American Journal of Gastroenterology
C ?2004 by Am. Coll. of Gastroenterology
Published by Blackwell Publishing
Increased Somatic Complaints and Health Care Utilization
in Children: Effects of Parent IBS Status and Parent
Response to Gastrointestinal Symptoms
Rona L. Levy, Ph.D., M.P.H., William E. Whitehead, Ph.D., Lynn S. Walker, Ph.D., Michael Von Korff, Sc.D.,
Andrew D. Feld, M.D., J.D., Michelle Garner, M.S.W., and Dennis Christie, M.D.
University of Washington, Seattle, Washington
OBJECTIVES:Irritable bowel syndrome (IBS) runs in families. The aims of this study were (i) to exclude biased
perception by a mother with irritable bowel as the explanation for increased gastrointestinal (GI)
symptoms in their children, (ii) to determine whether non-GI as well as GI symptoms run in families,
and (iii) to determine whether parent IBS status and solicitous responses to illness exert
independent effects on children’s symptom reports, medical clinic visits, and school absences.
METHODS:Two hundred and eight mothers with irritable bowel and their 296 children (cases: average age
11.9 yr; 48.6% male) and 241 nonirritable bowel mothers and their 335 children (controls: 11.8 yr;
49.0% male) were interviewed. Other factors assessed were stress, mother’s and child’s
psychological symptoms, child’s perceived competence, pain coping style, age, and sex. Children
were interviewed apart from their parents.
RESULTS:Case children independently reported more frequent stomachaches (F(591) = 9.22; p = 0.0025)
and non-GI symptoms (F(562) = 21.03; p < 0.001) than control children. Case children also had
more school absences (F(625) = 26.53; p < 0.0001), physician visits for GI symptoms (F(602) =
8.09; p = 0.005), and non-GI clinic visits (F(602) = 27.92; p < 0.001) than control children.
Children whose mothers made solicitous responses to illness complaints independently reported
more severe stomachaches (F(590) = 11.42; p < 0.001), and they also had more school absences
for stomachaches (F(625) = 5.33; p < 0.05), but solicitous behavior did not significantly impact
non-GI symptom reporting, clinic visits, or school absences. Differences between cases and controls
remained significant after adjusting for potential moderators.
CONCLUSIONS:(i) Frequent GI complaints in children whose mothers have irritable bowel are not explained by the
mother’s biased perceptions; (ii) children of mothers with irritable bowel have more non-GI as well
as GI symptoms, disability days, and clinical visits; (iii) and parent IBS status and solicitous
responses to illness have independent effects on the child’s symptom complaints.
(Am J Gastroenterol 2004;99:1–10)
Many patients with irritable bowel syndrome (IBS) report
experiencing more symptoms, including many nongastroin-
miss more work or school due to illness (3); and they visit
tions in IBS patients is childhood learning through modeling
and reinforcement. (1, 5) However, the research exploring
how this learning has occurred has been methodologically
A series of experiments from our group supports the hy-
pothesis that childhood social learning contributes to the de-
more symptoms, to consult doctors more often, and to take
showed that: (i) adult IBS patients are more likely to recall
that their parents reinforced illness behavior by giving them
toys or gifts or special privileges when they were ill as a
child, and that their parents modeled illness behavior (1, 5);
(ii) the effects of modeling (emulating the behavior of others
who are role models) and reinforcement are not explained
by the IBS patient’s stress or psychological symptoms (5);
IBS to run in families, concordance rates between parents
and children, compared to those between identical twins, in-
an equal or greater influence on the development of IBS (6).
Using the automated database of a large health maintenance
organization, we also showed that the children of patients
with IBS are more likely than control children to be brought
2 Levy et al.
to the pediatrician’s office for diarrhea and abdominal pain.
(4) However, we also found (4) that they are brought to the
raising questions about the specificity of the environmental
influences that take place.
A significant limitation of our previous work and of the
studies by other investigators (7, 8) is that data were not col-
lected independently from both the parent and the child on
matic symptoms. Thus, it is possible that an IBS patient may
have biased perceptions about the illness behavior displayed
by their parents (1, 5–7) and/or their children (4). Our previ-
ous finding that the children of parents with IBS are brought
to the medical clinic more often for GI symptoms (4) might
also reflect the bias of what the parents, not children, attend
to, since it is the parents who decide when to bring a child to
the medical clinic.
The current study was designed to address this concern by
interviewing children separately from their parents to deter-
mine what symptoms they report. We assessed both GI and
non-GI symptoms of the child. A validated questionnaire (8)
was used to assess whether mothers reinforced illness behav-
ior by making solicitous responses to GI complaints in their
In addition to the variables of parental response and
parental IBS status, there is a body of research that supports
the possibility that parent and child psychological status and
family stress have an effect on illness expression in children.
