Sociodemographic, home environment and parental influences on total and device-specific screen viewing in children aged 2 years and below: an observational study.

OBJECTIVE: This study aimed to investigate total and device-specific screen viewing (SV) and its determinants in children aged 2 years and below.
DESIGN: Cross-sectional study conducted in February 2014.
SETTING: Well-child clinics in Singapore national polyclinics. PARTICIPANTS: Parents of children (Singapore citizens or permanent residents) aged 2 years and below were enrolled during routine clinic visits. Out of 794 eligible parent-child dyads, 725 (91.3%) provided informed consent and were included in the analysis. MAIN OUTCOME MEASURES: Device-specific information on SV and determinants was ascertained using interviewer-administered survey questionnaires. The prevalence and duration of aggregate and device-specific SV were reported. Associations with potential determinants were investigated using multiple logistic regression analysis. A p value less than 0.05 was considered statistically significant.
RESULTS: The prevalence of daily SV and SV ≥ 2 h/day constituted 53.5% and 16.3%, respectively. The majority of children aged 18-24 months (88.2%) engaged in daily SV. TVs and mobile devices were the most commonly used screen devices, followed by computers and video consoles. In multivariable analysis, younger child age, Chinese ethnicity and setting rules on time of SV were strongly and consistently associated with lower levels of any SV and SV ≥ 2 h/day. Parental knowledge of SV recommendations and less parental SV were additionally associated with lower levels of SV ≥ 2 h/day. The number of screen devices was not associated with children's SV.
CONCLUSIONS: In contrast to recommendations, SV prevalence in children aged less than 2 years is high and appears to increase steadily across age groups. TVs and mobile devices are most frequently used. Improving parental knowledge of SV recommendations, reducing parental SV and especially the implementation of strict rules on SV time could be successful strategies to reduce SV in young children. Published by the BMJ Publishing Group Limited. For permission to use (where not already granted under a licence) please go to http://www.bmj.com/company/products-services/rights-and-licensing/

This study investigates screen viewing in very young children considering television viewing, as well as other relevant up-to-date screen device types, such as computers, game consoles and mobile devices.

The study comprehensively assesses sociodemographic, parental and home environmental influences on screen viewing.

Our findings are based on a large multiethnic Asian population with a very high response rate.

The study provides evidence that is highly applicable to the development of early childhood interventions aimed at reducing screen viewing.

The main limitation is the fact that screen viewing as the main outcome was based on a proxy report and could not be determined objectively.

Background

Screen viewing (SV) has been associated with substantial detrimental health effects, including a higher risk of type 2 diabetes mellitus, cardiovascular diseases and premature mortality.1 2 In children, increases in the time spent on SV have been found to negatively affect cognitive and language development,3 4 social interaction,5 6 attention deficits,4 7 8 as well as behavioural problems.9–11 SV in children has also been linked to obesity,12 metabolic risk and impaired vision.13–15

For children above the age of 2 years, SV recommendations, for instance from the American Academy of Pediatrics, state not to engage in more than 2 h of television (TV) or screen time per day.16 However, recent evidence-based recommendations for very young children (here defined as children below the age of 2 years) have discouraged SV altogether.17 18 Despite these recommendations, the use of screen devices in young children seems to have become common practice in recent years. For instance, in Japan, 86% of children spend over 1 h of TV time a day by 18 months of age.19 In Australia, the average child between 0 and 4 years spends 2.5 h on TV20 per day. Similarly, in the USA, 90% of children regularly watch TV or videos by age 2, with the average duration exceeding 1.5 h a day.21 22

Such sedentary behaviours have been shown to persist across age, indicating that children who spent more time watching TV, watch more TV later in life and that these early life behaviours could be associated with adverse health consequences, such as increased body mass index and serum cholesterol levels, as well as decreased cardiorespiratory fitness later in life.23–26 The increasing amount of time children spent on SV and the potential long-term implications of SV highlight the public health importance of preventing the adoption of SV behaviours early in life.

