Oiling a neglected wheel : An investigation of adolescent internalising problems in rural South Australia

Introduction: Despite a paucity of research, adolescents living in rural areas appear to have a heightened risk for developing a mental health problem compared with their urban counterparts. The main objectives of this study were to contribute to building an evidence base of prevalence rates and determinants of internalising problems of adolescents in rural South Australia. A multidimensional Process Model was used as theoretical framework to enable an investigation of the various determinants from individual, family and community domains; specifically, the contribution of self-esteem, parental acceptance and elements of social capital at an individual level (ie participation in the local community and proactivity in a social context represented structural social capital, and feelings of trust and safety, and neighbourhood connections represented cognitive social capital). Methods: In this cross-sectional prospective study, a total of 388 Year 9 (2nd year of secondary school) students (208 females, 180 males) aged 13–15 years (mean age = 14.2 years) participated from 11 high schools within the Country Health South Australian area. These adolescents completed a battery of self-reported measures online at school. Results: The results demonstrated that the adolescents experienced a ‘normal’ level of self-esteem and a ‘moderate’ level of perceived parental acceptance. The level of social capital was considered ‘low’ and the adolescents experienced a ‘moderate’ level of internalising symptoms. Based on the mean score of the Revised Child Anxiety & Depression Scales (RCADS), 25% of the adolescents experienced internalising symptoms ranging in severity from mild to severe, with no significant differences between males and females. Approximately 13% were considered above the clinical threshold, with 4% reporting experiencing severe


Introduction
Research efforts investigating the mental health of Australians living in rural areas, in particular rural adolescents, have been scant [1][2][3][4][5] .An overview of rural health research in Australia in the period 1990-1999 revealed only a small number of published articles on adolescent mental health 6 despite the fact that, as Kilkkinen et al 7 highlighted, children and adolescents represent a higher proportion of the population in rural and remote areas compared with urban areas, and research needs to target these priority groups.The Australian National Health Policy also supported the need for greater research attention on the mental health of rural and remote Australians 8 , and the rural health policy framework, 'Healthy Horizons', recognises 'mental health, suicide and attempted suicide as among the highest priorities for action in rural health' 9 .
This paucity of research investigating rural adolescent mental health in Australia in particular, is of concern for a number of reasons: • It has been empirically established that, in contrast to the overall health gains of world populations in recent decades, the burden of mental illness is mounting.
• Mental and behavioural health disorders are common, being present at any point in time in approximately 10% of the population, and affecting more than 25% of all people at some time during their lives 10 .
• In 2002, depression accounted for 4.5% of the worldwide total burden of disease.It is also responsible for the greatest proportion of burden attributable to non-fatal health outcomes, accounting for almost 12% of total years lived with disability worldwide 10 .
• It is estimated that by 2030, depression will be one of the leading causes of disease along with HIV/AIDS and heart disease 11 .
Moreover, it has been reported that adolescents living in rural areas appear to be at a heightened risk for developing a mental health problem, compared with their urban counterparts.This is apparently due to a number of psychosocial factors: • stressful life events, perceived control of life events 12 • lack of access to mental health services • the effect of the economic downturn on rural areas, and severe drought in a number of geographical regions.
In addition, the suicide rates are consistently higher in rural towns than in metropolitan areas; for example, the suicide rate of males aged 15-24 years in rural towns has been reported as approximately twice that of their city counterparts 13 .For young Indigenous males, the suicide rate is up to 40% higher than the national Australian average 14 .
In response to such issues, The Australian Rural Mental Health Study 15 embarked on building an evidence base for the adult rural population.'Proxy' (ie alternative) measures of child health and wellbeing have also been included in this study.The establishment of headspace 16 , a federally funded youth mental health foundation service, in a number of rural areas of Australia has included two centres in South Australia.However, Boyd 1 suggests that while we are facing a period of mental health reform with such strategies: …although assurances have been made that the needs of rural youth will be considered as part of a national strategy to transform youth mental health services, rural communities remain without vision or direction as to how this might be achieved.
One possible way to achieve direction in order to address the needs of adolescents from rural communities may be to enhance the evidence base of this population.This may be achieved by investigating not only prevalence rates of adolescent mental health problems, but also by examining the individual, family 7 and community factors, or the determinants 17 , of such problems.For it has been suggested that many large population-based studies have failed to include an assessment of such important factors associated with mental health problems 18 .
A number of studies have reported that family processes, in particular the nature and quality of adolescents' relationships with their parents, contribute to the onset and maintenance of a range of negative developmental outcomes 19,20 [24][25][26] , social capital has been shown to play a role in the incidence and prevalence of mental health outcomes 27,28 .
The following variables from the Process Model were selected for use in this study (Fig1): 1. Individual Process: self-esteem.
2. Family Process: parental acceptance.Information sheets and consent forms were provided to students and their parents and distributed by teachers.A list of South Australian organisations were also included to assist any parents who had possible concerns regarding their adolescent's behaviour, health, emotional wellbeing or issues regarding parenting.Students were also provided with a similar list.All participants were invited to enter a raffle for one of two $25 department store shopping vouchers.Once consent had been granted, students completed an online questionnaire at school under the supervision of a teacher, which took approximately 50 min to complete.

