Psycho-social wellbeing among in-school adolescents of Sri Lanka; prevalence of selected attributes and associated factors

Adolescence period is associated with psycho-social issues which affect their day to day activities and go beyond their adolescent age group. It has been recorded that prevalence of psycho-social issues are rising globally. The aim of this study was to estimate the prevalence and factors associated with the selected attributes of psycho-social well-being among 13 to 17-year in-school adolescents of Sri Lanka using the Global School Health Survey (GSHS), conducted in Sri Lanka. GSHS, was a cross-sectional study conducted among 3,650 adolescents of aged 13–17, attending government schools. Data was collected using a self-administered questionnaire. Psychosocial wellbeing included questions of four selected outcomes (i.e. bullying, felling alone, sleep disturbances due to a worry, suicidal ideation and practices) and 10 potential associated factors of these. The sample was selected through a two-staged cluster sampling technique. In determining the independent associated factors, multivariable analysis was done with multiple logistic regression.


Abstract Background
Adolescence period is associated with psycho-social issues which affect their day to day activities and go beyond their adolescent age group. It has been recorded that prevalence of psycho-social issues are rising globally. The aim of this study was to estimate the prevalence and factors associated with the selected attributes of psycho-social well-being among 13 to 17-year in-school adolescents of Sri Lanka using the Global School Health Survey (GSHS), conducted in Sri Lanka.
Methods GSHS, was a cross-sectional study conducted among 3,650 adolescents of aged 13-17, attending government schools. Data was collected using a self-administered questionnaire. Psychosocial wellbeing included questions of four selected outcomes (i.e. bullying, felling alone, sleep disturbances due to a worry, suicidal ideation and practices) and 10 potential associated factors of these. The sample was selected through a two-staged cluster sampling technique. In determining the independent associated factors, multivariable analysis was done with multiple logistic regression.

Results
Most of the students who responded belonged to the 13-15 age group (66.5%) and with 51.1% were females. More than one third of the participants (37.7%, 95%CI 36.1-39.5) had experienced bullying 30 days prior to the study. 'Feeling loneliness' and 'experiencing anxiety' had been experienced by 9.0% (95%CI 8.0-10.0) and 4.7% (95%CI 3.9-5.4), respectively. The prevalence of planning an attempt of suicide was 6.3%, while 6.6% had at least once attempted suicide. Multivariable analysis revealed that lacked parental support and being engaged in a ght were independently associated with being bullied, feeling lonely and considering attempting suicide. Mutual association of the four outcomes with each other indicate that being bullied was signi cantly associated with experiencing loneliness and feeling anxiety. In addition, experiencing loneliness was independently associated "considering an attempt of suicide" (OR = 2.9, 95%CI = 2.0 to 4.3). Feeling anxiety was also associated with considering attempting suicide (OR = 3.4, 95%CI = 2.1 to 5.4).

Conclusion
This study demonstrated that the prevalence of psycho-social wellbeing is among school going adolescents is a major public health problem in Sri Lanka. School-based interventions to address these issues should be designed with the goal of increasing adolescents' awareness of possible preventive measures of unintentional injuries.

Page 4/18
Background Adolescence includes the transitional period from childhood to the adulthood and often de ned as from 10 to 19 years (1,2). Prevalence of adverse psycho-social health issues is rising among these young people globally (3)(4)(5). One tenth of children and youth have some mental health problem that not only signi cantly interfere with their day today life but also go beyond the adolescence (3). Similarly, attributes of social disintegration, such as isolation and loneliness have been associated with a higher mortality (5,6). Furthermore, many of the psychological problems reported in the adulthood are often found to have been originated in the adolescence (4,7).
Presence of depressive symptoms, suicidal behaviours, loneliness, isolation are some manifestations of low psycho-social wellbeing among adolescents. It had been documented even before the 21st century that about 15% of adolescents experience depression to a moderate to severe degree (8). Depression and suicidal behaviours are inter-related. Suicidal ideations are strongly associated with suicidal attempts (8).
In an Estonian study with [13][14][15] year school children, self-assessed depressive feelings and suicidality were found to be 26.1% and 12.8% respectively (8). Prevalence of negative social attributes extend across a vast range exceeding 30% of the adolescents in some settings (1,9).
There are several associated factors negatively in uencing the above manifestations, including bullying, substance abuse, smoking and peer-violence. These factors are associated with many mental health problems like anxiety, depression and suicidal ideation (10,11). Additionally these factors are likely to be persistent or repeatedly occur, thus augmenting their negative effect over time (12). The prevalence of these factors varies from school grade to school grade. In some circumstances, even teachers are unaware of these factors. As an example, in a study done in Pakistan, more than half of the teachers did not have adequate knowledge on bullying (13). To the contrary favourably associated factors that reduce the risk of the above manifestations include family bond and the practice of life skills (14,15).
Sri Lanka, an island in the Indian Ocean with a population of 21.7 million, has been recently graduated to an upper-middle income level (16). Out of the total population, around 16.2% are between 10-19 years (17). Documented literature on adolescent psycho-social wellbeing is not commonly found in Sri Lanka. However, the limited available ndings imply that psycho-social problems among adolescents are far from being rare (18,19). Among the adolescents in Sri Lanka about 70% are schooling (20). Hence schools have become ideal settings for the exploration of adolescent issues.

