The association of chronotype, sleep duration and trajectories of health-risk behaviors among college students: a cohort study

doi:10.21203/rs.3.rs-1622136/v1

Download PDF

Research Article

The association of chronotype, sleep duration and trajectories of health-risk behaviors among college students: a cohort study

https://doi.org/10.21203/rs.3.rs-1622136/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Purpose: To describe the trajectories of health-risk behaviors (HRBs) among college students through four consecutive surveys and explore the relationship between chronotype, sleep duration and different trajectories of HRBs.

Methods: We used a data sample of 1,042 college students from the College Student Behavior and Health Cohort Study. Students reported sleep parameters, including chronotype (Morning and Evening Questionnaire-5, MEQ-5) and sleep duration. The behavior scale was used to evaluate four HRBs (smoking, alcohol use, low physical activity, smartphone addiction). The latent class growth analysis (LCGA) was used to estimate the trajectory of self-reported HRBs. Multivariate logistic regression models were used to study whether sleep parameters (chronotype and sleep duration) correlated with HRBs' trajectories.

Results: Four unique trajectories of behaviors were identified: unhealthy group (7.4%), increasing group (21.3%), decreasing group (10.3%) and healthy group (61.0%). Compared with the normal sleep, results from logistic regression analyses indicated that long sleep (>9 h) was associated with the decreasing group and the unhealthy group(p<0.05), while short sleep (<7 h) was associated with the increasing group and the unhealthy group(p<0.05). Compared with the M-types, the E-types were positively correlated with the unhealthy group, the increasing group, and the decreasing group (p<0.05).

Conclusion: E-types and aberrant sleep duration (short and long sleep duration) were significantly associated with the trajectory of HRBs. Findings underscore the need for targeted screening and prevention of modifiable sleep behaviors with the aim of improving HRBs in college students.

Chronotype

Sleep duration

Health-risk behaviors

Latent class growth analys

Lifestyle behaviors are leading contributors to global disease burden. Among the top 20 risk factors are smoking, alcohol, drug use, and poor diet [1]. HRBs, such as smoking, alcohol consumption, and physical inactivity often accumulate during adolescence and are relatively persistent throughout the life course [2]. Unfortunately, many adolescents do not meet the guidelines for specific health behaviors, prevalence rates for unhealthy behaviors among adolescents vary from 5.0–88.5% [3]. HRBs contribute to the leading causes of mobility and mortality among youth and adults, which cause serious public health problems [4]. In addition, HRBs are also environmental risk factors leading to chronic noncommunicable diseases [5].

Research shows that HRBs commonly occur at the same time or cluster together [6–7], rather than being isolated from other HRBs that occur alone. For example, adolescents who smoke regularly have also been found to engage in high alcohol consumption and risky sexual behavior [8]. Using data from the Health Survey for England, the research found that in 2008, the proportion of men with four multiple risk behaviors in the 16–24 age group was 6.7%, and the proportion of women was 5.7% [9]. Some studies try to describe the changes in HRBs with the age of adolescents [10], but most trajectory models focus on describing a single risk behavior. According to the risk accumulation hypothesis, exposure to multiple HRBs may lead to long-term health disadvantages [11], highlighting the need for greater knowledge about developmental trajectories of HRBs.

College life represents an important transition period for young adults characterized by increased social involvement. During this critical period, lifestyles acquired will affect the future life of young college students [12]. Chronotype is used to reflect the preference of the individual's internal circadian rhythm and activity time [13]. Generally, it is divided into three types: moring-types(M-types), neutral types-types(N-types) and evening typrs(E-types) through questionnaires [14]. Sleep health is a multidimensional concept that refers to sleep duration, sleep continuity, and quality or satisfaction [15]. E-types and aberrant sleep duration are common sleep problems among college students [16–18]. Recent studies have discovered that chronotype and sleep duration are associated with adolescent alcohol use, low physical activity, and mobile phone use [19–21]. Unfortunately, most of these studies are cross-sectional. Multiple sleep parameters, including chronotype and sleep duration, are associated with HRBs, suggesting different aspects of disrupted sleep health affect HRBs in adolescents. Despite the increasing number of studies on HRBs, few studies have explored the impact of different sleep dimensions on the developmental trajectories of multiple HRBs, especially among adolescents.

