DOI: https://doi.org/10.21203/rs.3.rs-1887144/v1
Aim: This study aimed to determine the prevalence and associated factors of co-occurring health risk behaviors among Bangladeshi adolescents.
Subject and Methods: A cross-sectional study involving 637 adolescent students in three high schools and three colleges in Dhaka city, Bangladesh, was conducted from November 2019-February 2020. Participants were selected using a multistage sampling technique. A self-administered questionnaire was used to collect data and statistical package for social sciences, version 22.0 was used to perform data analysis.
Results: The participants comprised of 57.1% male, and their mean age was 15.65 ± 1.58 years. Among the participants, 2.4 % (n=15) reported one health risk behavior, while 0.3 % (n = 2) reported no risk behaviors. The prevalence of co-occurring health risk behaviors among the participants was 26.4%. Logistic regression analysis showed that co-occurrence of health risk behaviors was associated with participants aged with 17-18 years (OR = 3.11; 95% CI = 1.95-4.95, p < .001), college students (OR = 2.74, 95% CI = 1.90-3.94, p < .001), participants having secondary level of fathers’ education (OR=2.63; 95% CI=1.68-4.09, p < .001), participants having secondary level of mothers’ education (OR = 1.84; 95% CI = 1.12-3.02, p = .015), living apart form family (OR = 2.22; 95% CI = 1.37-3.59, p = .001), poor academic performance (OR=2.06; 95% CI = 1.26-3.36, p = .004), and reporting truancy (OR = 1.81; 95% CI = 1.20-2.72, p = .005).
Conclusion: The findings emphasize the need for effective interventions to prevent risk behavior practices among adolescents.
During the last few decades, health risk behaviors have emerged as major determinants of premature and preventable mortality and morbidity among adolescents (Ford et al., 2011; Selivanova & Cramm, 2014). A health risk behavior is characterized as an activity conducted by an individual with a frequency or intensity that raises the risk of disease or injury, whether or not the individual is conscious of the harm (Steptoe, 2000). These behaviors include poor dietary intakes, physical inactivity and sedentary behavior, substance abuse (smoking, alcohol, and drugs), unsafe sexual activities, and reckless driving (Selivanova & Cramm, 2014). Several of them, such as reckless driving, unsafe sexual activity, and tobacco or alcohol use, can contribute to leading causes of death and morbidity (McPherson et al., 2013; Spengler et al., 2014). In addition, poor dietary intakes, smoking, alcohol use, physical inactivity, and sedentary lifestyles result in non-communicable disease outcomes such as cardiovascular disease, type 2 diabetes, and other illnesses (Ezzati & Riboli, 2013; Mewton et al., 2019).
Adolescents are generally considered healthy. According to the World Health Organization (WHO), about 1.1 million adolescents aged 10–19 died in 2016 and more than half of these deaths occurred in low- and middle-income African and Southeast Asian countries, mostly due to preventable risk behaviors (WHO, 2018a). Evidence suggest that risk behaviors appear to be clustered among adolescents (Sychareun et al., 2011). Risk behaviors such as poor dietary consumptions, low level of physical activity, sedentary behaviors, smoking, and alcohol consumption are commonly co-occurring among adolescents (Kipping et al., 2012; Hale & Viner, 2016; Leão et al., 2017).
Adolescents between the ages of 10 and 19 constitute 22% of the overall population in Bangladesh (Worldometer, 2020). Although adolescents are usually considered to be healthy, during this period morbidity and mortality may occur, often due to health risk behaviors. Bangladesh has a dearth of data on adolescent risk behaviors. Moreover, previous research almost exclusively focused on single risk behaviors, like physical inactivity (Khan et al., 2017), sedentary behaviors (Khan & Burton, 2016), tobacco use (Islam et al., 2016). However, adolescents demonstrate concurrent numerous risk behaviors that are inadequately understood, particularly from a Bangladeshi viewpoint. Therefore, the present study aimed to investigate the prevalence and associated factors of co-occurring health risk behaviors among Bangladeshi adolescents.