With technological advancements, screen devices have become pervasive in everyday life. Nowadays, comprehensive assessments of SV patterns in children have to consider TV, computer and video consoles, as well as mobile screen devices, especially smartphones and tablets.16 Their portability and ease of use anywhere and anytime could have important implications on the way these screen devices are being used. Studies in older children and adolescents have already highlighted their widespread use.27 28 However, little is known about current SV patterns in young children, especially with regard to the use of different screen devices,29 30 despite the common perception that modern mobile screen devices are more interactive and possibly beneficial for children's health. According to a recent systematic review of the available evidence in children aged 3 years and below, the majority of studies reported TV and video viewing. Only five studies have reported computer use and no studies have explored mobile device use.31 Equally important, little is known about the time when children begin to be exposed to these devices on a regular basis, their sociodemographic profile and potentially modifiable determinants. This knowledge is important to understand when and how best to prevent excessive SV behaviour in children, as well as later in life.

This study aimed to investigate overall and device-specific SV patterns in children aged 2 years and below. It also investigated differences in SV behaviour across age, sociodemographic factors, screen home environment, as well as parental knowledge and practice-related determinants of early childhood SV.

Methodology

Participants and procedures

This was a cross-sectional study conducted in February 2014 among parents bringing their healthy children, aged 2 years and below, to two National Healthcare Group Polyclinics (Singapore has a total of 18 polyclinics) for their well-child health visits, which may include developmental assessment and/or vaccinations. The take-up rate of this programme for children up to 2 years is more than 97%, providing a good representation of the children in the region at a given point in time.32 Parents satisfying the eligibility criteria were invited to participate in the study. Participants who were not Singapore citizens were excluded and verbal consent from the participants was obtained. After consent was obtained, trained interviewers administered survey questionnaires in a standardised form in the participants’ preferred language (English, Chinese, Malay and Tamil) during clinic visits.

This study was approved by the Singapore National Health Care Group Domain Specific Review Boards (DSRBs). Participation was entirely voluntary and informed consent was received from all the parents or caregivers.

Measurements

The study questionnaire was developed following a comprehensive literature review and refined through pilot studies. Interviewers were trained to ensure consistent administration of the survey questionnaires.

SV behaviour

Parents were asked to provide information about the amount of screen time their child was exposed to during weekdays and weekends. Parental recall of their child's SV time has been shown to be reliable, correlating well within previous observational studies.33 SV behaviour was ascertained in four broad device categories: TV/digital video disc, computer, video game consoles and mobile devices (including smartphones, tablets and other portable devices). To determine total SV, the amount of time spent on individual device categories was added up for weekdays and weekends with weights corresponding to 5/7 and 2/7, respectively. Subsequently, total SV time during the entire week was calculated. Each device-specific SV time was determined in the same way but only including information for respective devices.

Two SV statuses are presented, (1) ‘any SV per day’, and (2) ‘SV ≥2 h/day’ on any device, as well as for different screen devices. The second SV status is based on previously cited SV recommendations that indicate that even older children should not engage in more than 2 h of SV per day. We therefore chose this cut-off to indicate ‘heavy’ SV exposure.

Independent variables

Sociodemographic information of the parent was recorded, including ethnicity, parents’ education, parental marital status and household income, together with the child's age and gender. The screen environment at home was determined by assessing the number of functioning screen devices at home. The number of screen devices was recorded separately for each device type. Parental knowledge of professional SV recommendations for children under the age of 2 years was categorised into three categories: minimise SV altogether, less than 2 h of SV time, and ≥2 h of SV time. Parental SV was determined and summarised in the same way as for children. Total parental SV time was categorised according to tertiles.

Parental rule setting practices focused on whether parents set rules on the time (duration per day) and the programme (content) of their child's SV, respectively. These variables were combined (parents who set rules on both time and type of programme vs parents who did not set rules on time and programme) because of the strong association between both variables. The exact daily SV time permitted according to parental rules was further assessed and categorised into four groups: ‘no SV rules’, ‘SV time ≥2 h’ rule, ‘SV time of less than 2 h’ rule and ‘no SV time’ rule.

Data analysis

For categorical variables, counts and percentages were reported while for continuous variables, the median and interquartile ranges (IQRs) were reported. To assess association between categorical variables, the χ2 or Fisher exact tests were used where appropriate. The Kruskal-Wallis test was used to assess association between continuous and categorical variables. Multiple logistic regression was used to model two representations of SV status, ‘any SV per day’ and ‘SV of ≥2 h/day’ (representing ‘heavy users’), where sociodemographic factors, parental knowledge and practices were the independent variables. In these analyses, both rule setting characteristics (time and programme) were combined as ‘parents that set rules on time and type of programme’. To further explore the association with specific types of rules and the strictness of rules on time (with regard to the permitted duration of SV per day), multiple logistic regression analysis was performed separately for ‘rules on time’, ‘rules on programme’ and the strictness of rules on time in terms of the permitted SV time. The OR, its 95% CI and p values were reported. A p value <0.05 was considered statistically significant. The statistical analysis was carried out using the IBM Statistical Package for Social Sciences and Stata V.12.

