Behaviour Research and Therapy 58 (2014) 65e74

Contents lists available at ScienceDirect

Behaviour Research and Therapy

journal homepage: www.elsevier.com/locate/brat

Cognitive-behavioral therapy for hypochondriasis/health anxiety:

A meta-analysis of treatment outcome and moderators
Bunmi O. Olatunji a, *, Brooke Y. Kauffman b, Sari Meltzer a, Michelle L. Davis b,
Jasper A.J. Smits b, Mark B. Powers b
a
Department of Psychology, Vanderbilt University, United States
b
Department of Psychology, University of Texas-Austin, United States

a r t i c l e i n f o a b s t r a c t

Article history: The present investigation employed meta-analysis to examine the efficacy of cognitiveebehavioral
Received 18 March 2014 therapy (CBT) for hypochondriasis/health anxiety as well as potential moderators that may be associated
Received in revised form with outcome. A literature search revealed 15 comparisons among 13 randomized-controlled trials
12 May 2014
(RCTs) with a total sample size of 1081 participants that met inclusion criteria. Results indicated that CBT
Accepted 16 May 2014
Available online 24 May 2014
outperformed control conditions on primary outcome measures at post-treatment (Hedges's g ¼ 0.95)
and at follow-up (Hedges's g ¼ 0.34). CBT also outperformed control conditions on measures of
depression at post-treatment (Hedges's g ¼ 0.64) and at follow-up (Hedges's g ¼ 0.35). Moderator an-
Keywords:
Hypochondriasis
alyses revealed that higher pre-treatment severity of hypochondriasis/health anxiety was associated
Health anxiety with greater effect sizes at follow-up visits and depression symptom severity was significantly associated
Cognitive-behavioral therapy with a lower in effect sizes at post-treatment. Although effect size did not vary as a function of blind
Moderation assessment, smaller effect sizes were observed for CBT vs. treatment as usual control conditions than for
CBT vs. waitlist control. A dose response relationship was also observed, such that a greater number of
CBT sessions was associated with larger effect sizes at post-treatment. This review indicates that CBT is
efficacious in the treatment of hypochondriasis/health anxiety and identifies potential moderators that
are associated with outcome. The implications of these findings for further delineating prognostic and
prescriptive indicators of CBT for hypochondriasis/health anxiety are discussed.
© 2014 Elsevier Ltd. All rights reserved.

The central feature of hypochondriasis is a preoccupation with
the inaccurate belief that one has, or is in danger of developing, a
serious medical condition based on misinterpretations of benign
(or minor) bodily sensations (American Psychiatric Association
[APA], 2000). Research has shown that up to 9% of patients in
general medical practice clinics and up to 5% of the general popu-
lation meets diagnostic criteria for hypochondriasis (Creed &
Barsky, 2004; Gureje, Ustun, & Simon, 1997). The heightened
prevalence in medical practice may reflect the fact that hypo-
chondriasis is also characterized by a strong “disease conviction”
that persists despite appropriate medical evaluation. This preoc-
cupation with medical illness often focuses on specific signs or
symptoms (e.g., sore throat), diseases (e.g., cancer), organs (e.g.,
heart), or vaguely defined somatic phenomena (e.g., “my aching
veins”). Typically, patients with hypochondriasis attribute
* Corresponding author. Department of Psychology, Vanderbilt University, 301
Wilson Hall, 111 21st Avenue South, Nashville, TN 37203, United States.
E-mail address: olubunmi.o.olatunji@vanderbilt.edu (B.O. Olatunji).
unwanted bodily sensations to possible disease (e.g., “this headache
means I have a brain tumor”) and are highly concerned with their
authenticity. Perhaps the most readily observable sign of hypo-
chondriasis is the persistent attempt to seek reassurance about the
feared symptoms or illness. Individuals with this condition may
repeatedly contact doctors, seek additional tests, scour Internet
sites and medical texts, and seek reassurance from significant
others about bodily sensations which have been appropriately
evaluated and judged to be benign. This preoccupation with disease
can be disruptive to social, occupational, and family functioning,
and is associated with substantial economic costs (Katon & Walker,
1998).
Although hypochondriasis has historically been viewed as a
somatoform disorder (APA, 2000), the validity of this categorization
has not been without debate. It has been argued that hypochon-
driasis may be best conceptualized as an anxiety disorder (Olatunji,
Deacon, & Abramowitz, 2009). This view suggests that health
anxiety represents a continuous dimension (ranging from no health
anxiety to severe health anxiety) with ‘hypochondriasis’ at its