We hypothesize that sleep parameters are associated with multiple trajectories of HRBs. We carried out an epidemiological survey among college students, with the aim of exploring the trajectories of HRBs and examining the prospective associations between chronotype, sleep duration and the trajectories of HRBs.

Participants

Participants were enrolled in the College Student Behavior and Health Cohort Study, which is an ongoing cohort designed to track behaviors, physical and mental health in college students. This research used data from the first four follow-up studies. We needed to fit the trajectories of HRBs, so participants who completed the HRBs questionnaire less than three times were excluded. Finally, 1,042 college students who had both completed the MEQ-5 and behavioral questions over the course of two years.

This research was approved by the ethics committee of Anhui Medical University (No.20170291). All participants obtained written informed consent. Figure 1 show the selection criteria for participants.

Sociodemographic Data

Sociodemographic data were collected by the questionnaire, including gender, age, self-reported family economic status, self-reported academic performance, parents' education level and any siblings.

Chronotype

During Wave 1, the MEQ-5 [22] was used to evaluate the chronotype of college students. The scale is composed of 5 items and has good reliability and validity. Cronbach's α in this study was 0.68. In this research, it was divided into M-types (18 ~ 25 points), N-types (12 ~ 17 points) and E-types (4 ~ 11 points).

Sleep duration

During Wave 1, by asking participants "During the past month, how many hours of actual sleep did you get at night? ". Based on the National Sleep Foundation's updated sleep duration recommendations [23], students (aged 18 to 25 years) who were sleeping at least 7 h/night, responses were categorized into three groups: short sleep (< 7 hours), normal sleep (7–9 hours), and long sleep (> 9 hours).

Health-risk behaviors

During Wave 1 ~ Wave 4, the four kinds of HRBs self-reported by college students were evaluated to understand the cumulative health risks. By asking participants "During the past month, how many days did you smoke cigarettes? ", smoking was dichotomized (frequency of days of more than once as yes) for analysis. "During the past month, how many days have you drunk at least 1 glass of wine? ", drinking was dichotomized (frequency of days of more than once as yes) for analysis. Physical activity was assessed according to the short version of the International Physical Activity Questionnaire (IPAQ) [24]. The frequency and duration of physical activity of college students in the past week that were less than 600 MET were classified as low physical activity. The Self-rating Questionnaire for Adolescent Problematic Mobile Phone Use (SQAPMPU) was used to evaluate college students, smartphone addiction [25]. The questionnaire has 13 items, with good reliability and validity, and Cronbach's α coefficient is 0.87. If the adolescents' total score was ≥ P75 of the participants, it was determined as smartphone addicts.

Statistical Analysis

SPSS 23.0 was used for data processing and analysis, and the inspection level was α = 0.05.

First, to identify distinct subgroups of participants with different longitudinal trajectories, Mplus version 7.4 was used for the LCGA. This analysis is capable of identifying homogeneous subgroups in a larger heterogeneous population. By running the LCGA model continuously, the number of classifications was increased by one, and the optimal number of classifications was obtained. The model fitting criteria included adjusting the Bayesian information criterion (ABIC), Akaike information criterion (AIC), Bayesian information criterion (BIC), Lo-Mendell likelihood ratio test (LMR), and bootstrap likelihood ratio test (BLRT). The closer the ABIC and BIC values are to zero, the better fitting effect of the model is. The best model should have the minimum ABIC AIC and BIC, and then the combined judgment is based on relative entropy, classification probability, and classification interpret ability. Generally, relative entropy > 0.7 means that the model is in the acceptable range, and the proportion of each classification group should be > 5% of the total population.

Second, the χ² test was used to compare the differences between the demographic variables in the trajectories of different HRBs. Ultimately, the multivariate logistics regression model was used to analyze the correlation between the baseline chronotype, sleep duration and the trajectories of HRBs.