A cross-sectional study was carried out with adolescent students between November 2019 and February 2020, using a self-administered questionnaire. It was conducted in Dhaka, the capital city of Bangladesh, and also has the largest concentration of educational institutions. Inclusion criteria were (a) being a student of 8th to 12th grade of the chosen educational units; (b) having adolescent age range; (c) having willingness to participate in the study; and (d) being present in the classroom on data collection day.
A multi-stage sampling technique was employed to draw the sample from selected high schools and colleges located in Dhaka city. Initially, three high schools and three colleges were conveniently selected. Each grade level in the selected schools and colleges is subdivided into several sections based on the school/college standards. Then, two sections were selected randomly for each grade level from those available in high schools and colleges. A total of 32 sections was selected. All students enrolled in the selected sections were welcome to participate. The number of students in each section varies by school and college. The survey questionnaire was distributed to 713 students, 681 of whom responded, resulting in a 95.5 percent response rate. Incomplete surveys (n = 44) were removed, remaining data on 637 participants for the final analysis.
Participants were asked to report their age, sex, academic level, father’s education level, father’s occupation, mother’s education level, mother’s occupation, monthly family income, family size, living with whom and religious status. The study divided participants into three age groups: 13–14, 15–16, and 17–18. Monthly household income was divided into four categories: <30,000 Bangladeshi Taka (BDT), 50,000-100,000 BDT, and 100,000 BDT or above. Family size was divided into three categories: 1–4, 5–8, and > 8.
Health-related behaviors
The present study investigated dietary behaviors (breakfast, vegetables, and fruits intakes), physical activity, screen-based sedentary behavior (screen time), sleep risk behavior, and substance use (i.e., smoking, marijuana, alcohol). Health-related behaviors (excluding substance use) were recorded for 7 days to reduce memory recall bias.
Dietary behaviors
Dietary behaviors were assessed in three domains: breakfast intake, vegetables intake, and fruits intake. Breakfast intake was assessed by asking the question, “During the past 7 days, did you take your breakfast daily?” with responses: 1= yes and 2= no. Skipping breakfast was defined missing breakfast at least one day before seven days of the survey. Vegetable intake was assessed by the question, “During the past 7 days, how many times per day did you usually eat vegetables?” and fruits intake by the question, “During the past 7 days, how many times per day did you usually eat fruits?”. Inadequate vegetables intake was defined as having <3 times/day and inadequate fruits intake was defined as having <2 times/day (El-ammari et al., 2019; Ziaei et al., 2019).
Physical activity and Screen-based sedentary behavior
The present study measured leisure time physical activity by asking, “During the past 7 days, on how many days were you physically active (moderate to vigorous-intensity physical activity) for a total of at least 60 minutes per day?” and allowed response options: 1 = 0 days to 8 = 7 days. Insufficient physical activity was defined as “less than the WHO recommended of 60 minutes of moderate to vigorous physical activity per day.” (WHO, 2018d). Screen-based sedentary behavior was defined as spending 2 or more hours of screen-related activities each day while not in school/college or doing homework (Al-haifi et al., 2016).
Sleep risk behavior
The participants were asked to report their sleep and waking times for the last seven days, divided into weekdays and weekends. Sleep duration was estimated by subtracting sleep onset from wakeup (Min et al., 2018). Combining weekday and weekend sleep durations with weights of 6/7 and 1/7 yielded the mean daily sleep duration. Sleep risk was defined as less than 7 hours per day or more than 11 hours per day on average (Mewton et al., 2019).
Substance use
Participants were asked “During the past 7 days, how many cigarettes did you smoke per day?” Currently smoking was defined smoking ≥1 cigarettes per day. Participants were asked if they had ever used marijuana or alcohol. Both questions had the response options: 1 = yes, 2 = no.
Academic performance and truancy
Academic performance was evaluated by the last exam results and classified as follows: good (GPA 4.50-5), medium (GPA 3.50-4.50), and bad (GPA 3.5). Truancy was measured by ‘In the previous 30 days, how many days did you miss school/college without permission (from parents or teachers)?’ and defied as having 3 or more days of unexcused absences from school/college.
Co-occurring health risk behaviors
In the present study, co-occurring health risk behaviors were defined as five or more risk behaviors.