Walker and her colleagues have found consistent evidence,
based on patient, mother, and teacher reports, that child ab-
dominal pain patients have higher levels of anxiety and de-
pression than well patients. (9, 10) Higher levels of depres-
sion at the initial clinic visit significantly predicted failure
to recover from symptoms by the time of a 5-yr follow-up
assessment(11). Family stress has also been linked to higher
levels of somatic complaints and longer symptom duration
in children with recurrent abdominal pain (RAP). (12–15)
Finally, several investigators have found that parents of chil-
dren with RAP were characterized by high levels of anxiety,
depression, and somatization. (10, 14, 16, 17) Furthermore,
higher levels of parent somatization symptoms were asso-
ciated with symptom maintenance in their children several
months following the medical evaluation for RAP (14).
Therefore, the primary aims of this study were as follows.
(i) To test whether, when they are interviewed apart from
their parents, the children of IBS parents report more GI
symptoms and other somatic symptoms than control chil-
dren, and also whether they miss more school due to GI
complaints and are brought to the doctor more often for GI
symptoms. Increased symptom reports by the child when in-
terviewed separately would exclude biased perception and
reporting by a mother with IBS as a possible explanation for
observed intergenerational transmission patterns; (ii) To de-
termine whether the children of mothers who respond solici-
tously to GI complaints (as determined by questionnaire) re-
the effects of solicitous behavior are independent of the ef-
fects of mother’s IBS status; (iii) to determine whether the
effects of mother’s IBS status and the effects of solicitous
responses to GI complaints are unique to GI symptoms, or
whether they result in an increase in other types of somatic
complaints; (iv) finally, we also sought to determine whether
the associations of parent IBS status and solicitous responses
to the child’s illness complaints remain significant after con-
trolling statistically for the effects of family stress, psycho-
logical symptoms in the mother, psychological symptoms in
the child, the child’s perceived self-worth (competence), the
child’s coping behavior, gender, and age.
IBS (case) families were identified by reviewing the au-
tomated database of a large health maintenance organiza-
tion to identify all women who had received a diagnosis of
IBS or abdominal pain at a clinic visit during the preceding
2 yr and who had one or more children between the ages of
8 and 15. Women with a diagnosis of IBS received a letter
explaining the study and notifying them that they would be
telephoned and invited to participate unless they returned a
postcard refusing participation. Women with a diagnosis of
were not contacted by telephone unless they returned a post-
card confirming that they had symptoms consistent with IBS
and were interested in participating. The recruitment method
differed for women with a diagnosis of abdominal pain but
them would satisfy diagnostic criteria for IBS. All potential
study subjects identified by these two methods were further
screened by trained telephone interviewers to insure that the
mother met Rome I criteria for IBS diagnosis (18).
Control families were selected by identifying women who
did not have a clinic visit during the preceding 2 yr for IBS,
abdominal pain, constipation, or diarrhea; and who also had
children within the 8–15 yr old age range, and who received
their care at the same clinics as IBS patients. Controls were
further screened by telephone interview to insure that neither
the mother nor any adult family member met Rome I criteria
controls were ascertained in the telephone interview that: (i)
the mother was the child’s legal guardian and the child had
2 yr, (ii) no interviewed child had a developmental disability
requiring full-time special education, and (iii) that no family
member had a diagnosis of ulcerative colitis or Crohn’s dis-
ease. Subjects were given the choice of being interviewed in
their home or at their usual clinic.
resulting in the inclusion of multiple children from some
Increased Somatic Complaints and Health Care Utilization3
children in the same family would be more similar to each
other than to the members of other families.
Procedures and Measures
All procedures in this study, including informed consent
for adults and assent for children, were approved by the
Institutional Review Boards of Group Health Cooperative,
the University of Washington, and the University of North
from their parents, and the following questionnaires were ad-
ministered orally by a trained child interviewer. These ques-
tionnaires were all developed and validated for oral adminis-
tration to children.
cludes questions on how much the child was bothered by
GI symptoms and nonspecific somatic symptoms such as
headache in the preceding 2 wk. The instrument consists
of 21 symptoms from the Children’s Somatization Inven-
tory which has been demonstrated to have high reliability
and validity (19). For this study, we added an additional
GI symptom (vomiting) and eight cold symptoms.
measures children’s beliefs about their own abdominal
pain (20). The duration, frequency, and intensity of pain
episodes are assessed by four items each. In addition,
four items each assess the expected duration of the
condition of having recurrent abdominal pain and the
seriousness of this condition. Items are summed to yield
a perceived severity score. Alpha reliabilities for the
severity measure across administrations in a sample of
156 patients in a pediatric gastroenterology clinic ranged
from 0.90 to 0.92 (20). Two-wk and 6-month test–retest
reliabilities for the perceived severity score were 0.80
perceived severity of pain, the CPBQ includes subscales
that assess children’s appraisals of their ability to cope
with their pain. The latter subscales were not used in
analyses for this study.
how children cope with abdominal pain (21). The PRI
and accommodative, each with, subscales representing
specific strategies for coping with pain (e.g., seeking so-
cial support, resting, messaging the painful area, mini-
mizing pain, self-encouragement, distraction, behavioral
disengagement). Confirmatory factor analysis was used
to derive and cross-validate the factor structure in RAP
patients and school children. Across three samples, me-
dian alpha reliability exceeded 0.70 for all but one of the
to 0.71 (median = 0.59). Construct validity is supported
by the significant relation of subscales to subsequent out-
comes including pain, disability, and somatic symptoms.