General information

The 1,042 college students included in the survey had an average age of 18.81 years, of which 654 were females, accounting for 61.9%. More than average students lived in rural areas, and 24.4% of students self-reported that their family economy was poor. Among all students participating in the survey, 77.1% of students reported that their fathers had received junior high school and above education, and 56.1% of students reported that their mothers had received junior high school and above education. In addition, in a total of 1,135 participants, there were no differences between demographic characteristics (age, gender, family economy, academic performance, parental education level) and follow-up loss data at the time of baseline analysis (Table 1).

Trajectories of health-risk behaviors

Model fit statistics for models with a varying number of classes are presented in Table 2. Based on the model fitting statistics and the interpret ability of the categories, the best-fitting model was classified into four categories because the BIC values and aBIC values of the four categories were both lows (P<0.05), and the classification probability was ≥0.05 at the same time.

Figure 2 shows the trajectories of the four latent classes identified. The first trajectory (represented by the diamond marker) is characterized by relatively healthy behaviors, so it is named the healthy group, which accounts for approximately 61.0%. The second trajectory, which accounts for 10.3%, is named the decreasing group. This trajectory shows a gradual decline in HRBs over time. The third trajectory is named the increasing group because it initially rises significantly and then declines slightly over time. 21.3% of college students belong to this group. The last trajectory (represented by the triangle marker) was described as the unhealthy group because individuals following this trajectory had high levels of all four reports. The proportion of people belonging to the unhealthy group is 7.4%.

Sample characteristics of different health-risk behavior trajectories

Table 3 displays the characteristics and group differences of 1,042 college students. In the baseline sleep duration, the detection rates of short sleep, normal sleep, and long sleep were 35.5%, 57.2%, and 7.3%, respectively. This research also shows that the proportion of M-types (18.7%) was higher than that of E-types (10.1%) among college students. The detection rates of HRBs in different groups are related to gender, father’s education level, self-reported academic performance and family economy. Students with healthy group trajectory are more likely to be female who self-report good academic performance. Those male students whose fathers have a low education level were more likely to be classified into unhealthy group. The difference is statistically significant (P<0.05). Compared to normal sleep, there were fewer cases of the healthy group in short sleep and long sleep (P<0.05). Compared to M-types, there were more cases of the unhealthy group in E-types. The difference is statistically significant (P<0.05).

Association of chronotype , sleep duration and trajectories of health-risk behaviors

Taking healthy group as the reference group,, we used multiple logistic regressions to analyse the relationship between sleep parameters and the trajectory of HRBs. Measures of gender, self-reported academic performance, self-reported family economy and father’s education level were included as covariance due to known associations with HRBs. After controlling those covariance, compared with normal sleep, the short sleep were positively correlated with the unhealthy group and the increasing group, the OR (95% CI) for those were 1.76 (1.01, 3.08), 1.45 (1.05, 1.99), respectively. The long sleep were positively correlated with the unhealthy group, the decreasing group, the OR (95% CI) for those were 5.82 (2.80, 12.09), 3.06 (1.55, 6.06), respectively. Compared with the M-types, the E-types were positively correlated with the unhealthy group the increasing group, the decreasing group, the OR (95% CI) for those were 4.36 (1.72, 11.08), 2.88 (1.57, 5.27), 2.61 (1.28, 5.32), respectively(Figure 3).

To our knowledge, this is the first research to analyze the trajectory of multiple HRBs in college students, and further analyze its relationship between chronotype, sleep duration. Our study found that the trajectories of multiple HRBs over time can be divided into four patterns: healthy group, increasing group, decreasing group and unhealthy group. Aberrant sleep duration, including short sleep duration and long sleep duration were both associated with high-risk trajectories. The results of the study also identified a significant association between E-types and the trajectory of high-risk trajectories of HRBs.