Data analysis
Data were analyzed using the Statistical Package for the Social Sciences (SPSS) software for windows, version 22.0 (IBM SPSS Statistics, New York, United States). Data analysis involved descriptive statistics as well as inferential statistics. Descriptive analysis like frequency, percentage, mean and standard deviation was carried out. The information was presented using frequency tables and figures. Chi-square test (χ2) was used to compare the differences between groups. The logistic regression analysis was performed to determine the factors predicting co-occurring health risk behaviors and reported as an odds ratio (OR) with 95% confidence interval (CI), and the directions of relationship among all examined health risk behaviors were investigated applying Pearson correlation analyses. The association of variables was considered statistically significant if the p-value was less than or equal to 0.05.
The mean (± SD) age of the total participants was 15.65 ± 1.58 years with the range of 13–18 years. Of the total, 57.1% (n = 364) were males and 42.9% (n = 273) were females. Regarding academic level of participants, a majority i.e., 57.5% (n = 366) were high school students (8–10 grades), while 42.5% (n = 272) were college students (11–12 grades). The details of the descriptive statistics were shown in Table 1.
| Characteristics | Total (%) | Co-occurrence of HRBs | X2 | p-value | |
|---|---|---|---|---|---|
| No (%) | Yes (%) | ||||
| Gender | |||||
| Male | 364 (57.1) | 261 (71.7) | 103 (28.3) | 1.618 | 0.203 |
| Female | 273 (42.9) | 208 (76.2) | 65 (23.8) | ||
| Academic level | |||||
| High school | 366 (57.5) | 300 (82.0) | 66 (18.0) | 30.822 | < 0.001 |
| College | 271 (42.5) | 169 (62.4) | 102 (37.6) | ||
| Father’s education level | |||||
| Primary | 66 (10.4) | 46 (69.7) | 20 (30.3) | 19.380 | < 0.001 |
| Secondary | 143 (22.4) | 87 (60.8) | 56 (39.2) | ||
| Higher secondary | 138 (21.7) | 103 (74.6) | 35 (25.4) | ||
| Graduate or higher | 290 (45.5) | 233 (80.3) | 57 (19.7) | ||
| Father’s occupation | |||||
| Farmer | 39 (6.1) | 25 (64.1) | 14 (35.9) | 5.564 | 0.135 |
| Govt./private employee | 276 (43.3) | 215 (77.9) | 61 (22.1) | ||
| Business | 300 (47.1) | 214 (71.3) | 86 (28.7) | ||
| No work | 22 (3.5) | 15 (68.2) | 7 (31.8) | ||
| Mother’s education level | |||||
| Primary | 115 (18.1) | 83 (72.2) | 32 (27.8) | 6.159 | 0.104 |
| Secondary | 211 (33.1) | 147 (69.7) | 64 (30.3) | ||
| Higher secondary | 154 (24.2) | 112 (72.7) | 42 (27.3) | ||
| Graduate or higher | 157 (24.6) | 127 (80.9) | 30 (19.1) | ||
| Mother’s occupation | |||||
| Housewife | 543 (85.2) | 395 (72.7) | 148 (27.3) | 1.475 | 0.225 |
| Employed | 94 (14.8) | 74 (78.7) | 20 (21.3) | ||
| Monthly family income (BDT) | |||||
| < 30,000 | 252 (39.6) | 186 (73.8) | 66 (26.2) | 3.285 | 0.350 |
| 30,000 - <50,000 | 117 (18.4) | 88 (75.2) | 29 (24.8) | ||
| 50,000 - <100,000 | 161 (25.3) | 111 (68.9) | 50 (31.1) | ||
| ≥ 100,000 | 107 (16.8) | 84 (78.5) | 23 (21.5) | ||
| Living with whom | |||||
| With family | 555 (87.1) | 421 (75.9) | 134 (24.1) | 11.036 | 0.001 |
| Apart from family | 82 (12.9) | 48 (58.5) | 34 (41.5) | ||
| Academic performance | |||||
| Poor | 216 (3.9) | 146 (67.6) | 70 (32.4) | 8.686 | 0.013 |
| Medium | 262 (41.1) | 194 (74.0) | 68 (26.0) | ||
| Good | 159 (25) | 129 (81.1) | 30 (18.9) | ||
| Truancy | |||||
| No | 540 (79.1) | 384 (76.2) | 120 (23.8) | 8.173 | 0.004 |
| Yes | 133 (20.9) | 85 (63.9) | 48 (36.1) | ||
| Note: Co-occurring health risk behaviors defined as five or more risk behaviors | |||||
Prevalence of health risk behaviors
As regards specific health risk behaviors, Fig. 1 shows the evolution of percentages for each risk behavior. Nearly half of the participants (n = 47.3%) reported to skip their breakfast. Inadequate vegetables and fruits intake were reported by 67.5% (n = 430) and 85.7% (n = 546) participants, respectively. Around 65% (n = 415) of participants had insufficient physical activity level. Sleep risk behavior (i.e., sleep duration < 7 hours or > 11 hours/day) was reported by 26.4% (n = 168) participants and screen-time exposure > 2 hours/day was reported by 41.4% (n = 264). Around 10% (n = 63) reported currently smoking. Some of participants, 6.9% and 12.9%, respectively had ever used marijuana and alcohol.
