Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
www.jahonline.org
Original article
A B S T R A C T
Purpose: To assess the contribution of life events (LEs) on psychosomatic complaints in adolescents/youths
taking into account a set of socioeconomic variables.
Methods: We tested a conceptual model implemented with structural equation modeling on longitudinal data
from a representative sample of adolescents/youths and parents. Psychosomatic complaints were measured by
the Health Behaviour in School-aged Children scale and hypothesized to be affected by: (a) contextual factors at
distal level: nancial resources, home life and social support (KIDSCREEN), and parent baseline mental health
(SF-12); (b) triggering factors: LEs (Coddington Life Events Scales, with two typologies: desirability and familiarity); (c) intermediate factors: same as distal level but measured at follow-up; (d) immediate cause: mental health
at proximal level (Strengths and Difculties Questionnaire at baseline and follow-up); and (e) gender.
Results: The structural model yielded a good t (Comparative Fit Index .95, TuckerLewis Index .93, Root
Mean Square Error .04). Boys showed more psychosomatic complaints than girls ( .40, p .05). Girls
reported experiencing more LEs (p .05). Only undesirable LEs showed a signicant direct negative effect on
psychosomatic complaints, which became nonsignicant when mediated by home life and mental health.
Undesirable LEs had a remaining indirect effects on psychosomatic complaints (indirect .10, p .05) via
Home Life and Mental health, which were protective factors ( .41 and .15, p .05).
Conclusions: The experience of undesirable LEs increases the probability of psychosomatic complaints, but
the nal effect would be determined by previous levels of home life and mental health stability.
2011 Society for Adolescent Health and Medicine. All rights reserved.
* Address correspondence to: Jordi Alonso, M.D., Ph.D., Health Services Research Unit, IMIM-Hospital Del Mar, Barcelona Biomedical Research Park, Doctor
Aiguader, 88 08003 Barcelona, Spain.
E-mail address: jalonso@imim.es (J. Alonso).
The authors declare they have no conicts of interest.
1054-139X/$ - see front matter 2011 Society for Adolescent Health and Medicine. All rights reserved.
doi:10.1016/j.jadohealth.2010.11.260
200
Table 1
Summary of the time sequence of contextual variables for Berntsson model of psychosomatic complaints
Distal level
Intermediate level
Proximal level
Economic factors: international and national political systems, social, and cultural systems
Socioeconomic level of the family: education, profession, family structure, and employment
Familys economic resources
Family activities
Parents sense of coherence
Support from society and relatives
Health status of parents
School satisfaction
Contacts with peers
Mental stability
Social competence
201
Table 2
Variables and assessment instruments used from the KIDSCREEN follow-up study
Observed variables
Instruments
Recall period
Respondent
Item example
Psychosomatic
complaints
HBSC
Follow-up
Previous 6
months
Adolescents/youths
Life events
Financial resources
Adolescents/youths
Adolescents/youths
Adolescents/youths
Parents/proxy
Parents/proxy and
adolescents/youths
202
Figure 1. Etiological model of relationships among risk factors that might affect psychosomatic complaints in the KIDSCREEN follow-up study.
Statistical analysis
The nal dependent outcome was psychosomatic complaints
in adolescents/youths as measured at follow-up (in 2006). A
structural equation modeling was used to: (a) test whether intermediate variables acted as partial or full mediators for the
inuence of LEs on psychosomatic complaints; (b) establish the
magnitude of the relationship among variables; and (c) jointly
assess which factors most strongly inuenced psychosomatic
complaints in adolescents/youths [24].
One structural model for each LE category was constructed to
predict the direct and indirect effect of variables in the model. All
models took into account the time sequence of the different
variables: distal, triggering, intermediate, and proximal. As both
continuous and ordinal variables were included, the estimator of
choice was Unweighted Least Squares Estimation with mean and
variance corrections (ULSMV), to provide robust p-values and
standard errors [25]. Model t was assessed using Comparative
Fit Index (CFI), TuckerLewis Index (TLI), and Root Mean Square
Error (RMSEA) [26]. Conventional cut-off criteria indicating good
t are CFI and TLI greater than .90, whereas RMSEA typically used
a cut-off value of .05. The model was implemented and estimated using Mplus 5.2 [27].