This research is the first to examine the interdependence trajectories of four HRBs:smoking, alcohol use, low physical activity, smartphone addiction. As expected, the vast majority of people belong to the healthy group (61.0% of sample), while the unhealthy group only account for 7.4% of the total population. Similar percentages of adolescents in high-risk groups were found in the Mobile Youth Study, in which approximately 11% of children aged 12 to 18 exhibited increasing high risk in all three areas, including substance use, conduct problems, and sexual risk taking [26]. However, the study by Zhang et al. through the follow up of 1,974 children aged 7 to 9, the fitting results of four life style (screen time, physical activity, sleep duration and beverage intake) trajectories showed that persistent healthy groups accounted for 39.9%, while 17.7% of children were persistent unhealthy groups [27]. The large differences in grouping may be due to the differences in HRBs included in the studies and the fact that the subjects came from different ages, changing one behaviour will affect the prevalence of others [28].

The study found that about 10.1% of college students reported E-types, which was below the 20.5% of students from 8 major Lebanese universities who reported E-types [29]. Compared with M-type, Evening-type was associated with three high-risk HRBs classes. Children are typically M-types, gradually becoming E-types after entering puberty, and teenagers around the age of 20 sleep at the latest [30]. Some previous studies have found that those with later bedtimes had some lifestyle habits that might delay bedtimes, increase depressive symptoms, and lead to health hazards [31]. During adolescence, there is a shift in biologic preference to later sleep timing, which often conflicts with socially prescribed scheduling demands, such as early school start times [32]. It has been speculated that nocturnal time types may affect gene expression, metabolism, immunoendocrine function, and behavior due to the chronic mismatch between internal physiological time and externally imposed time for work and social activities [33]. Previous studies have found that a later phenotypes may be more associated with externalized behaviors, such as risk-taking and substance use [34]. Since these behaviors are more likely to occur at night times with less parental supervision, and the E-types provides greater opportunities for participation in such behaviors [35], it should not be surprising that the association with E-types and relatively high-risk trajectories. Furthermore, evening chronotype may contribute to HRBs through impairments in emotion regulation, executive function, response inhibition, and decision making [36-37].

Of particular note in this research is the association between sleep duration and the trajectory of HRBs. We observed that about 35.5% of students had short sleep, 7.3% of students reported long sleep, which was different from the detection rate of Shandong Adolescent Behavior and Health Cohort [29]. Using data from the Fairfax County Youth Survey, Owens found that students with sleep insufficiency were significantly more likely than respondents without sleep insufficiency to report drug use and delinquency [38]. Based on data from the Project for the Health of Adolescents in Southern Taiwan, the results indicated that both short and long sleep duration were significantly associated with having suicidal and drinking alcohol every week [39]. Previous studies have mainly found the effect of sleep duration on HRBs through cross-sectional studies. However, our research found that not only short but long sleep also affected HRBs trajectories in a cohort study. Regarding the mechanism that leads to the association between sleep duration and HRBs, based on the evidence from a systematic review, suggests that comparing various sleep duration against an “optimal” sleep duration allows non-linear relationships to be observed. U-shaped relationships between sleep duration and risk-taking where both short and long sleep duration are associated with heightened risk-taking [40]. Epidemiological studies have demonstrated that short sleep and sleep disturbance are associated with subsequent impairment in multiple domains of psychosocial and cognitive functioning, delinquency, impulsivity, substance use, and depression [41-42].

Strengths and limitations/implications and sleep research perspectives

These findings contribute to the literature is that we use a latent class growth model to fit the trajectory of the HRBs among college students. The combined analysis of relationship between multiple HRBs rather than single HRBs and chronotype, sleep duration is more in line with the needs of real-world research, which can reflect the dynamic changes and influencing factors of HRBs in adolescence. Second, we used a cohort study to carry out a follow-up investigation and collected data on various HRBs of college students many times, which makes the results more convincing. As far as we know, this is the first research to report the association between E-types, aberrant sleep duration and longitudinal HRBs. The highlight of this research is to explore the influence of circadian rhythm on multiply HRBs, so as to reveal the influence of circadian rhythm on human health.

