European Psychiatry 22 (2007) 69e74
Epidemiology of speci?c phobia subtypes: Findings from
the Dresden Mental Health Study
Eni S. Becker a,*, Mike Rinck a, Veneta Tu¨rke b, Petra Kause b, Renee Goodwin c,
Simon Neumer d, Ju¨rgen Margraf e
a Radboud University Nijmegen, Clinical Psychology, Behavioural Science Institute, PO Box 9104, 6500 HE, Nijmegen, Netherlands
b University of Technology Dresden, Clinical Psychology and Psychotherapy, Dresden, Germany
c Columbia University, Department of Epidemiology, New York, NY, USA
d Regionscenter for barne og ungdomspsykiatrie, Oslo, Norway
e University of Basel, Clinical Psychology and Psychotherapy, Basel, Switzerland
Received 12 September 2006; accepted 14 September 2006
Available online 8 December 2006
This study determined the prevalence, age of onset, comorbidity, and impairment associated with speci?c phobia subtypes in the community.
Data were drawn from the Dresden Mental Health Study (N ¼ 2064), a representative community-based sample of young women in Dresden,
Germany. The lifetime prevalence of any speci?c phobia was 12.8%, with subtypes ranging in prevalence between 0.2% (vomiting, infections)
and 5.0% (animals). There were signi?cant differences in the mean age of onset of speci?c phobias. Signi?cant differences in comorbidity pat-
terns also emerged between subtypes. No signi?cant differences were found in level of impairment associated with the subtypes. The ?ndings
suggest that speci?c phobias are common among young women and that they differ in prevalence, associated comorbidity, and mean age of
onset. These data suggest signi?cant differences in the phenomenology and clinical signi?cance of speci?c phobia subtypes.
Ó 2006 Elsevier Masson SAS. All rights reserved.
Keywords: Speci?c phobias; Subtypes of phobias; Onset; Comorbidity
in life, and that they predict the subsequent onset of depressive
and substance use disorders in adolescence and adulthood .
For a long time, speci?c phobias (SPs) were considered
Speci?c phobias all share a ‘‘marked and persistent fear of
a common, but inconsequential psychological problem. How-
clearly discernible circumscribed object or situations’’ accord-
ever, increasing evidence has shown that SPs are clinically
ing to the DSM-IV . DSM-IV differentiates among four
signi?cant and relatively understudied disorders. First, epidemi-
types of SPs: animal type, natural environment type, blood-
ologic studies show that SPs are the most common mental disor-
injection-injury type, and situational type. There is an addi-
ders in women [12,17], with lifetime prevalence rates doubling
tional ?fth category, ‘‘other’’, including phobias with fear of
those of men . Second, studies have documented signi?cant
noise and costumes. These categories were arrived at by clus-
rates of impairment, distress, and comorbidity associated with
tering according to content of fears, but there is little data
SPs . Although comorbidity rates are assumed to be moder-
about the coherence and validity of these clusters. There is
ate compared to other mental disorders, there are not enough
some evidence that the various subtypes have different ages
data to draw ?rm conclusions. Third, longitudinal epidemio-
of onset. However, data suggest that phobias might be hetero-
logic studies suggest that SPs have a mean age of onset early
geneous even within DSM-IV categories. Unfortunately, rela-
tively few studies have speci?cally examined subtypes, and
even fewer have differentiated between single SPs. Moreover,
* Corresponding author. Tel.: þ31 (0) 24 361 2668; fax: +31 (0) 24 361 5594.
E-mail address: firstname.lastname@example.org (E.S. Becker).
interpretation of such comparisons is often aggravated by the
0924-9338/$ - see front matter Ó 2006 Elsevier Masson SAS. All rights reserved.
E.S. Becker et al. / European Psychiatry 22 (2007) 69e74
inclusion of different SPs in subtypes, e.g. some studies of sit-
axis I disorders according to DSM-IV, for lifetime and point-
uational phobias include only claustrophobia , others only
prevalence. The F-DIPS is an earlier, almost identical version
driving phobia . In sum, previous community-based studies
of the DIPS , based on the ADIS-L . The following
have mainly reported general categories of phobia prevalence
disorders can be diagnosed: all anxiety disorders, all affective
and associated morbidity. Yet, relatively little is known about
disorders, the research-diagnosis mixed anxiety-depression, hy-
SP subtypes in the community.
pochondriasis, somatization disorder, conversion disorder and
Furthermore, several methodological features of previous
pain disorder, substance abuse and dependence, bulimia, an-
studies have limited the ability to generalize the ?ndings. First,
orexia, and some children’s disorders (separation anxiety, atten-
?ndings from clinical studies can not be generalized to the
tion-de?cit and disruptive behavior disorders, elimination
population, especially when only a small percentage of those
disorders). Furthermore, there is a socio-demographic section,
with SPs seek treatment. Second, very few epidemiological
a screening for psychosis, a screening for the general medical
studies have looked at comorbidity of SPs , and there is
condition and medication, a short section on family history of
even less information available on the comorbidity of SP sub-
psychological disorders, and a section about treatment for psy-
types. Studies with clinical samples are dif?cult to interpret in
chological disorders. The F-DIPS has good reliability (Kappa
terms of comorbidity, since it is over represented in clinical
for anxiety disorders 0.64, affective disorders 0.71, somatoform
samples [15,20]. Yet, comorbidity is of great importance, be-
disorders 0.66, Yule for substance abuse 0.85 and for eating dis-
cause it informs us of the relations between different disorders,
orders 0.94; ). Interviewers were either psychologist, physi-
and it may also provide valuable information about differences
cians, or psychology students in their last years of training.
between subtypes. More importantly, SPs are often the earliest
All underwent an extensive one-week training. All interviewers
manifestation of psychological disorders, and prevention strat-
attended supervision bi-weekly. Specially trained supervisors
egies will bene?t from knowing more about comorbidities.
proofread every interview.
