Screen time and sleep problems in South Indian preschoolers: a community-based cross-sectional study

Abstract

Objectives

To estimate the prevalence of sleep problems among children aged 2–5 years residing in South India, assess its association with screen time and identify a predictive screen time threshold.

Design

Population-based cross-sectional study.

Setting

Field practice areas in rural and urban centres of a medical college in South India.

Participants

In total, 523 children aged 2–5 years were selected by simple random sampling.

Primary outcome measures

Sleep problems were assessed using the validated bedtime problems, excessive daytime sleepiness, awakenings during the night, regularity of sleep and snoring sleep screening tool. Sociodemographic and behavioural factors, including screen time, were also examined. The optimal predictive screen time cut-off was identified using receiver operating characteristic (ROC) analysis.

Results

Sleep disturbances were reported in 39.6% of children (95% CI 35.5% to 43.8%). The most common sleep problems were irregular sleep (22.2%), bedtime resistance (20.8%) and night awakening (19.9%). Multivariate logistic regression showed strong associations between sleep problems and screen use in bed (adjusted OR (AOR) = 3.8; 95% CI 2.4 to 6.1), excess screen time (AOR=3.3; 95% CI 1.8 to 6), smaller family size (AOR=3.1; 95% CI 1.5 to 6.1), reduced physical activity (AOR=2.6; 95% CI 1.6 to 4.2), shorter birth spacing (AOR=1.8; 95% CI 1.1 to 2.8), lower socioeconomic status (AOR=1.8; 95% CI 1.2 to 2.8) and maternal screen time>2 hours/day (AOR=1.6; 95% CI 1.04 to 2.6). ROC analysis identified ≥2.4 hours per day of screen time as the optimal threshold for predicting sleep problems (area under the curve=0.800; sensitivity, 73.9% and specificity, 77.2%).

Conclusion

In this large population-based study, two of the five preschool children experienced sleep problems, with excess screen time, particularly screen use in bed, being the key contributing factor. This is one of the few Indian studies to establish an ROC-derived screen time threshold for identifying sleep problems. These findings can guide targeted parental advice and early preventive strategies to promote healthy sleep in preschool children.

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• This study adds to the limited community-based evidence from India to assess sleep problems and their relationship with screen time in children aged 2–5 years.

• Large diverse sample from urban and rural field practice areas, increasing the generalisability to similar low- and middle-income settings.

• Analyses were adjusted for key sociodemographic and behavioural factors, with receiver operating characteristic analysis identifying a predictive screen time threshold.

• The cross-sectional study design limits causal inference but provides novel population-based evidence for targeted interventional strategies.

Introduction

During early childhood, sleep is essential for healthy physical growth and the development of cognitive abilities.¹ Unhealthy sleep patterns are a growing global public health concern in young children.² Sleep disturbances during this sensitive developmental window may contribute to poor cognition, a reduced attention span and an increased risk of childhood obesity.³ Among various contributing factors, screen time (television viewing and other digital devices) has emerged as a significant determinant of sleep problems in children.⁴

A meta-analysis by Janssen et al reported that prolonged screen time among children aged <5 years was significantly associated with sleep-related issues, such as delayed bedtime, poor sleep quality and frequent night awakenings.⁵ Continuous or background television exposure reduces parent–child interaction and can indirectly disrupt children’s sleep routines and emotional regulation.⁶ Blue light exposure from electronic devices, especially before bedtime, reduces melatonin production and disrupts the circadian rhythm, thereby affecting the sleep cycle in young children.⁷ Excessive screen time also reduces physical activity, compounding its negative effects on sleep.⁸ The coronavirus disease 2019 pandemic has amplified these behavioural patterns, owing to lockdown, playschool/daycare centre closures and restrictions in outdoor play.⁹

