Study design

In this cross-sectional study, a web-based questionnaire was used to collect data. The study results are reported according to the Strengthening the Reporting of Observational Studies in Epidemiology guidelines [25].

Sampling and participants

Participants were recruited using registered monitors of a research company and stratified sampling by sex, age, and region of 1st- to 3rd-year junior high school students and their mothers living in Japan. The sample size was calculated using G*Power version 3.1.9.6 (Mac OS X 10.7 to 14, Heinrich Heine Universität, Düsseldorf, Germany). With an effect size f ² of 0.25, α error probability of 0.05, and a power (1-β error probability) of 0.8, the estimated number of participants was approximately 1,800, accounting for the possibility of missing responses owing to inappropriate answers. Participants received an e-mail containing the URL of the survey. The informed consent document included all required information posted on the first page of the online questionnaire, and those who consented to the study could proceed to the questionnaire. An online survey was conducted from August 21–22, 2023. Written informed consent was obtained from both children and their mothers (opt-out procedure). This study was approved by the Daito Bunka University Research Ethics Committee (approval number: DHR23–007).

Eligibility criteria

The study included children aged 12–15 years (junior high school students, 1st to 3rd grade) and their mothers who consented and could participate with their children. Those who did not provide consent to participate in the study or lacked Internet access were excluded.

Measurements

Statistical analysis

Based on the median total score, HL-14 scores were divided into two levels (high and low), while HL-SAC was divided into three levels (high, moderate, and low). Continuous variables with a non-normal distribution were analyzed using the Mann–Whitney U test. Categorical variables were analyzed using the chi-squared test. Responses to the EQ-5D-Y were converted into index values using the Japanese value set [27]. Multiple linear regression analysis was performed using the total EQ-5D-Y score as the dependent variable. To assess multicollinearity, the correlation coefficient and variance inflation factor between each independent variable were confirmed. The independent variables included socioeconomic factors, the presence of disease, HL-SAC, and HLS-14. SAS software Version 9.4 (SAS Institute, Cary, NC, USA) was used for data analysis.

Association between adolescent’s HL and health-related behavior

The mean total HLS-14 score obtained in this study was similar to that from a national survey of individuals aged 40–49 years using the same scale [26]. The mean HL-SAC score (SD) for the total sample was 26.2 (5.0) points, with female individuals having a higher mean HL-SAC score than male individuals.

In a report from seven European countries [31], the HL-SAC scores of 25,425 adolescents aged 13–15 years indicated that the percentage of low HL ranged from 9.0% to 18.4%, and the percentage of high HL ranged from 13.1% to 35.0%. Additionally, a report from 10 European countries, including 14,590 adolescents aged 15 years, showed that 13.3% had low HL, 67.2% had moderate HL, and 19.5% had high HL. The mean (SD) total score was 30.78 (5.34), with the highest score being 33.93 (4.78) and the lowest being 29.88 (5.32) points [20].

These results suggest that the findings in Japan (low: 0.5%, high: 2.8%) show a lower bias toward low and high HL levels. The scores in Japan were higher than those observed in the European samples. Furthermore, the mean scores for theoretical knowledge and critical thinking in the present study were lower than those reported in European countries, indicating that critical thinking and the application of health information are particularly low in Japan. Results from the European Consortium Health Literacy Questionnaire, which assessed 1,054 Japanese adults aged 20–69 years, suggested that the skills required to “evaluate” and “apply” health information are challenging to acquire [32].

In this study, parental factors (education level and HL level), health-related behaviors (physical activity, sleep conditions), and parental communication (frequency of family meals, conversation during meals) were found to be associated with HL-SAC. Previous studies have also reported that sociodemographic characteristics and social class factors such as sex, race, family affluence, parental education levels, and economic status are linked to children’s HL and health outcomes [15,20,21,33]. HL is based on lifestyle habits during early childhood, and studies have been reported that examine the association between parental HL levels and children’s health behaviors, with a particular focus on food and nutrition, oral hygiene, and medical conditions [34]. Additionally, the quality of communication with parents and family support plays a crucial role in children’s health. Previous studies have shown that the quality of parental communication and family support are significant predictors of HL-SAC [21]. In this study, the frequency of family meals and communication with parents were found to be associated with HL. Previous studies have highlighted the importance of the school and community environment, including factors such as school satisfaction, teacher support, neighborhood livability, and available facilities [21,24]. These factors are relevant to child health and should be considered in future research.

Association of HL-SAC and -health-related HRQoL

The percentage of children who responded “no problem” to dimensions related to mental health was low across the five dimensions. Conversely, more than 80‒90% reported “no problem” for “looking after oneself” and “doing usual activities.” A Japanese EQ-5D-Y report involving 3,636 children and adolescents aged 8‒15 years [28], a multinational study including 2,809 children and adolescents aged 8‒19 years from European and African countries [22], and a study of 1,229 adolescents aged 12–18 years from Spain [35] all showed low scores for the dimension related to mental health. Moreover, Japanese children had lower average mental health scores than did adolescents. In adults, age, sex, and chronic diseases tend to affect EQ-5D-Y scores [36], but in children and adolescents, there is no difference in scores by sex or age, and it has been shown that the scores are influenced by some childhood diseases and household income [28]. This study did not show an association with household income, but did show an association with parental HL levels. Many studies have reported an association between socioeconomic status and HL levels, and parental attributes have been reported to affect children and adolescent’s HL [20–21]. Although no direct association was found in this study, such background may have had an influence. Additionally, insufficient sleep time and poor sleeping conditions were linked to lower mental health scores. A previous study showed that sleeplessness negatively affected both HRQoL and mental health [37]. Increasing parental HL is expected to foster healthy lifestyle habits in children, ultimately improving HRQoL.

Limitations

This study has some limitations. The EQ-5D-Y was used because the research participants were a relatively healthy general population without severe disabilities or illnesses. However, previous studies have shown that its sensitivity is low, it cannot assess psychosocial health, the three-level option has weak detection power for moderate health disorders, and it lacks discrimination. Additionally, the ceiling effect is high, with scores tending to cluster at the upper end. This study included participants registered with a survey company and may have represented a population with high e-HL, which may have caused selection bias. The findings of cross-sectional studies using the HL-SAC among middle school students are not available for Japan; therefore, further research should be conducted to compare the results across populations and countries.