Results

Study population

All 1061 parent–child dyads visiting the assessment sites during the assessment periods were approached. Of them, 794 met the inclusion criteria and 725 (91.3%) provided consent, participated in the survey and provided complete information on the outcome variables of SV time.

Child and parental characteristics of the study population are presented in table 1. Findings show that 55.3% of children were male and the majority (58.6%) were of Chinese ethnicity with the remainder being mostly of Malay or Indian ethnicity. The median age of the children was 7 (IQR 10) months with 44.7% being younger than 6 months and 55.3% being 6–24 months old. The majority of parents were married (98.1%) where 79.5% of fathers and 76.0% of mothers had attained an educational level of postsecondary school and above. The majority of households (81.1%) had a monthly income of SG$3000 (where SG$3000 is approximately US$2011) or more.

Table 1

Participant, child and household characteristics (n=725)

	N per group	Per cent of total
Sociodemographic characteristics
Male	401	55.3
Female	324	44.7
<6 months	324	44.7
6–24 months	401	55.3
Age in months (median, IQR)	725	7.0 (10.0)
Chinese	425	58.6
Malay	196	27.0
Indian	82	11.3
Other	22	3.0
Parents not married	14	1.9
Parents married	711	98.1
Mother education (primary/secondary)	174	24.0
Mother education (postsecondary and above)	551	76.0
Father education (primary/secondary)	148	20.5
Father education (postsecondary and above)	575	79.5
Income, SG$<3000	134	18.9
Income, SG$3000–SG$6000	297	41.8
Income, SG$6001–SG$9000	170	23.9
Income, SG$>9000	109	15.4
Screen home environment
Number of TVs (median, IQR)	725	2 (1)
Number of mobile devices (median, IQR)	725	4 (2)
Number of computers (median, IQR)	725	2 (1)
Number of video consoles (median, IQR)	725	0 (1)
Parental knowledge of professional screen viewing recommendations
Recommend minimal viewing	91	12.7
Recommend viewing <2 h/day	439	61.3
Recommend viewing ≥2 h/day	186	26.0
Parental screen viewing practices
Parent screen time, hours/day (median, IQR)	713	8.0 (7.7)
Parental rule setting practices
Rules on time	594	82.9
Rules on programme content	583	81.4
Rules on time and programme	539	75.2
Rules on time,* no screen time	27	4.6
Rules on time,* less than 2 h	457	76.9
Rules on time,* ≥2 h	110	18.5

*Based on participants who set rules on screen viewing time (n=594).

The most common screen devices in households were mobile devices with a median of four devices per household, followed by TVs and computer. Video game consoles were considerably less common. With regard to the knowledge about professional SV recommendations, only 12.7% of parents were aware that SV in children below the age of 2 should be minimised and almost twice as many parents (26%) believed that the recommended duration was 2 h or more per day as recommended by guidelines. About 80% of parents set rules on time or programme of SV, and about 75% set rules on both time and programme. Although more than 10% of parents appeared to know that SV should be minimised according to professional SV recommendations, less than 5% among parents set rules to minimise SV. More than 75% of parents with rules on SV time set rules of less than 2 h SV per day (or 63% among all 725 parents).

SV behaviour

The prevalence of daily SV among children is presented in detail in table 2. The overall prevalence of daily SV was 53.5%, with TVs being most widely used (44.8%), followed by mobile devices (30.5%), computers (6.6%) and video game consoles (0.4%). The prevalence of daily SV of 2 h or more per day across all devices was 16.3%, and the prevalence of TV, mobile device, computer and video game console viewing was 11.4%, 3.9%, 1.0% and 0.0%, respectively.