Recent community-based Indian studies have reported increasing trends in sleep disturbance among preschool children, from 33.3% in 2011 to 48.3% in 2025.^{10 11} This increase parallels the rapid increase in screen time among children with easy access to television and mobile phones, which began as early as infancy.¹²,¹⁴ As per recommendations from the WHO and Indian Academy of Paediatrics, children aged between 2 and 5 years must limit their sedentary screen use to <1 hour per day, focusing on age-appropriate, educational and supervised content.¹⁵,¹⁷ Despite these guidelines, several Indian studies have reported high screen time among preschool children, ranging from 59% to 86% above the recommended daily limits.^{12 14 18}

Despite its critical role in early childhood development, sleep remains an underexplored area in paediatric public health, especially in community-based studies on low- and middle-income countries (LMICs). Although several international studies have reported associations between screen time and sleep disturbances, most are hospital based and conducted in high-income countries. To the best of our knowledge, this is one of the few population-based Indian studies aimed at examining the association between screen time and sleep problems in preschool children using the bedtime problems, excessive daytime sleepiness, awakenings during the night, regularity of sleep and snoring (BEARS) sleep screening tool and identifying the receiver operating characteristic (ROC) curve-derived screen time threshold for predicting sleep problems. Such evidence is scarce in LMICs and can guide culturally tailored parental advice and public health interventions.

Therefore, our study aimed to estimate the prevalence of sleep disturbance and its association with screen time in children aged 2–5 years in southern India. ROC curve analysis was performed to predict sleep disturbances caused by excessive screen usage.

Methods

Study design and setting

This community-based, cross-sectional study was conducted among the population served by rural and urban health centres affiliated with the Department of Community Medicine at a medical college in South India. These centres function as demonstration sites for undergraduate and postgraduate training and provide primary healthcare to the population. These sites were selected because of their well-defined catchment populations and routinely updated family and child registers. Establishing community links and trained field staff at these sites ensured data reliability, accessibility and feasibility.

Study population

Inclusion criteria

Children aged 2–5 years residing in the selected urban or rural areas for at least 6 months were included in this study.

Exclusion criteria

Children diagnosed with neurodevelopmental disorders or chronic illnesses affecting sleep, such as epilepsy or autism spectrum disorder, were excluded. The diagnosis was based on caregiver reports and confirmed through medical records when available. Children whose caregivers declined to participate or who could not be contacted for follow-up were also excluded.

Sample size

The sample size was determined from a community-based South Indian study that reported the prevalence of sleep problems among preschool children using the BEARS screening tool as 33.3%.¹⁰ Assuming a 95% CI (Z=1.96) and an absolute precision of 4.25%, the minimum required sample size was 473. After accounting for 10% loss to follow-up in the 7-day screen time diary, the sample size was 521. Data were collected for 523 children.

Sampling technique

Data were collected from rural and urban areas under our medical college and research institute, where a register of all children aged <5 years, including their age and address, is routinely maintained. Our study sampling frame comprised children aged 2–5 years recorded in these registers. The eligible population comprised 950 children from a rural health and training centre, which covers 9 panchayats (village administrative units), and 1050 children from an urban health and training centre, which covers 19 urban subdivisional regions. Each eligible child was assigned a unique serial number, and the required sample size was selected using simple random sampling. Random numbers were generated using www.random.org to identify participants.

Recruitment was conducted through household visits conducted by health inspectors and public health nurses. Each household selected was approached individually. In cases where the household was locked, three repeat visits were attempted at different times of day to ensure the maximum possibility of contact. The next household on the sampling list was approached only when it remained unavailable after three attempts. This ensured systematic inclusion and minimised selection bias.

Data collection

The investigator and coinvestigators used a structured questionnaire to collect data only after the mother or primary caregiver provided written informed consent. The questionnaire was pilot tested among ten caregivers of preschool children in a nearby community field area that was not included in the main study sites to ensure clarity and cultural appropriateness. Information on various background demographic characteristics, including age, sex, family type, number of children, birth spacing and going to daycare/balwadi¹⁹ (government-run preschool centre), was collected. Socioeconomic status (SES) was assessed using the Modified Prasad classification (2022 update). Classes I and II were grouped as the upper socioeconomic strata and classes III–V as the lower socioeconomic strata.²⁰ All households possessed at least one digital device; hence, device ownership did not influence the participant’s eligibility.