Prevalence of co-occurring health risk behaviors
The prevalence of co-occurring health risk behaviors investigated in this study is shown in Fig. 2. Among the participants, 0.3% (n = 2) did not have any health risk behaviors, while 2.4% (n = 15) had only one. About 12% (n = 74) of participants reported two risk behaviors, and 27.3% (n = 174) reported three risk behaviors occurring at a time. More than 26% had five or more co-occurring health risk behaviors.
Correlation between health risk behaviors
Table 2 shows the degree of association among the health risks behaviors investigated in this study. The skipping breakfast was significantly and positively associated with inadequate vegetable intake (r = .120, p < .01) and life-time alcohol use (r = .106, p < .01), but negatively associated with > 2 hours/day screen time (r=-.123, p < .01). Inadequate vegetable intake was significantly and positively associated with inadequate fruits intake (r = .129, p < .01). Inadequate fruits intake habit was significantly and negatively associated with currently smoking (r=-.105, p < .01) and life-time alcohol use (r=-.098, p < .05). >2 hours/day screen time had a negative correlation with currently smoking (r=-.159, p < 0.01), life-time alcohol use (r=-.162, p < .01) and life-time marijuana use (r=-.162, p < .01). Sleep risk was positively associated with life-time marijuana use (r = .090, p < .05). Currently smoking was significantly and positively associated with life-time alcohol use (r = .454, p < .01) and life-time marijuana use (.366, p < 0.01). The life-time alcohol use was significantly and positively associated with life-time marijuana use (r = .450, p < .01).
| Health risk behaviors | (1) | (2) | (3) | (4) | (5) | (6) | (7) | (8) | (9) |
|---|---|---|---|---|---|---|---|---|---|
| (1) Skipping breakfast | 1 | ||||||||
| (2) Inadequate vegetables intake | .120** | 1 | |||||||
| (3) Inadequate fruits intake | .000 | .129** | 1 | ||||||
| (4) Insufficient PA | .046 | .069 | − .044 | 1 | |||||
| (5) > 2 hours/day screen time | − .123** | − .060 | − .016 | − .074 | 1 | ||||
| (6) Sleep risk | .054 | − .034 | .051 | − .063 | − .024 | 1 | |||
| (7) Currently smoking | .076 | .017 | − .105** | .077 | − .159** | .052 | 1 | ||
| (8) Life-time alcohol use | .106** | .036 | − .098* | − .014 | − .162** | − .007 | .454** | 1 | . |
| (9) Life-time marijuana use | .027 | .070 | − .013 | − .035 | − .135** | .090* | .366** | .450** | 1 |
| ** Correlation is significant at the 0.01 level (2-tailed). | |||||||||
| *Correlation is significant at the 0.05 level (2-tailed). | |||||||||
Associations with the co-occurring health risk behaviors
The proportion of co-occurring of health risk behaviors was significantly higher among (ⅰ) participants from college vs. high school (37.6% vs. 18.0%, p<.001), (ⅱ) participants having fathers’ education level primary vs. graduate or higher (30.3% vs. 19.7%, p<.001), (ⅲ) participants living without vs. with family (41.5% vs. 24.1%, p=.001), (ⅳ) participants those academic performance poor vs. good (32.4% vs. 18.9%, p=.013), and (v) participants reporting truancy vs. no truancy (36.1% vs. 23.8%, p=.004). (Table 1)
Predictors of co-occurring health risk behaviors
Table 3 shows the results of an ordinal logistic regression analysis used to determine the factors that predict co-occurring health risk behaviors among the participants. College students were found to be 2.74 times more likely than high school students to have co-occurring health risk behaviors (OR = 2.743, 95% CI = 1.909–3.942, p < 0.001). Additional factors that significantly influence the co-occurring health risk behaviors includes: having secondary level of fathers’ education (OR = 2.63; 95% CI = 1.689–4.099, p < 0.001), having secondary level of mothers’ education (OR = 1.843; 95% CI = 1.124–3.022, p = 0.015), living apart form family (OR = 2.225; 95% CI = 1.376–3.598, p = 0.001), poor academic performance (OR = 2.06; 95% CI = 1.264–3.362, p = 0.004), and reporting truancy (OR = 1.81; 95% CI = 1.201–2.720, p = 0.005).
