Profiles of 24-h movement behaviors and physical fitness among preschool children: a latent profile analysis

Abstract

Objective

This study aims to identify the 24-h movement behavior patterns of preschool children using Latent Profile Analysis based on Compositional Data Analysis (CoDA), and to examine their associations with physical fitness.

Methods

The study employs a cross-sectional design. A total of 329 healthy children aged 4-6 years were selected. Accelerometers (ActiGraph wGT3-BT, Pensacola, FL, USA) were used to measure light physical activity (LPA), moderate-to-vigorous physical activity (MVPA), and sedentary behavior (SB), while sleep was assessed through parent and teacher questionnaires. The assessment of physical fitness was conducted in accordance with the “Chinese National Physical Fitness Test Standards” (Preschooler Section). To address the multicollinearity problems among components of physical activity (PA), CoDA was first applied, subsequently, Latent Profile Analysis was utilized to categorize 24-h movement behavior patterns, while a Generalized Ordered Logit Model (GOLM) was applied to investigate their associations with physical fitness.

Results

Three distinct behavioral patterns emerged from the analysis: the “brown bear group” (moderate PA and SB, high SP, N = 176, 53.5%), the “cheetah group” (high PA/MVPA, low SB, moderate SP, N = 102, 31%), and the “koala group” (low PA, high SB, lower SP, N = 51, 15.5%). After adjusting for potential confounding factors, it was found that compared with the “koala group”, the “brown bear group” and the “cheetah group” exhibited higher levels of physical fitness, with the probability of improving their physical fitness rating being 3.69 times and 6.36 times that of the “koala group,” respectively.

Conclusion

This study highlights the significant impact of active and healthy activity patterns on the physical fitness of preschool children, providing a foundation for formulating personalized preventive and interventional approaches in early childhood.

List of abbreviations

PA

Physical activity

LPA

Light physical activity

MVPA

Moderate to vigorous physical activity

MPA

Moderate physical activity

VPA

Vigorous physical activity

SB

Sedentary behavior

GOLM

Generalized ordered logit model

BMI

Body Mass Index

SES

Socioeconomic status

CoDA

Compositional Data Analysis

AIC

Akaike information criterion

BIC

Bayesian information criteria

aBIC

adjusted Bayesian information criteria

LMR

Lo-Mendell-Rubin correction likelihood ratio

BLRT

Bootstrap likelihood ratio test

1. Introduction

Physical fitness is defined as a set of physical characteristics that enable individuals to perform daily physical activities, it is a crucial indicator for evaluating both present and future health, it comprises multiple dimensions, including strength, cardiorespiratory function, coordination, and speed/agility.¹ During the preschool years, motor skills and physiological systems rapidly develop. This period is a time of high plasticity, making it a critical stage for the improving physical fitness.2, 3, 4 Recent data from China (2000-2020) indicate a concerning decline in cardiopulmonary function and muscle strength, highlighting a worrisome trend in their overall physical fitness status.⁵ Notably, the prevalence of overweight and obesity is increasing, with forecasts suggesting that by 2030, around 25.1% of Chinese children will have been affected by these issues.⁶ Aside from factors like genetics, insufficient physical activity (PA), prolonged sedentary behavior (SB), and insufficient sleep are critical contributors to the decline in children’s physical fitness.7, 8, 9 Consequently, investigating the relationship between preschool children’s daily movement patterns and physical fitness is of paramount importance.

Previous studies have typically examined the impact of PA, SB, and sleep on physical fitness in isolation.10, 11, 12 However, scholars have pointed out that these behaviors are not independent but are interconnected, collectively forming 24-h movement behaviors that influence health outcomes synergistically.^13,14 This highlights the necessity of adopting a holistic approach in the investigation of the association between 24-h movement behaviors and physical fitness.¹³ For instance, Pedisic emphasized the critical need for balancing the components of 24-h movement behaviors,¹⁵ a fact that laid the foundation for the development of Canada’s 24-h movement behavior guidelines and the integrated movement behavior model.¹⁶ Furthermore, due to the “unit-sum constraint” inherent in 24-h movement behaviors, traditional linear regression analysis may introduce multicollinearity problems.^15,17 To address these limitations, Chastin introduced Compositional Data Analysis (CoDA) into the field of PA research.¹⁸ This approach has since been widely adopted both domestically and internationally to explore the relationship between 24-h movement behaviors and physical fitness.^19,20 Notably, Lemos¹⁹ and Song²¹ have pointed out that moderate-to-vigorous physical activity (MVPA) and SB are key protective and risk factors for physical fitness, they also suggested that enhancing MVPA by converting time spent on SB and light physical activity (LPA) could significantly improve preschool children’s physical fitness.

