Discussion of the results

Utilization of primary health care services among adolescents (I)

The contribution of health and psychosocial factors was investigatedamong adolescents frequently and non-frequently using health care services. Although the adolescents mostly sought medical help for somatic reasons, psychosocial background factors predicted frequent primary health care use. There were gender differences in these health care utilization-related factors among adolescents. Female gender associated with frequent primary health care utilization. Furthermore, a high level of alcohol consumption and mental health symptoms in girls, and increased self-reported somatic complaints in the YSR in boys were associated with frequent primary health care utilization. Attending upper secondary school was related to less frequent primary health care use.

Emergency room services utilization among adolescents and young adults (II)

Known individual health and psychosocial characteristics and unobserved heterogeneity in emergency room use and visit rate were investigated among adolescents and young adults.

Individual patterns measured as “unobserved heterogeneity” were observed in emergency room services visits. The proportion of individual patterns in emergency room services utilization was larger among females than males, indicating the presence of an individual style effect, especially among females. A lower school performance at baseline associated with emergency room services utilization among males, and not studying or working at the follow-up among females. Female gender, a lower parental socioeconomic status, and childhood psychosocial problems increased the utilization rate of emergency room services.

In addition, mental health problems among males and being single among females associated with an increased emergency room services visit rate.

Mental health research participation and sample selection among adolescents (III)

Compared to the previous literature, attrition in this study was high. Due to the high drop-out rate, sample selection was analyzed in relation to known background characteristics of the study population. The study results refer to higher attrition among males and older adolescents. In addition, a lower school performance, lower parental employment, family disintegration such as divorce, and symptoms of depression and anxiety were associated with both consent and participation. However, the effect of mental health aspects was smaller than the effect of age and gender. (III)

6.1.2 Primary health care services utilization (I)

The association between female gender and frequent health care utilization is in line with the health statistics (Gissler et al. 2006; Nguyen et al. 2011) and previous study findings for young adults (Kestilä et al. 2007). Higher rates of utilization of health care services among females can be partly explained by gynecological reasons (Gissler et al 2006). Nevertheless, there might also be behavioral gender differences that are related to care seeking.

According to previous research, psychologically distressed young adults seldom seek professional help for their distress, despite their frequent use of health care services. The association between mental health problems and frequent health care visits parallels previous studies among adults (Kestilä et al. 2007; Smits et al. 2009). Nevertheless, these findings suggest a more significant association between mental health problems and frequent health care use in girls compared to boys.

In this study, hazardous alcohol use was associated with frequent primary health care use among adolescent girls. There have been no relevant previous studies on adolescent risky health behavior and frequent primary health care use. Among young adults, alcohol consumption and current or childhood adverse experiences associated with psychological distress, which in turn associated with frequent use of health services (Kestilä et al. 2007).

The finding of increased somatic complaints in the YSR subscale and a lower paternal occupational status among frequent primary health care users was similar to previous literature concerning adolescents. In a study by Vila et al. (2012), frequent adolescent primary health care visitors came from lower socioeconomic circumstances, and they had more past and current physical complaints, and a history of mental health consultations.

Homlong Rosvold and Haavet (2013) investigated school drop-out figures for adolescents, and observed higher drop-out rates among those with frequent attendance of school health services and youth health clinic use. These findings were consistent with the study of Vila et al. (2012), in which frequent primary care users had more days off school.

While previous studies have concentrated on days out of school, these current analyses included self-reported knowledge of school performance and school type. Attendance of an upper secondary school associated with significantly less frequent health care use when compared to comprehensive school attendance. Nevertheless, there were no associations between school performance and frequent primary health care use.

In line with the previous literature, female gender, mental health symptoms, a lower socioeconomic status, and school problems were related to frequent health care utilization among adolescents. The most interesting finding in this study was that there was an association between frequent health care services utilization and increased alcohol use among females. Surprisingly, no previous studies have been published on adolescent risky behavior, including alcohol or other substance use.

6.1.3 Utilization of emergency room services (II)

Emergency room utilization by females was greater than that of males, which is in line with the previous literature (Newton et al. 2010). Furthermore, among females, not studying or working and being single associated with increased emergency room services utilization.

There were also an association between a lower parental socioeconomic status and the

utilization of emergency room services. Recurrent emergency room visits by adolescents have been suggested to be related to socioeconomic deprivation (Newton et al. 2010; Rudge et al. 2013). The previous literature also suggests that dropping out from school, unemployment, and health problems often occur concurrently, and are associated with social exclusion (Mohajer and Earnest 2010). Socioeconomic distress, such as a low educational level and low income, has also been related to frequent emergency room use in adults (Sun, Burstin and Brennan 2003). Therefore, frequent health care services utilization might be associated with social exclusion. One might assume that an increased need for health care services could be a consequence of various psychosocial problems.

Among adults, chronic illnesses and the use of other health resources were related to higher emergency room utilization (Sun, Burstin and Brennan 2003). Surprisingly, in our study, females with no chronic somatic illness in childhood had higher emergency room visit rates. Mental health disorders, both current and previous disorders among males and previous disorders among females, associated with emergency room visits. In previous studies among adolescents, mental health problems have predicted higher emergency room utilization (Newton et al. 2010; Ballard et al. 2013). In adult studies, frequent emergency room visits also associated with psychological distress and mental health problems (Sun, Burstin and Brennan 2003; Jelinek et al. 2008; Stockbridge, Wilson and Pagán 2014).