http://dx.doi.org/10.1016/j.brat.2014.05.002
0005-7967/© 2014 Elsevier Ltd. All rights reserved.
66 B.O. Olatunji et al. / Behaviour Research and Therapy 58 (2014) 65e74
clinical endpoint (Marcus, Gurley, Marchi, & Bauer, 2007). This
argument is based largely on empirical observations that symp-
toms of hypochondriasis overlap with certain anxiety disorders:
namely, panic disorder (PD) and obsessive-compulsive disorder
(OCD). Like those with hypochondriasis, patients with PD are
hypervigilant to benign, arousal-related body sensations and often
erroneously attribute them to organic causes such as heart attacks,
strokes, and other serious medical conditions (Abramowitz, 2005;
Barsky, Barnett, & Cleary, 1995). Similarities have also been
observed between hypochondriasis and OCD in terms of preoccu-
pation with health and disease, and the repetitive and pervasive
nature of such preoccupation (Abramowitz, 2005; Fallon, Javitch,
Hollander, & Liebowitz, 1992). Much like PD and OCD, cognitive-
behavioral models (Abramowitz, Schwartz, & Whiteside, 2002;
Warwick & Salkovskis, 1990) posit that hypochondriasis is an
extreme form of health anxiety that emerges from the misinter-
pretation of benign and normally occurring experiences that lead to
anxiety and the use of safety behaviors which paradoxically
maintains the anxiety (Abramowitz & Moore, 2007; Abramowitz,
Olatunji, & Deacon, 2007).
The cognitive-behavioral model of hypochondriasis is depicted
in Fig. 1 (Abramowitz, Deacon, & Valentiner, 2007). This model
suggests that dysfunctional beliefs about bodily symptoms and
illness play a significant role in the development of hypochon-
driasis. Such beliefs often increase the likelihood of having cata-
strophic cognitions when exposed to benign bodily symptoms or
health-related information. Once concerned about the possibility
of acquiring an illness, the person becomes vigilant for any signs of
being ill and is motivated to reduce their worry by gaining certainty
Fig. 1. Cognitive-behavioral model of hypochondriasis outlined by Abramowitz, Deacon, et al. (2007).
about their health status. It is important to note that this process in
hypochondriasis is driven largely by mistaken beliefs about ill-
nesses. According to the cognitive-behavioral model, mistaken
beliefs in hypochondriasis are maintained (despite contradictory
information and repeated reassurance of good health from medical
professionals) by maladaptive strategies used to cope with health-
related anxiety. These strategies include attempts to prevent the
feared illness, avoidance, and attempts to attain certainty about
health status. These safety behaviors prevent individuals with hy-
pochondriasis from acquiring information that would disconfirm
their mistaken beliefs about illnesses. This cognitive-behavioral
model of hypochondriasis has been empirically supported and
proven to be clinically useful (Taylor & Asmundson, 2004). Based in
part on these observations, the newly published DSM-5 has
replaced the diagnosis of hypochondriasis with “illness anxiety
disorder” (APA, 2013).
Based on the findings of early, predominantly psychodynamic
interventions, hypochondriasis has historically been regarded as
resistant to psychological treatment (Taylor, Asmundson, & Coons,
2005). In fact, treatments were initially considered to be of limited
value for hypochondriasis (Ladee, 1966). Extreme interventions
were considered and some clinicians even employed prefrontal
lobotomies for the treatment of hypochondriasis (Bernstein,
Callahan, & Jaranson, 1975). Fortunately, the emergence of a
cognitive-behavioral model of hypochondriasis has paved the way
for the application of a more appropriate approach to treatment.
The cognitive-behavioral approach is derived largely from the
observation that symptoms of hypochondriasis d at both a topo-
graphical and functional level d overlap remarkably with certain
 B.O. Olatunji et al. / Behaviour Research and Therapy 58 (2014) 65e74
anxiety disorders (Olatunji et al., 2009). As with other anxiety
disorders, this approach emphasizes the role of dysfunctional be-
liefs in maintaining hypochondriasis. This approach also suggests
that avoidance and health-related safety behaviors prevent
dysfunctional beliefs from being disconfirmed, and thereby
perpetuate excessive health anxiety and somatic preoccupation
(Warwick & Salkovskis, 2001). This functional analysis of hypo-
chondriasis would then suggest that cognitive behavioral therapy
(CBT) may provide much needed symptom relief. CBT here refers to
the class of interventions that are based on the basic premise that
emotional disorders are maintained by cognitive and behavioral
factors, and that psychological treatment leads to changes in these
factors through cognitive and behavioral techniques (Beck &
Emery, 2005).
Research on the effects of CBT for hypochondriasis has produced
encouraging results. In one study, CBT was found to be superior to
no treatment in reducing health anxiety, the need for reassurance,
and the frequency of checking behavior (Warwick, Clark, Cobb, &
Salkovskis, 1996). A subsequent study found that patients
receiving cognitive therapy showed significantly more reduction in
symptoms of hypochondriasis at mid-treatment and at post-
treatment compared to those receiving a stress management
intervention (Clark et al., 1998). Given that research has shown that
patients with excessive health anxiety have a preference for psy-
chological treatments over drug treatments (Walker, Vincent,
Furere, Cox, & Kjernisted, 1999), a better understanding of the ef-
ficacy of CBT may provide a solid empirical foundation for making it
the gold standard treatment for hypochondriasis. Meta-analysis is
the primary means by which researchers have synthesized the
results from multiple treatment trials examining the efficacy of
various treatments. Although the use of meta-analytic data is not
without limitations, this approach has offered very preliminary
insight into delineating the general effectiveness of CBT in the
treatment of hypochondriasis. An initial meta-analytic review of
the literature concluded that CBT is the most effective treatment for
hypochondriasis (Taylor et al., 2005). However, this investigation
examined only 5 studies utilizing CBT and included uncontrolled
comparisons in the analysis. Thomson and Page's (2007) more
recent Cochrane review, which included only six studies, also
concluded that CBT was the most effective psychological treatment
for hypochondriasis. However, several limitations of the existing
trial data at the time of the publication were noted. For example,
the reliance on wait-list control conditions in many trials makes it
difficult to rule out the possibility that improvements after CBT
were due to nonspecific factors.
Although much remains unknown about the incremental ef-
ficacy of CBT for hypochondriasis, even less is known about reli-
able predictors of treatment outcome. For example, some research
suggests that comorbid psychiatric disorders are poor prognostic
indicators (Barsky, Fama, Bailey, & Ahern, 1998; Noyes et al., 1994).
However, other studies report that comorbid anxiety and
depression is either related to good outcome (House, 1989) or
unrelated to outcome (Greeven et al., 2009). Research has also
shown that younger age is associated with good outcome (Barsky,
1996; House, 1989) while other studies have failed to observe this
association (Speckens, Spinhoven, van Hemert, Bolk, & Hawton,
1997). Taylor et al. (2005) suggest that the inconsistent findings
with regards to predictors of treatment outcome in hypochon-
driasis may be because some variables predict outcome for some
treatments but not others. It may also be the case that the
inconsistent findings reflect the fact that the variables are weak
predictors of outcome. Yet another possibility is that individual
studies lack the power necessary to detect predictors of treatment
outcome. To the extent that this is the case, meta-analysis of many
trials may prove useful in clarifying some of these discrepancies.
67
Clarifying the substantive predictors of treatment outcome may
have important implications for treatment planning in CBT. For
example, if it is the case that comorbidity predicts worse treat-
ment outcome, this may be justification for the use of adjunctive
treatments to CBT that more directly target the comorbid condi-
tions. This may also have implications for the time course of CBT
for hypochondriasis. Among those with more severe comorbid
conditions, implementation of CBT may be difficult, thereby
interfering with effective treatment. Treatment of comorbid dis-
orders before initiating CBT for hypochondriasis in such cases may
then maximize outcome.
In addition to clarifying predictors of treatment outcome,
examining the extent to which sample-specific characteristics (e.g.,
gender) and study-specific characteristics (e.g., number of treat-
ment sessions) predict outcome in CBT for hypochondriasis may
also prove to be informative. Indeed, it has been observed that the
question towards which all outcome research should ultimately be
directed is the following: “What treatment, by whom, is most
effective for this individual with that specific problem, and under
which set of circumstances?” (Paul, 1967, p. 111). Although no one
study can address such a complicated question (Beutler, 1991),
examination of moderators in the context of meta-analysis of
treatment outcome studies examining the efficacy of CBT for hy-
pochondriasis may be an important preliminary step in beginning
to address this question. Accordingly, the present investigation
employs a meta-analytic approach to examine the efficacy of CBT
for hypochondriasis. This investigation is especially timely given
the limitations of the prior two meta-analyses CBT for hypochon-
driasis, one of which consisted of only 5 CBT studies and included
uncontrolled comparisons (Taylor et al., 2005) and the other which
was limited to only 6 studies (Thomson & Page, 2007). In addition,
this investigation builds on prior work by examining moderators of
treatment outcome. Four questions derived from the existing
literature were examined.
1. Do CBT treatments outperform control conditions on hypo-
chondriasis/health anxiety outcome measures at post-
treatment and at follow-up?
2. Do CBT treatments outperform control conditions on depression
severity measures at post-treatment and follow-up?
3. Does higher pre-treatment hypochondriasis/health anxiety
severity and higher pre-treatment depression scores predict
lower between-group effect sizes at post-treatment and follow-
up?
4. Does comparator treatment/condition, participant age, per-
centage of female participants, percentage of comorbidity,
number of sessions, inclusion of treatment integrity checks and/
or blind assessors (i.e., clinical trial quality) moderate treatment
effect sizes?
Method
Study selection
Randomized controlled trials (RCTs) of CBT (see inclusion
criteria below) for hypochondriasis/health anxiety were selected
using a comprehensive search strategy. A search was conducted in
PsycINFO and MEDLINE (1966 to March 2014). The searches
included the following terms: “cognitive behavioral” or “cognitive
behavioral therapy,” and “clinical trial” or “trial” alone and in
combination with “hypochondriasis” or “health anxiety.” These
words were searched as key words, title, abstract, and Medical
Subject Headings. In addition, we contacted authors of CBT trials for
emerging publications.
68 B.O. Olatunji et al. / Behaviour Research and Therapy 58 (2014) 65e74
Fig. 2. Study selection and reasons for exclusions.
As depicted in Fig. 2, the initial search strategies produced 2988
potential articles. Examination of the abstracts identified 25 relevant
and non-repeating studies. Inclusion criteria for the meta-analysis
were as follows: (a) participants who met full DSM-III-R, DSM-IV,
or DSM-IV-TR criteria for hypochondriasis or had clinical levels of
health anxiety; (b) adequate control condition (psychological pla-
cebo, wait-list control, or pill placebo); and (c) more than one session
of CBT. Treatments were classified as CBT if they included cognitive
techniques (e.g. cognitive restructuring, behavioral experiments,
etc.), behavioral techniques (e.g. in-vivo exposure, imaginal expo-
sure, etc.), or a combination of these strategies. Acceptance-based
therapies (e.g., ACT), which emphasize engaging in valued behav-
iors despite anxiety, were included if they also utilized cognitive or
behavioral techniques. Exclusion criteria for the meta-analysis
were: (a) single case studies; (b) treatment conditions aimed at
augmentating a psychological treatment; (c) studies with insuffi-
cient data, unless study authors were able to provide such data; and
(d) studies with redundant data. Of the 25 identified studies,14 were
excluded based on these criteria. Two additional studies were then
identified after reviewing the references of the remaining studies.
Table 1 shows the 15 comparisons within the 13 studies with a total
sample size of 1081 participants that met the final inclusion criteria
and were included in the meta-analyses.
Software
All analyses were completed with Comprehensive Meta-
Analysis (Bornstein & Rothstein, 1999).
Procedure
Data on the following variables were collected when reported in
each of the 13 included studies: treatment conditions (13 studies),
control type (13 studies), treatment dose (number of sessions; 13
studies), number of participants (13 studies), mean age (11 studies),
and percentage of female participants (12 studies). Treatment
integrity variables, including inclusion or exclusion of treatment
integrity checks and blind assessors, were collected categorically and
assumed to be absent if not reported in the manuscript. Dependent
variables were labeled as either primary or secondary outcome
measures as defined by the authors in the original publication.1
Control conditions were classified into three categories: pla-
cebo, treatment as usual, or wait-list. Treatments categorized as
placebo were treatments that did not include the cognitive and
behavioral components of CBT, including: attention control, relax-
ation, problem-solving, and pill placebo. Wait-list (WL) was defined
as a control condition in which participants did not receive any
treatment for hypochondriasis/health anxiety symptoms for a
specified amount of time. Five of the 13 studies utilized treatment
as usual as the control condition, four of the 13 studies had wait-list
as the control condition, three of the 13 studies had a psychological
placebo as the control condition, and one study employed a pill
placebo as the control condition.
Effect size calculation
Between-group effect sizes for each study were computed using
Hedges's g (Rosenthal, 1991). Studies with multiple outcomes were
1
Primary and secondary outcome measures were defined in accordance with
how the measures were described in the articles with the exception of two articles
(Barsky, Ahern, Bauer, Nolido, & Orav, 2013 and Warwick, 1996). Insufficient data on
the primary outcomes were provided for Barsky et al. and primary outcomes were
not clearly stated in the Warwick (1996) article, therefore we defined the primary
and secondary outcome consistent with the population of articles. Primary mea-
sures were exclusively measures of hypochondriasis/health anxiety. Secondary
measures generally included measures of hypochondriasis/health anxiety, general
anxiety, and somatization, and depression. Subsequent analyses were also con-
ducted to examine the effects of CBT on depression exclusively.
 B.O. Olatunji et al. / Behaviour Research and Therapy 58 (2014) 65e74 69