Global School-Based Student Health Survey (GSHS) has been developed by the World Health
Organization with other collaborators with aims that include helping countries in identifying priorities and trends (21,22). Accordingly, GSHSs have been conducted globally by Ministries of Health and Education from 2003 (20). The last GSHS has been completed in Sri Lanka in 2016 and the previous in 2008 with the objective of assessing the prevalence of health behaviours and the protective factors (11,23). In the GSHS-2008, there were some eye-opening ndings. As examples nearly 10% of students between 13-15 years had stated that they had seriously considered attempting suicide during previous 12 months. Similarly, 7.6% had felt alone always or most of the time(23). In the GSHS-2016, participants of 16 and 17 years were additionally included in the study population. Therefore, ndings of GSHS-2016 would spread across a full range of 13 to 17 year period of the adolescence and would enable detecting trends when compared with the ndings of 2008.
In this background, this paper describes the ndings of a secondary analysis of the GSHS-2016 data in relation to factors associated with the selected attributes of psycho-social well-being among 13 to 17year in-school adolescents of Sri Lanka.

Methods
GSHS-2016 was completed as a descriptive cross-sectional study among a representative sample of inschool adolescents aged 13-17 years who were in grades 8 to12, in 25 districts of Sri Lanka. Being enrolled in a government-school was an eligibility-criteria. Data collection was done between 1st October 2016 to 30th November 2016. The survey included a two-stage cluster sampling method in a sampling frame of 40 schools including 3650 students. In stage I, 40 schools were selected and stage II, systematic sampling was used to select classes of those schools. Minimum sample size needed at the data collection stage was estimated to be 3125 with a 5% margin of error and an assumed response rate of 80%. At the end of the two rounds, the selected classes included 3650 potential participants.
The data collection was done using a standard self-administered questionnaire developed by WHO. The questionnaire was culturally adopted to the Sri Lankan context and translated to the two mainly practiced languages (i.e. Sinhalese and Tamil) in Sri Lanka. The participants had the option of selecting their choice of the language of the questionnaire. In addition to the demographic details of the generic questionnaire, the section on psychosocial well-being included questions of four selected domains (i.e. bullying, felling alone, sleep disturbances due to a worry, suicidal ideation and practices) and 10 potential associated factors of these. Data collection was facilitated and supervised by the principal investigator and was done by trained data collectors.
The data set was cleaned and edited for inconsistencies. Missing data were not statistically imputed. Analysis was done with Statistical Package of Social Sciences version 21. The prevalence of the outcome-domains and associated factors were described with the point estimates as well as con dence intervals. Bi-variable analysis was done using the uni-variable logistic regression. The effect measures of these associations were presented with their interval estimates. In determining the independent associated factors, multivariable analysis was done with multiple logistic regression. Beta coe cients and their con dence intervals were described for the associated factors.
Informed written consent was obtained from the from the parents/ guardians. Data collection was done on pre-planned days without interfering with the routine activities of the schools. Ethics approval was obtained from the Ethics Review Committee of Faculty of Medicine, Colombo (Approval number-EC-16-184).

Results
The overall response rate of the selected participants was 89% with 3262 usable data records. One hundred and sixteen (n=388) either did not give consent to be included in the study or was absent from school on the day of the data collection.