Despite the aforementioned studies offering new perspectives for understanding 24-h movement behaviors, they predominantly employ a “variable-centered” approach to examine its associations with health outcomes. This perspective often overlooks the diverse nature of daily movement behaviors, resulting in a biased comprehension of children’s health.²² To overcome this limitation, researchers recommend “individual-centered” investigative strategies that seek to more accurately capture the variability and distinction in individual activities and behaviors.^23,24 Latent profile analysis is one such “individual-centered” analytical method, which categorizes individuals with similar behavioral traits into distinct subgroups by identifying individual behavior patterns.²⁵ Compared to the “variable-centered” research paradigm, this approach not only identifies different behavior patterns based on multidimensional behavioral patterns within a 24-h period but also reveals and emphasizes group diversity, facilitating a more nuanced depiction of subgroup traits and uncovering potential disparities among different groups.^24,26,27 This approach provides a theoretical basis for personalized health interventions.

Latent profile analysis has been widely applied to investigate daily movement behavior patterns and health outcomes in children and adolescents.28, 29, 30 Studies by Verswijveren et al. and Alosaimi et al. have demonstrated the existence of three distinct behavioral patterns in this population the highly active profile (high PA, low SB), the sustained sedentary profile (high SB, low PA), and the mixed profile (high or low levels of both PA and SB).^29,31 Among these, the sustained sedentary profile typically faces a higher risk of obesity and poorer cardiorespiratory fitness compared to the other groups.^29,30,32 However, the studies mentioned above have overlooked the sleep factor^29,30 and failed to fully consider the integrative nature of 24-h movement behaviors.³³ Some studies that recognize the holistic nature of 24-h movement behaviors face limitations, such as reliance on subjective reports^32,34,35 and not accounting for the compositional data properties of the 24-h movement behaviors,³⁶ which may lead to classification bias.¹⁸ To improve classification accuracy, it is particularly important to combine objective measurement data when applying latent profile models, and accounting for the compositional characteristics of 24-h movement behaviors. To date, only three studies have explored the association between 24-h movement behavior profiles and obesity indicators by integrating multiple factors.37, 38, 39, 40 However, these studies have focused mainly on adults’ behavior patterns during work and leisure.^37,39 It is important to note that adults’ activity patterns are significantly different from those of preschool children, which may lead to different health outcomes.⁴¹ Studies investigating 24-h movement behavior profiles in preschool children remain scarce, therefore, more samples of preschool children are needed to provide supporting evidence.

To address this research gap, this study combines CoDA and latent profile analysis to investigate,37, 38, 39 the 24-h movement behavior patterns of Chinese preschool children and their association with physical fitness. It aims to provide theoretical guidance and practical insights for developing personalized intervention strategies for preschool children’s physical fitness and implementing public health policies. Specifically, the objectives of this study are as follows: 1) To identify 24-h movement behavior patterns of Chinese preschool children through a combination of CoDA and latent profile analysis; 2) To analyze the relationship between 24-h movement behavior patterns and physical fitness using logistic regression models. Given that variables such as age,⁴² sex,⁴³ BMI,⁴⁴ and Family Socioeconomic Status (SES) are closely related to PA and physical fitness levels in preschool children,⁴⁵ this study includes these variables as covariates to control for their potential interference with the relationship between 24-h movement behavior patterns and physical fitness.

2. Methods

2.1. Participants and procedure

We recruited 380 preschoolers aged 4 to 6 years (M = 4.43, SD = 0.92; 43.2% girls) from 4 kindergartens in central-southern China. Prior to the data collection, informed consent was obtained from the preschoolers’ parents, as a result, 367 preschoolers participated in the study. To ensure data completeness, kindergarten teachers, parents, and research assistants provided supervision and support throughout the process. The inclusion criteria consisted of preschool children free from illness or discomfort, with informed consent provided by their guardians. The exclusion criteria included: preschool children with illness and discomfort; children who did not complete the one-week PA test or physical fitness test; and those with missing data and outliers. In the end, 329 children completed the study protocol and were included in the data analysis (see Fig. 1). During the data collection period, no participants reported any health issues. The Ethics Committee of Hengyang Normal University approved the entire process (Ethics approval number: NO.2021003).

Fig. 1.

Sample recruitment and exclude process diagram.

