Screen exposure time of children under 6 years old: a French cross-sectional survey in general practices in the Auvergne-Rhône-Alpes region

We conducted a descriptive cross-sectional survey of parents of children under 6 years of age in the Auvergne-Rhône-Alpes region in January 2019. This region, the second largest in France in terms of population size, is quite representative of the whole France considering that it contains urban areas with high population density and rural areas.

General practices were drawn at random from the list of GPs of the Auvergne-Rhône-Alpes Regional Health Agency, and parents of children who received medical care from the selected GPs were requested to participate to the study. Random selection was made using Stata software. The objectives and modalities of the study were presented to the GPs by telephone call before obtaining their oral agreement to participate. No response after three telephone calls was considered a refusal.

Thirty questionnaires per GP were distributed. They were offered to parents by the medical secretaries or left directly in the waiting room. A collection box was made available for the participants to deposit the completed questionnaires. An information note was given to each parent, explaining the objectives of the study, the conditions of anonymization and the confidentiality of the data collected. The GPs participating in the study were contacted twice during the 4 weeks of inclusion.

All parents with at least one child under 6 years of age who agreed to participate in the survey were eligible for inclusion. Parents with several children under the age of 6 were asked to complete the questionnaire for the youngest child. Exclusion criteria were parents of children older than 6 years or those with difficulties with the French language.

Ethical approval for the survey was obtained on the 5th of June 2019 from the Ethics Committee of the ‘Collège National des Généralistes Enseignants’ (CNGE) under number 16051998. The study was performed according to good research practices and the Declaration of Helsinki. Participants received an explanatory letter and an anonymous questionnaire by post. Written Informed consent was obtained from the participants.

We designed a questionnaire with 32 questions divided into 3 parts. It was based on the latest position statements of the French Paediatric Society (SFP) [24]. To evaluate its comprehension and acceptability, the questionnaire was pretested with 30 parents with different sociodemographic characteristics. It was composed of the following 3 parts:

The primary endpoint was the child’s ADST regardless of the type of device (TVs, computers, tablets, smartphones, video game consoles), reported by the parents in minutes and distinguishing weekdays from weekends.

The objective was to include 500 questionnaires. This sample size was considered to be large enough to have representative data and accurate estimations. We assumed that for a 4-week inclusion period, approximately 20 to 30 questionnaires would be collected per GP. Thus, we needed the participation of at least 25 GPs to reach 500 questionnaires. Of the 6201 general practitioners in the Auvergne-Rhône-Alpes region, 150 were selected after randomisation. In case of refusal of a general practitioner drawn at random, this one would be replaced by a doctor of the same gender (the next one in the list) in order to keep a gender balance.

The study sample is described by numbers and percentages for categorical data and by means ± standard deviations and medians and interquartile ranges for continuous data. Daily screen time was graphically analysed, and normality was assessed using Shapiro–Wilk’s test.

Daily screen time (weekdays and weekend days) was analysed using the Mann–Whitney test for two group comparisons and the Kruskal–Wallis test for 3 (or more) groups. Relationships between daily screen time and continuous data were analysed using Spearman’s correlation coefficient.

Analyses were performed using Stata (version 15, StataCorp, College Station).

A total of 26 GPs agreed to participate in the study. Their characteristics are described in Table 1. Among the 780 questionnaires distributed, 531 were collected, and 486 were analysed. Thirty-six patients who did not meet the inclusion criteria and 9 who partially completed the study were excluded (Fig. 1). A total of 81.2% (n = 394) of the questionnaires were completed by mothers; the mean age of the respondents was 34 years (± 5.1). Among the children, 46.1% (n = 224) were girls; the mean age was 3.7 years (±1.5) (Table 2). A total of 449 households (93.4%) reported owning at least one television set (Fig. 2). In addition, 431 families (93.7%) reported owning at least one smartphone, 393 reported owning at least one computer (84.2%) and 208 reported owning a tablet (44.4%).

The ADST for children under 2 years of age on weekdays was 26 minutes (±44.0), with a median IQR of 0 [0–30] (see Table 2 and Fig. 3) and that on weekends was 30 minutes (±46.9), with a median IQR of 0 [0–50]. The ADST of children older than 2 years on weekdays was 66 minutes (±82.1), with a median IQR of 50 [20–90] and that on weekends was 103 minutes (±91.3), with a median IQR of 90 [45–120]. Both daily weekday and weekend average screen times reported by the parents were significantly higher for boys than for girls (weekday with a median IQR of 48 [20–90], weekend days with a median IQR of 30 [5–60], P = 0.02, respectively) (Fig. 3). The ADST increased significantly with the age of the child (p < 0.001).