Table 2

Total and device-specific screen viewing (n=725)

	No screen viewing		Up to 2 h of screen viewing		≥2 h of screen viewing
	N	Per cent	N	Per cent	N	Per cent
Total screen time	337	46.5	270	37.2	118	16.3
TV	400	55.2	242	33.4	83	11.4
Mobile device	504	69.5	193	26.6	28	3.9
Computer	677	93.4	41	5.7	7	1.0
Video game console	722	99.6	3	0.4	0	0.0

Figure 1 presents the prevalence of daily total SV and by device, as well as SV 2 h or more per day across 6-month age groups. Among children younger than 6 months, any daily SV activity was reported in 29%, and a higher prevalence was reported with increasing age, being 88.2% at age 18–24 months. Differences between age groups and SV status across all devices and each device (apart from video consoles) were statistically significant. In all age groups, TV viewing was the most common form of SV behaviour, followed by mobile devices, computers and video game consoles. The difference in prevalence of SV between TV and mobile device was less pronounced at age 18–24 months as compared with the younger age groups.

Figure 1

Prevalence of screen viewing behaviour overall and according to device type according to 6-month age groups (n=725). *Difference across age groups, p<0.001.

In those with any SV activity (table 3), median total SV was 60.0 (IQR 95.7) minutes per day among children aged less than 6 months. It was 80 (IQR 120) minutes per day in those aged 18–24 months (p=0.0004). SV aggregated across devices was significantly higher at an older age (p=0.0004), but when investigating different devices a statistically significant difference was only observed for mobile devices.

Table 3

Screen viewing according to age groups and median screen viewing time among children who are exposed to screen devices

	<6 Months (n=324)				6 to <12 (n=164)				12 to <18 (n=144)				18–24 Months (n=93)
	N	Per cent	Median	IQR	N	Per cent	Median	IQR	N	Per cent	Median	IQR	N	Per cent	Median	IQR	p Value*
Total screen time	94	29.0	60.0	95.7	102	62.2	60.0	90.0	110	76.4	70.0	122.1	82	88.2	80.0	120.0	0.0004
TV	79	24.4	45.0	81.4	84	51.2	51.4	97.1	95	66.0	60.0	98.6	67	72.0	60.0	72.9	0.5820
Mobile device	42	13.0	23.2	49.3	46	28.0	23.6	22.1	72	50.0	32.1	45.0	61	65.6	40.0	30.0	0.0004
Computer	7	2.2	21.4	32.9	14	8.5	30.0	32.9	11	7.6	30.0	45.0	16	17.2	43.9	75.0	0.0876
Video consoles	1	0.3	5.7	0.0	0	0.0	–	–	1	0.7	60.0	0.0	1	1.1	10.0	0.0	0.3679

*For comparison of median screen viewing across age groups.

Sociodemographic, screen home environmental, parental knowledge and practice correlates of SV in children

In unadjusted analysis, increased child age, Malay and Indian ethnicity and lower maternal educational level were significantly associated with higher prevalence of any daily SV behaviour among the sociodemographic variables (table 4). The screen home environment was not significantly associated with SV behaviour and a borderline significant association between lack of parental knowledge of professional SV recommendations and higher prevalence of any daily SV behaviour was found. With regard to practices, parental SV was not significantly associated with daily SV behaviour, but the absence of rule setting practices were significantly associated with higher prevalence of any SV activity among children.

Table 4

Unadjusted and adjusted associations between any screen viewing (SV) activity and sociodemographic, home environment, parental knowledge and practices