Measures
Self-esteem: The Rosenberg Self-Esteem Scale (RSES) was used to assess self-esteem 32 .The RSES is a 10-item selfreport uni-dimensional measure of global self-esteem, consisting of statements related to overall feelings of selfworth or self-acceptance.Scores range from 0 to 30, with higher scores indicating higher self-esteem.Scores between 15 and 25 are considered to be within normal range; scores below 15 suggest low self-esteem.Example questions include: 'I feel that I am a person of worth' and 'I feel at least on an equal plane with others'.In this study the internal consistency alpha (α) = 0.86.
Parental Acceptance: Parental acceptance was assessed using the Acceptance subscale from the revised Child Report of Parent Behaviour Inventory (CRPBI) 33 .This revised measure contains ten items designed to assess the adolescents' perceptions of the extent to which they feel accepted by their parent/s.Adolescents responded on a 3 point Likert-type scale from 1 (= not like her/him) to 3 (= a lot like her/him) as to how well items described their  35 .In this study α = 0.87.

Results
Prior to analyses, variables were examined for accuracy of data entry, missing values and fit between the distributions and the evaluation of assumptions of multivariate analysis.All assumptions were considered to be met.Based on the total mean score of the RCADS, 25% of adolescents experienced anxious and/or depressive symptoms ranging in severity from mild to severe (Table 2).Approximately 4% of adolescents reported experiencing severe symptoms.
The relationship between parental acceptance and selfesteem was moderate and positive (Table 3), indicating that adolescents who reported a higher level of parental acceptance also reported a greater level of self-esteem.A moderate, negative relationship was found between parental acceptance and internalising symptoms, indicating that adolescents who reported that they felt accepted by their parents experienced lower symptoms of anxiety and depression.A moderate and negative relationship was found between self-esteem and the total internalising score, signifying that the greater level of self-esteem, the less internalising difficulties experienced by the adolescents.
The relationship between social capital and self-esteem were smaller, yet still significant and positive, indicating that adolescents who reported a greater sense of social capital within their communities, experienced higher self esteem.
The relationship between parental acceptance and social capital was very small and negative, yet significant, indicating that adolescents who perceived their parents to accept them reported higher social capital.The relationship between social capital and internalising problems was not significant.
Results from a linear regression demonstrated that social capital was significantly related to greater self esteem, F (1, 386), = 16.80,β = .22,p = .000.Approximately 4.2% of the variance in self-esteem was explained by its relationship with social capital (R 2 = .04).
A simultaneous regression analysis was conducted to evaluate how much of the variance in internalising problems could be explained by the level self-esteem, parental acceptance and social capital.The linear contribution of the three predictors was significantly related to internalising problems F (3, 384) = 58.33,p = .000.Approximately 33% of the variance in internalising problems could be accounted for by the combination of predictors, R 2 = .33.
Self-esteem and parental acceptance were the significant contributors to the prediction of internalising problems in the adolescent sample (β = .52,p = .000).Social capital was not a significant predictor of internalising problems.

Discussion
The present study contributes a small step towards building an evidence-base for the prevalence rates and determinants of internalising problems of adolescents from rural South Australia.Moreover, employing a multidimensional theoretical framework enables an investigation of the various determinants of mental health problems from individual, family and community domains.

Implications
Although the results from the present study represent adolescents from South Australia only, a number of implications are worth noting.For instance, based on the total mean score of the RCADS, the adolescents experienced a 'moderate' level of internalising symptoms.Further, approximately 13% of adolescents were considered to be As a consecutive National Mental Health Survey is approaching, it is timely to highlight the importance of including as many Australians as possible from rural and remote areas, and in particular young people.This will produce a more accurate evidence-based representation of Australia's adolescent population, which will more effectively inform policy and ultimately facilitate implementation of relevant strategies.
Because this study was cross-sectional, causality cannot be demonstrated.Nonetheless, the findings demonstrate the importance of parental acceptance and may have implications for preventative and intervention programs for adolescents and their families.Not only did higher levels of parental acceptance predict lower levels of internalising problems, but they also predicted higher levels of selfesteem.Such findings may be applied to educate parents regarding the benefits of parental acceptance.To begin with, it is important to inform parents that 'acceptance' does not equate to agreeing with or accepting all of their adolescent's behaviours, but more to do with recognising, validating and demonstrating understanding of their experiences, thoughts and feelings, even during times of conflict.
Such an approach shows implicit and explicit interest and concern in the adolescent's perspective (which is especially important during times of conflict or disagreement), without creating a possible impression of disappointment.Thus, an adolescent is likely to internalise this acceptance, leading to a positive inner sense of self.This is particularly necessary when faced with the challenges inherent in the developmental years when an increase in autonomous functioning requires belief in one's self-worth.
The results of this study showed that higher levels of selfesteem decreased the probability of experiencing internalising symptoms.Teaching parents relevant strategies to understand and display acceptance may not only help to strengthen the relationship with their adolescent, but further enhance their self-esteem, which may help to prevent or delay possible anxious or depressive symptoms from emerging.

