OBJECTIVES: To assess mental health and health-related quality of life (HRQoL) of children and adolescents in Spain and to investigate the existence of a social gradient in mental health and HRQoL.
METHODS: Within the Spanish National Health Survey (2006), the parents’ version of the Strengths and Difficulties Questionnaire was administered to a population aged 4 to 15 years, and the parents’ version of the modified KIDSCREEN-10 Index was given to a population aged 8 to 15 years. Sociodemographic data and information on family structure, socioeconomic status, health status, and discrimination were collected. Regression models were developed to analyze associations of socioeconomic status with mental health and HRQoL.
RESULTS: A total of 6414 children and adolescents aged 4 to 15 years participated. Mean Strengths and Difficulties Questionnaire score was 9.38 (SD, 5.84) and mean KIDSCREEN-10 Index score (n = 4446) was 85.21 (SD, 10.73). Children whose mothers had a primary school education (odds ratio [OR]: 1.37; 95% confidence interval [CI]: 1.29–1.46) or a secondary education (OR: 1.21; 95% CI: 1.14–1.29) presented poorer mental health than those whose mothers had a university degree. Children from disadvantaged social classes (IV–V) showed slightly poorer HRQoL scores (OR: 0.98; CI: 0.97–0.99) than the remaining children.
CONCLUSIONS: There is a social gradient in the mental health of children and young adolescents in Spain. No social gradient was found for HRQoL, although children from families of disadvantaged social classes had slightly worse HRQoL scores than their counterparts from more advantaged classes.
- CI —
- confidence interval
- ENSE —
- Encuesta Nacional de Salud de España (Spanish National Health Interview Survey)
- HRQoL —
- health-related quality of life
- KS —
- OR —
- odds ratio
- SDQ —
- Strengths and Difficulties Questionnaire
- SES —
- socioeconomic status
- TDS —
- Total Difficulties Score
What's Known on This Subject:
The importance of and interest in childhood mental problems have increased worldwide. There are few population studies on child and adolescent mental health and health-related quality of life (HRQoL).
What This Study Adds:
A social gradient was found in childhood mental health according to maternal education level and social class, but none was found in HRQoL, although children from disadvantaged social classes had somewhat lower HRQoL scores than their more advantaged counterparts.
Mental health problems are prevalent in childhood and adolescence and represent a considerable burden of disease. They are an important health issue that affects young people’s relationships with family, in school, and with society in general and are associated with intensive use of health and social services. Mental health disorders affect 10% to 20% of the child and adolescent population.1,2 Moreover, more than half of the mental health problems in the general population arise in childhood, and there is a continuity between these conditions and those present in adult life.3–5
Considerable effort has been made to standardize assessment of mental health in the general population of children and adolescents.6–10 The factors associated with pediatric mental health problems include gender, age, psychiatric disease in the child’s parents, immigrant status, single-parent family, family difficulties, and lower family socioeconomic status (SES), among others.1,11–13 Most population studies on child and adolescent mental health have reported poorer mental health in children with a low family education level or social class.11,13–16
Health-related quality of life (HRQoL) is an important health outcome measure that is being used increasingly in the infantile and juvenile age groups. Its interest lies in the capacity to reflect the multidimensionality of health in both general populations and disease-specific populations.17,18 Use of HRQoL as an outcome measure of health status can be useful in daily clinical practice, in effectiveness studies, and as an indicator of the quality of care received.19,20 Furthermore, identification of individuals with poorer HRQoL enables early detection of health needs. In the pediatric age, at a time when chronic disease is less prevalent than later in life, systematic use of this measure in community-based studies can enable recognition of neglected groups who may be disadvantaged from the perspective of health and social care.21–25 Children from disadvantaged social groups generally achieve poorer scores regarding both mental health and HRQoL.
Social inequalities in child health refer to the health differences between populations of children defined according to their social condition, economic status, demography, or geography. Health inequities are determined socially and are unfair and modifiable differences. Recent interest in this line has focused on the concept of a social gradient in health; that is, the greater the social disadvantage, the poorer is health. Social gradients in child health have been described for a number of diseases.26
In 2006, the Spanish National Health Interview Survey (ENSE, Encuesta Nacional de Salud de España) incorporated sections on mental health and HRQoL in the Children’s Questionnaire. The ojectives of this study are to determine the status of mental health and HRQoL in the Spanish infantile-juvenile population, to analyze the associated socioeconomic and psychosocial factors, and to specifically investigate the existence of a social gradient in mental health and HRQoL in this population.
