The bidirectional longitudinal association between health-related quality of life and academic performance in adolescents: DADOS study

Well-being is as a higher-order construct that integrates health (e.g., mental or physical functions), health-related (e.g., work or habits), and non-health-related domains (e.g., autonomy or integrity), which depends on complex interactions between individual and contextual factors [1]. Indeed, prior research has shown that individuals with high well-being are usually successful across multiple domains of life, including work and social relationships, and show better mental and physical health, as well as longer longevity [2]. Well-being encompasses some related but distinguishable constructs such as health-related quality of life (HRQoL), which could be defined as a broad multidimensional concept comprising the individual's perceptions of how life is going in terms of health and health-related domains [1]. HRQoL is considered a relevant health indicator, since it allows to capture the burden of disease from the perspective of an individual, and has been proposed as a predictor of mortality and morbidity [3]. From a public health perspective, assessing HRQoL during adolescence is of paramount importance, since quality of life in this period stablishes the basis for quality of life and health status in adulthood [4]. Thus, identifying factors related to adolescents’ HRQoL is of relevance to advance in the understanding of this multidimensional construct.

It has been recently suggested that HRQoL could be positively related with academic performance [5], which refers to educational goals that students have to reach in a particular period of time [6]. Importantly, academic performance may strongly shape a person’s life chances in terms of work and health. In fact, higher academic performance during adolescence has been associated with better earnings [7] and health status later in life [8]. While the relationship of adolescents’ health with HRQoL [9, 10] and academic performance has been widely investigated [6], how the psychological construct of HRQoL and academic performance are associated is understudied and remains to be clarified.

Regarding the overall construct of well-being and its association with academic performance, a recent systematic review conducted by Amholt et al. [11] concluded that findings from previous studies were inconsistent, since both positive and null associations were reported. Moreover, the studies included in this review, mostly with a cross-sectional design, focused on analyzing the unidirectional relationship between subjective well-being and students’ academic performance [11]. Interestingly, few previous research investigated the plausibility of a bidirectional association between these two constructs in adolescents, which suggested that the direction of this association is still unknown since divergent results were reported [12‐14]. For instance, Wu et al. [12] investigated the bidirectional association between the main components of subjective well-being (i.e., life satisfaction, and positive and negative affect) and a standardized index of academic performance, showing that only life satisfaction and positive affect at baseline were positively correlated with adolescents’ academic performance 14 months later. Conversely, Steinmayr et al. [13] reported that adolescents’ grade point average (GPA) at baseline was positively associated with changes in life satisfaction at 1-year follow-up, while there was no association in the other direction. Additionally, Ng et al. [14], who examined the association between life satisfaction and GPA in two waves with a 5-month interval in a sample of adolescents, suggested a positive reciprocal causal relationship. Regarding the specific construct of HRQoL, there is only one study that has examined its bidirectional association with academic performance in adolescents, which reported controversial results and gender differences [15]. Specifically, Bortes et al. [15] found a positive association between HRQoL at baseline and GPA 24 months later, and a negative association between GPA at baseline and HRQoL at follow-up in adolescent girls, while no associations were found among boys.

The scarce literature examining the bidirectional longitudinal association between HRQoL and academic performance evidenced two major limitations. First, most previous studies included only a global indicator of academic performance, without examining the association of each single academic performance indicator with HRQoL. Moreover, in this context, prior evidence did not include academic abilities, which reflect competences and content acquired in specific areas of knowledge. Second, most of these studies present a relatively short-term longitudinal design. Therefore, given the paucity of knowledge and the lack of conclusive findings, more studies are necessary to address these gaps in the literature and to clarify the directionality of the association between HRQoL and academic performance in adolescents. Thus, the present study intended to explore the bidirectional association between HRQoL and academic performance in adolescents between two timepoints with a 24-month interval. Since the influence of sex in this association remains unclear, a secondary aim was to analyze whether this association varies between boys and girls.

The characteristics of the adolescents at baseline and at 24-month follow-up are shown in Table 1. Participants’ HRQoL score was 50.1 at baseline and 48.7 at follow-up. Academic grades ranged from 6.9 to 8.1 at baseline and from 6.3 to 8.2 at follow-up. Regarding academic abilities, the overall score was 49.0 at baseline and 59.8 at follow-up.

Bidirectional longitudinal associations between HRQoL and academic performance based on the cross-lagged panel models after adjustment for sex, pubertal stage, waist circumference, socioeconomic status, and parents’ education level are shown in Table 2. The headings used in the table are represented as pathways in Fig. 1. At baseline, after multiple comparisons correction, only GPA was cross-sectionally and positively associated with HRQoL (p = 0.010). HRQoL at baseline was not associated with any of the academic performance indicators 24 months later (all p > 0.05). Nevertheless, all the academic grades and the overall score of academic abilities at baseline were positively associated with HRQoL at follow-up (all p < 0.05).

Sensitivity analyses were conducted with the sample reduced to those with complete HRQoL and academic performance data at both timepoints (n = 184), and results were largely similar (Tables S2 and S3). In addition, stratified analyses by sex for the cross-lagged path in which academic performance at baseline predicts HRQoL at follow-up (β_CL-2) are presented in Fig. 2. In girls, math, language, physical education, and GPA at baseline were positively associated with HRQoL at follow-up, while in boys, all the academic grades indicators (except physical education), numeric ability, and the overall score of academic abilities at baseline were positively associated with HRQoL 24 months later.