Despite the above advantages, there is no denying that this research has limitations in the following aspects: although we have adopted a self-reporting scale that has been fully verified and used, reporting bias is still unavoidable. In addition, this research was conducted in two specific geographic regions of China, and the research results may not be applicable to young people who have been extrapolated to various parts of China, which needs to be verified by national representative research in the future.

Our findings indicated that sleep parameters, including both E-types and aberrant sleep duration, were common among college students, and they were independently associated with increased odds of HRBs’ trajectory. Insight into the changing patterns and predictors of multiple HRBs in adolescents may contribute to more successful and targeted prevention.

HRBs, health-risk behaviors; MEQ-5, Morning and Evening Questionnaire-5; LCGA, latent class growth analysis; M-types, Moring-types;N-types, Neutral -types; E-types, Evening-types; IPAQ, the International Physical Activity Questionnaire; SQAPMPU, The Self-rating Questionnaire for Adolescent Problematic Mobile Phone Use; ABIC, Bayesian information criterion; AIC, Akaike information criterion; BIC, Bayesian information criterion; LMR, Lo-Mendell likelihood ratio test; BLRT, bootstrap likelihood ratio test.

Ethics approval and consent to participate

The design and data collection were reviewed and approved by the Ethics Committee of Anhui Medical University, China(No.20170291). All participants wrote informed consent for inclusion prior to the administration of the survey. All methods were carried out in accordance with relevant guidelines and regulations.

Consent for publication

Not applicable.

Availability of data and materials

The datasets that were generated analyzed for the current study are not publicly available as the author does not have permission to share the data. However, it will be available form the corresponding author on reasonable request.

Competing interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

This study was supported by National Natural Science Foundation of China (82073578, 82173542).

Authors' contributions

Chenfang Wang and Fangbiao Tao contributed to the design of the study, conceived the survey and statistical analysis, and wrote the paper. Shuman Tao was involved in the feldwork. Tingting Li and Yang Xie played a role in participant recruitment and data collection. Xiaoyan Wu and Fangbiao Tao proposed the research design. All authors reviewed the manuscript.

Acknowledgments

The authors sincerely thank the consent of data access from College Student Behavior and Health Cohort Study, carried out by the Institute of Anhui Medical University, and all staff andstudents involved in this study for their full support.

Authors' information

¹Department of Maternal, Child and Adolescent Health, School of Public Health, Anhui Medical University, Hefei, China, ²MOE Key Laboratory of Population Health Across Life Cycle, Hefei, China, ³NHC Key Laboratory of Study on Abnormal Gametes and Reproductive Tract, Hefei, China, ⁴Anhui Provincial Key Laboratory of Population Health and Aristogenics, Anhui Medical University, Hefei, China, ⁵The Second Hospital of Anhui Medical University, Hefei, China