Assessment of sleep problems

The BEARS sleep screening tool, a validated instrument designed for primary care and community health settings, was used to obtain information on sleep problems.²¹ The BEARS tool has been used in Indian settings and culturally adapted in several countries, including Spanish and Persian versions, supporting its cross-cultural applicability.^{10 22 23} Local language experts translated the questionnaire into the local language Tamil (online supplemental material 1). The forward–backward translation method was used and culturally adapted. The tool was administered during home visits through interviews with the child’s primary caregiver. All investigators/coinvestigators received standardised training from a paediatric sleep expert on the administration of the BEARS questionnaire. Periodic supervision and cross-checks were conducted to ensure uniformity and reduce interviewer bias. The entered data were double-checked by an independent data manager to ensure accuracy. A child was considered to have sleep problems if any one of the five domains was affected, consistent with the original description by Owens and Dalzell and the subsequent Indian application by Ravikiran et al.^{10 21} Problems in the individual domains were also recorded. The BEARS tool has been widely used and is well suited for identifying sleep disturbances in children aged 2–5 years, especially in community-based studies.

Assessment of screen time

Children’s screen time was measured using a prospective observational method. The mother or primary caregiver was provided with a screen time diary to monitor and record the child’s screen time over a 7-day period. Caregivers were instructed to record all types of exposure to screen, such as television and mobile devices. The diary was collected during follow-up home visits. To ensure accuracy and compliance, health inspectors and public health nurses made periodic telephone calls to the caregivers during the observational period. At the end of 7 days, the average screen time was calculated and considered the child’s screen time.

In addition, information was collected on maternal screen time and contextual screen use behaviours, including whether the child typically watched screens alone (television or digital media without active parental presence or supervision) and whether the screen was used before going to sleep. According to the WHO criteria, children with a screen time of >1 hour per day were classified as having excess screen time.

Patient and public involvement

Patients and/or the public were not involved in the design, or conduct or reporting or dissemination plans of this research.

Statistical analyses

IBM SPSS Statistics (V.29.0; IBM Corp, Armonk, NY, USA) was used for data analysis, and R software (V.4.5.1, released 13 June 2025; R Foundation for Statistical Computing, Vienna, Austria) was used to generate the figures. Sociodemographic and behavioural characteristics are expressed as percentages and frequencies. Spearman’s rank correlation coefficient was used to determine the correlation between maternal and child screen time.

Associations between sleep problems and sociodemographic and behavioural characteristics were assessed using χ² tests for the categorical variables, and the Mann–Whitney U test was used for the continuous variables. For 2×2 tables with small cell counts (expected frequency<5), Fisher’s exact test was used instead of the χ² test. A multivariate logistic regression model was used to identify the factors associated with sleep problems, and adjusted ORs (AORs) and 95% CIs were calculated. The cut-off values for categorisation were based on established recommendations. Physical activity of at least 3 hours/day was the minimum requirement per the WHO guidelines for children aged between 2 and 5 years.¹⁵ Maternal screen time was classified as ≤2 or > 2 hours/day, based on evidence linking >2 hours/day to adverse sleep and mental health outcomes and aligned with public health recommendations for recreational use.²⁴,²⁶ Maternal screen time and physical activity duration were additionally analysed as continuous variables in a sensitivity model to assess the robustness of associations. Multicollinearity was assessed using the variance inflation factor (VIF), and all variables with VIF values of <2 were included in the final model.

Effect modification was analysed by testing the interaction terms between excess screen time and covariates. Stratified analyses were performed to identify the significant interactions. ROC curve analysis was performed to assess the predictive value of total screen time for sleep problems. Optimal cut-off values using the Youden index, sensitivity and specificity were reported. The p value<0.05 was considered statistically significant. No imputation was necessary, as there were no missing data, and the dataset was complete.