4. The Self-Perception Profile for Children (SPPC), which
measures children’s perceived global competence as well
(e.g., academic, athletic, and social) (22, 23) The SPPC
has a 4-point response format and six items on each sub-
ming responses to items on that subscale. Harter reported
that internal consistency exceeded 0.78 for Global Self-
From the Child Symptom Checklist, we treated the seven
GI questions as a distinct subscale for bothersomeness of GI
during the past 2 wk by (i) nausea or upset stomach, (ii)
pain in your stomach or abdomen (stomachache), (iii) food
making you sick, (iv) feeling bloated, (v) constipation, (vi)
diarrhea (loose, runny bowel movements), and (vii) vomiting
(throwing up). The five choices for answering each question
ranged from “not at all” to “a whole lot.”
treated as a separate subscale so that we could test whether
the effects of parent IBS status and reinforcement are unique
to the reporting of GI symptoms. The items on this subscale
relate to (i) headaches; (ii) dizziness or faintness; (iii) pain in
heart or chest; (iv) feeling low in energy or slowed down; (v)
pain in back, knees, or arms (three items); (vi) sore muscles;
(vii) trouble getting your breath; (viii) hot or cold spells; (ix)
numbness or tingling; (x) weak or heavy feeling (two items);
(xi) heart beating too fast; and (xii) blurred vision.
From the Children’s Pain Beliefs Questionnaire, we in-
cluded only the subscales for condition seriousness and con-
dition frequency. A sample question from the condition se-
riousness scale was “My stomachaches mean that I’m very
sick.” A sample question from the condition frequency scale
was “I almost always have a stomachache.”
PARENT QUESTIONNAIRES. Mother was asked to com-
plete the following questionnaires for each eligible child.
1. Child Symptoms Checklist (see above): Parent version,
ered by GI and other nonspecific symptoms in the past
2 wk, plus questions on the frequency of school absences
and medical clinic visits for GI and cold symptoms in the
preceding 3 months (19).
2. Illness Behavior Encouragement Scale (IBES): Parent
version, one subscale of which measures solicitous re-
sponses to GI symptom reports (i.e., reinforcement of
illness behavior) (8, 24). The IBES, was modeled after
Whitehead’s Social Learning Scales (25). Psychometric
properties of the instrument were assessed in samples of
pediatric patients (n = 58) and school children (n = 151).
The scales demonstrated good internal consistency: coef-
ficient alphas ranged from 0.75 to 0.88.
psychological symptoms in the child (26).
4 Levy et al.
4. Children’s Pain Beliefs Questionnaire (see above): Par-
ent form (which measures the parent’s beliefs about their
child’s pain (20). In addition, mothers were asked to com-
plete the following questionnaires only once, no matter
how many children they had.
5. The Family Inventory of Life Events (FILE) is a parent-
events experienced by family members in the previous
year. McCubbin et al.(27) report that internal consistency
reliability is 0.81 and 4-wk test–retest reliability is 0.80.
Agreement between family members on the occurrence
of events supports the validity of the FILE.
6. The anxiety, depression, and somatization subscales of
the SCL-90R, which measures the mother’s psycholog-
ical symptoms. Derogatis (28) reports test–retest coeffi-
very high convergent validity with MMPI scales measur-
ing similar content areas. Derogatis and Cleary (29) also
demonstrated a theoretically consistent factor structure
for the SCL-90R in a sample of 1,002 patients.
7. The Pain Response Inventory (see above) was reworded
to assess the mother’s characteristic responses to her own
8. Illness Behavior Encouragement Scale (IBES) (8): Solic-
itous responding to GI complaints was measured by the
your child has a stomachache or abdominal pain, how of-
ten do you tell your child he/she doesn’t have to finish
all of his/her homework?” and “...how often do you give
your child special privileges or let him/her do things that
he/she isn’t usually allowed to do (like staying up late or
watching more TV)?” To simplify the interpretation of
the data, we divided subjects into low, middle, and high
scorers on this scale. This was done by pooling the data
for case and control children, calculating the frequency
distribution of scores, and dividing the distribution into
three equal parts.
HEALTH CARE UTILIZATION DATA. The automated
database of Group Health Cooperative was used to calcu-
late the total number of ambulatory visits made by the child
for any illness and also the total number of visits for a lower
the date of the interview back 2 yr in time and forward 1 yr
Although symptom reports were obtained from both the
child and the mother independently, reports of medical clinic
visits and school absences were obtained only from moth-
ers because we believed decisions regarding clinic visits and
school absences would be made by the mother and would not
reflect the child’s perceptions or responses to illness; thus,
obtaining the child’s report of these events would only be
an indirect way of asking about behaviors (decisions) of the
mother. We also sought to minimize the response burden on
DATA ANALYSIS. The study aims were addressed first by
analyses of variance. These were followed by analyses of
covariance to determine whether the effects of solicitousness
and parent IBS status remained significant after statistically
adjusting for moderator variables.