Forouzanfar MH, Alexander L, Anderson HR, et al. Global, regional, and national comparative risk assessment of 79 behavioural, environmental and occupational, and metabolic risks or clusters of risks in 188 countries, 1990-2013: a systematic analysis for the Global Burden of Disease Study 2013. Lancet. 2015;386: 2287-2323
Wijbenga L, de Winter AF, Almansa J,et al. Multiple health risk behaviors and mental health from a life course perspective: the Dutch TRAILS study. Prev Med. 2022;154: 106870.
De Winter AF, Visser L, Verhulst FC, et al. Longitudinal patterns and predictors of multiple health risk behaviors among adolescents: the TRAILS study. Prev Med, 2016;84: 76-82.
Noble N, Paul C, Turon H, et al. Which modifiable health risk behaviours are related? A systematic review of the clustering of smoking, nutrition, alcohol and physical activity health risk factors. Prev Med. 2015;81: 16-41.
Prüss-Ustün A, van Deventer E, Mudu P, et al. Environmental risks and non-communicable diseases. BMJ. 2019;364: l265.
Héroux M, Janssen I, Lee DC, et al. Clustering of unhealthy behaviors in the aerobics center longitudinal study. Prev Sci. 2012;13: 183-195.
Poortinga W. The prevalence and clustering of four major lifestyle risk factors in an English adult population. Prev Med. 2015;44:124-128.
Huang DY, Lanza HI, Murphy DA,et al. Development of risk behaviors adolescence: potential pathways to co-occurrence. Int J Behav Dev, 2012;36: 247-257.
David Buck, Francesca Frosini. Clustering of unhealthy behaviours over time implications for policy and practice. Available at: https: //www.makingeverycontactcount.org/media/1045/018. Accessed August 2012.
Hautala D, Sittner K, Walls M. Latent trajectories and profiles of commercial cigarette smoking frequency from adolescence to young adulthood among north American indigenous people. Nicotine Tob Res. 2020;22: 2066-2074.
Kuh D, Ben-Shlomo Y, Lynch J, et al. Life course epidemiology. J Epidemiol Community Health. 2003;57: 778-783.
Ge Y, Xin S, Luan D, et al. Association of physical activity, sedentary time, and sleep duration on the health-related quality of life of college students in Northeast China. Health Qual Life Outcomes. 2019;17: 124.
Montaruli A, Castelli L, Mulè A, et al. Biological rhythm and chronotype: new perspectives in health. Biomolecules, 2021;11: 487.
Park H, Lee HK, Lee K. Chronotype and suicide: the mediating effect of depressive symptoms. Psychiatry Res, 2018;269: 316-320.
Buysse DJ. Sleep health: can we define it? Does it matter? Sleep. 2014;37: 9-17.
Sirtoli R, Balboa-Castillo T, Fernández-Rodríguez R, et al. The association between alcohol-related problems and sleep quality and duration among college students: a multicountry pooled analysis. Int J Ment Health Addict. 2022;27: 1-18.
Becker SP, Jarrett MA, Luebbe AM, et al. Sleep in a large, multi-university sample of college students: sleep problem prevalence, sex differences, and mental health correlates. Sleep Health. 2018;4: 174-181.
Naja F, Hasan H, Khadem SH, et al. Adherence to the mediterranean diet and its association with sleep quality and chronotype among youth: a cross-sectional study. Front Nutr. 2022;8: 805955.
McKnight-Eily LR, Eaton DK, Lowry R, et al. Relationships between hours of sleep and health-risk behaviors in US adolescent students. Prev Med. 2011;53: 271-273.
Haynie DL, Lewin D, Luk JW, et al. Beyond sleep duration: bidirectional associations among chronotype, social jetlag, and drinking behaviors in a longitudinal sample of US high school students. Sleep, 2018;41: zsx202.
Mireku MO, Barker MM, Mutz J, et al. Night-time screen based media device use and adolescents' sleep and health-related quality of life. Environ Int. 2019;124: 66-78.
Adan A, Almirall H. Horne & Östberg morningness-eveningness questionnaire: a reduced scale. Pers Individ Diff, 1991;12: 241a53.
Hirshkowitz M, Whiton K, Albert SM, et al. National Sleep Foundation's updated sleep duration recommendations: final report. Sleep Health. 2015;1: 233-243.
Puciato D, Borysiuk Z, Rozpara M. Quality of life and physical activity in an older working-age population. Clin Interv Aging. 2017;12: 1627-1634.
Tao S, Fu J, Wang H,et al. The development of self-rating questionnaire for adolescent problematic mobile phone use and the psychometric evaluation in undergraduates. Chin J Sch Health. 2013;34: 26-29.
Mustanski B, Byck GR, Dymnicki A, et al. Trajectories of multiple adolescent health risk behaviors in a low-income African American population. Dev Psychopathol, 2013;25: 1155-1169.
Zhang A, Fang J, Wan Y, et al. Joint trajectories of life style indicators and their links to psychopathological outcomes in the adolescence. BMC Psychiatry, 2021;21: 407.
Nauha L, Jurvelin H, Ala-Mursula L, et al. Chronotypes and objectively measured physical activity and sedentary time at midlife. Scand J Med Sci Sports. 2020;30: 1930-1938.
Najem J, Saber M, Aoun C, et al. Prevalence of food addiction and association with stress, sleep quality and chronotype: a cross-sectional survey among university students. Clin Nutr, 2020;39: 533-539.
Randler C, Fassl C, Kalb N. From lark to owl: developmental changes in morningness-eveningness from new-borns to early adulthood. Sci Rep, 2017;7: 45874.
Merikanto I, Lahti T, Puusniekka R, et al. Late bedtimes weaken school performance and predispose adolescents to health hazards. Sleep Med, 2013;14: 1105-1111.
Carskadon MA, Acebo C, Jenni OG. Regulation of adolescent sleep: implications for behavior. Ann N Y Acad Sci, 2004;1021: 276-291.
Roenneberg T, Merrow M. The circadian clock and human health. Curr Biol. 2016;26: 432-443.
Karan M, Bai S, Almeida DM, et al. Sleep-wake timings in adolescence: chronotype development and associations with adjustment. J Youth Adolesc, 2021;50: 628-640.
Randler C, Bilger S. Associations among sleep, chronotype, parental monitoring, and pubertal development among German adolescents. J Psychol. 2009;143: 509-520.
Pieters S, Burk WJ, Van der Vorst H, et al. Prospective relationships between sleep problems and substance use, internalizing and externalizing problems. J Youth Adolesc. 2015;44: 379-388.
Wong MM, Robertson GC, Dyson RB. Prospective relationship between poor sleep and substance-related problems in a national sample of adolescents. Alcohol Clin Exp Res. 2015;39: 355-362.
Owens J, Wang G, Lewin D, et al. Association between short sleep duration and risk behavior factors in middle school students. Sleep. 2017;40: zsw004
Yen CF, King BH, Tang TC. The association between short and long nocturnal sleep durations and risky behaviours and the moderating factors in Taiwanese adolescents. Psychiatry Res. 2010;179: 69-74.
Short MA, Weber N. Sleep duration and risk-taking in adolescents: a systematic review and meta-analysis. Sleep Med Rev. 2018;41:185-196.
Kenney SR, Paves AP, Grimaldi EM, et al. Sleep quality and alcohol risk in college students: examining the moderating effects of drinking motives. J Am Coll Health. 2014;62: 301-308.
Liu X, Liu ZZ, Liu BP, et al. Associations between sleep problems and ADHD symptoms among adolescents: findings from the Shandong Adolescent Behavior and Health Cohort (SABHC). Sleep. 2020;43: zsz294.