Results

A total of 523 children participated in this study, with 294 (56.2%) residing in urban and 229 (43.8%) in rural areas. Urban families had slightly older fathers and shorter birth spacing than that in rural families. Maternal employment was significantly higher in urban households than in rural households, and independent screen use and longer maternal screen time were more frequent among urban children than among rural children. Importantly, 73% of the children exceeded the WHO-recommended daily screen time limit of 1 hour, indicating a high burden of excess screen time across both locations (table 1).

Table 1. Sociodemographic and behavioural characteristics of children aged 2–5 years based on the place of residence (N=523).

Variable	Rural (n=229), n (%)	Urban (n=294), n (%)	P value
Gender
Male	117 (51.1)	143 (48.6)	0.578
Female	112 (48.9)	151 (51.4)
Total children in the family
≤2	196 (85.6)	266 (90.5)	0.084
>2	33 (14.4)	28 (9.5)
Birth spacing
≤2	54 (23.6)	96 (32.7)	0.023
>2	175 (76.4)	198 (67.3)
Going to daycare/balwadi
Yes	141 (61.6)	192 (65.3)	0.378
No	88 (38.4)	102 (34.7)
Maternal occupation
Housewife	203 (88.6)	232 (78.9)	0.003
Employed	26 (11.4)	62 (21.1)
Fathers’ occupation
No formal education	2 (0.9)	5 (1.7)	0.070
School (Primary–higher secondary)	153 (66.8)	168 (57.1)
Graduate to above	74 (32.3)	121 (41.2)
Fathers’ age	Median 32 IQR 30–35	Median 33.6 IQR 30–37	0.03
Type of family
Nuclear	117 (51.1)	218 (74.1)	<0.001
Others	112 (48.9)	76 (25.9)
SES Modified BG Prasad’s classification
Classes 1 and 2	93 (40.6)	173 (58.8)	<0.001
Classes 3–5	136 (59.4)	121 (41.2)
Physical activity
≤3 hours/day	117 (51.1)	218 (74.1)	<0.001
>3 hours/day	112 (48.9)	76 (25.9)
Watches screen alone
Yes	147 (64.2)	230 (78.2)	<0.001
No	82 (35.8)	64 (21.8)
Screen used in bed
Yes	63 (27.5)	93 (31.6)	0.307
No	166 (72.5)	201 (68.4)
Maternal screen time
≤2 hours/day	98 (42.8)	96 (32.7)	0.017
>2 hours/day	131 (57.2)	198 (67.3)
Excess screen time (WHO guidelines 2019)
Yes	158 (69)	224 (76.2)	0.066
No	71 (31)	70 (23.8)

SES based on the modified Brahm Govind (BG) Prasad classification.

Excess screen time is defined as screen used >1 hour/day, according to the WHO guidelines (2019).

SES, socioeconomic status.

Sleep problems were observed in 39.6% (95% CI 35.5 to 43.8%) of children. Among the five BEARS domains, irregular sleep patterns (22.2%) and bedtime resistance (20.8%) were the most common, followed by night awakening (19.9%). Excessive daytime sleepiness (9.4%) and snoring (1.1%) were less frequently reported (figure 1).

Figure 1. Prevalence of sleep problems in children aged 2–5 years, as assessed using the BEARS sleep screening tool. BEARS, bedtime problems, excessive daytime sleepiness, awakenings during the night, regularity of sleep and snoring.

Children with sleep problems had higher total screen time, as indicated by the box plot, which also showed a wider IQR and multiple outliers. These patterns suggest higher and more variable screen times among children with sleep disturbances. When examined based on device type, television and mobile screen time showed similar patterns, with higher median use and wider variability among children with sleep problems than among those without sleep problems (figure 2A). Children with a higher bedtime screen use had more sleep problems than those with a lower bedtime screen use. Spearman’s rank correlation revealed a moderate, statistically significant positive association between bedtime screen use and sleep problems (ρ=0.408, p<0.001). Although both groups had outliers, a wider IQR and higher upper values in the sleep problems group indicated greater variable bedtime screen use (figure 2B).