2.2. Measures

2.2.1. 24-h movement behaviors

We used the ActiGraph wGTX3-BT accelerometer to objectively measure the PA and SB data of preschool children, for which the validity in monitoring children’s activity has been established.⁴⁶ Prior to the data collection, all parents and kindergarten teachers received detailed training on using the accelerometer, including instructions on how to wear the device and important precautions. Research assistants guided the children to wear the accelerometer continuously on the upper right iliac crest for 7 days (i.e., 5 weekdays and 2 weekend days), with the device being removed only during showering, swimming, and nighttime sleep. To ensure data accuracy, parents and teachers were required to record the device’s wear and non-wear time daily, as well as any circumstances that might affect wearing (e.g., discomfort). During data processing, continuous 20 min periods of zero counts were considered invalid non-wear data.⁴⁷ Following the recommendations of Choi et al.,⁴⁸ the criteria required at least 8h of wear per day and at least 3 consecutive days of wear (including at least 2 weekdays and 1 weekend day). Movement behaviors were classified based on the preschool children’s PA intensity standards proposed by Butte et al.⁴⁹ and using a 15-s sampling interval. ActiLife software (version 6.13.3) was used to categorize behaviors into SB (<239 counts/min), LPA (240-2119 counts/min), moderate physical activity (MPA, 2120-4449 counts/min), and vigorous physical activity (VPA, ≥4450 counts/min), with MVPA defined as the sum of MPA and VPA.

Sleep duration was assessed by guardians and kindergarten teachers through questionnaires, in which they recalled the children’s daytime and nighttime sleep durations on weekdays and weekends. This method has been applied and validated in similar populations.⁵⁰

2.2.2. Physical fitness

The physical fitness assessments were conducted following the “National Physical Fitness Testing Standards” (Preschooler Section) set by the General Administration of Sport of China.⁵¹ The assessment includes two components: body morphology and physical fitness. Body morphology indicators include children's height (in meters) and weight (in kilograms). Height was measured using a vertical stadiometer (SHANGHE, SH-E50, China) at the school clinic, and weight data were obtained from the most recent records provided by the respective kindergartens. Body Mass Index (BMI) was subsequently calculated using the following formula: weight/height.² Physical fitness comprises five indicators: agility, coordination, flexibility, muscular strength, and balance. The 10-m shuttle run reflects agility. The standing long jump reflects lower limb muscular strength and explosiveness. The tennis ball throw reflects upper limb and abdominal muscle strength. The continuous two-foot jump reflects coordination and lower limb muscle strength. The sit-and-reach test reflects flexibility, and the walking balance beam test reflects balance ability. The physical fitness assessment includes 8 items, each worth 5 points, for a total of 40 points. The specific scoring standards are shown in (see Table 1). The tests were administered by two researchers, trained through a standardized system, along with the school kindergarten teachers, to ensure the scientific accuracy of the testing process.

Table 1.

Comprehensive rating standards for physical fitness level.

Physical fitness level	Score
Excellent	>31 score
Good	28-31score
Qualified	20-27 score
Not qualified	<20 score

2.2.3. Covariates

Based on previous studies and preliminary analyses,¹² age has been found to be closely related to PA participation and physical fitness levels among preschool children, and physical fitness levels are known to differ by sex.^42,43 Additionally, BMI levels may influence PA participation.⁴⁴ High-income and highly educated families may attach more importance to children’s PA, thus indirectly impacting their children's physical fitness levels.⁴⁵ Given this, we statistically adjusted for the aforementioned variables by including them as covariates. Demographic information (children’s age and sex) was kindergarten teacher-reported, while SES data (parents’ educational level, family income, and occupational status) were parent-reported via a questionnaire. The SES index was computed using principal component analysis and categorized into three levels: low, medium, and high.

2.2.4. Statistical analysis

All statistical analyses were conducted using the “Compositions”¹⁴ packages in R software (version 4.4.3) as well as Mplus 8.3. The statistical significance level was set at 5%.

To ensure the integrity of the data, we cross-integrated and processed the data based on accelerometers, objective data of PA and subjectively reported data of sleep, ensuring that the dataset met the requirements for compositional data analysis. The data were then transformed using the isometric log-ratio (ilr) technique from the Compositions package. Specifically, the time data for SB, LPA, MVPA, and sleep were converted into three ilr coordinates. This transformation preserves the relative information of time-use data, allowing it to be used as an input vector in standard statistical models. It provides an effective solution to overcomes the inherent limitations of the original dataset, thereby enhancing the precision and interpretability of the statistical analysis.