Table 1
Studies included in the meta-analysis.

Study Conditions N Mean age # of sessions Primary outcome measure Secondary outcome measure

Sorensen et al. (2011) CBT vs. WL 56 Not provided 16 HAMA, HAI BDI, BAI, HAMD
Seivewright et al. (2008) CBT vs. TAU 49 Not provided 4.3 HAI HADS-D
Hedman et al. (2011) CBT vs. Psych PL 81 39.05 12 HAI IAS, WI, BAI, ASI, MADRS-S, QOLI
Greeven et al. (2007) CBT vs. Pill PL 75 40.32 7.3 WI IAS-HA, IAS-Illness Behavior, SCL-90,
MADRS, BAS
Barsky and Ahern (2004) CBT vs. TAU 187 43.21 6 WI HCQ, HAI, SSI, SAS
Visser and Bouman (2001) Exposure vs. WL 42 37.7 12 IAS-HA, IAS-Illness Behavior, BDI, MOCI, SCL-Total
IAS-Body, SAS, SCL-90(Som)
Visser and Bouman (2001) CT vs. WL 40 36.7 12 IAS-HA, IAS-Illness Behavior, BDI, MOCI, SCL-Total
IAS-Body, SAS, SCL-90(Som)
Bourgault-Fagnou and CBT vs. WL 39 68.94 6 WI SHAI, SSI, STAI-S, STAI-T, GDS
Hadjistavropoulos (2013)
Bourgault-Fagnou and ECBT vs. WL 36 68.53 6 WI SHAI, SSI, STAI-S, STAI-T, GDS
Hadjistavropoulos (2013)
Warwick (1996) CBT vs. WL 36 37 16 HAI BDI, BAI
Sumathipala et al. (2008) CBT vs. TAU 150 35 3 GHQ-30 BSI,# of visits,# of complaints
Speckens et al. (1995) CBT vs. TAU 79 37.1 16-Jan Intensity-Mean, Intensity-Max IAS-HA, IAS-Illness Behavior, WI,
HADS-A, HADS-D
Barsky et al. (2013) CBT vs. Psych PL 89 51.84 9 HAI SSI, MCQ, Beliefs, SCL-90, FSQ
McManus, Surawy, Muse, CBT vs. TAU 74 42.6 8 HAC BAI, BDI, FFMQ
Vazques-Montes, and
Williams (2012)
Buwalda, Bouman, and CBT vs. Psych PL 48 41 6 GIAS BDI
Van Duijn (2006)

Note. CBT ¼ Cognitive Behavior Therapy, WL ¼ Waitlist, TAU ¼ Treatment As Usual, PL ¼ Placebo, NP ¼ Not Provided, HAMA: Hamilton Anxiety Rating Scale, HAI: Health
Anxiety Inventory, BDI: Beck Depression Inventory, BAI: Beck Anxiety Inventory, HAMD: Hamilton Depression Rating Scale, HADS-D: Hospital Anxiety and Depression Scale-
Depression, IAS: Illness Attitude Scale, WI: Whitely inventory, ASI: Anxiety Sensitivity Inventory, MADRS: Montgomery Asberg Depression Rating Scale, QOLI: Quality of Life
Inventory, IAS-HA: Illness Attitude Scale-Health Anxiety, IAS-Illness Behavior: Illness Attitude Scale-Illness Behavior, SCL-SOM: Symptom Check List-Somatization scale, BAS:
Brief Anxiety Scale, HCQ: Hypochondriacal Cognitions Questionnaire, SSI: Somatic Symptoms Inventory, SAS: Somatosensory Amplification Scale, IAS-Body: Illness Attitude
Scale-Body, MOCI: Maudsley Obsessive Compulsive Inventory, SCL-Total: Symptom Check List-Total, SHAI: Short Health Anxiety Inventory, STAI-S: State Trait Anxiety In-
ventory State-Scale, STAI-T: State Trait Anxiety Inventory-Trait Scale, GDS: Geriatric Depression Scale, GHQ-30: General Health Questionnaire-30, BSI: Bradford Somatic
Inventory, HADS-A: Hospital Anxiety and Depression Scale-Anxiety, MCQ: Metacognitions Questionnaire, SCL-90: Symptom Checklist-90, FSQ: Functional Status Question-
naire, HAC: Health Anxiety Composite, FFMQ: Five Facet Mindfulness Questionnaire, GIAS: Groningen Illness Attitude Scale.