Figure 2. (A) Comparison of total, television and mobile screen time between children with and without sleep problems. (B) Comparison of bedtime screen use between children with and without sleep problems.

A moderate positive correlation was observed between maternal and child’s screen time (Spearman’s ρ=0.47, p<0.001), as shown in the scatterplot. The line of best fit illustrates that a higher screen time of mothers was associated with greater screen time in children (figure 3).

Figure 3. Scatterplot showing the association between maternal screen time and child screen time.

Factors associated with sleep problems in children

Young children were more likely to have sleep problems; with every 1-month increase in age, the odds of having sleep problems decreased by 4% (AOR=0.96; 95% CI 0.94 to 0.97). Multivariate logistic regression identified that children from smaller families (AOR=3.1; 95% CI 1.5 to 6.1), those with shorter birth spacing (AOR=1.8; 95% CI 1.1 to 2.8) and those from lower SES (AOR=1.8; 95% CI 1.2 to 2.8), were significantly associated with sleep problems. Behavioural factors strongly associated with sleep problems included less physical activity (AOR=2.6; 95% CI 1.6 to 4.2), mothers’ screen time >2 hour (AOR=1.6; 95% CI 1.04 to 2.6), screen use in bed (AOR=3.8; 95% CI 2.4 to 6.1) and excess screen time (AOR=3.3; 95% CI 1.8 to 6). Screen time had the strongest association with sleep problems, making it an important behavioural factor. Multicollinearity analysis showed VIF values between 1.02 and 1.62 for all variables, suggesting low collinearity; therefore, all predictors were included in the final regression model (table 2).

Table 2. Associations between sleep problems and sociodemographic and behavioural factors among children aged 2–5 years.

Variable	Sleep problem n (%)	No sleep problem n (%)	OR (95% CI)	P value	AOR (95% CI)*	P value
Place of residence
Rural	83 (36.2)	146 (63.8)	1
Urban	124 (42.2)	170 (57.8)	1.1 (0.9 to 1.2)	0.169	–	–
Gender
Female	104 (39.5)	159 (60.5)	1
Male	103 (39.6)	157 (60.4)	1 (0.7 to 1.4)	0.987	–	–
Age of the child in months	Mean 40.8 SD 10.6	Mean 42.9 SD 12.2	0.98 (0.97 to 0.99)	0.039	0.96 (0.94 to 0.97)	<0.001
Total children in the family
>2 (ref)	17 (27.9)	44 (72.1)	1
≤2	190 (41.4)	272 (58.9)	1.8 (1.1 to 3.2)	0.047	3.1 (1.5 to 6.1)	0.02
Birth spacing
>2	139 (37.3)	234 (62.7)	1
≤2	68 (45.3)	82 (54.7)	1.3 (0.9 to 2.0)	0.08	1.8 (1.1 to 2.8)	0.016
Going to daycare/balwadi
Yes	121 (36.3)	212 (63.7)	1
No	86 (45.3)	104 (54.7)	1.4 (1.1 to 2.1)	0.045	–	–
Maternal occupation
Employed	34 (38.6)	54 (61.4)	1
Housewife	173 (39.8)	262 (60.2)	1.1 (0.6 to 1.7)	0.843	–	–
Type of family
Nuclear	130 (38.8)	205 (61.2)	1
Others	77 (41.0)	111 (59.0)	1 (0.8 to 1.2)	0.629	–	–
SES
Classes 1 and 2	96 (36.1)	170 (63.9)	1
Classes 3–5	111 (43.2)	146 (56.8)	1.1 (0.9 to 2.1)	0.097	1.8 (1.2 to 2.8)	0.005
Physical activity
>3 hours/day	40 (21.3)	148 (78.7)	1
≤3 hours/day	167 (49.9)	168 (50.1)	3.6 (2.4 to 5.5)	<0.001	2.6 (1.6 to 4.2)	<0.001
Watches screen alone
No	28 (19.2)	118 (80.8)
Yes	179 (47.5)	198 (52.5)	3.8 (2.4 to 6.0)	<0.001	–	–
Screen used in bed
No	102 (27.8)	265 (72.2)	1
Yes	105 (67.3)	51 (32.7)	5.3 (3.5 to 8)	<0.001	3.8 (2.4 to 6.1)	<0.001
Maternal screen time
≤2 hours/day	50 (25.8)	144 (74.2)	1
>2 hours/day	157 (47.7)	172 (52.3)	1.4 (1.2 to 1.6)	<0.001	1.6 (1.04 to 2.6)	0.035
Excess screen time
No	20 (14.2)	121 (85.8)	1
Yes	187 (49.0)	195 (51)	5.8 (3.4 to 9.6)	<0.001	3.3 (1.8 to 6.0)	<0.001