			Unadjusted model				Multivariable model*
	N	Per cent SV	OR	Lower 95% CI	Upper 95% CI	p Value	OR*	Lower 95% CI	Upper 95% CI	p Value*
Sociodemographic characteristics
Age, <6 months	324	29.0	1.0				1.0
Age, >6 months	401	73.3	6.72	4.85	9.32	<0.001	8.07	5.57	11.69	<0.001
Chinese	425	46.6	1.0			<0.001†	1.0			<0.001†
Malay	196	65.3	2.16	1.52	3.06	<0.001	2.72	1.73	4.28	<0.001
Indian	82	58.5	1.62	1.00	2.61	0.049	2.53	1.39	4.61	0.002
Others	22	63.6	2.01	0.82	4.88	0.125	3.19	1.13	9.05	0.029
Male	401	55.1	1.0				1.0
Female	324	51.5	0.87	0.65	1.16	0.338	0.93	0.65	1.32	0.686
Mother education (primary/secondary)	174	60.9	1.0				1.0
Mother education (postsecondary)	551	51.2	0.67	0.48	0.95	0.025	0.70	0.42	1.16	0.167
Father education (primary/secondary)	148	59.5	1.0				1.0
Father education (postsecondary)	575	51.8	0.73	0.51	1.06	0.098	1.00	0.60	1.66	0.993
Income, <3000	134	58.2	1.0			0.278†	1.0			0.929†
Income, 3000–6000	297	54.2	0.85	0.56	1.28	0.439	0.96	0.57	1.62	0.882
Income, 6001–9000	170	52.4	0.79	0.50	1.25	0.309	1.11	0.59	2.08	0.755
Income, >9000	109	45.9	0.61	0.37	1.01	0.056	0.94	0.47	1.90	0.866
Parents married	711	53.3	1.0				1.0
Parents not married	14	64.3	1.58	0.52	4.75	0.418	1.85	0.52	6.62	0.343
Screen home environment
Up to 1 TV	309	49.5	1.0			0.176†	1.0			0.111†
2 TVs	236	56.8	1.34	0.95	1.88	0.093	1.57	1.03	2.40	0.036
3+ TVs	180	56.1	1.30	0.90	1.89	0.159	1.29	0.79	2.10	0.315
Up to 2 mobile devices	157	47.8	1.0			0.265†	1.0			0.326†
3 Mobile devices	199	55.3	1.35	0.89	2.06	0.160	1.47	0.89	2.45	0.135
4+ Mobile devices	369	55.0	1.34	0.92	1.94	0.128	1.26	0.77	2.07	0.352
Up to 1 computer	312	58.3	1.0			0.078†	1.0			0.892†
2 Computers	242	49.8	0.71	0.51	0.99	0.046	1.10	0.72	1.70	0.656
3+ Computers	171	50.0	0.71	0.49	1.04	0.078	1.02	0.61	1.70	0.939
0 Video consoles	504	53.0	1.0				1.0
1+ Video consoles	221	54.8	1.07	0.78	1.48	0.659	1.04	0.69	1.56	0.863
Parental knowledge of professional recommendations
Recommend minimal viewing	91	41.8	1.0			0.058†	1.0			0.145†
Recommend viewing <2 h/day	439	55.1	1.71	1.08	2.71	0.021	1.45	0.85	2.46	0.172
Recommend viewing >2 h/day	186	55.4	1.73	1.04	2.87	0.034	1.84	1.00	3.39	0.049
Parental SV practices
Parent screen time (lowest tertile)	238	55.9	1.0			0.357†	1.0			0.430†
Parent screen (medium tertile)	240	55.0	0.96	0.67	1.38	0.846	0.88	0.57	1.36	0.568
Parent screen time (highest tertile)	235	49.8	0.78	0.55	1.12	0.185	0.75	0.48	1.16	0.196
Parental rule setting practices
No rules on time and programme	178	60.7	1.0				1.0
Rules on time and programme	539	51.2	0.68	0.48	0.96	0.029	0.65	0.43	0.98	0.041

*Adjusted for all other co-variables.

†Overall p value.

When controlling for all other variables, older children (OR 8.07, 95% CI 5.57 to 11.69), and children with Malay (OR 2.72, 95% CI 1.73 to 4.28), Indian (OR 2.53, 95% CI 1.39 to 4.61) or other ethnicities (OR 3.19, 95% CI 1.13 to 9.05) had a significantly increased odds of daily SV. Among home environmental variables, parents who reported having two TVs were significantly more likely to report SV in their children (OR 1.57, 95% CI 1.03 to 2.40). Although the association between knowledge of SV recommendations and children engaging in any SV was insignificant, parents who believed that guidelines recommend more than 2 h of SV had a significant greater odds of engaging in any SV (OR 1.84, 95% CI 1.00 to 3.39). Similarly, children of parents who specified screen time rules had a significant reduced odds of any SV activity (OR 0.65, 95% CI 0.43 to 0.98).

As presented in table 5, observed associations with regard to 2 or more hours of SV per day were similar for sociodemographic variables. However, associations with parental knowledge, parental SV and parental rule setting were more consistently and more strongly associated with viewing 2 or more hours per day as compared with any daily SV. The screen home environment had devices with significant associations in the unadjusted analysis. Specifically, a greater number of TVs, mobile devices and video game consoles was significantly associated with higher prevalence of SV time of 2 or more hours. However, they become insignificant in the adjusted analysis. Different from table 4, incorrect parental belief on SV recommendations and higher parental SV time were significantly associated with higher prevalence of SV time of 2 or more hours. Moreover, rule setting practices were significantly associated with lower SV. These significant findings persisted after adjustment.