Limitations and suggestions
There are a number of limitations worth noting.First and foremost, South Australia is merely one region in Australia and no rural district in Australia is homogenous 3,4 Heterogeneity within rural settings must be recognised as differences do exist.For instance, different rates of suicide have been acknowledged according to the degree of geographical remoteness 13 .Further, the findings from this study relied on adolescent self-reports and cross-sectional data.Thus, longitudinal data are required to establish causal relationships.
In the present study, social capital did not predict lower internalising problems.Perhaps such a result reflects the low social capital scores reported by the adolescents, rather than not being beneficial to mental health.It would be of interest to examine the outcomes of studies which report higher levels of social capital to explore the effects on mental health outcomes.However, social capital is such a multifaceted concept that it is not likely to be represented by any single measure 37 .
Exploring potential comparisons between the prevalence rates and determinants of mental health problems of rural and urban adolescents 3,38 is also important.This is especially noteworthy when considering that potential differences between rural and urban populations may have been overlooked with the omission of rural populations included in national surveys 3,30,31 .
Additionally, it may be useful to examine the role of other factors included in the Process Model 21 in the developmental outcomes of rural adolescents.For example, emotion regulation, parental autonomy granting, and/or a range of health indicators may be explored.By adopting a strengthsbased approach, the determinants of mental health may also be investigated.This may include examining protective factors that could cushion the effects of adversity specifically to conditions which rural groups are more likely to encounter (eg the impact of drought) 39 .
Finally, progressing from investigating what factors contribute to the onset of a particular developmental outcome, to examining how and why, or the underlying processes involved would be worthwhile.For instance, examining the interplay between various theoreticallyderived individual processes as mediators of the relationships between specific family and/or community process factors.For instance, the potential role emotional security plays in the relationship between parental punitiveness and GAD; whether other aspects of social capital (eg social trust) account for the relationship between self-efficacy and PD; or the impact that self-control plays in the relationship between perceived teacher or peer support and depressive symptoms.

Conclusions
It has been suggested that in many rural and remote communities in Australia, mental health issues may be something of a 'sleeping giant' 3 .The challenge facing health professionals is to provide practical and evidence-based programs designed to address and/or strengthen a range of individual, family, and community factors -before the 'giant' begins to stir.Although the pathways to internalising problems are complex, it is anticipated that the knowledge gained from the current study may assist in preventing the development of mental disorders in rural youth, where local mental health services are often limited or unavailable.

Figure 1 :
Figure 1: Selected variables to be investigated in the present study (self-esteem, parental acceptance, dimensions of social capital, and internalising problems).
above the clinical threshold, with 4% reporting experiencing severe symptoms.Although it is difficult to compare these findings with national figures (because young people were not included in the 2007 National Study, and anxiety disorders and young people from rural areas were not included in the 1997 National Study) such results must be of concern.
Cognitive Social Capital Feelings of trust and safety -Example questions included: 'If someone's car breaks down outside your home, do you invite them into your home to use your phone?' and 'Does your local community feel like home?' Neighbourhood connection -Example questions included: 'If you were caring for a child and needed to go out for a while, parent/s.Example questions include: 'My Mother/Father is a person who ...makes me feel better after talking over my worries with her/him' and '...tells me how much s/he loves me'.Higher scores indicated greater levels of parental acceptance as perceived by the adolescent.In this study α = .89.Proactivity in a social context -Example questions included: 'Have you ever picked up other people's rubbish in a public place?' and 'If you need information to make a life decision, do you know where to find that information?' would you ask a neighbour for help?' and 'When you go shopping in your local area are you likely to run into friends and acquaintances?' Responses ranged from: 'Not at all/Not much' (=1) to 'Yes, definitely/Frequently' (=4).Total structural, cognitive and combined scores were calculated.Possible combined total scores ranged from 12 to 48.Both structural and cognitive total scores each ranged from 6 to 24.Higher scores indicated greater social capital as reported by the adolescents.and severe: 95-141.Adolescents scoring within the moderate symptom category indicate symptoms above the clinical threshold