p<0.05 in an adjusted analysis is considered significant.

^*Estimates derived from the multivariate logistic regression model.

AOR, adjusted OR; SES, socioeconomic status.

For sensitivity analysis, maternal screen time and duration of physical activity were entered as continuous variables in the multivariate logistic regression model, whereas other predictors were retained in categorical form. The direction and strength of the associations remained consistent with those in the primary analysis. Each additional hour of physical activity was associated with 31% lower odds of sleep problems (AOR=0.69; 95% CI 0.54 to 0.87; p=0.002), whereas each additional hour of maternal screen time showed a borderline increase in risk (AOR=1.12; 95% CI 0.99 to 1.25; p=0.065). The overall model fit remained good (Hosmer–Lemeshow p=0.70 and Nagelkerke R² = 0.34), confirming the robustness of the findings.

Effect modifier analysis

Interaction terms were used in the logistic regression model to assess whether sociodemographic and behavioural factors modified the effect of excess screen time on sleep problems. No significant effect modification was observed for any variable, except birth spacing (p=0.019). In the stratified analysis, excess screen time had a strong association with sleep problems in children with birth spacing <2 years (AOR=7.4; 95% CI 2.4 to 22.3) compared with those with birth spacing >2 years (AOR=2.3; 95% CI 1.2 to 4.7) (table 3).

Table 3. Stratified analysis of excess screen time and sleep disturbances based on birth spacing.

Group	Unadjusted OR	95% CI	P value	AOR	95% CI	P value
Overall	5.8	3.4 to 9.6	<0.001	3.08	1.6 to 5.9	0.001
Birth spacing <2 years	13.2	4.8 to 36.2	<0.001	7.4	2.4 to 22.3	<0.001
Birth spacing >2 years	4.2	2.3 to 7.6	<0.001	2.3	1.2 to 4.7	0.014

AOR, adjusted OR.

Children with screen time of >1 hour had ten times higher odds of having bedtime resistance (OR=10.1; 95% CI 4.0 to 25.5) and excessive daytime sleepiness (OR=9.7; 95% CI 2.3 to 40.6) than that in children with screen time <1 hour. Irregular sleep (OR=4.3; 95% CI 2.2 to 8.4) and frequent night awakenings (OR=3.8; 95% CI 1.9 to 7.3) were also more common among children with excessive screen time than among their counterparts. Owing to its low prevalence, snoring did not reveal a statistically significant relationship (p=0.198, Fisher’s exact test) (table 4).

Table 4. Association between screen time >1 hour per day and specific sleep problems (BEARS domains).

