the study; collection, management, analysis and interpretation of data; and preparation, review or approval of the manuscript. The authors have no conflict of interest to declare. C.S.B. and P.C.P. were involved in study concept and design, in analysis and interpretation of data, as well as in obtaining funds for the study; C.S.B., P.C.P., M.M., S.J., M.P. and C.T. were involved in acquisition of data; C.S.B. drafted the manuscript; C.S.B., P.C.P., C.E.B., B.W.M. and J.M. were involved in critical revision of the manuscript for important intellectual content; C.S.B. and C.M. conducted statistical analysis; M.M. and P.C.P. provided administrative, technical and material support. The authors acknowledge contributions by research staff including Laura Burr, Stephanie Clairmont and Kate Smith.