Sample Selection and Data Sources
The study information was obtained from the 2006 ENSE Children’s Questionnaire (0–15 years) and the Household Questionnaire. The ENSE is a nationally representative health interview survey conducted every 4 years on behalf of the Spanish Ministry of Health, Social Policy, and Equality. The fieldwork was performed by the National Statistics Institute (Instituto Nacional de Estadística) between June 2006 and June 2007. Multistage stratified sampling was conducted, with the first-stage units being census tracts and the second units being the main family households. Within each household with children or adolescents (aged 0–15 years), 1 was randomly selected for the Children’s Questionnaire. The sample is independent and representative of each autonomous community (an organizational division of the Spanish territory) and is not proportional; hence, the results were weighted to obtain national indicators.
The information was collected by personal interview. The ideal respondent for the Children’s Questionnaire was the person best informed about matters related to the child’s health and health care, usually the child’s mother or father.
The sample size for the total ENSE was estimated at 31 300 households, distributed in 2236 census tracts. The response obtained was 96% of the theoretical sample (including 31% of substitutions) and was >90% in all the autonomous communities. The methods used by ENSE and the database can be accessed through the Web site of the Ministry of Health, Social Policy, and Equality.27 The current study includes a sample of children and adolescents aged 4 to 15 years in the mental health study and a population aged 8 to 15 years in the HRQoL study.
The dependent variables were mental health and HRQoL of the pediatric population. Mental health was assessed by using the parents’ version of the Strengths and Difficulties Questionnaire (SDQ).28 The SDQ is a 25-item questionnaire that detects probable cases of mental disorder and behavioral disorder in children and adolescents aged 4 to 16 years. It includes 5 scales, containing 5 items: 4 scales measure negative aspects of mental health (emotional symptoms, behavioral problems, hyperactivity, and peer problems), and the single prosocial conduct scale records positive behavior. Each item has 3 possible responses (not true, somewhat true, or certainly true), and scoring is from 0 to 10. The sum of the scores on the 4 scales related to negative aspects yields the Total Difficulties Score (TDS-SDQ), with a range of 0 to 40. The higher the TDS-SDQ score, the poorer the child’s mental health is.29
Evaluation of HRQoL used the shortest parent-reported version of the KIDSCREEN (KS) instrument, the KS-10 Index. KS is a generic HRQoL questionnaire that has been transculturally developed in 13 European countries for the population of children and adolescents aged 8 to 18 years. The instrument includes 3 versions: KS-52 (52 items and 10 dimensions), KS-27 (27 items and 5 dimensions), and the KS-10 Index, which contains 10 items that score as an index.30 The validity and reliability of KS have been proven in the European population.31–33 The questions on the KS-10 Index present a recall period of 1 week and a 5-point Likert response scale. In the current study, the version of the KS-10 Index from the European Eurobarometer study34 was computed to facilitate comparison with that study and with future versions of ENSE. One of the KS-10 items has been replaced by a general health question, while maintaining the psychometric properties of the original version as much as possible. The modified KS-10 Index scores were transformed to a scale of 0 to 100: the higher the score, the better the HRQoL.
The independent variables analyzed included age (3 categories: 4–7, 8–11, and 12–15 years), gender, place of residence (17 autonomous communities and 2 autonomous cities analyzed jointly), place of birth (Spain or outside Spain), and family SES with regard to social class and maternal educational level. Social class was categorized according to the occupation of the home’s main wage earner based on the classification proposed by the Spanish Society of Epidemiology.35 Social Class I includes managerial and senior technical staff and liberal professionals; class II includes intermediate occupations and managers in commerce; class III includes skilled nonmanual workers; class IV includes skilled (IVa) and partly skilled (IVb) manual workers; and class V includes unskilled manual workers. For the study analysis, social classes were grouped into 3 categories: I + II (advantaged), III (middle), and IVa + IVb + V (disadvantaged). The educational level referred to the highest level of schooling completed by the mother, categorized into 3 groups: primary school or less, secondary school, and university degree. The family structure was included in the analysis (single-parent family, 2-parent family, or other), as well as the number of family members: 2 to 3, 4 to 5, or >5 members. The variables related to health and the use of health services included activity limitations (yes/no) in the 6 months preceding the interview, declaration of any chronic condition (yes/no) by using a list of relatively common chronic diseases occurring in the pediatric population, and hospitalizations (yes/no) in the previous year. Lastly, a single question on perceived discrimination was included, asking whether the child had experienced discrimination owing to his or her gender, ethnic group, country of origin, social class, or religion.