categorized as above (see Table 1) and then combined within each
domain. When the necessary data were available, Hedges's g was
calculated directly using the following formula: g ¼ X T  X C =SP
where X T is the mean of the treatment group, X C is the mean of the
comparison group, and SP is the pooled standard deviation. If these
data were not provided, Hedges's g was estimated using conversion
equations for significance tests (e.g., t, F) (Rosenthal, 1991). All effect
sizes were corrected for small sample sizes according to Hedges
and Olkin (1985). Therefore, a smaller sample size reduces the
estimated effect size helping control for different sample sizes
across studies. These controlled effect sizes may then be inter-
preted conservatively with Cohen's convention of small (0.2), me-
dium (0.5), and large (0.8) effects (Cohen, 1988). Hedges's g also
may be computed directly from Cohen's d with the following for-
mula: g ¼ d (1  3/4 (n1 þ n2)  9). When there were multiple
outcomes per domain they were combined according to Borenstein,
Hedges, and Rothstein (2007). The overall mean effect size for all of
the studies combined was computed using the following formula:
P P
g¼ wj gj = wj where wj is the weight for each study and gj is the
effect size for each study. Effect sizes were calculated with random
effects models. The random effects analysis estimates the overall
effect size assuming the studies included are only a sample of the
entire population of studies and/or that the studies are
heterogeneous.
Results
points (post and follow-up) on primary outcome measures were
included in this test of heterogeneity. The test was significant,
Q(14) ¼ 58.03, p < .001, suggesting that random effects analyses
were most appropriate for this study. This significant heterogeneity
also suggests it may be appropriate to employ moderator analyses
to identify potential sources of study variability.
Question 1: does CBT outperform the control conditions on primary
and secondary outcome measures at post-treatment and follow-up?
Fig. 3 depicts the relative advantage of CBT compared to control
conditions on primary outcomes at post-treatment in an analysis of
9 studies and 11 comparisons (Hedges's g ¼ 0.95 [SE ¼ 0.15, 95% CI:
0.66e1.22, p < .001]). The analysis involving outcomes at follow-up
included 7 studies and 7 comparisons and indicated that CBT out-
performs control conditions on primary outcomes at follow-up,
although the effect size was significantly smaller than that
observed at post-treatment (Hedges's g ¼ 0.34 [SE ¼ 0.14, 95% CI:
0.06e0.62, p ¼ .016]).
A post-treatment analysis of secondary outcome measures
where CBT outperformed control conditions included 9 studies and
11 comparisons (Hedges's g ¼ 0.56 [SE ¼ 0.11, 95% CI: 0.35e0.78,
p < .001]). A follow-up analysis of 7 studies and 7 comparisons
revealed that CBT also outperformed control conditions on sec-
ondary outcomes at follow-up (Hedges's g ¼ 0.20 [SE ¼ 0.10, 95% CI:
0.02e0.38, p ¼ .028]).1