Sleep problem	≤1 hour/day n (%)	>1 hour/day n (%)	OR (95% CI)	P value
Bedtime problems	5 (3.5)	104 (27.2)	10.1 (4.0 to 25.5)	<0.001
Excessive daytime sleepiness	2 (1.4)	47 (12.3)	9.7 (2.3 to 40.6)	<0.001*
Night awakenings	11 (7.8)	93 (24.3)	3.8 (1.9 to 7.3)	<0.001
Irregular sleep	9 (6.4)	107 (28.0)	4.3 (2.2 to 8.4)	<0.001
Snoring	0 (0)	6 (1.6)	–	0.198*

Excessive screen time defined as >1 hour/day per the WHO guidelines (2019). Sleep problems classified using BEARS domains. Reference group: ≤1 hour screen time.

^*P values based on Fisher’s exact test (two-sided).

BEARS, bedtime problems, excessive daytime sleepiness, awakenings during the night, regularity of sleep and snoring.

ROC curve analysis was performed to assess the predictive capacity of total screen time in children with sleep problems. The area under the curve was 0.800 (95% CI 0.760 to 0.840; p<0.001), indicating a good predictive value. The optimal threshold value of ≥2.375 hours/day of screen time was identified using Youden index (0.51), with a sensitivity of 73.9% and a specificity of 77.2% (figure 4).

Figure 4. ROC curve analysis of total screen time predicting sleep disturbances in children aged 2–5 years. ROC, receiver operating characteristic.

Discussion

This large community-based study identified a high prevalence of sleep problems (39.6%) among children aged 2–5 years, as assessed using the BEARS sleep screening tool. Behavioural factors, such as excess screen time (AOR=3.3; 95% CI 1.8 to 6) and screen use in bed (AOR=3.8; 95% CI 2.4 to 6.1), were associated with sleep problems. Other contributing factors included reduced physical activity, short birth spacing, smaller family size and longer maternal screen time. ROC curve analysis identified a threshold of ≥2.4 hours per day of screen time as predictive of sleep problems, highlighting the need for early interventions for this modifiable risk factor.

The observed prevalence (39.6%) of sleep problems falls within the global and Indian estimates of 30%–50%.¹¹²⁷,³⁰ Irregular sleep patterns, bedtime problems and night awakenings were the most common, comparable with the findings of Ravikiran et al in rural Karnataka.¹⁰ In contrast, a study from Egypt reported a higher prevalence of irregular sleep (60.5%), bedtime problems (58.9%) and night awakening (31%), highlighting the influence of cultural and lifestyle factors on sleep problems.³¹ Beyond identifying these sleep issues, our study focused primarily on the influence of environmental factors, particularly screen time, on sleep disturbance. Children with sleep problems had higher total screen time, particularly with increased bedtime screen use, than that in children without sleep problems. A systematic review by Janssen et al⁵ reported similar problems, whereas a randomised controlled trial by Pickard et al³² demonstrated that avoiding screens an hour before bedtime significantly improved sleep quality in children. Our study observed a moderate correlation between maternal and child screen time (r=0.5), consistent with the finding from a longitudinal study by Sim et al,³³ establishing the role of parental modelling on child’s screen use, an association that remains underexplored in Indian studies.

Multivariate regression analysis revealed that shorter birth spacing, smaller family size, lower SES, reduced physical activity and screen-related behaviours, such as exceeding WHO-recommended screen time limits, use of screen in bed and higher maternal screen time, were significantly associated with sleep problems in children. These results are consistent with those of a systematic review by Newton et al³⁴ and a study by Staples et al,³⁵ which highlighted the influence of bedtime screen use, parental modelling and small birth spacing on sleep problems. Indian studies by Kaur et al¹² and Jain et al³⁶ reported similar associations. Our study identified birth spacing as an effect modifier, with a significant association of screen time and sleep problem with shorter birth spacing, indicating that overburdened parents are more likely to rely on screen to manage their children, which may indirectly affect their sleep. The literature supports the idea that shorter intersibling intervals can strain parental time and caregiving resources, thus influencing child outcomes.³⁷