| Characteristics | B | S.E | OR | 95% CI | p-value | |
|---|---|---|---|---|---|---|
| Sex | ||||||
| Male | 0.233 | 0.184 | 1.26 | (0.88–1.81) | 0.204 | |
| Female | Ref. | |||||
| Academic level | ||||||
| High school | Ref. | |||||
| College | 1.009 | 0.185 | 2.74 | (1.90–3.94) | < 0.001 | |
| Father’s education level | ||||||
| Primary | 0.575 | 0.306 | 1.77 | (0.97–3.23) | 0.060 | |
| Secondary | 0.967 | 0.226 | 2.63 | (1.68–4.09) | < 0.001 | |
| Higher secondary | 0.329 | 0.245 | 1.38 | (0.85–2.24) | 0.180 | |
| Graduate or higher | Ref. | |||||
| Father’s occupation | ||||||
| Farmer | 0.182 | 0.567 | 1.20 | (0.39–3.64) | 0.748 | |
| Government/private | -0.498 | 0.480 | 0.60 | (0.23–1.55) | 0.300 | |
| Business | -0.149 | 0.475 | 0.86 | (0.33–2.18) | 0.753 | |
| No work | Ref. | |||||
| Mother’s education level | ||||||
| Primary | 0.490 | 0.291 | 1.63 | (0.92–2.88) | 0.092 | |
| Secondary | 0.611 | 0.252 | 1.84 | (1.12–3.02) | 0.015 | |
| Higher secondary | 0.462 | 0.272 | 1.58 | (0.93–2.70) | 0.089 | |
| Graduate or higher | Ref. | |||||
| Mother’s occupation | ||||||
| Housewife | 0.327 | 0.270 | 1.38 | (0.81–2.35) | 0.226 | |
| Employed | Ref. | |||||
| Monthly family income (BDT) | ||||||
| < 30,000 | 0.259 | 0.276 | 1.29 | (0.75–2.22) | 0.347 | |
| 300,00 - <50,000 | 0.185 | 0.318 | 1.20 | (0.64–2.24) | 0.560 | |
| 50,000 - <1,00,000 | 0.498 | 0.291 | 1.64 | (0.93–2.90) | 0.087 | |
| ≥ 1,00,000 | Ref. | |||||
| Living with whom | ||||||
| With family | Ref. | |||||
| Apart from family | 0.800 | 0.245 | 2.22 | (1.37–3.59) | 0.001 | |
| Academic performance | ||||||
| Poor | 0.724 | 0.249 | 2.06 | (1.26–3.36) | 0.004 | |
| Medium | 0.410 | 0.247 | 1.50 | (0.92–2.44) | 0.097 | |
| Good | Ref. | |||||
| Truancy | ||||||
| No | Ref. | |||||
| Yes | 0.592 | 0.209 | 1.80 | (1.20–2.72) | 0.005 | |
| Note. B = Regression coefficient; S.E. =Standard Error; OR = Odds Ratio (OR); CI = Confidence Interval. | ||||||
Adolescence is a transition period and during this period, people may engage in behaviors that are harmful to their health. In this study, 0.3% of study participants reported no health risk behavior. The study further showed that nearly 26% of participants reported five or more health risk behaviors occurring at a time. According to a study of Brazilian high school students (14–19 years old), about 10% of participants did not engage in any health risk behaviors, while 58.5% reported they were exposed to at least two risk behaviors at the same time (Brito, Hardman & De Barros, 2015). The Australian Child and Adolescent Survey on Mental Health and Wellbeing reported that about 32.0% of adolescents (13–17 years old) in Australia had one or more of behaviors (Mewton et al., 2019). According to studies, adolescents engage in health-risk behaviors as a result of peer pressure and curiosity (Mewton et al., 2019; Spengler et al., 2014). The findings of this study suggest to take appropriate steps to promote healthy behaviors among this vulnerable group. Parental monitoring is effective in reducing the risk behavior practices among adolescents (Yu et al., 2006). A number of studies have showed that parental monitoring can help reduce risk behaviors practices among adolescents (Hindelang et al., 2001; Kipping et al., 2012; Yu et al., 2006).