Heterogeneity Question 2: does CBT outperform the control conditions on measures

of depression at post-treatment and follow-up?
To test the assumption that the effect sizes of this study were
from a homogeneous sample, a heterogeneity analysis was con- An analysis of measures of depression at post-treatment
ducted. A total of 13 studies and 15 comparisons across pooled time included 9 studies and 11 comparisons and showed that CBT
70 B.O. Olatunji et al. / Behaviour Research and Therapy 58 (2014) 65e74
Fig. 3. Effect size estimates (Hedges' g) for the efficacy of CBT compared to control
conditions on primary symptom reduction.
outperformed control conditions (Hedges's g ¼ 0.64 [SE ¼ 0.12, 95%
CI: 0.41e0.86, p < .001]). The pooled follow-up analysis included a
total of 4 studies and 4 comparisons, where CBT also outperformed
control conditions on depression measures (Hedges's g ¼ 0.35
[SE ¼ 0.13, 95% CI: 0.01e0.61, p ¼ .008]).
Question 3: does higher pre-treatment hypochondriasis/health
anxiety severity and higher pre-treatment depression scores predict
lower effect sizes?
The analysis for pre-treatment severity included 4 studies and 4
comparisons with 218 participants and revealed that higher pre-
treatment severity was not significantly associated with higher
effect size at post treatment (b ¼ 0.00, p ¼ .400). The analysis of
pre-treatment severity at follow up visits included 3 studies and 3
comparisons with 325 participants and revealed that higher pre-
treatment severity was significantly associated with higher effect
size at follow-up (b ¼ 0.03, p < .050). The analysis for pretreatment
depression included 5 studies and 6 comparisons with 292 par-
ticipants where higher pre-treatment severity was significantly
associated with lower effect size at post treatment (b ¼ 0.09,
p < .050).2 Insufficient data was provided to analyze pre-treatment
depression severity at follow-up visits.
Question 4: does effect size vary as a function of control type, blind
assessment, number of sessions, percentage of females, mean age,
and comorbidity?
The following analyses were completed using fully random ef-
fects categorical moderator analyses. The 4 studies and 6 compar-
isons employing wait-list control conditions (Hedges's g ¼ 1.10
[SE ¼ 0.14, 95% CI: 0.83e1.37, p < .001]) showed larger effect sizes
than the 5 treatment as usual control conditions on primary out-
comes across pooled time points (post and follow-up) (Hedges's
g ¼ 0.46 [SE ¼ 0.13, 95% CI: 0.20e0.71, p < .001]). Comparison be-
tween other control types and placebo-controlled studies was not
2
Studies that used the Health Anxiety Inventory were used to examine the
relationship between higher pre-treatment hypochondriasis/health anxiety and the
effect size for CBT. Studies that used the Beck Depression Inventory were used to
examine the relationship between higher pre-treatment depression and the effect
size for CBT. These were the most common measures used.
possible as few studies employed these control conditions (1 pill
placebo and 3 psychological placebos).
There was no significant difference in effect size for post-
treatment outcomes for 7 comparisons and 6 studies with blind
assessors (Hedges's g ¼ 1.11 [SE ¼ 0.19, 95% CI: 0.73e1.49, p  .001])
and 4 comparisons and 3 studies which did not report having blind
assessors (Hedges's g ¼ 0.66 [SE ¼ 0.18, 95% CI: 0.31e1.01, p < .001]).
This comparison (4 comparisons and 4 studies) was not significant
for follow-up outcomes either for the studies with blind assessors
(Hedges's g ¼ 0.72 [SE ¼ 0.19, 95% CI: 0.10 to 0.65, p ¼ .155]) and
the 3 comparisons and 3 studies without blind assessors (Hedges's
g ¼ 0.46 [SE ¼ 0.25, 95% CI: 0.03 to 0.95, p ¼ .068]). Treatment
integrity was reported in 12 studies, but a comparison was not
possible because only 1 study failed to report treatment integrity.
The following analyses were completed using unrestricted
maximum likelihood meta regressions. There was a significant
relation between effect size and number of sessions (11 compari-
sons and 9 studies; b ¼ 0.07, p ¼ .038), with more sessions asso-
ciated with larger effect sizes at post-treatment (Fig. 4). The relation
between effect size and number of sessions at follow-up was not
significant (7 comparisons and 7 studies; b ¼ 0.01, p ¼ .858).
There was no significant relation between effect size and female
percentage at post-treatment (10 comparisons and 8 studies;
b ¼ 0.00, p ¼ .822) nor at the follow-up analysis (7 comparisons
and 7 studies; b ¼ 0.01, p ¼ .373). Additionally, there was not a
significant relation between effect size and mean age at post-
treatment (9 comparisons and 7 studies; b ¼ 0.00, p ¼ .767) or at
follow up (6 comparisons and 6 studies; b ¼ 0.02, p ¼ .465). An
insufficient amount of studies provided information on comor-
bidity percentage; therefore, that analysis was not possible.
Publication bias: “the file drawer problem”
The omission of file drawer studies is a problem because it can
lead to biased meta-analytic estimates if their psychometric prop-
erties (i.e., reliability and validity) differ from psychometric prop-
erties of published studies (Rosenthal, 1991). Rosenthal's fail-safe N
was an ingenious response to this file drawer problem in the
integration of research (Orwin, 1983; Rosenthal, 1979). This
formula allowed for a direct assessment of the threat posed by
sampling bias in the literature search (Orwin, 1983). Rosenthal
suggested the following equation to compute a fail-safe
2
N:X ¼ KðKZ  2:706Þ=2:706 where K is the number of studies in
the meta-analysis and Z is the mean Z obtained from the K studies
(Rosenthal, 1991). Rosenthal also suggested that findings may be
considered robust if the required number of studies (X) to reduce
the overall effect size to a non-significant level exceeded 5K þ 10,
which in this study would be 85 (Rosenthal, 1991). The analyses
conducted on the 15 comparisons amongst the 13 studies revealed
that it would require more than 166 current or future unpublished
studies with an effect size of 0 in order to bring the overall effect
size of the primary analyses within the non-significant range. This
suggests that the findings in this meta-analysis are robust as the
fail-safe-N (166) is much larger than the convention for robust
cutoff or 5K þ 10 (85).
Discussion
Traditionally, hypochondriasis was considered resistant to psy-
chotherapy, possibly because of the shortcomings of psychody-
namic and psychoanalytic conceptualizations, which historically
dominated the treatment of this condition. These models were
removed from behavior, cognition, and biology, and the treatments
derived from them produced little enduring benefit. Within the last
two decades, however, a model of hypochondriasis as “health
 B.O. Olatunji et al. / Behaviour Research and Therapy 58 (2014) 65e74
Regression of Number of Sessions on Hedges's g
2.00
1.80
1.60
1.40
Hedges's g
1.20
1.00
0.80
0.60
0.40
0.20
0.00
3.13 4.53 5.94 7.34
Number of Sessions
Fig. 4. The relationship between CBT for hypochondriasis/health anxiety effect size and number of sessions.
anxiety” has been advanced that draws from the cognitive (i.e.,
dysfunctional beliefs, body vigilance, anxiety sensitivity, intoler-
ance of uncertainty) and behavioral (i.e., avoidance, safety-seeking)
processes implicated in the development of other anxiety disorders
(Olatunji et al., 2009). This model, depicted in Fig. 1, has informed
the application of CBT to the treatment of hypochondriasis by
helping patients recognize and modify faulty beliefs about illness
and eliminate behavioral responses that prevent the self-correction
of faulty beliefs (Taylor & Asmundson, 2004). The present investi-
gation employed meta-analysis to examine the efficacy of this
approach to the treatment of hypochondriasis/health anxiety. The
findings showed that CBT outperformed control conditions on
primary symptom outcome measures at post-treatment showing a
large effect size. This finding is consistent with prior meta-analyses
demonstrating that CBT is effective in reducing symptoms of hy-
pochondriasis (Taylor et al., 2005; Thomson & Page, 2007). How-
ever, the present study expands this prior research by including a
number of RCTs that have been published since these previous
meta-analyses, and thus adds to the evidence base of CBT for hy-
pochondriasis/health anxiety.
The present investigation also found that CBT outperformed
control conditions on primary hypochondriasis/health anxiety
symptom outcome measures at follow-up showing a small effect
size. Although the sustained efficacy (the ability to produce lasting
symptomatic changes) of CBT-based approaches has been ques-
tioned (e.g., Eddy, Dutra, Bradley, & Westen, 2004), these meta-
analytic findings suggest that treatment gains attributed to CBT
(relative to controls) are observed after treatment is completed.
While these findings are encouraging, additional research is needed
to adequately determine the extent to which CBT produces longer
lasting symptom changes for patients with hypochondriasis/health
anxiety. This will require future studies to include substantially
longer follow-up intervals so more definitive inferences can be