These findings highlight the influence of behavioural factors on sleep problems in children. Excess screen time, particularly in bed, disrupts the circadian rhythm by reducing melatonin production induced by the blue light emitted by screens, thereby delaying sleep onset and negatively affecting sleep quality.^{38 39} The fast-paced and interactive content can further overstimulate children both cognitively and emotionally.⁴⁰ Maternal screen time influences children’s screen behaviour through modelling.^{41 42} In households with short birth spacing, parents use screens as a soothing strategy to manage caregiving demands.¹⁴ Low levels of physical activity may further reduce the opportunities for adequate energy expenditure, resulting in sleep problems.⁴³ Taken together, the interplay of these biological, behavioural and social factors provides a plausible explanation for sleep disturbances in this cohort. In our study, the ROC-derived cut-off of 2.4 hours/day provides balanced sensitivity (73.9%) and specificity (77.2%) for predicting sleep problems. This threshold exceeds the WHO recommendation and likely reflects high baseline screen time and a focus on sleep-specific outcomes. Although the cut-off from our study may have high clinical value for identifying children at high risk of sleep problems, adherence to that WHO recommendations is essential for the overall development in early childhood.

These findings highlight the potential for targeted interventions to mitigate screen-related sleep problems in preschoolers. Strategies, such as limiting recreational screen use, particularly before bedtime; promoting active play; ensuring screen-free bedtime routines and encouraging parental modelling of healthy screen habits, may be effective. Evidence from recent interventional studies, including those from India and other settings, demonstrates that structured, multicomponent programmes combining caregiver education, behavioural counselling and digital-use monitoring can significantly reduce screen time and improve sleep outcomes among young children.⁴⁴,⁴⁶ Integrating similar caregiver-focused educational modules into routine maternal and child health services, well-baby clinics and preschool programmes can provide a scalable public health approach to promote healthy sleep in early childhood.

To the best of our knowledge, this is one of the few large population-based studies, including 523 preschoolers from both urban and rural settings, enhancing the representativeness of our findings. A validated and widely used paediatric sleep screening tool was used to assess children across multiple sleep domains. Screen time was recorded using a 7-day parental diary, which provided an accurate estimate. Standardised training and periodic supervision of the investigators ensured data quality and minimised interviewer bias. The use of multivariate regression and ROC curve analysis strengthened the analysis, providing a screen time threshold for sleep problems. Behavioural and familial factors, such as maternal screen time, family size, birth spacing and physical activity, were also considered, offering a comprehensive understanding of their influences on sleep.

Despite the methodological rigour, this study has certain limitations. Because this was a cross-sectional study, causal relationships between screen time and sleep problems could not be inferred. Sleep and screen time were parent-reported, introducing potential recall and social desirability biases. Although validated internationally, BEARS is a screening tool rather than a diagnostic tool, and formal psychometric testing of the Tamil-adapted version has not yet been undertaken. Data on screen content and the exact timing of exposure before bedtime were not collected, representing additional unmeasured confounders. Finally, as this study was conducted in a single-state setting, generalisability is limited.

As screen time significantly affects sleep during the crucial developmental period of 2–5 years, it is essential to integrate sleep and screen time counselling into routine maternal and child health services, well-baby clinics and preschool health programmes. Parents should be counselled on the importance of avoiding screen use in bed, promoting physical activity and modelling healthy screen behaviours. Public health policy should prioritise setting clear, enforceable screen time limits in daycare and preschools. Further longitudinal and randomised controlled trials are required to establish causality and evaluate the effectiveness of targeted interventions in reducing sleep-related issues in young children.

Conclusion

This large community-based study in South India found that nearly 40% of children aged 2–5 years experienced sleep problems, highlighting an important public health issue. These problems are strongly associated with excessive screen time, particularly when screens are used in bed. Limiting preschoolers’ screen use should be a public health priority, supported by parental modelling, the promotion of active play and a screen-free bedtime routine.

Data availability statement

Data are available on reasonable request.