Mean scores on the TDS-SDQ and modified KS-10 Index were calculated for the overall sample and for each category of the independent variables and were compared by using the t test or analysis of variance, depending on the nature of the variable. To facilitate interpretation of the scores, effect sizes between groups divided according to education level and social class were calculated for the TDS-SDQ and the KS-10. Effect sizes of 0.2–0.5, 0.51–0.8, and >0.8 were considered small, medium, and large, respectively.36 To analyze the influence of context variables on mental health and HRQoL, we had contemplated the use of multilevel analysis, considering the autonomous community of residence as level 2; however, after building the respective empty models, the intraclass coefficient was found to be <3%, and the multilevel approach was rejected. Poisson regression models were used to analyze associations of maternal education level and social class with mental health and HRQoL in the overall sample, because it was the method that best fitted the outcome variables. The final models were adjusted by progressively including variables that were statistically significant in the bivariate analysis while following the proposed conceptual model. That is, the basic (crude) models were initially calculated, and subsequently, the sociodemographic, family, and discrimination variables were included. Health status variables were excluded in the adjusted models for HRQoL. Statistical significance was set at 0.05.
To achieve sample representativeness to national and regional levels in the results, weights were used for each individual. Statistical analyses were performed with SPSS version 18 for Windows (SPSS Inc, Chicago, IL) and the Stata module to estimate variance of complex-sample survey data (version 11; Stata Corp, College Station, TX).
A total of 6414 children and young adolescents aged 4 to 15 years were included in the mental health analysis (93.8% of the sample), and 4446 aged 8 to 15 years were included in the HRQoL study (94.8% of the sample). The characteristics of the sample are shown in Table 1. The sample was mainly composed of young people born in Spain (90.3%), living in 2-parent families (89.9%), in social class IV to V (51.6%), and with a secondary school maternal educational level (45.4%). Mean score was 9.38 (SD, 5.84; interquartile range: 5.0–13.0) on the TDS-SDQ and 85.21 (SD, 10.73; interquartile range: 80.0–92.5) on the modified KS-10 Index. Several statistically significant differences were found for mental health and HRQoL in the bivariate analysis (Table 2).
Effect sizes for mental health and HRQoL according to the family level of education and social class are shown in Table 3. The largest effect size for TDS-SDQ scores was observed between children from families with a primary level of education and those from families with a university degree (0.52).
The multivariate analysis of mental health and HRQoL is shown in Table 4. There was a statistically significant social gradient for mental health in both maternal level of education and social class. Higher scores (poorer mental health) were seen in children from more disadvantaged families: children of mothers with a primary school education (odds ratio [OR]: 1.37; 95% confidence interval [CI]: 1.29–1.46) and those of mothers with a secondary school education (OR: 1.21; 95% CI: 1.14–1.29) had worse mental health status than those whose mothers had university degrees. Similar results were seen for social class: children from social class III (middle; OR: 1.15; 95% CI: 1.08–1.22) and those from the more disadvantaged classes (IV–V; OR: 1.31; 95% CI: 1.24–1.38) had worse mental health status than those from advantaged social classes. This social gradient persisted after adjusting for several variables related to sociodemographics, family, health, and discrimination. Maternal primary education (OR: 0.99; 95% CI: 0.97–1.01), and secondary education (OR: 1.00; 95% CI: 0.98–1.01) showed nonsignificant association with HRQoL. Nevertheless, children from families in social class IV to V presented slightly poorer HRQoL scores than their counterparts in social class I to II (OR: 0.98; 95% CI: 0.97–0.99).
In this study, a social gradient was found for childhood mental health according to the maternal education level and family social class. Children born to disadvantaged families presented poorer mental health. This social gradient was not found for HRQoL; however, children from families of social classes IV to V showed slightly poorer HRQoL scores than their counterparts from more advantaged classes. The minimal differences in HRQoL according to social class, although statistically significant, do not seem to be relevant from the clinical and epidemiologic viewpoints.