Table 5

Unadjusted and adjusted associations between screen viewing (SV) time per day ≥2 h and sociodemographic, home environment, parental knowledge and practices

			Unadjusted model				Multivariable model*
	N	Per cent SV	OR	Lower 95% CI	Upper 95% CI	p Value	OR	Lower 95% CI	Upper 95% CI	p Value
Sociodemographic characteristics
Age, <6 months	324	7.7	1.0				1.0
Age, >6 months	401	23.2	3.61	2.26	5.77	<0.001	4.88	2.87	8.31	<0.001
Chinese	425	9.7	1.0			<0.001†	1.0			<0.001†
Malay	196	28.1	3.65	2.33	5.72	<0.001	3.93	2.26	6.85	<0.001
Indian	82	19.5	2.27	1.20	4.28	0.011	4.14	1.88	9.12	<0.001
Others	22	27.3	3.51	1.30	9.47	0.013	5.39	1.70	17.08	0.004
Male	401	17.5	1.0				1.0
Female	324	14.8	0.82	0.55	1.23	0.339	0.80	0.50	1.28	0.353
Mother education (primary/secondary)	174	21.8	1.0				1.0
Mother education (postsecondary)	551	14.5	0.61	0.40	0.94	0.024	0.68	0.37	1.27	0.227
Father education (primary/secondary)	148	21.0	1.0				1.0
Father education (postsecondary)	575	15.1	0.67	0.43	1.06	0.089	0.90	0.49	1.66	0.737
Income, <3000	134	19.4	1.0			0.224†	1.0			0.894†
Income, 3000–6000	297	17.5	0.88	0.52	1.49	0.636	0.95	0.49	1.84	0.877
Income, 6001–9000	170	15.3	0.75	0.41	1.36	0.346	1.19	0.52	2.69	0.684
Income, >9000	109	10.1	0.47	0.22	0.99	0.048	0.91	0.34	2.45	0.859
Parents married	711	16.2	1.0				1.0
Parents not married	14	21.4	1.41	0.39	5.15	0.600	1.57	0.38	6.47	0.533
Screen home environment
Up to 1 TV	309	12.3	1.0			0.031†	1.0			0.263†
2 TVs	236	17.8	1.54	0.96	2.49	0.074	1.60	0.91	2.82	0.105
3+ TVs	180	21.1	1.91	1.17	3.13	0.010	1.40	0.74	2.64	0.298
Up to 2 mobile devices	157	10.2	1.0			0.005†	1.0			0.092†
3 Mobile devices	199	13.1	1.32	0.68	2.57	0.405	1.24	0.58	2.67	0.581
4+ Mobile devices	369	20.6	2.29	1.29	4.06	0.005	1.98	0.98	4.01	0.056
Up to 1 computer	312	18.0	1.0			0.090†	1.0			0.403†
2 Computers	242	12.0	0.63	0.39	1.01	0.057	0.93	0.52	1.68	0.815
3+ Computers	171	19.2	1.09	0.67	1.75	0.737	1.41	0.74	2.68	0.292
0 Video consoles	504	14.3	1.0				1.0
1+ Video consoles	221	20.8	1.58	1.05	2.38	0.029	1.06	0.64	1.76	0.827
Parental knowledge of professional recommendations
Recommend minimal viewing	91	13.2	1.0			0.002†	1.0			0.021†
Recommend viewing <2 h/day	439	13.2	1.00	0.51	1.95	0.995	0.70	0.33	1.48	0.347
Recommend viewing >2 h/day	186	24.7	2.16	1.08	4.32	0.029	1.46	0.66	3.22	0.349
Parental SV practices
Parent SV time (lowest tertile)	238	13.0	1.0			0.004†	1.0			0.005†
Parent SV time (medium tertile)	240	12.9	0.99	0.58	1.69	0.972	0.79	0.43	1.46	0.455
Parent SV time (highest tertile)	235	23.0	1.99	1.23	3.23	0.005	1.90	1.07	3.36	0.028
Parental rule setting practices
No rules on time and programme	178	21.9	1.0				1.0
Rules on time and programme	539	14.3	0.59	0.39	0.91	0.017	0.55	0.33	0.90	0.019

*Adjusted for all other co-variables.

†Overall p value.