Mplus 8.3 was used to perform latent profile analysis using the ilr coordinates. This model, based on the principles of finite mixture modeling, can identify distinct homogeneous subgroups of individuals based on their composition of PA, SB, and Sleep.⁵² The model incorporated the covariance of the ilr coordinates, as the ilr coordinate system is based on a regular covariance matrix, differing only in rotation. It was assumed that within-group variance and covariance remained consistent across typologies. The model was fitted using the Bayesian Information Criterion (BIC), the adjusted (aBIC) and the Akaike Information Criterion (AIC), with lower values indicating better model fit. The Lo-Mendell-Rubin likelihood ratio test (LMR) was used to compare the fit of different models. If the model fit significantly improved with the addition of a profile (p < 0.05), the k-profile model was considered superior to the k-1-profile model.⁵³ Additionally, entropy was used to assess the determinacy of the classification, with values ranging from 0 to 1. A value closer to 1 indicates higher confidence in the classification. To further validate the model, both statistical significance and sample size were considered, with each profile requiring a sample size of at least 10% of the total sample.⁵⁴ Once the optimal model was determined, the modal assignment method was used to allocate each child to the profile with the highest probability. In the descriptive statistical analysis, data for each comparison metric were weighted by the individual’s posterior probability (confidence). Weighted averages, standard deviations, and weighted proportions were calculated.

Finally, we used the GOLR model to investigate the relationships between physical movement patterns (predictor variables) and physical fitness indicators (outcome variables). Age, sex, BMI, and SES were adjusted for potential confounders.

3. Results

3.1. Participant characteristics

Table 2 presents the characteristics of the participants. A total of 329 preschool children (187 boys and 142 girls) provided accelerometer data. The average age was 4.43 (±0.92) years, and the mean physical fitness score was 25.46 (±5.97). Among the participants, 51 (15.5%) scored in the excellent range (>31 points), 86 (26.1%) in the good range (28—31 points), 140 (42.5%) in the passable range (20—27 points), and 52 (15.8%) in the poor range (<20 points). The mean component values of the 24-h movement behaviors were as follows: sleep 668.8 min, SB 516.7min, LPA 209.3 min, and MVPA 45.2 min (see Table 2).

Table 2.

Descriptive characteristics.

Variable	M/N	SD/%
Age (years)	4.43	0.92
Height (score)	3.83	1.12
Weight (score)	4.28	1.31
BMI (score)	15.76	1.69
FIT (score)	25.46	5.97
24-h movement behaviors
MVPA/min/day	45.2	14.8
LPA/min/day	209.3	29.5
SB/min/day	516.7	41.9
Sleep/min/day	668.8	35.2
Sex
boys	187	56.8%
girls	142	43.2%
SES
Low	104	31.6%
Middle	101	30.7%
Hight	124	37.7%
Physical fitness
Excellent	51	15.5%
Good	86	26.1%
Qualified	140	42.5%
Not qualified	52	15.8%

Note:BMI: Body Mass Index; FIT: Physical Fitness; MVPA: Moderate-to-Vigorous Physical Activity; LPA: Light Physical Activity; SB: Sedentary Behavior; SES: Socioeconomic Status; M: Mean; N=Number of samples; SD: Standard Deviation.

3.2. Latent profile analysis of 24-h movement behaviors in preschool children

Latent profile analysis was conducted using models with 1 to 5 profiles. According to the model selection criteria and statistical tests, three behavior patterns were found to represent the best model. From the perspective of statistical model fit, the three-profile model demonstrated the best balance. When the number of profiles increased from two to three, the information criteria (AIC, BIC, aBIC) decreased significantly, indicating a substantial improvement in the model’s explanatory power. Additionally, the LMR remained significant (P = 0.012), confirming that the three-profile model was superior to the two-profile model. At this point, the entropy value (Entropy = 0.75) was also higher than that of the two-profile model (Entropy = 0.71). In contrast, for the four-profile model, although AIC, BIC, aBIC continued to decrease and entropy slightly increased (0.79), the LMR statistical test was non-significant (P = 0.11). Furthermore, the probability of one of the profiles was below 10% of the total sample size (5.6%), suggesting that adding a fourth profile did not result in a substantial model improvement (see Table 3).

Table 3.

Statistical indicators for models with 1—5 typologies.

profile	Loglikelihood	AIC	BIC	aBIC	Entropy	LMR	BLRT	Proportion (%)
1	260.82	−509.64	−486.72	−505.75
2	343.88	−667.78	−629.58	−661.30	0.71	0.012	<0.001	73.3/26.7
3	390.05	−752.10	−699.62	−743.03	0.75	<0.01	<0.001	53.5/31/15.5
4	405.39	−774.78	−707.02	−763.12	0.79	0.11	<0.001	32.4/52.8/5.6/9.1
5	413.25	−782.51	−698.47	−768.26	0.74	0.41	0.012	39.4/39.1/8.9/5.3/7.2

Note:AIC: Akaike information criteria; BIC: Bayesian information criteria; aBIC: adjusted Bayesian information criteria; LMR: Lo-Mendell-Rubin correction likelihood ratio; BLRT: bootstrap likelihood ratio test.