The main findings of the present study indicated that HRQoL at baseline did not predict academic performance 24 months later, while all the academic grades and the overall score of academic abilities at baseline predicted HRQoL at follow-up in adolescents. Results of the stratified analyses by sex were largely similar. Specifically, in girls, math, language, physical education, and GPA at baseline predicted HRQoL 24 months later, while in boys, all the academic grades indicators (except physical education), numeric ability, and the overall score of academic abilities at baseline predicted HRQoL at follow-up. The current study contributes to the existing literature by exploring how HRQoL and a wide range of academic performance indicators are related to each other across time.

To our knowledge there is only one previous study analyzing the bidirectional association between HRQoL and academic performance in youth [15], which hampers further comparisons with other studies. In this study, Bortes et al. [15] analyzed a sample of 723 Swedish adolescents (15 years at baseline) showing a reciprocal relationship between HRQoL and GPA in girls, but not in boys. Importantly, the bidirectional associations found in girls did not point in the same directions, that is to say, HRQoL at baseline was positively associated with GPA 24 months later, while GPA at baseline was negatively associated with HRQoL at follow-up. Although these results differ from those found in the present study, our findings partially concur with prior evidence analyzing the bidirectional association between the broad higher-order construct of well-being and academic performance. Particularly, Steinmayr et al. [13] examined the bidirectional association between subjective well-being (including cognitive and affective components) and GPA in 290 German adolescents. In line with our results, the authors concluded that GPA at baseline was positively associated with changes in life satisfaction (the cognitive component of well-being) at 1-year follow-up. Similarly, one study conducted in 807 Chinese children (9 years) revealed that academic performance positively predicted subjective well-being 18 months later [27]. Likewise, Ng et al. [14] suggested that adolescents’ GPA exerted a positive effect on their life satisfaction 5 months later. However, contrarily to our findings, they also indicated that adolescents’ life satisfaction, albeit to a lesser extent, exerted a positive effect on their subsequent GPA. Also in contrast to our results, Wu et al. [12] reported that life satisfaction together with positive affect positively predicted adolescents’ academic performance (measured by a single index) 14 months later. Collectively, more longitudinal studies with longer designs and including the specific construct of HRQoL and multiple academic performance indicators are needed to elucidate the directionality of the HRQoL-academic performance association.

The reasons underlying why academic performance may positively predict HRQoL 24 months later cannot be elucidated in the present study; however, we suggest some mechanisms that may explain this association. First, based on the self-determination theory developed by Ryan and Deci in the seventies and eighties, academic performance may improve HRQoL through fulfilling the need for competence, which is considered a basic need that is essential for personal well-being [28]. Second, we speculate that health status could play an important role in this association. For instance, better academic performance may increase self-esteem with positive effects on well-being [27]. In addition, it is likely that adolescents with better academic performance are more health literate, which could lead them to take better care of themselves and have a better perception of their overall health status, positively influencing their HRQoL. Lastly, both academic performance and HRQoL may be influenced by diverse school-related variables. In fact, academic engagement, which is widely acknowledged as an important determinant of successful academic performance, has also shown to predict well-being [29], which may partially explain the academic performance-HRQoL association. In addition, although academic performance seems to be strongly linked to cognition, it also involves non-cognitive skills such as school environment, motivation, effort, attitude, interest, family support, personality, autonomy, self-perception, social acceptance, or teaching influence [30, 31], which in turn, may influence HRQoL.

Regarding stratified analyses by sex, some differences in both school performance and psychological-related issues between adolescent girls and boys could partially explain the sex-specific results found in the present study. In the case of physical education, which positively predicted HRQoL 24 months later in girls (but not in boys), we speculate that since girls tend to be more aware of their personal appearance and are more frequently dissatisfied by their body image than boys [32], girls could be specially interested in the health-related contents addressed in this subject, which in turn may positively influence their HRQoL. Meanwhile, the divergent results found for natural sciences, social sciences, and numeric ability, which were positively associated to HRQoL at follow-up in boys, but not in girls, could be explained not only by sex individual differences in terms of abilities, but also by the biological and cultural transmission of gender stereotypes [33]. In this sense, on the one hand, a previous study has stated that girls tend to achieve higher grades in subjects including emotional involvement (e.g., subjects related to humanities and social sciences) because of teachers' expectancies [33]. On the other hand, the belief that boys should get higher grades in mathematics than girls may increase their self-efficacy and self-concept towards this ability [33], which in turn may positively influence their HRQoL. Collectively, these issues and the fact that girls’ HRQoL declines more than boys’ over time [4, 34] may partially explain the sex-differences found in this longitudinal study.

Our findings have several important implications from an educational and public health perspective. In this sense, according to the OECD [35], ‘‘academic achievement that comes at the expense of students’ well-being is not a full accomplishment” (p. 4). Therefore, based on our results, policy makers should consider factors that influence adolescents’ academic performance to improve their HRQoL. In this context, education and health professionals should not only focus on study skills and pedagogy to improve adolescents’ academic performance, but also for instance on the promotion of healthy habits. Specifically, the promotion of compliance with healthy lifestyles (e.g., sleep, physical activity or diet) which have been positively associated with academic performance [36, 37] and the avoidance of unhealthy ones (e.g., screen time) which have been negatively associated with academic performance [38] could improve both their educational outcomes and their HRQoL.