The mental health results found in this study are in keeping with those of previous European and international studies.1,12,14,16 The youngest children and those from single-parent, immigrant, and disadvantaged families had worse mental health scores. Similar findings were obtained in the recent Catalan Health Survey,37 which also showed a clear social gradient as determined by family social class and mother’s educational level. The adult population in families with a low level of education or disadvantaged social class presents higher morbidity rates and chronic physical and mental conditions.38 These circumstances could limit their economic resources and the available time for parenting activities and caring for their children.39 The lower income might also limit the children’s access to educational and social activities.
Effect sizes found for the association of maternal level of education and social class and mental health were low to moderate, although they represent at least the minimal important difference according to international consensus.36 Moreover, policy makers and clinicians should be aware of the magnitude of social inequalities in mental health. The mental health of >1 million Spanish children aged 4 to 15 years would improve by a medium to large difference on the TDS-SDQ (effect size, 0.52) if they achieved the same mental health status as children from families in the highest educational category.
With respect to HRQoL, the results of this study concur in great part with those reported by other authors21,34,40–43; however, we found minimal differences in HRQoL according to social class and no differences according to maternal education level. The absence of differences according to maternal level of education found in the current study may be partly explained by differences in the study methods and instruments used. For example, the European Health Behavior in School-aged Children 2005/2006 Survey included different age ranges and used the self-administered version of the KS-10 Index and the Family Affluence Scale to indicate family SES.41 In the German BELLA Study,42 conducted between 2003 and 2006, the age investigated (7–17 years), the HRQoL instrument used (self-administered German Questionnaire for Measuring Health-Related Quality of Life in Children and Adolescents, the KIND-R questionnaire), and the indicator of family SES (Winkler Index) all differed from those used in the current study. Similarly, von Rueden et al43 investigated HRQoL in 7 European countries in a sample with a greater representation of adolescents, using the full self-administered version of the KS (KS-52) and the Family Affluence Scale, in addition to educational level. All these studies reported a socioeconomic gradient in the HRQoL of children and adolescents. The differences relative to the current study may have also originated in part from our use of parent proxy-reporting for the child’s HRQoL. It is known that agreement between parents and children is lower when subjective aspects, such as well-being, are being assessed than when more observable aspects are analyzed, such as behavior. The childhood living conditions could be another factor related to the absence of social gradients in HRQoL documented in the current study, and, to some extent, the characteristics of the Spanish health care system. Health care in Spain is financed by taxes and provides universal coverage to the population. Although several social inequalities persist, as is also the case in Northern European countries, health care accessibility would be less unequal than in countries such as the United States, for example, where 13% of uninsured children with serious mental health problems are more prone to present unmet needs and poorer health outcomes than their insured counterparts.15,16
One of the main limitations of this study, as was mentioned, is that the information was obtained from the child’s main caregiver, which could imply reporter bias. Nonetheless, in surveys on health in children, it is a universally accepted practice to collect information from the main caregiver, usually the child’s mother. Another limitation derives from the cross-sectional design of the ENSE, which does not allow directionality to be determined in the associations found. The regression models showed acceptable fit, although there were only marginal differences between the basic (crude) models and the adjusted models. This fact could be related to the independent effect of social determinants on child mental health. Confounding and interaction effects were ruled out. The inclusion of other variables that were absent in this study, such as parental psychiatric morbidity, may improve these models and should be considered in future studies. Finally, although differences were found in mental health and HRQoL according to gender in the current study, these differences seemed to have no influence on SES inequalities.
One of the strengths of the study resides in the use of reproducible and comparable instruments and survey methods. Mental health and HRQoL were assessed with the parents’ version of the SDQ and KS-10, respectively. Both instruments are well recognized and widely used internationally, so the study can contribute to international research by enabling comparisons with results from other countries.
Future studies should address the mechanisms producing these social inequalities in health. The addition of sections on mental health and HRQoL in the ENSE Children’s Questionnaire provides a valuable resource for comprehensive monitoring of health in the target age group. Detection of poor scores could alert to a loss of health (physical, mental, and social) and enable implementation of measures to avoid health problems later in life. It will also facilitate evaluation of the policies needed for an effective approach to the social gradient in mental health, by early interventions on the social environment closest to the child.