3.3. Profile characteristics and naming of 24-h movement behaviors of preschool children

Latent profile analysis of 24-h movement behaviors in 329 preschool children identified three distinct groups (see Fig. 2 and Table 4 for details). The “brown bear group” (53.5%) demonstrated moderate levels of PA, with an average of 43.4 min of MVPA, 204.2 min of LPA, 516.1 min of SB, and the longest sleep duration among the three groups (676.3 min). The “cheetah group” (31%) had the highest MVPA (60.8 min), the lowest SB time (484.5min), and a slightly shorter sleep duration (663.6 min), though still within the recommended range. The “koala group” (15.5%) was characterized by the lowest MVPA (25.8 min) and the highest SB time (562.6 min), with no compensatory increase in sleep duration (658.7 min). Significant differences were observed in the daily duration of each movement behavior among the three groups (P < 0.01) (see Table 4).

Fig. 2.

Potential profile of 24-h movement behaviors of preschool children.

Table 4.

Description and statistics of the potential profile of 24-h movement behaviors of preschool children.

Variable	Brown Bear Group		Cheetah Group		Koala Group		F
	N = 176 (53.5%)		N = 102 (31%)		N = 51 (15.5%)
	M	SD	M	SD	M	SD
MVPA/min/day	43.4	6.4	60.8	10.3	25.8	4.7	437.16∗∗
LPA/min/day	204.2	23.8	231.1	21.8	192.9	29.2	33.60∗∗
SB/min/day	516.1	36.3	484.5	32.8	562.6	32.1	71.16∗∗
Sleep/min/day	676.3	35.8	663.6	34.6	658.7	39.7	10.79∗∗

Note:N=Number of samples; M: Mean; SD: Standard Deviation; ∗∗p < 0.001.

3.4. The association between 24-h movement behavior profiles and physical fitness levels in preschool children

We used the generalized ordered logit regression (GOLR) model to investigate the association between 24-h movement behavior profiles and physical fitness levels. After controlling for potential confounding variables. including age, sex, BMI, and SES, the results indicated that the effect of 24-h movement behavior profiles on physical fitness levels was statistically significant (see Table 5) (P < 0.05).

Table 5.

The generalized ordered logical regression model fits statistical indicators.

Medel	Df	Loglikelihood	X2	Df	Pr (>Chisq)	P
Full model	10	−404.88	65.066	7	1.46e-11	<0.01
Null model	3	−437.41

Note:Df: Degree of Freedom; X2: Chi-square: P: P value.

The results indicated that preschool children in the “brown bear group” and the “cheetah group” had significantly higher levels of physical fitness compared to children in the “koala group”. Specifically, the odds of having an elevated physical fitness level were 3.69 times higher for the “brown bear group” (OR = 3.69, 95% CI:1.93-7.15, P < 0.01) and 6.36 times higher for the “cheetah group” (OR = 6.36, 95% CI:3.08-12.36, P < 0.01), when referenced against the “koala group” (see Table 6).

Table 6.

The association between 24-h movement behavior profiles and physical fitness.

Variable	β	SE	t	P	OR	95% CI
Koala Group	Reference	-	-	-	-	-
Bear Group	1.30	0.33	3.92	<0.01∗	3.69	1.93-7.15
Cheetah Group	1.85	0.30	4.95	<0.01∗∗	6.36	3.08-12.36
Age	0.67	0.11	5.97	<0.01∗∗∗	1.96	1.58-2.46
Sex	−0.26	0.21	−1.22	0.22	0.77	0.50-1.17
BMI	0.06	0.06	0.94	0.34	1.06	0.93-1.20
SES2	−0.18	0.27	−0.68	0.49	0.83	0.48-1.41
SES3	0.27	0.27	1.05	0.31	0.75	0.437-1.30

Note: OR: odds ratio; CI: confidence interval; BMI: body mass index; SE: Standard Error; SES: Socioeconomic Status; ∗∗∗P < 0.001,∗∗P < 0.01,∗P < 0.05.

4. Discussion

This study utilized compositional data analysis (CoDA) and latent profile analysis to comprehensively identify, for the first time in preschool-aged children, the diversity of 24-h movement behavior patterns. Although the overall activity behavior of preschool children is moderate, significant heterogeneity was observed in their daily activity behaviors, identifying three distinct patterns: the “brown bear group”, and the “cheetah group”, the “koala group”. This finding suggests that an individual-centered approach is better suited to capture their daily activity behaviors and acknowledge inter-individual differences.^25,27 Furthermore, after controlling for potential confounding factors, the children in the “brown bear group” and “cheetah group” exhibited higher physical fitness levels, with the odds of having improved fitness being 3.69 and 6.36 times greater, respectively, than that of the “koala group”.