Independent variables in these analyses were (i) group
(case children vs control children) (ii) solicitousness (IBES
(8) scores divided into three equal ranges), and (iii) in some
analyses, we also compared the mother’s report of symptoms
to the child report of his/her own symptoms to test the hy-
pothesis that the mother’s perception of the child’s symptoms
may differ from the child’s.
Dependent variables in the analyses for aims 1 and 2 were
three different measures of GI symptoms—the bothersome-
ness of GI symptoms on the Child Symptom Checklist (19),
the seriousness of stomachaches reported on the Children’s
Pain Beliefs Questionnaire (20), and the frequency of stom-
achaches reported on the Children’s Pain Beliefs Question-
absences for GI symptoms and other illnesses as reported
by the mother, and the number of medical clinic visits for GI
Cooperative. Dependent variables for aim 3 were the both-
ersomeness of non-GI symptoms as reported on the Child
Symptom Checklist (19), the number of school absences for
non-GI symptoms as reported by the mother, and the num-
ber of clinic visits documented in the automated medical
records of Group Health Cooperative. These dependent vari-
ables were tested in separate analyses of variance.
Moderating variables, which were entered as covariates in
the analyses, included (i) family stress measured by the Fam-
ily Inventory of Life Events (27); (ii) mother’s psychological
distress measured on the SCL-90R (28); (iii) child’s psycho-
logical symptoms measured on the Child Behavior Checklist
(26); (iv) child’s self-perception of his/her competence mea-
sured on the global scale of the Children’s Self-Perception
Inventory (22, 23); (v) the child’s coping style measured on
three scales of the Pain Response Inventory (21) (active, pas-
sive, and accommodative coping); (vi) the child’s age; and
(vii) the child’s gender.
Because multiple children were included from some fami-
variance between families. The data were therefore analyzed
covariates, variables that were measured only at the family
level (family stress and mother’s psychological symptoms)
were treated as random effects, while covariates that were
unique to each child were treated as fixed effects. The alpha
level for inferring significance was p < 0.05, and estimated
“p-values” are reported.
achaches in the past 3 months and medical clinic visits for
stomachaches in the past 3 months—were severely skewed
(most subjects had none). These variables could not be trans-
formed to achieve a normal distribution. Consequently, the
Increased Somatic Complaints and Health Care Utilization5
main effects for groups (case vs control) and parent solici-
tousness were retested by nonparametric tests and were con-
sidered to be statistically significant only if a Mann-Whitney
trols and/or if a Kruskal Wallis test confirmed a significant
differences between low, middle, and high tertiles of subjects
on the ARCS.
ing made a medical clinic visit for IBS (77.4% of cases) or
these two groups. Participation rate was 55.1% for eligible
igible families in which the mother consulted for abdominal
pain, and 48.8% for controls.
Characteristics of the Sample
Participants were 296 case children from 208 families, and
dren were well matched with respect to demographic charac-
teristics (Table 1).
Gastrointestinal Symptoms Reported by the Child
When the dependent variable was the bothersomeness of GI
symptoms reported by the child for the preceding 2 wk on
the Child Symptom Checklist, and the model tested was lim-
ited to the effects of parent IBS status (case vs control) and
solicitousness (low, middle, high), children of parents with
IBS were more likely to report bothersome GI symptoms
with the relationships not being significantly influenced by
parental solicitousness (F(591) = 0.33). The difference be-
tween cases and controls is shown in Figure 1. After adjust-
symptoms, child’s psychological symptoms, child’s sense of
0.05). Child reports of bothersome GI symptoms were mod-
erated by parent reports of child psychological symptoms
(F(157) = 13.64; p < 0.0005; higher levels of psycholog-
ical symptoms were associated with more bothersome GI
Characteristics of the Sample
Number of children
Age (mean ± SD)
Gender (percent male)
11.89 ± 2.62
11.84 ± 2.49
case and control children was significantly greater when symptoms
were reported by the mother than when symptoms were reported by
the child (p < 0.001 for the interaction).
symptoms) and child reports of passive coping strategies
(passive coping was associated with more bothersome GI
symptoms, F(157) = 13.16, p < 0.0005).
As noted above, the participation rate was lower for case
families in which the mother had consulted for abdominal
for IBS (21.3% vs 55.1%); this suggested the possibility that
the abdominal pain families might be more self-selected and
therefore less representative of families in which the mother
has IBS. To test for this, we repeated the above analyses
after eliminating the 67 case children recruited from moth-
ers with abdominal pain. The results were essentially iden-
tical: the main effect for parent IBS status was significant
(F(528) = 16.98; p < 0.001), but the main effect for solici-
tousness was not (F(528) = 1.21; p = 0.30). The differences
between groups remained significant after adjusting for the
moderating variables (F(503) = 3.97; p < 0.05). Significant
covariates were the child’s psychological symptoms and the
child’s use of passive coping mechanisms.