made regarding the durability of CBT for hypochondriasis/health
anxiety. It is important to note that the overall effect size for CBT
was significantly larger at post-treatment (g ¼ .95) compared to
follow-up (g ¼ 0.34). This suggests that the therapeutic effects of
CBT for hypochondriasis/health anxiety may diminish after the
acute treatment phase. Future research aimed at identifying stra-
tegies that may be employed during and after CBT (i.e., booster
sessions) in order to better sustain treatment gains for patients
with hypochondriasis/health anxiety may prove valuable. Such
approaches may be vital in reducing the health care costs associated
with excessive doctor visits and requests for unnecessary medical
tests that is often observed among those with hypochondriasis/
health anxiety.
71
8.75 10.15 11.55 12.96 14.36 15.77 17.17
Although the present findings suggest that the therapeutic ef-
fects of CBT for hypochondriasis/health anxiety may diminish after
the acute treatment phase, they also suggest that the acute thera-
peutic effects of CBT for patients with hypochondriasis/health
anxiety may generalize to other symptoms. More specifically, CBT
outperformed control conditions on secondary outcome measures
at post-treatment. Although secondary outcome measures did
include assessment of symptoms of hypochondriasis/health anxi-
ety, they also included assessment of psychiatric symptoms broadly
defined, as well as specific symptoms of general anxiety, somati-
zation, and depression. Prior research has shown that hypochon-
driasis/health anxiety does demonstrate significant associations
with general psychopathology (Gropalis, Bleichhardt, Witthoft, &
Hiller, 2012; Weck, Neng, Richtberg, & Stangier, 2012). However,
the present study suggests that CBT that is specifically targeted
towards the reduction of hypochondriasis/health anxiety may also
significantly reduce other symptoms of distress.
The present study also examined if CBT treatments out-
performed control conditions on reducing depressive symptoms
specifically at post-treatment. The experience of depression is quite
common among those with hypochondriasis/health anxiety (Noyes
et al., 1994), so much so that some have speculated that hypo-
chondriasis is a “masked” form of depression (Lesse, 1980). The
present findings revealed that CBT for patients with hypochon-
driasis/health anxiety also resulted in significant reductions in
symptoms of depression. These findings are consistent with prior
work showing that CBT that specifically targets anxiety-related
disorders also outperform control conditions in reducing symp-
toms of depression (Hofmann & Smits, 2008; Olatunji, Davis,
Powers, & Smits, 2013). Research has shown that symptoms of
hypochondriasis/health anxiety may emerge before those of
depression (Noyes et al., 1994; Simon Gureje, & Fullerton, 2001),
suggesting that depressive symptoms may represent a response to
the distress and functional impairment associated with hypo-
chondriasis/health anxiety (Abramowitz, 2004). Given this
functional relationship, directly targeting symptoms of hypochon-
driasis/health anxiety may be expected to also lead to reductions in
symptoms of depression.
The present investigation also examined the extent to which
initial pre-treatment hypochondriasis/health anxiety severity and
pre-treatment depression scores predicted CBT effect sizes. Prior
longitudinal research suggests that severity of hypochondriasis
may be of prognostic significance (Buwalda & Bouman, 2008;
Hartman et al., 2009). Although it is intuitive to predict that
higher pre-treatment hypochondriasis/health anxiety severity will
predict lower effect sizes, the opposite was true in the present
72 B.O. Olatunji et al. / Behaviour Research and Therapy 58 (2014) 65e74
study. That is, higher pre-treatment hypochondriasis/health anxi-
ety was found to be associated with an increase in CBT effect size at
follow-up visits. This finding does replicate those of a recent study
that found that health anxiety at baseline was positively associated
with symptom improvement after internet-based CBT for severe
health anxiety (Hedman, Andersson, et al., 2013; Hedman,
Lindefors, et al., 2013). The present findings also compliment
those of Nakao, Shinozaki, Ahern, and Barsky (2011) who found that
more baseline anxiety predicted larger improvements associated
with a CBT intervention for hypochondriasis. Contrary to conven-
tional wisdom, the finding that psychological treatment might be
more efficacious for high-severity than for low-severity patients
has also been observed in the treatment of other disorders
(Driessen, Cuijpers, Hollon, & Dekker, 2010; Smits et al., 2013;
Smits, Minhajudin, Thase, & Jarrett, 2012).
The question remains as to why CBT is more effective (relative to
controls) for patients with more severe hypochondriasis/health
anxiety. Is this merely regression toward the mean? According to
Driessen et al. (2010), evidence of moderation is most likely to be
found when efficacious treatments are compared with stringent
controls and the sample contains both more and less severe
patients. In such cases, nonspecific treatments may be sufficient for
low-severity patients, but high-severity patients will require a
treatment that has specific effects beyond the simple provision of
treatment as usual in order to fully benefit. Should CBT be proven to
be more efficacious relative to control conditions for patients with
more severe symptoms of hypochondriasis/health anxiety than for
those with less severe symptoms, this does not mean that treat-
ment guidelines should be revised to recommend CBT as a mono-
treatment for patients with severe symptoms. Indeed, the moder-
ated finding of symptom severity should be interpreted with
caution given that the effect was small (b ¼ 0.03) and the clinical
significance of such an effect is not sufficiently apparent. The
question of the incremental efficacy of CBT for this group of patients
requires more head-to-head comparisons between CBT and other
“bona fide” psychological treatments before more definitive in-
ferences can be made.
Unlike pre-treatment hypochondriasis/health anxiety severity,
pre-treatment depression symptom severity was significantly
associated with a decrease in the advantage of CBT over control
conditions at post-treatment. This finding is consistent with recent
research showing that depressive symptoms at baseline were
negatively related to improvement after internet-based CBT for
severe health anxiety (Hedman, Andersson, et al., 2013; Hedman,
Lindefors, et al., 2013). As previously noted, depressive symptoms
often co-occur with hypochondriasis/health anxiety (Noyes et al.,
1994). The presence of severe depression may then impede
response to CBT for hypochondriasis/health anxiety through mul-
tiple mechanisms. For example, depressive symptoms may
decrease motivation and compliance with difficult exposure as-
signments that are often employed as homework in CBT, thereby
preventing meaningful reductions in symptoms of hypochondria-
sis/health anxiety. It is important to note that the finding showing
that pre-treatment depression symptom severity may predict
worse outcomes is not unique to hypochondriasis/health anxiety.
Indeed, prior research has shown that pretreatment levels of
depression also predict worse outcomes for patients with OCD
(Abramowitz, Franklin, Street, Kozak, & Foa, 2000; Stewart, Yen,
Egan, & Jenike, 2006), a disorder that is related to hypochondria-
sis/health anxiety (Abramowitz, 2005). Based on these findings,
there have been efforts to develop and implement a treatment
program specifically for depressed OCD patients (Abramowitz,
2004). Although the present findings did indicate that CBT for hy-
pochondriasis/health anxiety also reduces depressive symptoms,
the findings also suggest that patients with hypochondriasis/health
anxiety that present with severe depression may benefit from
supplementary interventions that specifically target depressive
symptoms.
Examination of the effect size as a function of treatment as usual
and waitlist controls revealed that CBT showed smaller effect sizes
when compared to treatment as usual control conditions than
when compared to waitlist controls. One interpretation of these
findings is that non-specific factors may influence treatment
outcome in CBT for hypochondriasis/health anxiety to some degree.
In fact, examination of the effect size for CBT compared to waitlist
controls and that of CBT compared to treatment as usual control
conditions in the present investigation would suggest that the
differences are far from negligible in magnitude. Examination of
the extent to which other sample-specific and study-specific
characteristics predicted outcome also revealed some important
findings. For example, the CBT effect size for hypochondriasis/
health anxiety did not vary as a function of blind assessment at post
treatment or follow-up.
More CBT sessions were also found to be associated with larger
effect sizes at post-treatment. Although a relationship between
number of CBT treatment sessions and treatment efficacy has not
been consistently observed in other disorders (Abramowitz, 1996;
Rosa-Alca
zar, Sa