Table 1 The difference of demographic characteristics between analysis data and missing data.
Characteristic	Included (n=1042)	Excluded (n=93)	χ²/t	P-value
Age	18.81±1.18	18.60±0.80	1.52	0.09
Gender			0.01	0.93
Male	397（38.1）	35（37.6）
Female	645（61.9）	58（62.4）
Living place
Rural	583（56.0）	50（53.8）	0.17	0.68
Urban	459（44.0）	43（46.2）
Any siblings
Yes	242（23.2）	26（28.0）	1.60	0.30
No	800（76.8）	67（72.0）
Self-reported family economy
Low	254（24.4）	19（19.4）	5.78	0.06
Medium	736（70.6）	64（68.8）
high	52（5.0）	10（11.8）
Self-reported academic performance			2.60	0.27
Low	220（21.1）	19（20.4）
Medium	644（61.8）	52（55.9）
high	178（17.1）	22（23.7）
Father^’s education level			3.79	0.15
Primary school and below	239（22.9）	18（19.4）
Middle school	500（48.0）	39（41.9）
Senior high school and above	303（29.1）	36（38.7）
Mother^’s education level			0.77	0.68
Primary school and below	457（43.9）	40（43.0）
Middle school	366（35.1）	30（32.3）
Senior high school and above	219（21.0）	23（24.7）
^aP-value < 0.05. ^bP-value < 0.001.