Final parameters for continuous variables are parameterized as
standardized regression coefcients [28]. Mental health stability
was treated as ordered categorical, with four ordered categories
assuming a proportional odds logistic model [29]. To simplify interpretation, mental health stability parameter estimates were reparameterized to reect changes in odds ratios [29].
There were marginally signicant differences between baseline and follow-up samples regarding mean age and socioeconomic level, which is a common result in longitudinal studies of
this type [30]. Additionally, attrition was compatible with a missing completely at random (MCAR) data pattern. Littles test [31]
for MCAR was nonsignicant (p .26), indicating that results
Table 3
Sample descriptives on baseline and follow-up. KIDSCREEN follow-up study
Observed variables
Total (n 331)
Mean/percentage
Baseline
Adolescents and youths
Financial resources
Social support and peers
Home life
Any mental health disorder: possible
and probable cases
Parents
Age
Mother responding
Parents mental health status
Follow-up
Adolescents and youths
Age
Psychosomatic complaints
Headache
Stomach ache
Backache
Feeling low
Irritable or bad-tempered
Feeling nervous
Sleeping difculties
Dizziness
LCUs global life events
LCUs undesirable life events
LCUs desirable life events
LCUs family life events
LCUs extrafamily life events
Social support and peers
Home life
Any mental health disorder: possible
and probable cases
Mental health stability
Worsened
Remained poor mental health
Improved
Stayed healthy
Parents
Parents mental health status
a
SD
52
54.5
52.22
17%
8.7
9.7
9.7
.4
42.22
77.5%
52.4
6.8
8
15.4
34.6
4.3
4.5
4.2
4.4
4.1
3.8
4.4
4.6
48.8
22.8
22.2
14.4
33
51.4
49.6
18.9%
2.8
5.2
.9
.8
1.1
.9
1
1.2
1
.8
53.5
36.3
25.2
22.9
42.2
8.8
9.1
.5
12.7%a
6.9%
9.4%
71%
48.3
10.3
203
204
a) Financial
resources
0.32 (0.04)
0.21(0.05)
0.87
0.60
0.20 (0.05)
a) Parents
mental health
status at
baseline
0.34
0.48 (0.05)
(0.04)
a) Home life
at baseline
Distal
level
a) Social
support and
peers at
baseline
0.37 (0.05)
b)Undesirable life
events
0.39 (0.05)
Triggering
factor
-0.21 (0.05)
0.83
0.82
0.67
c) Parents
mental health
status at
follow-up
c) Home life at
follow-up
0.25 (0.05)
c) Social
support and
peers at
follow-up
Intermediate
level
-0.44 (0.08)
1.54 OR increase for better pronostic.
0.83
0.09 (0.05)*
0.41 (0.05)
Proximal
level
d) Mental
health
stability
-0.15 (0.06)
0.73
Psychosomatic
complaints
e) Gender
0.40 (0.10)
for boys
Figure 2. Structural parameters for the model of undesirable life events. KIDSCREEN follow-up study. Note: Dashed lines represent nonsignicant paths. Double headed
arrows represent the residual variance of the endogenous variables. Only signicant standardized parameters and their standard errors are shown (p .05), * except (p
.08). Categorical variable parameter interpretation as odds ratios is shown in italics. Fit indexes: 2 22.87, df 15, p .09; Comparative Fit Index .95; TuckerLewis
Index .93; Root Mean Square Error: .04.
sidered in this study such as different social roles, coping mechanisms, and lifestyles could also explain the greater frequency of
complaints among boys [34]. However, this result should be
considered in future research.