nchez-Meca, Go
mez-Conesa, & Marín-Martínez,
2008), this finding suggests that there may be added benefit to
more CBT sessions for patients with hypochondriasis/health anxi-
ety. This finding does appear to be in line with recent work showing
that treatment adherence (i.e. the number of completed treatment
modules) in internet-based CBT for severe health anxiety signifi-
cantly predicted outcome (Hedman, Andersson, et al., 2013;
Hedman, Lindefors, et al., 2013). In addition to the number of ses-
sions, future research is needed to examine if the length of CBT
treatment sessions is associated with outcomes as there may be
some advantages to a more intensive approach to the treatment of
hypochondriasis/health anxiety.
Hypochondriasis is a chronic, disabling, prevalent and costly
disorder that has been generally viewed as refractory to treatment
(Barsky & Ahern, 2004). Although the specific and problem focused
nature of CBT has been the basis of concerns that it may not
generalize to “real-life” patients who frequently present with co-
occurring conditions (Westen, Novotny, & Thompson, 2004), the
present meta-analytic investigation suggests that CBT is effective in
reducing both disorder-specific and nonspecific (e.g., depression)
symptoms for those with hypochondriasis/health anxiety. The
present investigation also identified some moderators of outcome
that may have prognostic and prescriptive implications. A prog-
nostic variable is one that predicts outcome irrespective of the
treatment, whereas a prescriptive variable (often referred to as a
moderator) predicts a different pattern of outcomes between two
or more treatment modalities (Hollon & Beck, 1986). Important
prescriptive indicators were observed in the present study that may
guide future research. For example, the prescriptive indicators
showed that CBT is more effective than control conditions for pa-
tients with more severe symptoms of hypochondriasis/health
anxiety but less effective for patients with more severe symptoms
of depression. However, most RCTs that have been evaluated for the
treatment of hypochondriasis/health anxiety have been CBT
focused. One study did find that patients with hypochondriasis
who received CBT did significantly better on all measures relative to
the wait-list control group, and on a specific measure of health
anxiety compared with short-term psychodynamic psychotherapy
(Sorensen, Birket-Smith, Watta, Buemann, & Salkovskis, 2011).
Furthermore, the short-term psychodynamic psychotherapy group
did not significantly differ from the waiting-list group on any
outcome measures. More examination of the efficacy of other
psychological treatments for hypochondriasis/health anxiety will
 B.O. Olatunji et al. / Behaviour Research and Therapy 58 (2014) 65e74
be valuable in identifying more substantive prescriptive variables
that may inform individually tailored treatment.
Although the present meta-analysis represents a significant
improvement on prior work that was limited by a small number of
studies, some of which were uncontrolled (Taylor et al., 2005;
Thomson & Page, 2007), the use of only RCTs is perhaps another
limitation of the study as it resulted in the availability of a relatively
small number of studies for analysis. Thus, examination of some
moderators (i.e., effects of blind assessors) in the meta-analysis
were potentially underpowered. Accordingly, some significant
moderators had relatively small beta weights and these findings
should be interpreted with some caution. Another issue is that an
insufficient amount of studies provided information on variables
(i.e., comorbidity percentage) that would allow for a more
comprehensive examination of potential moderators of outcome.
This observation highlights the importance of more thorough
assessment and reporting of data in RCTs. Although the use of RCTs
does allow for greater confidence in the present findings, more
RCTs of CBT for hypochondriasis/health will also allow for the
identification of more reliable moderators and mediators of treat-
ment outcome. A recent study examining the mediating role of
putative mechanisms in internet-delivered CBT for severe health
anxiety found that reduced perceived risk of disease, less attention
to bodily symptoms, and reduced intolerance of uncertainty
significantly mediated improvement in health anxiety (Hedman,
Andersson, et al., 2013; Hedman, Lindefors, et al., 2013). The
mediating effects of these processes are consistent with the
cognitive-behavioral model represented in Fig. 1. Future research
delineating additional mediating mechanisms is crucial because
such mechanisms provide important insight into the processes that
should be targeted during treatment.
It is important to note that the very nature of RCTs may make it
difficult to detect meaningful moderators and mediators, especially
if stricter eligibility criteria and overly rigid delivery of CBT pro-
tocols eliminate the very processes that influence how patients
respond to treatment. The use of more effectiveness trials in com-
munity settings may complement RCTs in efforts to identify mod-
erators and mediators of treatment outcome. Indeed, a confluence
of efficacy and effectiveness trials may ultimately reveal what
treatment, by whom, is most effective for patents with hypochon-
driasis/health anxiety, and under which set of circumstances.
Although the newly published DSM-5 has replaced the diagnosis of
hypochondriasis with illness anxiety disorder (APA, 2013), many of
the underlying processes remain the same. Patients with illness
anxiety disorder experience heightened bodily sensations, are
intensely anxious about the possibility of an undiagnosed illness,
are not easily reassured, and engage in excessive safety behaviors to
minimize anxiety about health concerns. Given the overlap in
symptoms, CBT should prove to be equally as effective for illness
anxiety disorder as it is for hypochondriasis.
References3
Abramowitz, J. S. (1996). Variants of exposure and response prevention in the
treatment of obsessiveecompulsive disorder: a meta-analysis. Behavior Therapy,
27, 583e600.
Abramowitz, J. S. (2004). Treatment of obsessive compulsive disorder in patients
who have comorbid depression. Journal of Clinical Psychology, 60, 133e1141.
Abramowitz, J. S. (2005). Hypochondriasis: conceptualization, treatment, and
relationship with obsessiveecompulsive disorder. Annals of Clinical Psychiatry,
17, 211e217.
Abramowitz, J. S., Deacon, B. J., & Valentiner, D. P. (2007). The Short Health Anxiety
Inventory: psychometric properties and construct validity in a non-clinical
sample. Cognitive Therapy and Research, 31, 871e883.
3 health anxiety. PLoS ONE, 8(10), e77752.
*Indicates the studies used in the meta-analysis.
73
Abramowitz, J. S., Franklin, M. E., Street, G., Kozak, M., & Foa, E. B. (2000). Effects of
comorbid depression on response to treatment for obsessiveecompulsive dis-
order. Behavior Therapy, 31, 517e528.
Abramowitz, J. S., & Moore, E. L. (2007). An experimental analysis of hypochon-
driasis. Behaviour Research and Therapy, 45, 413e424.
Abramowitz, J. S., Olatunji, B. O., & Deacon, B. J. (2007). Health anxiety, hypo-
chondriasis, and the anxiety disorders. Behavior Therapy, 38, 86e94.
Abramowitz, J. S., Schwartz, S. A., & Whiteside, S. P. (2002). A contemporary con-
ceptual model of hypochondriasis. Mayo Clinic Proceedings, 77, 1323e1330.
American Psychiatric Association. (2000). Diagnostic and statistical manual of mental
disorders (4th ed. text revision). Washington, D.C: Author.
American Psychiatric Association. (2013). Diagnostic and statistical manual of mental
disorders (5th ed., rev.). Washington, DC: Author.
Barsky, A. (1996). Hypochondriasis: medical management and psychiatric treat-
ment. Psychosomatics, 37, 48e56.
* Barsky, A. J., & Ahern, D. K. (2004). Cognitive behavior therapy for hypochon-
driasis. A randomized trial. Journal of the American Medical Association, 291,
1464e1470.
* Barsky, A. J., Ahern, D., Bauer, A., Nolido, N., & Orav, J. (2013). A randomized trial of
treatments for high-utilizing somatizing patients. Journal of General Internal
Medicine, 28, 1396e1404.
Barsky, A., Barnett, M. C., & Cleary, P. D. (1995). Hypochondriasis and panic disorder:
boundary and overlap. Archives of General Psychiatry, 51, 918e925.
Barsky, A., Fama, J., Bailey, E., & Ahern, D. K. (1998). A prospective 4- to 5-year
study of DSM-III-R hypochondriasis. Archives of General Psychiatry, 55,
737e744.
Beck, A. T., & Emery, G. (2005). Anxiety disorders and phobias: A cognitive perspective
(15th annual ed.). New York: Basic Books.
Bernstein, I. C., Callahan, W. A., & Jaranson, J. (1975). Lobotomy in private practice:
long-term follow-up. Archives of General Psychiatry, 32, 1041e1047.
Beutler, L. E. (1991). Have all won and must all have prizes? Revisiting Luborsky
et al'.s verdict. Journal of Consulting and Clinical Psychology, 59, 226e232.
Borenstein, M., Hedges, L. V., & Rothstein, H. (2007). Meta-analysis d Fixed effects
versus random effects. Biostat Inc. (Tech. Rep).
Bornstein, M., & Rothstein, H. (1999). Comprehensive meta-analysis: A computer
program for research synthesis. New Jersey: Biosta.
* Bourgault-Fagnou, M. D., & Hadjistavropoulos, H. D. (2013). A randomized trial of
two forms of cognitive behaviour therapy for an older adult population with
subclinical health anxiety. Cognitive Behaviour Therapy, 42, 31e44.
Buwalda, F. M., & Bouman, T. K. (2008). Predicting the effect of psychoeducational
group treatment for hypochondriasis. Clinical Psychology and Psychotherapy, 15,
396e403.
* Buwalda, F. M., Bouman, T. K., & Van Duijn, M. A. J. (2006). Psychoeducation for
hypochondriasis: a comparison of a cognitiveebehavioural approach and a
problemsolving approach. Behaviour Research and Therapy, 45, 887e899.
Clark, D. M., Salkovskis, P. M., Hackmann, A., Wells, A., Fennell, M., Ludgate, J., et al.
(1998). Two psychological treatments for hypochondriasis. A randomised
controlled trial. British Journal of Psychiatry, 173, 218e225.
Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.).
Hillsdale, NJ: Erlbaum.
Creed, F., & Barsky, A. (2004). A systematic review of the epidemiology of somati-
sation disorder and hypochondriasis. Journal of Psychosomatic Research, 56,
391e408.
Driessen, E., Cuijpers, P., Hollon, S. D., & Dekker, J. J. (2010). Does pretreatment severity
moderate the efficacy of psychological treatment of adult outpatient depression?
A meta-analysis. Journal of Consulting and Clinical Psychology, 78, 668e680.
Eddy, K., Dutra, L., Bradley, R., & Westen, D. (2004). A multidimensional meta-
analysis of pharmacotherapy for obsessiveecompulsive disorder. Clinical Psy-
chology Review, 24, 1011e1030.
Fallon, B., Javitch, J., Hollander, E., & Liebowitz, M. (1992). Hypochondriasis and
obsessive compulsive disorder: overlaps in diagnosis and treatment. Journal of
Clinical Psychiatry, 52, 457e460.
Greeven, A., van Balkom, A. J., van der Leeden, R., Merkelbach, J. W., van den
Heuvel, O. A., & Spinhoven, P. (2009). Cognitive behavioral therapy versus
paroxetine in the treatment of hypochondriasis: an 18-month naturalistic
follow-up. Journal of Behavior Therapy and Experimental Psychiatry, 40, 487e496.
* Greeven, A., van Balkom, A. J., Visser, S., Merkelbach, J. W., van Rood, Y. R., van
Dyck, R., et al. (2007). Cognitive behavior therapy and paroxetine in the treat-
ment of hypochondriasis: a randomized controlled trial. American Journal of
Psychiatry, 164, 91e99.
Gropalis, M., Bleichhardt, G., Wittho € ft, M., & Hiller, W. (2012). Hypochondriasis,
somatoform disorders, and the anxiety disorders: sociodemographic variables,
general psychopathology, and naturalistic treatment effects. Journal of Nervous
and Mental Disease, 200, 406e412.
Gureje, O., Ustun, T. B., & Simon, G. E. (1997). The syndrome of hypochondriasis: a
cross-national study in primary care. Psychological Medicine, 27, 1001e1010.
Hartman, T. C., Borghuis, M., Lucassen, P., van de Laar, F., Speckens, A., & van Weel, C.
(2009). Medically unexplained symptoms, somatisation disorder and hypo-
chondriasis: course and prognosis. A systematic review. Journal of Psychoso-
matic Research, 66, 363e377.
Hedges, L. V., & Olkin, I. (1985). Statistical methods for meta-analysis. London: Aca-
demic Press.
Hedman, E., Andersson, E., Andersson, G., Lindefors, N., Lekander, M., Rück, C., et al.
(2013). Mediators in internet-based cognitive behavior therapy for severe
74 B.O. Olatunji et al. / Behaviour Research and Therapy 58 (2014) 65e74
* Hedman, E., Andersson, E., Andersson, G., Ljo
tsson, B., Rück, C., Asmundson, G.,
et al. (2011). Internet-based cognitive-behavioural therapy for severe health
anxiety: randomized controlled trial. British Journal of Psychiatry, 198,
230e236.
Hedman, E., Lindefors, N., Andersson, G., Andersson, E., Lekander, M., Rück, C., et al.
(2013). Predictors of outcome in Internet-based cognitive behavior therapy for
severe health anxiety. Behaviour Research and Therapy, 51, 711e717.
Hofmann, S. G., & Smits, J. A. (2008). Cognitive-behavioral therapy for adult anxiety
disorders: a meta-analysis of randomized placebo-controlled trials. Journal of
Clinical Psychiatry, 69, 621e632.
Hollon, S. D., & Beck, A. T. (1986). Cognitive and cognitive-behavior therapies. In
S. L. Gartield, & A. E. Bergin (Eds.), Handbook of psychotherapy and behavior
change (Vol. III); (pp. 443e482). New York: Wiley.
House, A. (1989). Hypochondriasis and related disorders: assessment and man-
agement of patients referred for a psychiatric opinion. General Hospital Psy-
chiatry, 11, 156e165.
Katon, W. J., & Walker, E. A. (1998). Medically unexplained symptoms in primary
care. Journal of Clinical Psychiatry, 59, 15e21.
Ladee, G. A. (1966). Hypochondriacal syndromes. Amsterdam: Elsevier.
Lesse, S. (1980). Masked depression-the ubiquitous but unappreciated syndrome.
Psychiatric Journal of the University of Ottawa, 5, 268e273.
Marcus, D. K., Gurley, J. R., Marchi, M. M., & Bauer, C. (2007). Cognitive and
perceptual variables in hypochondriasis and health anxiety: a systematic re-
view. Clinical Psychology Review, 27, 127e139.
* McManus, F., Surawy, C., Muse, K., Vazques-Montes, M., & Williams, M. G. (2012).
A randomized clinical trial of mindfulness-based cognitive therapy versus un-
restricted services for health anxiety (hypochondriasis). Journal of Consulting
and Clinical Psychology, 80, 817e828.
Nakao, M., Shinozaki, Y., Ahern, D. K., & Barsky, A. J. (2011). Anxiety as a predictor of
improvements in somatic symptoms and health anxiety associated with cog-
nitiveebehavioral intervention in hypochondriasis. Psychotherapy and Psycho-
somatics, 80, 151e158.
Noyes, R., Kathol, R. G., Fisher, M. M., Phillips, B. M., Suelzer, M. T., & Woodman, C. L.
(1994). Psychiatric comorbidity among patients with hypochondriasis. General
Hospital Psychiatry, 16, 78e87.
Olatunji, B., Davis, M., Powers, M., & Smits, J. (2013). Cognitive-behavioral therapy
for obsessiveecompulsive disorder: a meta-analysis of treatment outcome and
moderators. Journal of Psychiatric Research, 47, 33e41.
Olatunji, B. O., Deacon, B. J., & Abramowitz, J. S. (2009). Is hypochondriasis an
anxiety disorder? British Journal of Psychiatry, 194, 481e482.
Orwin, R. G. (1983). A fail safe N for effect size in meta-analysis. Journal for
Educational Statistics, 8, 157e159.
Paul, G. L. (1967). Strategy of outcome research in psychotherapy. Journal of
Consulting Psychology, 31, 109e118.
Rosa-Alca
zar, A. I., Sa