Characteristics of participants
Most of the students who responded belonged to the 13-15 age group (66.5%). The study group consisted of an equal proportion of males (48.9%) and females (51.1%) (table 1). Further analyses were performed on the 3173 participants who were within the 13 to 17 years (i.e. excluding the 88 participants who were studying between grades 8-12 but being outside these age limits).
Selected outcomes related to psycho-social well-being psycho-social well-being The table 2 shows the percentage of participants who experienced three selected outcomes in relation to psycho-social well-being, namely being subjected to bullying, feeling loneliness and experiencing anxiety.
The prevalence of experiences in bullying was similar between the two age categories (age 13-15 years and 16-17 years). In contrast, "feeling loneliness" was signi cantly higher in the older age group (12.0%, 95% CI 9.9-14.0) compared to the 13-15 category (7.6%, 95% CI 6.5-8.7). Signi cantly more students in the elderly group (i.e. age group 16-17 years) (6.3%, 95% CI 4.8-7.9) had experienced anxiety compared to their younger counterparts (3.7%, 95% CI 2.9-4.5). Table 3 depicts three outcomes in relation to psycho-social well-being, related to suicides: ideation, planning and attempting. Nearly one tenth of the participants have seriously considered attempting suicide. The prevalence of planning an attempt of suicide was 6.3%, while 6.6% had at least once attempted suicide. None of these outcomes showed signi cant difference between different sex and age groups.
Associations of psycho-social wellbeing Table 4 shows unadjusted associations of eleven selected characteristics in relation to four selected outcomes of the psycho-social wellbeing: being bullied, feeling loneliness, experiencing anxiety and considering attempting suicide.
It is observed that except for the age, other characteristics are statistically signi cantly associated with being bullied at 5% signi cant level. The highest effect measure was observed for "being physically attacked". A participant who has been physically attacked is 4.9 times likely to be subjected to bullying compared to one without. Except for the sex of the participant, other characteristics were signi cantly associated with experiencing loneliness. The highest effect measure was observed for "lacking parental support" (OR=2.5, 95% CI= 1.9-3.2). Ten characteristics had been associated with considering attempting suicide, with age and sex being the only exceptions. "Lacked parental support" and "having ever smoked" were associated with 2.7 times likelihood of considering attempting suicide compared to absence of those factors. Table 5 shows the adjusted effect measures of these associations. Adjustments have been done for the other 10 characteristics in this table as well as for the presence of these outcomes (as mentioned in table   6). Out of the characteristics mentioned in Table 5, ve were independently associated with being bullied: male sex (OR=1.7, 95% CI=1.4 to 2.1), lacking parental support (OR=1.3, 95% CI=1.1 to 1.5), lacking parental supervision (OR=1.6, 95% CI=1.2 to 1.9), being physically attacked (OR=3.0, 95% CI=2.5 to 3.7) and being in a ght (OR=2.1, 95%CI=1.7 to 2.5). Older age (OR=1.7, 95% CI=1.2 to 2.3) and lack of parental support (OR=1.8, 95% CI=1.3 to 2.5) were independently associated with experiencing loneliness while male sex was observed to be a protective factor (OR=0.6, 95% CI=0.4 to 0.8). Five factors were associated signi cantly with considering attempting to suicide. Those were: not having close friends (OR=1.6, 95% CI=1.01 to 2.7), not engaging in sports (OR=1.5, 95% CI=1.1 to 2.1), lacking parental support (OR=1.7, 95% CI=1.2 to 2.3), lacking parental supervision (OR=1.5, 95% CI=1.1 to 2.1) and being in a ght (OR=1.5, 95% CI=1.1 to 2.1).