According to the parents, 316 children (66%) had watched a TV and/or multimedia screen the day before the questionnaire. The previous day’s average screen time was 43 minutes (±57) across all age groups. Children’s ADSTs increased concomitantly with parents’ ADSTs, both on weekdays (correlation coefficient r = 0.31, p < 0.001) and weekends (r = 0.37, p < 0.001).

According to the parents, 456 children (94.4%) did not have a screen in their room, and 443 children (91.5%) never used a multimedia screen in their room. Children who had a TV in their room had higher weekday and weekend ADSTs than those who did not (124 ± 91.8, median 120 [60–150] vs. 56.0 ± 76.6, median 30 [10–60], P < 0.001 on weekdays and 187 ± 131, median 150 [120–240] vs. 86.0 ± 83.7, median 60 [30–120], P < 0.001 on weekends, respectively) (see Table 2 and Fig. 3). One hundred and two parents (25.4%) said they never discussed the content of the program their child watched, and for 72 (14.8%) of them, the child was regularly or always alone in front of a screen.

Concerning TV, 119 parents (24.7%) estimated that their child watched TV during the week before school, and 62 parents (12.7%) watched TV in the evening before bedtime. Furthermore, 53 children (10.9%) were regularly present when their parents watched TV, and 45 children (9.2%) were always present. In 63 households (12.9%), the television was always on during meals. Finally, 78.8% of the parents in our study wanted to discuss screen recommendations with their general practitioner.

The multivariate Poisson regression model adjusted for clinically relevant criteria showed that, compared to Secondary School parents, the higher the parents’ level of education, the less the children watched a screen (IRR = 0.79, 95%CI [0.66–0.94] (Tables 3 and 4 and Fig. 4). Compared to one TV at home (reference), on weekdays, children with no TV were less exposed (IRR = 0.67, 95%CI [0.45–1.01], p = 0.053), than children with three or more television screen (IRR = 1.38, 95%CI [0.99–1.92], p = 0.059). As well on weekdays as on weekends, when parent’s daily screen time was > 2 hours, children exposure increased, with IRR = 1.34, 95%CI [1.13–1.59], p = 0.001. Compared to parents who were not aware of any harmful effects of screen overexposure considered as reference, the children of parents who were aware of at least 3 adverse effects had lower ADSTs during the weekdays and weekends. Finally, children from single-parent families tend to be more exposed than others on weekdays with IRR = 1,23, 95% CI [0.9–1.65], p = 0.162 and weekends with IRR = 1.17, 95% CI [0.9–1.47], p = 0.165 and weekends.

Our study which focuses on parents of children followed in general practice, highlights that children of parents with Bachelor’s degree or above education level were less exposed to screens. On weekdays, those without TV at home had less screen time. On the other hand, the children of parents who were well informed about the possible adverse health consequences of overuse of screens had lower average screen time. On weekend days, the ADST of girls was less important than boys. The ADST tended to increase among children from single-parent families on weekdays and weekend days. Finally, there was no difference by rural or urban location.

We did not find similar studies among a population of patients followed in general practice. Several studies in primary care involved patients followed in paediatric clinics or offices [2, 27]. In our study, both weekday and weekend ADSTs were lower than those reported in American data from 2020 [1]. Indeed, before the COVID-19 pandemic and the implementation of the first containment, American children under 2 years of age spent an average of 49 minutes per day in front of screens. Those aged 2 to 4 years had a daily screen time of 150 minutes per day. It should be explain by the fact that our data predate the American data and that all studies tend to show a steady increase in screen time among children [12, 14, 28, 29]. Our results for children under 2 years of age are also lower than those of an Australian study published in 2016 that revealed a daily screen time of more than 2 hours for 40% of 18-month-olds [30]. This difference might be caused by our age range including children younger than 18 months with less screen time.