nchez-Meca, J., Go
mez-Conesa, A., & Marín-Martínez, F. (2008).
Psychological treatment of obsessiveecompulsive disorder: a meta-analysis.
Clinical Psychology Review, 28, 1310e1325.
Rosenthal, R. (1979). The “file drawer problem” and tolerance for null results.
Psychological Bulletin, 86, 638e641.
Rosenthal, R. (1991). Meta-analytic procedures for social research (rev. ed.). Thousand
Oaks, CA, US: Sage Publications, Inc.
* Seivewright, H., Green, J., Salkovskis, P., Barrett, B., Nur, U., & Tyrer, P. (2008).
Cognitive-behavioural therapy for health anxiety in a genitourinary medicine
clinic: randomised controlled trial. British Journal of Psychiatry, 193, 332e337.
Simon, G. E., Gureje, O., & Fullerton, C. (2001). Course of hypochondriasis in an
international primary care study. General Hospital Psychiatry, 23, 51e55.
Smits, J. A. J., Hofmann, S. G., Rosenfield, D., DeBoer, L. B., Costa, P., Simon, N. M.,
et al. (2013). d-cycloserine augmentation of cognitive behavioral group therapy
of social anxiety disorder: prognostic and prescriptive variables. Journal of
Consulting & Clinical Psychology, 81, 1100e1112.
Smits, J. A. J., Minhajuddin, A., Thase, M. E., & Jarrett, R. J. (2012). Outcomes of acute
phase cognitive therapy in outpatients with anxious versus nonanxious
depression. Psychotherapy & Psychosomatics, 81, 153e160.
* Sorensen, P., Birket-Smith, M., Wattar, U., Buemann, I., & Salkovskis, P. (2011).
A randomized clinical trial of cognitive behavioural therapy versus short-term
psychodynamic psychotherapy versus no intervention for patients with hypo-
chondriasis. Psychological Medicine, 41, 431e441.
* Speckens, A. E. M., van Hemert, A. M., Spinhoven, P., Hawton, k. E., Bolk, J. H., &
RooiJmans, H. (1995). Cognitive behavioural therapy for medically unexplained
physical symptoms: a randomized controlled trial. BMJ, 311, 1328e1332.
Speckens, A. E. M., Spinhoven, P., van Hemert, A. M., Bolk, J., & Hawton, K.
(1997). Cognitive behavioural therapy for unexplained physical symptoms;
process and prognostic factors. Behavioural and Cognitive Psychotherapy, 25,
291e294.
Stewart, S. E., Yen, C. H., Egan, S. D., & Jenike, M. (2006). Outcome predictors for
severe obsessiveecompulsive patients in intensive residential treatment. Jour-
nal of Psychiatric Research, 40, 511e519.
* Sumathipala, A., Siribaddana, S., Abeysingha, M. R. N., De Silva, P., Dewey, M.,
Prince, M., et al. (2008). A randomised controlled trial of cognitive behavioural
therapy versus structured care for medically unexplained symptoms: a ran-
domized controlled trial. British Journal of Psychiatry, 193, 51e59.
Taylor, S., & Asmundson, G. (2004). Treating health anxiety: A cognitive-behavioral
approach. New York: Guilford Press.
Taylor, S., Asmundson, G., & Coons, M. (2005). Current directions in the treatment of
hypochondriasis. Journal of Cognitive Psychotherapy: An International Quarterly,
19, 285e304.
Thomson, A. B., & Page, L. A. (2007). Psychotherapies for hypochondriasis. Cochrane
Database of Systematic Reviews, 4, 1e43.
* Visser, S., & Bouman, T. K. (2001). The treatment of hypochondriasis: exposure
plus response prevention vs. cognitive therapy. Behaviour Research and Therapy,
39, 423e442.
Walker, J., Vincent, N., Furere, P., Cox, B., & Kjernisted, K. (1999). Treatment pref-
erence in hypochondriasis. Journal of Behavior Therapy and Experimental Psy-
chiatry, 30, 251e258.
* Warwick, H. M. C. (1996). A controlled trial of cognitive-behavioural treatment of
hypochondriasis. British Journal of Psychiatry, 169, 189e195.
Warwick, H. M. C., Clark, D. M., Cobb, A. M., & Salkovskis, P. M. A. (1996). Controlled
trial of cognitive-behavioural treatment of hypochondriasis. British Journal of
Psychiatry, 169, 189e195.
Warwick, H., & Salkovskis, P. (1990). Hypochondriasis. Behaviour Research and
Therapy, 28, 105e117.
Warwick, H., & Salkovskis, P. (2001). Cognitive-behavioral treatment of hypo-
chondriasis. In V. Starcevic, & D. R. Lipsitt (Eds.), Hypochondriasis: Modern per-
spectives on an ancient malady (pp. 314e328). New York: Oxford University
Press.
Weck, F., Neng, J. M., Richtberg, S., & Stangier, U. (2012). Dysfunctional beliefs about
symptoms and illness in patients with hypochondriasis. Psychosomatics, 53,
148e154.
Westen, D., Novotny, C. M., & Thompson-Brenner, H. (2004). The empirical status of
empirically supported psychotherapies: assumptions, findings, and reporting in
controlled clinical trials. Psychological Bulletin, 130, 631e663.