The findings of this study revealed that college students were more likely than high school students to have co-occurring health risk behaviors. College students are more aged than high school students and previous research has found that as individuals age, they are more likely to engage in risky behaviors (Leão et al., 2017; Hale & Viner, 2016). Among the reasons for older adolescents were more likely to engage multiple health risk behaviors, it can highlight curiosity, freedom, being close to peers and friends, participating in festive celebrations and going to friends’ houses (de Moura et al., 2018; Garrido, Weiler & Taussig, 2018).
This study further revealed significant association between father’s education level and participant’s co-occurring of health risk behaviors. It is expected that educated parents would be more conscious of their sons/ daughters and thus will track them so that they would not be able to take health risk behaviors. Educated parents can also encourage them to develop healthier behaviors. In this study, participants who lived apart from their families were more likely than those who lived with their families to have co-occurring health risk behaviors. It could be due to a lack of parental supervision.
Engagement in health risk behaviors results in poor academic outcomes. Adolescents who participate in less risk behaviors achieve more success later in life, and those who practice healthy behaviors earlier in life can continue to do so throughout their lives (Bradley & Greene, 2013). Several studies have found association between academic performance (letter grades, exam scores, or other indicators of student achievement) and health risk behaviors (Busch et al., 2014; Rasberry et al., 2017; Wald et al., 2014). The findings of this study revealed a significant link between academic performance and co-occurring health risk behavior, as well as a link between truancy and co-occurring health risk behavior, implying that teachers and public health educators will benefit from each other's academic and health improvement goals. Collaboration between teachers and public health educators could help students achieve better educational and health outcomes.
Limitations
This study has several limitations that should be taken into account when comparing the results or generalizing the findings to other fields. The study is cross-sectional and does not allow the causality of any of the factors involved to be established. The participants came from urban areas and were students at selected high schools and colleges in the city of Dhaka. They may not be representative of those from other urban centers, non-urban areas or adolescents who do not attend school or college. The study used an anonymous, self-directed questionnaire, and compared to face-to-face interviews, self-reporting has limitations, including multiple biases (e.g., social desirability, memory recall, etc.). In addition, another important risk behavior, sexual risk behaviors cannot be investigated in the present study due to the social taboo and objections of the respected authorities.
It can be concluded that a sizable proportion of adolescents reported co-occurring health risk behaviors in Dhaka city of Bangladesh. The findings emphasize the need for effective interventions to prevent risk behavior practices among adolescents. Efforts should be made to to create awareness among adolescents, parents and teachers about risk behavior practices by public health bodies. This needs to be implemented in order to bring the prevalence of risk behaviors down to an acceptable level.
Funding
We received no specific grant from any funding source for conducting this study.
Ethical approval
This study involving human participants followed the 1964 Helsinki declaration and its later amendments. The Biosafety, Biosecurity, and Ethical Clearance Committee of Faculty of Biological Sciences, Jahangirnagar University, Savar, Dhaka-1342, Bangladesh approved the study protocol [BBEC, JU/M2019 (12)3].
Informed consent
All participants gave their informed consent to participate in the study.
Consent for publication
Not applicable
Conflicts of interest
The authors declare no conflict of interest.
Acknowledgments
Authors would like to thank all the students who voluntarily participate in the study. Authors would also like to thank the respective school/college authority for their support.