The number of recurrent emergency visits is reported to have increased due to alcohol-related injuries in adolescents (Linakis et al. 2009). Among adults, an association has been observed between problem drinking and increased emergency room visits (Cherpitel and Ye 2008; Jelinek et al. 2008). In this study, hazardous alcohol use in adolescence (age 13–18) among females related to emergency room service utilization, but not to the frequency of emergency room visits. No such relationships were observed in the alcohol use of young adults (aged 19–24 years).

According to some earlier studies (Callahan and Cooper 2010; Lau et al. 2014), the absence of private insurance decreased emergency room visit rates among adolescents and young adults. In Finland, health insurance is statutory, and most people use public health care services. There have so far been no studies on the effects of voluntary health insurance on emergency room use among adolescents or young adults.

No previous research on individual patterns of emergency room utilization was found.

In our study, this effect of individual patterns was larger among females than among males.

In practice, this study result means that for males, the utilization of emergency services is more random than for females. Therefore, one might assume that for some reason, females have adopted a different style of health care services utilization. There is a paradox in health utilization between genders: In general, females are healthier than males, but after the teenage years, females use health care services more often compared to males (Gissler et al.

2006; Nguyen et al. 2011). Biological gender differences might have an effect on experiencing somatic symptoms and symptom-related emotional stress. On the other hand, behavioral gender differences might have an effect on care seeking. Therefore, compared to males, females might be more interested in health-related issues and socially more active in sharing health-related concerns.

The existence of significant individual patterns means that if we know an individual’s past behavior, we might be able to predict the future utilization of services. In theory, an individual style of health care services utilization might also refer to differences in supply or access of the services. Thus, in this study, the supply health care services in one city was not investigated, but it was assumed to be constant during the follow-up period.

As in the previous literature, this study indicated that frequent emergency room utilization associated with female gender, mental health problems, and a lower parental socioeconomic status. Alcohol consumption among adolescent females predicted emergency room utilization.

6.1.4 Mental health research participation and sample selection (III)

Attrition, especially among males, has previously been noted in several Finnish general population-based adolescent mental health studies (Hurtig et al. 2005; Kaltiala-Heino, Fröjd and Marttunen 2010; Väänänen et al. 2011; Miettunen et al. 2014; Mäki et al. 2014).

A high parental professional level was associated with attending the study, which parallels the findings of previous research (Audrain et al. 2002; Hurtig et al. 2005; Goodman and Gatward 2008; Fröjd, Kaltiala-Heino and Marttunen 2011; Väänänen et al. 2011; Mäki et al. 2014). For example, Audrain et al. (2002) suggested that adolescents whose parents have a lower educational level participate less frequently than those with more highly educated parents.

Parental divorce or separation was slightly associated with participation in the follow-up study, which is in line with the previous literature. Drop-out from a follow-up study was associated with being adolescent and having other than an intact family in the Northern Finland Birth Cohort 1986 Study (Hurtig et al. 2005), and with being adolescent and not currently living with parents in the Adolescent Mental Health Cohort study (Väänänen et al. 2011).

Internalizing symptoms measured with the YSR predicted participation in both phases, while externalizing symptoms reduced participation in the second phase. Examining the data more closely, anxious and depressive symptoms increased participation, and this effect was more clearly seen among girls. These findings are consistent with a previous study on the relationship between social phobia and depression among adolescents (Adolescent Mental Health Cohort), in which quite a high drop-out rate (36.9%) was associated with depressive symptoms measured with the BDI (Väänänen et al. 2011). In this study, depressive symptoms measured with the BDI did not differ between participants and non-participants, although findings from the YSR did. Contrary to these study findings, another Finnish study suggested that internalizing problem behavior may be underrepresented among respondents in adolescent mental health surveys (Fröjd, Kaltiala-Heino and Marttunen 2011). One must keep in mind that adolescent girls show more internalizing symptoms, while adolescent boys show more externalizing symptoms (Ormel et al. 2012).

Different study findings might be partly explained by sample selection: compared to previous investigations, females were highly overrepresented in this study.

Adolescent self-reported alcohol use or smoking had no effect on participation.

Nevertheless, providing information on illegal substance use or mental health issues may

be difficult in a controlled situation under the supervision of adults. Studies concerning adolescent substance abuse are among the most problematic in terms of participation and parental consent (Smith, Boel-Studt and Cleeland 2009). Adolescents prefer to hide their adverse behavior or abuse, particularly from their parents, and the active demand for parental consent may reduce the participation of adolescents with problem behavior (Moolchan and Mermelstein 2002; Tigges 2003).

Cognitive functions were subjectively evaluated in this study using self-rated school performance in various school subjects. Good performance in general subjects among girls and in the native language among boys associated with participation in the follow-up. Poor performance in mathematics increased both consenting and participation in boys. Non-responders had lower cognitive skills compared to Non-responders in the study sample of Boys et al. (2003). Furthermore, de Winter et al. (2005) described non-respondents as having more learning difficulties and problems at school than respondents.

In summary, the study results were mostly in line with the previous literature. More precisely, school problems among adolescents and a lower educational level among their parents were associated with study attrition. In this study, internalizing symptoms of adolescents associated with participation, whereas the opposite results have been reported in previous research. Surprisingly, no association was recorded between alcohol or tobacco use and attrition. Furthermore, in comparison to the previous literature, the higher drop-out rate and over-representation of females in this study sample had to be considered as a possible confounder of the study results.