4.1. 24-h movement behavior patterns and characteristics of preschool children

This study represents the initial application of Latent Profile Analysis to model 24-h movement behavior patterns in preschool children. Therefore, comparisons are necessarily limited to similar cohorts of children and adolescents. Among preschool children, we identified three distinct activity behavior patterns: the “bear group”, the “cheetah group”, and the “koala group”. The findings indicate that the majority of preschool children (53.5%) belong to the “bear group”, which is characterized by a relatively balanced overall level of 24-h movement components. In particular, the mean time spent on MVPA (43.4 min), LPA (204.2 min), Sleep (676.3 min), and SB (516.1 min). Meanwhile, the so-called “cheetah group” which accounted for 31.0% of all participants, demonstrated the highest average of PA (MVPA: 60.8 min; LPA: 231.1 min) and the lowest average duration of sedentary time (SB: 484.5 min). These findings suggests that the majority of preschool children in this Chinese cohort adhere to a more active lifestyle. This is consistent with previous studies (Verswijveren et al. and Li et al.),^29,30 where majority of children were classified as “active breakers” or “moderately active”. This phenomenon may be closely related to the emphasis placed by families and the educational system on PA in preschool education. For example, some parents may value their children’s PA and be willing to invest significant time and effort in enabling their children to participate in after-school programs and weekend recreational sports, which may encourage preschool children to develop a relatively active lifestyle early on.⁵⁵ Furthermore, disparities in the quality of preschool education and the uneven distribution of educational resources across China result in children facing different opportunities for PA based on their preschool environment.⁵⁵ Therefore, the quality of preschool settings and the priority assigned to PA are likely the primary factors contributing to the observed heterogeneity in daily activity patterns among preschool children.

Although preschool children in the “brown bear group” and “cheetah group” displayed higher levels of activity, the “koala Group”, characterized by sedentary and inactive type (with only 25.8 min of MVPA and up to 562.6 min of SB), still accounted for 15.5%, which is a phenomenon worthy of attention. Deeply influenced by East Asian Confucian culture, society tends to place a high value on academic achievement.⁵⁶ For instance, children transition from early education to the formal schooling, to improve academic performance and rankings, some kindergartens may allocate more time to academic courses, ignoring physical education courses and activities.⁵⁷ At the same time, parents often place too much emphasis on their children’s academic performance, neglecting children’s participation in sports, this educational and cultural background often leads children to invest more time and energy in textbook learning, which may lead to reduced PA and a subsequent increase in SB.^36,58 Thus, personalized interventions for preschool children should be implemented early, by fostering healthy lifestyle habits, encouraging PA, and reducing SB, to improve overall child health and future well-being.

4.2. Association between 24-h movement behavior profiles and physical fitness in preschool children

This study found that preschool children in the “cheetah group” had significantly better physical fitness than those in other groups, this aligns with previous research.^29,30,32,38 Verswijveren et al. and Janda et al. using latent profile analysis, identified 24-h movement behavior patterns in children and adolescents and found that “active breakers” (high MVPA, low SB) play an important role in preventing overweight and obesity.^29,38 In addition, Costa et al.’s study also confirmed that the cardiorespiratory health levels of the “highly active group” were significantly better than those of the “consistently sedentary group”.³² This further validates the key role of regular MVPA, reduced SB, and adequate sleep in promoting children’s health. Higher levels of MVPA contribute to improved cardiovascular function and bone health,⁵⁹ enhance motor skills and coordination, and support overall growth and development in children^60,61 Similarly, low levels of SB help to maintain proper blood circulation and energy balance,⁶² while sufficient sleep contributes to hormonal regulation, appetite control, and energy restoration.⁶³ Additionally, the “cheetah group” participates in MVPA for longer durations compared to other groups, possibly due to a more active lifestyle fostered from an early age, this elevated frequency of physical interaction potentially provides children with diverse social exchange scenarios, which may further enhance preschool children’s social skills, cognitive development, and other psychological aspects,^64,65 thereby promoting emotional development.^66,67 Good emotional and psychological states can improve children’s self-efficacy and confidence, making them more willing to participate in physical activities, creating a positive cycle and thereby improving physical fitness.⁶⁵ However, the significant health advantage of the “cheetah group” is not merely the result of cumulative benefits of individual physiological indicators, but rather in the optimal coordination and balance of 24-h movement behaviors.¹⁵ Therefore, future health interventions should focus on the synergy among all components of 24-h movement behaviors to collectively promote the overall physical health development of preschool children.