One limitation of our study was a somewhat low response
rate at the follow-up (54%). Although this level of response is
typical in postal surveys [40], it is important to indicate that
despite attrition, the sample was representative of the Spanish
population in terms of age and gender when compared to census
data [12]. In addition, individuals lost in the follow-up sample
were shown to be compatible with the assumption of missing at
random data, which also points at good external validity of our
results. Second, the CLES questionnaire was used to collect LEs
which had occurred up to 3 years before the assessment date.
Likewise, the SDQ was originally developed for children aged 16,
whereas our sample included adolescents/youths aged up to 21.
However, a pilot test performed before the follow-up study
showed that using these questionnaires outside their original
age range was adequate [12]. Finally, some factors which were
signicant mediators in the Bertssons model, were not included
in our study, and could have had an effect on the results [9]. They
should be considered for inclusion in future studies.
Despite these limitations, the present study is one of the few
based on a general population sample of adolescents and youths.
Most research into LEs and psychosomatic complaints has been
carried out in adults, in clinical settings or in convenience samples of students. Our results add population representative information on ages that has been seldom studied in the published
data. We also found support for an etiological sequence and
simplied previous models in terms of number of variables studied while explaining a high degree of the variance.
One straightforward implication of the present study is that
intervention programs should aim to increase family support as
well as to improve the nancial and social resources available to
families. These factors have been shown to be nuclear in the
etiological process from undesirable LEs to psychosomatic complaints. Future research should assess whether other factors can
inuence the frequency of psychosomatic complaints, and to
conrm the increased vulnerability of boys.
Acknowledgments
The authors are grateful to CIBER-ESP and the Ministerio de
Educacin for providing a grant for E. Villalonga-Olives to prepare the present manuscript at the University Clinic HamburgEppendorf. This research was supported by FIS Expte PI042315,
DURSI-GENCAT (2005-SGR-00491), and Ministerio de Ciencia e
Innovacin FSE (JCI-2009-05486).
JMV, LR, and JA participated in the conception and design of
the study. EVO, CGF, ME, and JAP analyzed the data. EVO, JMV,
MH, MF, URS, LR, and JA participated in the drafting of the article.
All authors contributed to a critical revision of the manuscript
and made a substantial contribution to its content, and read and
approved the nal manuscript.
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
References
[31]
[1] Aro H, Hanninen V, Paronen O. Social support, life events and psychosomatic symptoms among 14-16-year-old adolescents. Soc Sci Med 1989;29:
1051 6.
[2] Greene JW, Walker LS. Psychosomatic problems and stress in adolescence.
Pediatr Clin North Am 1997;44:155772.
[3] Gerber M, Puhse U. Dont crack under pressure!Do leisure time physical
activity and self-esteem moderate the relationship between school-based
stress and psychosomatic complaints? J Psychosom Res 2008;65:3639.
[4] Fekkes M, Pijpers FI, Verloove-Vanhorick SP. Bullying behavior and associations with psychosomatic complaints and depression in victims. J Pediatr
2004;144:1722.
[5] Laftman SB, Ostbert V. The pros and cons of social relations: An analysis of
adolescents health complaints. Soc Sci Med 2006;63:61123.
[6] Ackerman BP, Kogos J, Youngstrom E, et al. Family instability and the
problem behaviors of children from economically disadvantaged families.
Dev Psychol 1999;35:258 68.
[7] McLoyd VC. Socioeconomic disadvantage and child development. Am Psychol 1998;53:185204.
[8] Stock C, Kucuk N, Miseviciene I, et al. Differences in health complaints
among university students from three European countries. Prev Med 2003;
37:535 43.
[9] Berntsson LT, Kohler L, Gustafsson JE. Psychosomatic complaints in schoolchildren: A Nordic comparison. Scand J Public Health 2001;29:44 54.
[10] Committee on Evaluation of Childrens Health, National Research Council,
Institute of Medicine of the National Academies. Childrens health, the
[32]
[33]
[34]
[35]
[36]
[37]
[38]
[39]
[40]
205