Association of rule setting type and strictness with SV

Table 6 presents unadjusted and adjusted associations of rule setting practices with any SV and SV of ≥2 h/day. Regardless of the SV outcome and adjustment, ‘rules on programme’ were not significantly associated with SV in young children. On the contrary, ‘rules on time’ were consistently associated with a lower likelihood of any SV and SV of ≥2 h/day. Moreover, stricter rules on SV time were consistently associated with lower levels of any SV and SV of ≥2 h/day.

Table 6

Unadjusted and adjusted associations of types of rules and strictness of rules with any screen viewing and ≥2 h of screen viewing

	Any screen viewing							≥2 h of screen viewing
	N	Per cent screen viewing	p Value	OR*	Lower 95% CI	Upper 95% CI	p Value*	N	Per cent screen viewing	p Value	OR*	Lower 95% CI	Upper 95% CI	p Value*
Rules on programme content
Yes	583	52.3		1.0				583	16.1		1.0
No	133	58.7	0.187	1.45	0.92	2.28	0.111	133	15.8	0.925	1.11	0.62	1.98	0.730
Rules on time
Yes	594	50.8		1.0				594	13.8		1.0
No	123	66.7	0.001	1.72	1.07	2.78	0.026	123	27.6	<0.001	2.33	1.37	3.98	0.002
Strictness of rules on time			<0.001†				0.003†			<0.001†				<0.001†
No rules on time	123	66.7		1.0				123	27.6		1.0
0 h of screen viewing	27	14.8	<0.001	0.09	0.02	0.36	0.001	‡
Less than 2 h of screen viewing	457	52.5	0.005	0.65	0.40	1.07	0.091	484‡	10.3	<0.001	0.29	0.17	0.52	<0.001
≥2 h of screen viewing	110	52.7	0.031	0.50	0.27	0.94	0.031	110	31.8	0.486	0.98	0.51	1.89	0.955

*Adjusted for age, gender, ethnicity, mother's education, father's education, parental marital status, income, number of TVs, number of mobile devices, number of computers, number of video game consoles, knowledge of screen viewing recommendations, parental screen viewing.

†Overall p value.

‡Combined categories due to n=0 screen viewing in rule category ‘0 screen viewing’.

Discussion

This study investigated SV patterns and their correlates in very young children, considering multiple different screen device types, including TV, computer, as well as mobile screen devices and video game consoles. Given the widespread use, especially of mobile screen devices in day-to-day life, this comprehensive approach is important to fully understand SV patterns. Based on a recent systematic review of the literature, this study is the first in young children to investigate SV patterns and their correlates in such detail.31

Our findings show that among children up to the age of 2, SV on a daily basis is very common. Even in children aged less than 6 months, about 30% were engaging daily in SV and by the age of 18–24 months, about 90% of children engaged in daily SV activities. We made similar observations with regard to SV time, so that by the age of 18–24 months the median SV time was almost 1.5 h/day and almost 1/3 of the children were engaging in 2 or more hours of SV time per day. Our findings are consistent with research from other countries that have previously reported high exposure to TV, computer and videos at young ages.22

Although TVs still appear to be the most common form of screen devices young children are exposed to, our findings also highlight that TVs are not the only relevant screen device children are using. Mobile devices have become widely available in the entire population and they now seem to reflect the second most important source of SV in very young children. Furthermore, we noted that the difference in the prevalence between TV and mobile devices is least pronounced at the age of 18–24 months when compared with younger age groups making SV attributable to mobile devices comparable with that with TVs. Moreover, in terms of the SV time, we noted a significant difference across age groups for total SV and mobile devices but not for TVs or other screen devices. Previous studies from the USA and Europe, targeting older children, have also reported on the frequent use of mobile devices. In addition, recent discussions in the scientific literature have addressed the potential health implications of the widespread use of modern mobile screen devices in very young children, but our study appears to be among the first to comprehensively quantify patterns and relevant correlates in this population.27 29 30

Besides age, we noted that Malay, Indian and other ethnic groups were considerably more likely to engage in any SV and 2 or more hours of SV per day as compared with Chinese. This is consistent with existing literature that also reported ethnic differences in SV in Western populations.31 34 On the other hand, we did not identify significant associations between daily SV or SV of 2 or more hours per day and other sociodemographic variables, such as child's gender, mother's or father's educational level, income and parental marital status, when taking potential confounders into consideration.31