Higher child ratings of the seriousness of abdominal pain
on the Pain Beliefs Questionnaire were significantly asso-
ciated with parental solicitousness toward the child’s GI
symptoms (F(590) = 11.42; p < 0.001). This is shown in
Figure 2. Post hoc tests showed that the group lowest on so-
licitousness was significantly different from both the mid-
dle group (p = 0.022) and the high solicitousness group
(p = 0.001), but the middle group was not significantly
different from the high group. After adjustment for the
same set of potential moderators, the effects of solicitous-
ness remained significant (F(157) = 9.47; p = 0.0025). The
relationship between solicitousness and seriousness of pain
was moderated by passive coping (greater solicitousness as-
sociated with more passive coping, F(157) = 36.82, p <
0.0001), accommodative coping styles (greater solicitous-
ness associated with more accommodative coping, F(157) =
30.46; p < 0.001), and younger age (more solicitousness
in younger children, F(157) = 5.76; p < 0.02). Repeat-
ing the analysis after eliminating 67 cases in which the
6 Levy et al.
ment of illness behavior) influenced how serious children reported
their stomach aches to be p < 0.001. Bars represent low, middle,
and highest thirds of scores on the solicitousness scale of the IBES.
Parent solicitous responses to GI symptoms (reinforce-
mother consulted for abdominal pain did not change the
0.0025), but abdominal pain frequency was not significantly
associated with parental solicitousness. After adjustment for
the same set of potential moderators, the effects of mother’s
icant (F(157) = 4.50; p < 0.05). The only significant covari-
ates were passive coping (mother having IBS was associated
with higher scores on passive coping, F(157) = 33.53; p <
0.0001). Repeating the analysis after eliminating 67 cases
in which the mother consulted for abdominal pain did not
change the outcome.
NON-GI SYMPTOMS REPORTED BY THE CHILD.
Children of parents with IBS were more likely to report
bothersome non-GI somatic symptoms than children of con-
trol parents (more bothersome non-GI symptoms in cases, F
of solicitousness (F(562) = 1.49; p = 0.226). After adjust-
ment for the same set of potential moderators, the effects
of mother’s IBS status on non-GI illness behavior remained
significant (F(562) = 5.48; p < 0.05). There were several
sociated with more non-GI symptoms, F(562) = 8.88; p =
0.003), passive coping style (more passive coping associ-
ated with more non-GI symptoms, F(562) = 17.81; p <
0.001), child’s psychological symptom index (more psycho-
logical symptoms associated with more non-GI symptoms,
F(562) = 6.90; p = 0.009), and age (more non-GI symptoms
in older children, F(562) = 5.42; p = 0.020).
GASTROINTESTINAL SYMPTOMS REPORTED BY
THE CHILD’S MOTHER. When the analysis of the both-
ersomeness of GI symptoms was repeated and the mother’s
report of the child’s symptoms was compared to the child’s
report using an analysis of variance, there was a significant
interaction between who provided the data and the effects of
parent IBS status (F(583) = 25.99; p < 0.001). This inter-
action is shown in Figure 1, where it can be seen that the
difference between cases and controls was greater when the
mother reported on the bothersomeness of GI symptoms as
toms. This analysis also showed a significant interaction be-
tween who provided the data and the effect of solicitousness
(F(583) = 4.60; p = 0.010), which was due to the fact that
that their children had more bothersome GI symptoms.
When a similar analysis of variance was used to compare
(F(577) = 13.00; p < 0.001), reflecting a greater difference
between cases and controls when the mother provided the
data than when the child reported on his/her own GI symp-
toms. However, the interaction between who provided the
data and solicitousness was not significant (F(577) = 1.34;
and parent IBS status was again significant (F(581) = 10.11;
p = 0.002) reflecting a bigger difference between cases and
controls when the mother provided the data. The interaction
between who provided the data and the effect of solicitous-
ness was again not significant (F(581) = 0.03).
SCHOOL ABSENTEEISM. Children with higher school
were more likely to have parents with IBS (F(625) = 26.53;
p < 0.0001), who were also more solicitous in response
to their child symptoms(F(625) = 5.33; p < 0.05). Sig-
nificant covariates were greater child psychological symp-
toms (F(157) = 10.92; p < 0.002) and passive coping styles
(F(157) = 5.83; p < 0.02), both of which were found to
amplify the difference in school absences between case and
control children. These findings (Fig. 3) were confirmed by
both parent IBS status (p < 0.001) and solicitousness (p < 0.05).
Gray bars represent controls and black bars represent cases.