Table 2 Fit indices for latent class growth analysis of health-risk behaviors.
Fit Indices
AIC	BIC		ABIC
10524.35		10558.99		10536.76
9631.14		9680.63		9648.87
9490.45		9554.78		9513.50
9367.77		9446.95		9396.13
9373.77		9467.79		9407.45
Abbreviations: AIC is Akaike’s information criterion, BIC is Bayesian information criterion, ABIC is adjustde the Bayesian information criterion, LMR is Lo-Mendell likelihood ratio test, BLRT is bootstrap likelihood ratio test.

Table 3 Participant characteristics of health-risk behaviors trajectory groups
Characteristic	N	Health-risk behaviors trajectory groups				χ² value
Characteristic	N	Healthy group （n=636）	Decreaing group （n=107）	Increasing group （n=222）	Unhealthy group （n=77）	χ² value
Gender						101.22^b
Male	397（38.1）	189（47.6）	65（16.4）	81（20.4）	62（15.6）
Female	645（61.9）	447（69.3）	42（6.5）	141（21.9）	15（2.3）
Living place						2.36
Rural	583（56.0）	351（60.2）	62（10.6）	121（20.8）	49（8.4）
Urban	459（44.0）	285（62.1）	45（9.8）	101（22.0）	28（6.1）
Any siblings						2.02
Yes	242（23.2）	141（58.2）	29（12.0）	51（21.1）	21（8.7）
No	800（76.8）	495（61.9）	78（9.8）	171（21.3）	56（7.0）
Self-reported academic performance						26.60^b
Low	220（21.1）	115（52.3）	34（15.5）	48（21.8）	23（10.4）
Medium	644（61.8）	389（60.4）	64（9.9）	143（22.2）	48（7.5）
high	178（17.1）	132（74.2）	9（5.0）	31（17.4）	6（3.4）
Self-reported family economy						12.68^a
Low	254（24.4）	149（58.7）	29（11.4）	50（19.7）	26（10.2）
Medium	736（70.6）	451（61.3）	77（10.5）	163（22.1）	45（6.1）
high	52（5.0）	36（69.2）	1（1.9）	9（17.3）	6（11.6）
Father^’s education level						13.74^a
Primary school and below	239（22.9）	131（54.8）	25（10.5）	55（23.0）	28（11.7）
Middle school	500（48.0）	311（62.2）	57（11.4）	106（21.2）	26（5.2）
Senior high school and above 303（29.1）		194（64.0）	25（8.3）	61（20.1）	23（7.6）
Mother^’s education level						8.49
Primary school and below	457（43.9）	265（58.0）	47（10.3）	112（24.5）	33（7.2）
Middle school	366（35.1）	231（63.1）	34（9.3）	76（20.8）	25（6.8）
Senior high school and above 219（21.0）		140（63.9）	26（11.9）	34（15.5）	19（8.7）
Chronotype						32.89^b
M-types	195（18.7）	121（62.1）	27（13.8）	35（17.9）	12（6.2）
N-types	742（71.2）	475（64.0）	60（8.1）	154（20.8）	53（7.1）
E-types	105（10.1）	40（38.1）	20（19.1）	33（31.4）	12（11.4）
Sleep duration						54.99^b
Short sleep	370（35.5）	211（57.0）	36（9.7）	94（25.5）	29（7.8）
Normal sleep	596（57.2）	391（65.6）	55（9.2）	120（20.1）	30（5.1）
Long sleep	76（7.3）	34（44.7）	16（21.1）	8（10.5）	18（23.7）
^aP-value < 0.05. ^bP-value < 0.001. Abbreviations: M-types is the morning types, N-types is neutral types, E-types is evening types.

No competing interests reported.

supp.docx

Download PDF

Version 1

posted

You are reading this latest preprint version

The association of chronotype, sleep duration and trajectories of health-risk behaviors among college students: a cohort study

Status:

Version 1

Abstract

Figures

Introduction

Materials And Methods

Participants

Sociodemographic Data

Chronotype

Sleep duration

Health-risk behaviors

Statistical Analysis

Results

Discussion

Conclusion

Abbreviations

Declarations

References

Tables

Additional Declarations

Supplementary Files

Status:

Version 1