Parental knowledge and parenting practices may have the potential to reflect targets for health promotion strategies to reduce or delay the update of SV in very young children. Our study found strong and consistent associations between child SV and parental rule setting practices. Findings indicate that children are less likely to engage in SV if parents set rules on SV time, and additionally if the rules on SV time are stricter. On the other hand, rules on programmes did not appear to be significantly associated with SV in children. A recent systematic review of the literature reported mixed and less consistent findings of such TV policies at home. However, this review summarised studies of very different methodology and, importantly, mostly targeted older children and adolescents. Only two of the included studies had focused on children below the age of 2, and they only concentrated on TV viewing.35–37 Our findings are similar to previous research in older children, however, that also indicated that the type of rules (time or programme) may have implications in terms of SV.37 In addition to rule setting practices, knowledge of professional SV recommendations and parental SV behaviour was significantly associated with SV in children, particularly with viewing 2 or more hours per day. These findings seem to be consistent with studies conducted in older children and adolescents that reported associations between children's and parental SV.27 31 38

Our study has a number of strengths, including a very high participation rate and detailed outcome assessments. Building on Singapore's National Childhood Immunization programme facilitated this study of a defined population of healthy children aged 2 and below. Some limitations, however, have to be acknowledged. First, the survey recruited from 2 out of the 18 polyclinics covering the whole of Singapore. Variations in the overall population reflected in the other centres may hence be unaccounted for. Second, this is a cross-sectional study which limits its ability to draw inferences about the direction of cause and effect. However, modern screen devices, in particular tablets, have only become widely used after the launch of the iPad in 2010. Hence, at this stage, longitudinal data on the use and determinants of modern screen devices are not available. Third, the study was conducted in February and may not reflect seasonal variations in screen time exposure, although seasonal variation in weather is limited in tropical Singapore. Fourth, our study relied on a proxy report. While questionnaires were developed on the basis of the existing literature and extensively piloted, they were not validated and we were not able to determine SV objectively. This could introduce measurement bias. Finally, we did not distinguish between the type of programme screen time was spent on or whether it was, for instance, spent for video calling with family members. This will be important to consider in future studies because research has indicated that health consequences may vary depending on the activity performed.39

Our findings have considerable public health implications because they highlight the importance of targeting SV practices at a very early age, ideally even during the first 6 months of life in order to develop appropriate practices to prevent the update of these activities in the first place. Our study also showed that TV viewing still seems to be the most common but not the only relevant screen device young children are exposed to. With increasing age, mobile devices take up an increasing part of total SV time, and by the age of 18–24 months are almost as prevalent as TVs. Different devices may require different intervention strategies and recommendations, given their portability, small size and multifunctional usability (eg, watching videos, playing games, reading, listening to audiobooks). We were able to identify a number of potentially modifiable factors that were strongly associated with SV practices in children, including parental knowledge of professional SV recommendations, parental SV and rule setting practices and, to a lesser degree, the number of devices in the household. These findings were particularly present in the case of heavy SV, which suggests that reducing SV behaviour in young children might be an achievable target, while replacing SV entirely could be more challenging. Modifiable factors offer opportunities for health promotion strategies, administered for instance by healthcare professionals during developmental assessments or teachers in childcare centres at a very young age. This could help to delay and reduce SV in children. Our findings are consistent with earlier publications that reported a high prevalence of regular TV viewing in very young children.19 20 22 40 However, our findings expand the existing literature considerably in various aspects. Importantly, we were able to investigate multiple up-to-date screen device types and a broad spectrum of sociodemographic, environmental, parental knowledge and practices related correlates of SV in very young children. In addition, literature reviews have shown that the bulk of the existing evidence comes from the USA and Europe, with a very limited number of studies from Asian countries, despite the rapid economic development and increasing concerns about the increase in non-communicable diseases.31

In conclusion, our study found a high prevalence of SV overall. Within the first 2 years of a child’s life, the prevalence of SV increased across age groups. By the age of 2 years, almost all children were exposed to screen devices daily and a substantial proportion of this young study population exceeded 2 h of SV, usually recommended as the maximum for much older children. While TV viewing remains the most common contributor to overall SV in children, our findings also confirm the increasingly important role of mobile screen devices. Strategies aimed at reducing the rapid uptake of regular SV practices in children are urgently needed to influence future behaviours and health. We were able to identify potential targets for such strategies, especially for SV or 2 or more hours per day, including parental knowledge regarding professional SV recommendations, parental SV practices and the implementation of strict rules on SV time in young children.