School absences for stomach aches are influenced by
Increased Somatic Complaints and Health Care Utilization7
test yielded Z = 5.05, p < 0.001; and for solicitousness, the
Kruskal Wallis test yielded X2= 16.77 (p < 0.001). After
adjustment for potential moderators, the effect of parent IBS
status remained significant (F(157) = 11.69; p < 0.001), and
the effects of solicitousness showed a trend (F(157) = 3.16;
p = 0.08). When the dependent variable was the number of
child disability days for any illness, there was no significant
effect for parent IBS status or solicitousness.
medical clinic visits during a 3-yr period were more likely to
have parents with IBS (F(602) = 8.09; p = 0.005) but not
parents who were more solicitous in response to their child
by the nonparametric Mann-Whitney test (Z = 5.829; p <
0.001). After adjustment for potential moderators, the effect
of parent IBS status remained significant (F(545) = 7.61;
(F(545) = 6.81; p = 0.009).
The results were similar when ambulatory health care vis-
its for any cause made by the child in 3 yr were analyzed:
parent IBS status was again significant (F(602) = 27.92;
p < 0.001) but there was no significant effect of solicitous-
ness. This finding, shown in the right panel of Figure 4, was
confirmed by a Mann-Whitney test (Z = 6.379; p < 0.001).
After adjustment for potential moderator variables, the effect
of parent IBS status again remained significant (F(545) =
12.08; p = 0.001). The significant covariates were overall
child psychological problems (F(545) = 34.30; p < 0.001),
mother’s overall psychological distress (F(545) = 4.58; p =
the child (F(545) = 13.70; p < 0.001). Children made more
clinic visits when they or their mother had more psychologi-
cal symptoms, and higher levels of passive coping and older
age were also associated with increased numbers of medical
than controls both for GI symptoms (left panel, p = 0.005) and for
all illnesses combined (right panel, p < 0.001).
Case children make significantly more health care visits
The importance of the findings reported in this study lies in
three new observations: First, by interviewing the child inde-
pendently of his or her mother, we have shown that previous
reports of more GI symptoms in the children of IBS patients
is not an artifact of the parent’s biased reporting: the find-
ings in this study add further evidence that the tendency to
report excess symptoms, sometimes referred to as “illness
behavior,” is transmitted to children by parents who have
IBS. Second, we have shown that the effects of parent IBS
status and solicitousness are not unique to the development
of functional GI disorders but are significantly associated
with children’s reports of more non-GI symptoms as well.
This last observation may contribute to the development of
multiple, extra-intestinal symptoms that are frequently seen
in association with IBS (2) and that account for two-third
of their excess health care costs in IBS patients (31). Third,
we have shown that parent IBS status and solicitousness ex-
ert independent effects on the child’s symptom reporting,
clinic visits, and illness-related school absenteeism, which
has implications for the design of treatment and prevention
Child Reporting of Symptoms
In previous studies, evidence for the occurrence of excess
symptom reporting, health care utilization, and disability
days in the children of IBS patients was based on the par-
ent’s report that the child had more GI symptoms (8) or on
for GI complaints (4). However, these indices of illness be-
attention to symptoms that she has herself. This distinction is
clearly important as we have shown that the differences be-
tween case and control children are substantially larger when
the mother reports on the child’s GI symptoms compared to
when the child reports on his/her own GI symptoms (Fig. 1).
children of mothers who had IBS missed more school days
for GI complaints, and were taken to the medical clinic more
often for investigation and management of GI symptoms.
Moreover, the contributions of parent IBS status and solici-
tousness were robust; they remained significant after adjust-
ing for the possible moderating effects of child age and sex,
family stress, psychological symptoms of the mother, psy-
chological symptoms of the child, the child’s self-confidence
(sense of competence), and the child’s style of coping with
We believe these data have implications for the etiology of
with recurrent abdominal pain continue to report abdominal
pain and related symptoms compatible with IBS as adults
clinical problem in its own right since it is a costly condi-
tion that affects 10–15% of school-aged children and causes
8 Levy et al.
dren, and results in significant health care costs (35, 36).
The analysis of non-GI symptoms reported by children sug-
gests that the effects of parent IBS status are not unique to
GI symptoms; instead, children of IBS mothers report more
symptoms of all types. This is consistent with previous re-
ports that adult IBS patients have more non-GI symptoms
and consult physicians more often for non-GI complaints as
compared to controls who do not have IBS (2). It is possi-
ble that mothers with IBS influence many types of somatic
complaints in addition to GI symptoms.
Social Learning versus Heritability
In a previous study of twins (6) we found evidence for both
a genetic and a social learning contribution to the intergen-
erational transmission of IBS symptoms: the concordance
than the concordance between dizygotic twins, but the con-
cordance between mothers and children was greater than the
concordance between monozygotic twins. Although the de-
sign of the current study does not allow for a definitive sep-
aration of genetic from environmental influences, the data
add further support to the plausibility of noninherited factors
contributing to this intergenerational pattern.
