Challenges in measurement of adolescent mental health: how are gender patterns affected when level of symptoms is analysed simultaneously with impairment?

Background: Adolescent mental health surveys in public health are sometimes questioned because of their main focus on self-reported symptoms, lacking data on impairment, e.g. the consequences on everyday life of the mental health problems. While public health studies typically reveal higher prevalence rates of internalising problems for girls than boys, there are indications that the gender pattern may change when self-reported data on symptoms are analysed simultaneously with impairment. The purpose is to determine how gender patterns of adolescent mental health solely based on symptoms are affected when level of symptoms is analysed simultaneously with impairment.

Methods: Questionnaire data on adolescent mental health were collected in schools by Statistics Sweden in the autumn of 2009 as part of a national total population study in grades 6 and 9 in Sweden. In this study only data from grade 9 students are used (n = 91 627; response rate = 80 per cent). Psychosomatic symptoms were measured with the Psychosomatic Problems scale including eight items. Impairment was measured with four items included in the SDQ impact supplement. The associations between these key constructs were analysed with logistic regression and contingency tables.

Results: When analysing variables on psychosomatic symptoms and impairment independently, the results are consistent with typical findings of gender patterns in adolescent internalising mental health. Girls report both more psychosomatic symptoms, and more negative consequences in everyday life, than boys. The gender patterns are, however, strongly affected when impairment is conditioned on level of psychosomatic symptoms. Except for the Home Life setting, in the settings of Friendships, Classroom Learning and Leisure Activities, the previously reported gender pattern favoring higher disturbances among girls becomes partly reversed implying that boys report more negative consequences than girls. Hence, while girls report a higher prevalence of psychosomatic symptoms, boys appear to suffer from such symptoms more than, or as much as, girls in three out of four everyday life settings.

Conclusions: The study confirms the insufficiency of solely including data on symptoms in the measurement of adolescent mental health. Regardless of the causes of the complex gender pattern shown in this study, the results highlight the importance of simultaneous inclusion of indicators of impairment as well as symptom counts and frequency in the measurement of adolescent mental health.