This study is the first nationwide one developed in Spain investigating mental health and HRQoL in children and young adolescents. The risk factors and social gradient related to mental health coincide with the data obtained from other Western countries: poorer mental health scores are seen in young people belonging to underprivileged families. No social gradient was found for HRQoL, although children from families in disadvantaged social classes had slightly poorer HRQoL scores than their counterparts from more advantaged classes.
The authors thank Belén Sanz, PhD, Research Fellow at the National School of Health, Instituto de Salud Carlos III, and Mónica Suárez, from the Ministry of Health, Social Policy and Equality, for their invaluable help in developing the database.
- Accepted April 24, 2012.
- Address correspondence to Luis Rajmil, PhD, Agència d'Informació, Avaluació i Qualitat en Salut (AIAQS), Catalan Agency for Health Information, Assessment and Quality, Carrer de Roc Boronat, 81-95 (2na Planta), 08005 Barcelona, Spain. E-mail: ;
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: No external funding.
- 1.↵Meltzer H, Gatward R, Goodman R, Ford T. The Mental Health of Children and Adolescents in Great Britain. London, United Kingdom: Social Survey Division of the Office for National Statistics on behalf of the Department of Health, the Scottish Health Executive and the National Assembly for Wales; 1999
- Friedman RM,
- Katz-Leavy JW,
- Manderscheid RW,
- Sondheimer DL
- Kessler RC,
- Berglund P,
- Demler O,
- Jin R,
- Merikangas KR,
- Walters EE
- Reef J,
- Diamantopoulou S,
- van Meurs I,
- Verhulst FC,
- van der Ende J
- Goodman SH,
- Hoven CW,
- Narrow WE,
- et al
- 9.Mata Zubillaga D, Torres Hinojal C, Carro Serrano A, Ortega García E, Suárez Rodríguez A. Uso del Cuestionario de Capacidades y Dificultades (SDQ) como instrumento de cribado de trastornos psiquiátricos en la consulta de pediatría de Atención Primaria. Bol Pediatr. 2009;49(209):259–262
- Rothenberger A,
- Becker A,
- Erhart M,
- Wille N,
- Ravens-Sieberer U,
- BELLA study group
- Hölling H,
- Kurth BM,
- Rothenberger A,
- Becker A,
- Schlack R
- Lipman EL,
- Offord DR,
- Boyle MH
- 15.↵Kataoka SH, Zhang L, Wells KB. Unmet need for mental health care among U.S. children: variation by ethnicity and insurance status. Am J Psychiatry. 2002;159(9):1548–1555
- Starfield B
- 27.↵Ministerio de Sanidad M, Política Social e Igualdad Encuesta Nacional de Salud de España 2006. Available at: www.msps.es/estadEstudios/estadisticas/encuestaNacional/encuesta2006.htm. 2011. Accessed November 28, 2011
- 29.↵SDQ. Information for researchers and professionals about the Strengths & Difficulties Questionnaires. Available at: www.sdqinfo.com. Accessed July 2, 2011
- The KIDSCREEN Group Europe
- Ravens-Sieberer U,
- Auquier P,
- Erhart M,
- et al.,
- European KIDSCREEN Group
- Ravens-Sieberer U,
- Erhart M,
- Rajmil L,
- et al.,
- European KIDSCREEN Group
- 34.↵The Gallup Organization. Flash Eurobarometer: Parents’ Views on the Mental Health of Ttheir Child: Analytical Report. Available at: www.ec.europa.eu/public_opinion/flash/fl_246_en.pdf. Accessed December 3, 2011
- Cohen J
- Borrell C,
- Rue M,
- Pasarín MI,
- Rohlfs I,
- Ferrando J,
- Fernandez E
- Belsky J,
- Bell B,
- Bradley RH,
- Stallard N,
- Stewart-Brown SL
- Ravens-Sieberer U,
- Wille N,
- Erhart M,
- et al.,
- BELLA study group
- von Rueden U,
- Gosch A,
- Rajmil L,
- Bisegger C,
- Ravens-Sieberer U
- Copyright © 2012 by the American Academy of Pediatrics