Independence of Solicitousness and Parent IBS Status
When we began this research, previously published studies
seemed to indicate that parent IBS status and solicitousness
were closely linked and would be indistinguishable in their
school absences. However, these data suggest that they have
differential effects: parent IBS status influenced children’s
perceptions of the frequency and bothersomeness of symp-
toms but did not affect their perception of the seriousness of
GI symptoms. On the other hand, parental solicitousness in-
fluenced the children’s perception of how serious their stom-
achaches were, but did not influence their perception of how
often stomachaches occurred. This distinction implies that
ment of recurrent abdominal pain in children should sep-
arately target how parents behave when they are ill them-
selves as well as how they respond to their child’s somatic
Passive Style of Coping with Pain
The only moderator variable found to be consistently associ-
measured on the Pain Response Inventory. Passive coping is
(e.g., “think to yourself there is nothing you can do, so you
eryone”), and catastrophizing (e.g., “feel like you can’t stand
it anymore,” and “think to yourself that something might be
really wrong with you”). These cognitions about abdominal
pain should therefore be specifically targeted in cognitive be-
havioral treatments for recurrent abdominal pain.
Parent IBS Status: What Are the Operating Mechanisms?
One of the goals of this study was to further advance un-
derstanding of the specific mechanisms operating under the
broad rubric of “environmental influences on behavior.”
While we directly measured solicitous responses to GI com-
plaints, which provides a good operational definition of “re-
inforcement,” we also measured a broad category of “parent
IBS status,” which we believe was a reasonable operational
definition of “modeling.” To be included in our case group,
parents had to report GI symptoms to their physicians to
receive a diagnosis of IBS or abdominal pain. Thus, by def-
inition, these were all parents who engaged in some form
of illness behavior, specifically health care seeking for, and
reporting of, GI symptoms. Thus, it is reasonable to suggest
that the mechanism responsible for the effects of parent IBS
status on the child’s symptom reporting might be modeling,
defined as the observation and then imitation of the behavior
of another (37). However, this interpretation must be consid-
on whether the child directly observed the mother displaying
necessity of obtaining consent from two (or more) family
members (i.e., the mother and each eligible child) was an
important factor in determining the level of participation. A
on illness-related absences but found that schools keep very
(38) we have shown that, to the extent that illness-related
these school data confirm the validity of the mother’s report.
Also, use of missed school and doctor’s visits as indices of
illness behavior is limited by the fact that these behaviors
may be influenced by mother’s perceptions and help-seeking
a strong influence in whether or not a child goes to school
or to the doctor’s office. Third, only mothers were studied
in this research. Other research has demonstrated that the
health status of the father may have as great an influence on
the illness behavior of the child as the mother (4).
Implications for Clinical Practice
Abdominal pain is the most common recurrent pain com-
plaint of childhood (39), and it is associated with high health
care costs (4) and school absenteeism (34). Moreover, a sub-
stantial number of children with functional abdominal pain
continue to have abdominal pain consistent with IBS and a
variety of other health complaints as they grow up (32, 33,
40). The goal of this and related research is to understand
the mechanisms responsible for functional GI disorders. The
Increased Somatic Complaints and Health Care Utilization9
weight of the findings of much of our research and that of
others suggests that understanding the mechanisms of social
learning may contribute to (i) a reduction of the intergener-
ational transmission of illness behavior by appropriate edu-
cation of parents with IBS and (ii) more effective cognitive
functional GI disorders. Key observations in this study that
and treatment protocols include (i) showing that the somatic
complaints, health care utilization, and disability days that
are transmitted from the parent to the child includes a broad
and is not restricted to the reporting of GI symptoms, and (ii)
effects that are distinct from each other and independent of
the effects of stress, psychological symptoms, and coping
Current State of the Literature and Contributions
of this Study
Children of IBS mothers are reported to have more GI symp-
but this could be due to biased perceptions or selective atten-
tion by parents and not reflect the behavior of the child.
This study interviewed children apart from their parents
and showed that they do in fact report more GI (and also
non-GI) symptoms than control children.
Social learning was previously assumed to have relatively
specific effects, primarily on GI symptom reports, disability
days, and clinic visits.
This study shows that social learning effects are non-
specific: case children report more non-GI symptoms, miss
school for non-GI complaints, and are taken to medical clin-
ics for non-GI indications more often than control children.
Parent IBS status and solicitousness were assumed to be
closely linked and perhaps indistinguishable in their effects.
This study shows that parent IBS status and solicitousness
ing the frequency of complaints and solicitousness affecting
primarily the perceived seriousness of complaints.
sible for intergenerational transmission by (i) more sophisti-
cated and more direct measures of modeling and reinforce-
ment and (ii) direct manipulation of these mechanisms to
This work is supported by NIH grants RO1 HD36069 and
Reprint requests and correspondence: Rona L. Levy, Ph.D.,
University of Washington, 4101 15th Avenue NE, Seattle, WA
Received June 10, 2004; accepted July 19, 2004.
1. Whitehead WE, Winget C, Fedoravicius AS, et al. Learned
illness behavior in patients with irritable bowel syndrome
and peptic-ulcer. Dig Dis Sci 1982;27:202–8.