Discussion
This is the rst documented journal article exploring the factors independently associated with four important attributes (i.e. being bullied, feeing lonely, experiencing anxiety and ideation of attempting suicide) among a national-wide sample of 13-17 year old school children. The prevalence of these attributes was notably high with nearly 38% being bullied, 9.3% considering attempting suicide, 9% feeling lonely and 5% experiencing anxiety. In general, there was a notable mutual co-existence of these attributes, highlighting their vicious impacts of the life of the school children. The present study revealed: Older age (for 2 attributes), male gender (for 2 attributes), lacking parental support (for 3 attributes), lacking parental supervision (for 2 attributes), being physically attacked (for 1 attribute) and engaging in a ght (for 4 attributes), not engaging in sports (for 1 attribute) and not having close friends (for 1 attribute) as independent risk factors.
The comparison of the prevalence of the attributes with the ndings of the 2008 survey depicts notable observations. The percentage of children subjected to bullying has remained more or less the same (around 37% in overall) with a slight increase among boys (23). The proportion of children who felt lonely (most of the time or always) remained the same (i.e. 7.6%) despite having an extended sample with two additional older years (23). In fact, the prevalence was relatively higher (12% versus 7.6%) among the children between 16-17 years compared to those between 13-15 years. Even though a slight decrease was observed in the proportion of those seriously considering the suicide (9.9% versus 9.3%), the burden was relatively higher (0.5% in excess) among the children who were 16-17 years. These provides two important inferences. One is that the burden of thee psycho-social attributes have remained unchanged in between 2008 to 2016. Second is that these remain issues with similar or higher magnitude even among the sub-group of 16-17 years, which were not included in the 2008 survey. Both these phenomena re ect the necessity of novel interventions in addressing these immediately.
It is important to review similar regional GSHS ndings with these Sri Lankan observations. As an example, the percentage of children 13-17 years who seriously considered suicide in past 12 months were: 4.9% in Bangladesh (2014) (29). This re ects that these problems are not unique to Sri Lanka, but persist as regional and global burdens. However with the far improved health standards and achievements as well as with the well-established public health infrastructure, Sri Lanka is at a more favourable position in addressing these (30,31).
The risk factors explored in the survey are known to be present with a notable magnitude in the study population. This is complemented by other published literature on the 2016 GSHS as well. As examples, the GSHS ndings show that the prevalence of violence-related factors are far from rare with 44.2% being in a ght and 35.1% being physically attacked (11). Additionally, the prevalence of alcohol, smoking, smokeless tobacco and substance abuse within the previous 30 days had been between 2-4%(32). The alcohol and tobacco related risk factors are in general reported at a lower level than compared to regional settings (25)(26)(27)(28), whereas violence related items like being engaged in a ght are reported at the same or a higher magnitude (24)(25)(26)(27)(28). The ndings of the present study re ect the co-existence of the psychosocial attributes. Each four attribute was signi cantly associated with at least two more attributes when adjusted for all the factors included in the study. This in in congruence with the global literature that these conditions have mutual associations (33)(34)(35). These imply the importance of screening and detection of school children of affected with these attributes and the value of early intervention as the presence of one attribute is independently associated with the occurrence of another.
Out of the 14 factors explored for each attribute (i.e. including their mutual associations), except for age and sex, all other nine were modi able. Out of these, factors related to the domains of lack of parental involvement, presence of violence and not engaging in sports activities were detected as independently associated with the psycho-social attributes. Even if the substance-use did not yield direct signi cant associations with these four attributes, these are well known to be associated with factors that increase their likelihood with reference to observations of the same study population (11,32). This highlights the need of a comprehensive package addressing the web of causation of these related outcomes. The initiatives like "promotion of life skills" and "health promoting school programme" have been implemented with added emphasis in recent years as collaborative efforts of Ministries of Health and Education, addressing these phenomena (36,37). The impact of these would hopefully help in reducing these burdens, which could be evaluated in the next GSHS.
There were several limitations of the study. Firstly, the study participants include only those who have been enrolled in government schools having excluded the students of the private sector. This was emphasized in interpreting data. Secondly the responses were given as stated by the participants without veri cation of the responses. Due to the sensitiveness of the questions asked, utilization of the selfadministered questionnaires was the best available option. However, before administration of the questionnaires, the importance of the provision of genuine responses were highlighted to the whole classroom. Thirdly the study did not capture non-school going adolescents of the selected age group. Therefore, generalizing the ndings to the entire adolescent population in Sri Lanka is not possible, as their circumstances may differ signi cantly. This was clearly mentioned in the eligibility criteria and was concerned in the interpretation of ndings.

Conclusions
The four psycho-social outcomes-being bullied, experiencing loneliness, feeling anxiety and considering attempting suicide are prevalent in notable proportions among the school children of 13-17 years of Sri Lanka as suggested by the ndings of GBSHS-2016. Nine out of eleven factors explored in the present study are signi cantly associated with at least one of these attributes. The presence of one attribute is signi cantly associated with the presence of at least two other attributes. High prevalence of the potential risk factors of these when combined with this co-existence of these attributes could potentially increase the burden and worsen the impact of these. Detecting and managing these attributes among school children and addressing the potential risk factors must be regarded as a local as well as a regional priority.