In addition, this study found that preschool children in the “brown bear group” (characterized by moderate PA levels and longer sleep duration) were also associated with higher physical fitness levels. This finding is inconsistent with the study by Janda et al.³⁸ which identified the “moderately active group” (with MVPA, LPA, and SB at median levels, and the longest sleep duration) as being at significantly increased risk of childhood obesity. This discrepancy may be attributed to differences in the PA indicators. In Janda et al.’s study, the average MVPA of the “moderately active group” was only 31.7 min (significantly lower than the 43.4 min in the “brown bear group”), suggesting that a 10-min difference in MVPA may be sufficient to reach a threshold for health effects. Research has confirmed that substituting 10 min of SB with MVPA is significantly associated with improvements in metabolic health.⁶⁸ At the same time, the activity patterns of “brown bear group” may also represent a type of regular and health-promoting behavior. Studies have shown that even with a moderate total amount of PA, children who engage in regular and well-distributed PA have better health outcomes than those who engage in high-intensity, irregular PA or prolonged sedentary periods.⁶⁹ In addition, the indirect health benefits associated with adequate sleep duration are also crucial. In the study by Janda et al. sleep lasted only 472.3 min, while the “brown bear group” had 676.3 min, which falls within the recommended range. The study found that maintaining 11–12 h of sleep significantly lowered the risk of obesity in preschool children, compared to sleeping less than 10 h or more than 13 h.⁷⁰ Adequate sleep can optimize the balance of neuroendocrine hormones such as growth hormone and cortisol, thereby directly affecting physical development and immune function, while also indirectly improving children’s physical performance and recovery capacity the following day.⁶³ In this study, the “brown bear group” demonstrated moderate levels of PA and sleep within the 11-12 h healthy range, potentially explaining their higher physical fitness levels. Although this study supports a significant association between the “brown bear group” and higher physical fitness levels in preschool children, these findings should be interpreted cautiously. Due to the cross-sectional design of this study, causal inference remains limited, thus, further longitudinal or interventional studies are needed to clarify these relationships.

This study did not explicitly present the physical fitness status of children in the “koala group”, however, it is clear that the “koala group” characterized by high SB and low SP, is the least healthy group among the three, demonstrating the lowest physical fitness levels, likely due to this group’s relatively low MVPA level (only 25.8 min). suggesting a failure to achieve the associated health benefits of sufficient MVPA. More importantly, prolonged SB significantly reduces key physiological indicators such as cardiovascular fitness and muscular strength.^8,12,19 Furthermore, with the increasing pressure of academic competition, the phenomenon of “involution” has become more prominent, leading to the premature introduction of teaching content in early childhood education, many kindergartens teach primary school knowledge before the expected developmental stage, and various subject and interest classes are continuously emerging. Traditional “mind-body dualism” educational concepts, along with the slogan “winning at the starting line” have long trapped parents in cultural comparisons and anxiety, resulting in the long-term absence of PA in children's education, thereby further compressing PA time.⁷¹ Meanwhile, the current kindergarten physical education curriculum is affected by the issue of “four priorities and four negligence”, characterized by excessive focus on academic learning (cultural knowledge) and neglect of diverse coordinated development. This curriculum’s lack of exploratory and scientific elements.⁷² Subsequently diminishes the enthusiasm for physical activity participation. This deficit of exploratory and scientific modes further restricts children social interactions and participation in physical activities, exacerbating the deterioration in physical fitness.⁷³ Thus, the current educational model—defined by prioritizing intellectual development over physical fitness, focusing on outcomes rather than process, and valuing competition above health—may be a key factor contributing to the decline in physical fitness among preschool children.

4.3. Practical implications

This study reveals that the daily activity behavior patterns of preschool children vary significantly. Consequently, future intervention should accurately identify the behavioral and characteristics of different groups in order to provide personalized developmental guidance and support. Meanwhile, families, kindergartens, and society need to collaborate to promote the overall physical fitness development of preschool children.

4.3.1. Addressing differences: accurately identifying movement behavior patterns and developing personalized health intervention strategies

Brown bear group: Exercise intensity and variety. The PA level of in this group is moderate, so intervention should focus on increasing duration and intensity of PA, adopting multi-combination exercise programs, and encouraging an early transition of the “brown bear group” to the “cheetah group”.⁷⁴ Cheetah group: Maintain a high level of activity and manage it scientifically to avoid overloading. Children in this group exhibit excellent physical activity performance, and exercise interventions should ensure that activities are reasonable and scientific, preventing overexertion and potential exercise injuries. At the same time, they can serve as role models to encourage other children to participate actively in exercise. Koala group: Reduce SB and stimulate interest in healthy exercise. Interventions should focus on decreasing sedentary time and gradually increasing PA. Using digital tools such as smartwatches and activity trackers to monitor activity in real-time,^75,76 combined with organizing group games and parent-child competitions, can awaken children’s motivation to exercise and help them develop good exercise habits.