2. Whitehead WE, Palsson O, Jones KR. Systematic review
of the comorbidity of irritable bowel syndrome with other
disorders: What are the causes and implications? Gastroen-
3. Drossman DA, Li ZM, Andruzzi E, et al. United-
States Householder Survey of Functional Gastrointestinal
Disorders—Prevalence, sociodemography, and health im-
pact. Dig Dis Sci 1993;38:1569–80.
tional transmission of gastrointestinal illness behavior. Am
J Gastroenterol 2000;95:451–6.
5. Whitehead WE, Crowell MD, Heller BR, et al. Modeling
adult illness behavior. Psychosom Med 1994;56:541–50.
6. Levy RL, Jones KR, Whitehead WE, et al. Irritable bowel
syndrome in twins: Heredity and social learning both con-
tribute to etiology. Gastroenterology 2001;121:799–804.
J Clin Gastroenterol 1987;9:324–30.
8. Walker LS, Zeman JL. Parental response to child illness
behavior. J Pediatr Psychol 1992;17:49–71.
9. Walker LS, Garber J, Greene JW. Psychosocial correlates
of recurrent childhood pain: A comparison of pediatric pa-
tients with recurrent abdominal pain, organic illness, and
psychiatric disorders. J Abnorm Psychol 1993;102:248–58.
10. Walker LS, Greene JW. Children with recurrent abdominal
pain and their parents: More somatic complaints, anxiety,
11. Walker LS, Heflinger CA. Quality of life predictors in pe-
diatric abdominal pain patients: Findings at initial assess-
12. Greene JW, Walker LS, Hickson G, et al. Stressful life
events and somatic complaints in adolescents. Pediatrics
13. Walker LS, Greene JW. Negative life events, psychosocial
resources, and psychophysiological symptoms in adoles-
cents. J Clin Child Psychol 1987;16:29–36.
14. Walker LS, Greene JW. Negative life events and symptom
resolution in pediatric abdominal pain patients. J Pediatr
15. Walker LS, McLaughlin FJ, Greene JW. Functional illness
and family functioning: A comparison of healthy and so-
maticizing adolescents. Fam Process 1988;27:317–25.
nal migraine or depression predispose children to headache
and stomach-ache? Headache 1992;32:353–5.
17. Routh DK, Ernst AR. Somatization disorder in relatives of
children and adolescents with functional abdominal pain. J
Pediatr Psychol 1984;9:427–37.
18. Thompson WG, Longstreth GF, Drossman DA, et al. Func-
tional bowel disorders and functional abdominal pain. Gut
19. Walker LS, Garber J, Greene JW. Somatization symptoms
in pediatric abdominal pain patients: Relation to chronicity
10 Levy et al.
and validation of the Children’s Pain Belief Question-
Gainsville, Florida, 1997.
idation of the pain response inventory for children. Psychol
Denver, Colorado: University of Denver, 1985.
24. Van Slyke DA. Maternal influences on children’s pain be-
havior. Dissertation Abstracts Int 2001;62(2B):1103.
25. Whitehead WE, Busch CM, Heller BR, et al. Social learn-
ing influences on menstrual symptoms and illness behavior.
Health Psychol 1986;5:13–23.
Department of Psychiatry, 1991.
27. McCubbin H, Patterson J, Wilson L. Family inventory of
life events and changes (FILE). In: McCubbin HI, Thomp-
Madison: University of Wisconsin System, 1996:103–178.
28. Derogatis L. SCL-90-R: Administration, scoring, and pro-
cedures manual. Minneapolis MN: National Computer Sys-
for the primary symptom dimensions of the SCL-90. Br J
Soc Clin Psychol 1977;16:347–56.
30. Suzuki S, Sheu C-F. Using PROC MIXED in hierarchical
linear models: Examples from two- and three-level school-
effect analysis, and meta-analysis research. 2002.
31. Levy RL, Von Korff M, Whitehead WE, et al. Costs of
care for irritable bowel syndrome patients in a health main-
tenance organization. Am J Gastroenterol 2001;96:3122–
32. Christensen MF, Mortensen O. Long-term prognosis in
children with recurrent abdominal pain. Arch Dis Child
33. Stickler GB, Murphy DB. Recurrent abdominal pain. Am J
Dis Child 1979;133:486–9.
34. Walker LS, Guite JW, Duke M, et al. Recurrent abdominal
pain: A potential precursor of irritable bowel syndrome in
adolescents and young adults. J Pediatr 1998;132:1010–5.
Scientific Publishers, 1975.
36. Apley J, Naish N. Recurrent abdominal pain: A field study
Dis Child 1958;46:337–40.
37. Martin G, Pear J. Behavior modification: What is it and how
to do it. Englewood Cliffs, NJ: Prentice Hall, 1978.
38. Garner MD, Levy RL, Whitehead WE, et al. Measuring
the impact of gastrointestinal illness on school attendance.
124th Ed. 2003:A532.
ment. New York: Guilford, 1990.
40. Apley J, Hale B. Children with recurrent abdominal pain:
How do they grow up? Br Med J 1973;3:7–9.
Query Download full-text
Q1 Author: Please provide more details in Ref. .