Our study took place just before the Covid 19 pandemic. It can be assumed that the Covid pandemic changed children’s screen use behaviours, at least for older children. During the first wave of the pandemic, some studies assessed screen time among preschoolers. In an international study, parents reported an average increase of nearly 1 h of screen time per day in 3- to 7-year-old children [31]. This increase was largely due to their use for entertainment purposes. Another study, by Fitzpatrick et al., found an increase in screen time specially before bedtime [32]. In this study, children’s age and parental use of multimedia screens were factors associated with increased screen time, but teleworking parents were less likely to have overexposed children. Not surprisingly, older children with online schooling requirements spent more time in front of a screen at first containment. In an other international survey in children under the age of three, this increase in screen time was also confirmed during the first lockdown [33].

Our study also found that some environmental factors may influence children’s screen time.

In similar studies, the number of TV screens in the home is associated with children’s screen time [34, 35]. In the same way, living in a single-parent family is a contributing factor to increased screen time. However, although there is an upward trend in our study, the association is not significant. One of the explanations for this is that screens can be a means of distracting or calming children when the parent is engaged in certain tasks, which would seem to be even more valid in a single-parent family [28, 36].

In studies, children’s screen time is commonly associated with parental screen time [29, 35]. Lauricella et al. found that the amount of time parents used a multimedia technology (computers, tablets, smartphones) was associated with the amount of time the child used that same technology [37]. In other words, parental behaviours and habits with respect to multimedia screens strongly influence those of the children. Birkin et al. and Matta et al. have established that the regulation of their own use of new technologies by parents allows to avoid the reproduction of harmful behaviours by their children. This can be done by setting up parental rules [35, 38].

Parental education level is associated with children’s screen time [14, 38]. Atkin et al. found that children of mothers with low levels of education were more likely to exceed 2 h/day of screen time [14] and for Kiliç et al. the frequency of tablet use and ownership among children was inversely related to maternal education and household income [28]. But for Paudel et al. review, the association between educational status and children’s mobile screen media use is not really demonstrated [39].

Some studies showed that children’s mobile screen use increases when parents perceive beneficial effects and educational value [37]. Similarly, negative parental beliefs about screen-based mobile media are associated with decreased screen use [40]. In addition, a French study conducted in 2019 showed that parents were mainly misinformed about the risk of obesity [25]. Finally, our data demonstrate that the more knowledge parents had about the harmful effects associated with overuse of screens, the more children’s screen time decreased. This finding is particularly interesting because there are relatively few data in the literature on this aspect.

The number of parents included reached 500, giving enough statistical power to the study. The final sample of GPs was representative of the region’s GPs for sex, age, and practice location. However, the results must be interpreted considering a number of limitations.

The survey was systematically proposed to parents of children under 6 years of age, but this did not exclude a possible selection bias by the secretaries or GPs offering the questionnaires. Furthermore, parents who were involved and aware of this subject could easily agree to answer the questionnaire. Another selection bias was related to the nonparticipation of populations with a language barrier. The effect of non-participating population is difficult to interpret regarding the data available in the literature. This population may include people with low education level, and/or migrants. We could make the assumption that this exclusion underestimates the screen time of our population. Indeed, in her study about children under 3 years of age, Duch showed a positive correlation between screen time and ethnic minority status [41]. This, together with the low rate of participating GPs, limits the extrapolation of the results to the entire French paediatric population.

The questionnaires were anonymously completed by the patients before or after the consultation and placed in a collection box. This method allowed the parents to be as honest as possible in their answers, but as this was a declarative survey, it does not exclude a possible social desirability bias, which could lead to a minimisation of the ADST.

For our primary endpoint, we used the parents’ global memory to determine children’s ADSTs on weekdays and weekends. A memory bias is possible because of the difficulty of providing synthetic and global data for a usual practice. This bias can be balanced by the fact that the parents’ assessments of the previous day’s screen time across all age groups was lower than their overall estimates of daily weekday and weekend average screen times. In addition, our study did not assess daily exposure times for every type of screen.

An Australian study found that health-related habits in families crystallise most easily in early childhood [42]. Thus, educational measures regarding sensible screen use should be implemented in early childhood to promote appropriate use. In our study, 45% of the parents had established rules for the use of screens, and the children of parents who were aware of several harmful effects had less screen time. If we add that nearly 80% of the parents would like more information on the subject, these data should encourage primary care practitioners, especially GPs, to screen children at risk of overexposure and to sensitize their families to implement preventive measures.