4.3.2. Collaborative governance: cooperation between home, kindergarten and community to optimize children's physical fitness intervention

Based on the results of this study, health intervention strategies should focus on the distribution of children’s PA throughout the day,⁷⁷ particularly by reducing SB and increasing MVPA for the “koala group.” This requires the collaborative efforts of schools, families, and the community. Studies have shown that multi-component interventions yield superior health benefits compared to single-component ones,^78,79 However, current physical fitness intervention practices face severe systemic challenges, such as parental cognitive bias, insufficient school provision, lack of social equity, and imbalanced coordination in public health service delivery.⁸⁰ Therefore, this study recommends: (1) Strengthening the government's macro-level guidance and regulatory investment in interventions for the physical fitness of preschool children, promoting the formation of healthy activity behavior patterns through public health policies and resource support; (2) Recognizing the importance of parents’ early guidance and implementing comprehensive interventions for preschool children’s physical fitness as early as possible; (3) Strengthening the scientific selection of factors contributing to preschool children’s physical fitness and the development of a system for it; (4) In early childhood education, developing PA frameworks and curriculum interventions based on the developmental characteristics of preschool children as early as possible; (5) Actively using digital technology to promote international exchange and cooperation in research on preschool children’s physical fitness interventions. By promoting the development of a “home-kindergarten-community” collaborative intervention strategy, this approach not only enhances children’s physical fitness but also helps them develop healthy lifestyles, laying the foundation for future comprehensive health and development.

5. Study strengths and limitations

This study’s strength lies in its innovative integration of compositional data analysis with latent profile analysis, which comprehensively captures the characteristics of 24-h movement behavior time combinations and identifies latent profiles that represent preschool children’s movement patterns, demonstrating considerable methodological value. Furthermore, the application of accelerometers to assess PA ensures the accuracy and objectivity of the data.

However, the present study has several limitations. First, a cross-sectional design was employed, which does not allow for a clear determination of the causal association between 24-h movement behaviors and physical fitness outcomes. Therefore, future longitudinal studies are needed. Second, due to the limited language expression and behavioral cognition of preschool children, attempts to subdivide behaviors into categories and collect data through questionnaires or statistics may lead to data distortion because of potential caregiver reporting bias. Consequently, this study did not further differentiate SB into screen time and non-screen time, which limits our ability to perform a thorough analysis of the influence of these activity patterns on children’s health. Future research ought to investigate the impact of these factors on specific health outcomes. Although accelerometers provide relatively objective activity data, nighttime wear may affect children’s sleep quality. Consequently, relying on both objective measurements and subjective reports (i.e., collecting sleep time via parent and teacher reports) may introduce recall and integration biases, thus potentially compromising the accuracy of the reported sleep duration. We recommend that future studies use more accurate objective instruments to measure sleep duration.

6. Conclusions

This study innovatively combines compositional data analysis (CoDA) with latent profile analysis, revealing three distinct 24-h movement behavior patterns in preschool children. It also confirms significant differences in physical fitness levels across these activity behavior patterns, with the “cheetah group” particularly exhibiting higher levels of physical fitness. Based on these findings, promoting physical fitness among preschool children requires increasing MVPA time, minimizing prolonged SB, and maintaining adequate sleep, thereby encouraging the adoption of a lifestyle characterized by more movement and less sedentary time. The findings of this study provide significant insights into the relationship between 24-h movement behavior patterns and physical fitness in preschool children, and contribute theoretical guidance to the development of preschool education practices and public health policies.

Author contributions

Y.B: Software, Methodology, Visualization, Data Curation, Writing-Original Draft. Y. L: Conceptualization, Methodology, Resources, Writing-Review & Editing. Y.Z.Y: Methodology, Data Curation, Writing-Review & Editing; L.F: Methodology, Data Curation, Writing-Review & Editing. L. P: Methodology, Data Curation, Writing-Review & Editing. L.X.M: Investigation, Data Curation, Writing-Review & Editing. C.Y.F: Investigation, Data Curation, Writing-Review & Editing.

Ethics statement

All authors of this study adhered to all ethical standards during the research process, followed ethical principles, ensured the protection of participants’ privacy and declared no conflicts of interest.

Data availability statement

The datasets required for the analysis presented in this study are not accessible to the public. However, scholars who interested may reach out to the first author to seek permission to use the datasets, statistical code, or additional resources utilized in this analysis.

Funding

This study was funded by the Philosophy and Social Science Foundation of Hunan (No. 25YBA243).

Conflicts of interest

All authors of this study state that throughout this entire investigation process, no ethical standards were violated. The study adhered strictly to ethical principles, focused on protecting the privacy of participants and their families, and there were no conflicts of interest involved in the research process.