Against this background, the goals of the current study
In addition to the diagnoses, the severity of the disorders
were threefold. First, the study determined the prevalence of
was rated on a scale ranging from 0 to 8 (0 ¼ no severity to
SP subtypes in the community among women. Second, the
8 ¼ maximum severity). Furthermore, the interviewer rated
study assessed the mean age of onset of each subtype. Third,
the degree of distress and of impairment due to the diagnoses
the study investigated associations between SP subtypes and
on scales ranging from 0 (no distress or impairment) to 8 (se-
psychiatric comorbidity and impairment.
vere distress and impairment). Interviewers were extensely
trained in administering those scales. There were several ques-
tions to pinpoint the onset of the disorder. Due to the early on-
set of SP, the exact date was hard to determine. Participants
The Dresden Study of mental health is a prospective epide-
not remembering the onset were not included in the onset anal-
miological study designed to collect data on the prevalence,
yses. If participants claimed always to have suffered from
incidence, course, and risk factors of mental disorders. The re-
a certain fear, the age of onset was set to 3 years.
sults presented here are from the baseline survey, which was
conducted from July 1996 to September 1997.
2.3. Analytic strategy
Data were analyzed using the Statistical Package for Social
Sciences, SPSS, Windows, German Version 8.0. Statistical
The sample was drawn from the Dresden government regis-
tests included analyses of variance and Chi-square tests. The
try of residents. 5204 women were located and eligible (age 18e
bivariate comorbidities were obtained by estimating odds ra-
24, German) for the study. From this sample, 2064 ?lled out
tios using the ‘‘risk’’ procedure in the SPSS software package.
questionnaires and took part in the structured interview of men-
Cumulative incidence (in percent) curves are presented to il-
tal disorders, and 998 ?lled out questionnaires only, yielding
lustrate age of onset.
a response rate of 58.8%. The demographic data of these women
did not differ from the data of those who did not reply. The re-
sponse rate is somewhat lower than in other studies. One reason
might be that women of that age group usually have the lowest
3.1. Socio-demographic data
response rate (Wittchen, personal communication). Further-
more, the data were collected shortly after the German reuni?-
The majority of the participants had a partner or spouse
cation, when people were still very mistrustful regarding
(66.6%), but very few were married (4.5%), separated or di-
surveys. And third, participants could not be reimbursed. Data
vorced (0.4%). About half of the women were living with their
of the 2064 interviewed participants are reported here.
parents; about a quarter with a partner or spouse and about
14% lived alone. 7% had children of their own. A few women
2.2. Diagnostic assessment
were still attending school (3.6%); about 40% were students at
the university. Very few dropped out of school without a degree
The diagnostic assessment was based on the F-DIPS .
(0.4%), consistent with mandatory German school law
The F-DIPS is a structured interview allowing the diagnosis of
(school-leaving certi?cate). A minority of the participants
E.S. Becker et al. / European Psychiatry 22 (2007) 69e74
went to a ‘‘Hauptschule’’ (3.7%), the lowest level of school
(e.g., situation phobias). This is due to some participants re-
education, approximately one third went to the medium level
porting situational phobias that did not ?t into the given sub-
of schooling (Realschule und Polytechnische Oberschule),
groups (e.g., fear of ships). Animal phobia (mean 6.2) and
and 58% ended schooling with a degree that allows them to
environmental phobia (mean 6.4), start early in life. Interest-
attend university (Abitur). Almost half of the young women
ingly, phobia of heights (mean 9.1) has a later onset than the
were working, 31.3% of the whole sample full-time, 16%
other environmental phobias. Most SPs begin around the age
part time. Approximately 5% were unemployed. There were
of 8, although situational phobias (mean 13.4) start much later.
no differences concerning sociodemographic data between
Here phobia of lifts starts earlier around the age of 9 years. A
the different SP subtypes or fears.
closer look at physical phobias shows that phobia of infections
starts later than phobia of doctors or of vomiting. Fig. 1 shows
3.2. Prevalence rates of SPs
the age curves for SP subgroups. The age curve for the cate-
gory ‘‘other’’ shows a discontinuity. This is due to phobias
As shown in Table 1, 12.8% of the young women ful?lled
of noise and costumes starting very early in life, whereas the
the criteria of a lifetime prevalence (LP) of at least one SP. The
other phobias like exam phobia start much later. The curves
point prevalence (PP; disorder was present in the last 7 days up
show that animal and environmental phobias cluster together;
to the interview) for all SPs amounted to 9.9%, and the 12-
blood phobia being just slightly later. Phobia of heights lies in-
month prevalence to 10%. Animal phobias were reported
between, and situational phobias start much later.
most frequently (PP 4.3%). They were followed by blood pho-
Duration needs careful interpretation, since the participants
bias (1.8%), physical phobias (fear of seeing a doctor PP 1%,
are still very young and therefore phobias starting later in life
vomiting PP 0.1%, and fears of contagion 0.1%), and heights
will necessarily have shorter durations. Nevertheless, the mean
(1.6%). Heights are usually clustered with other environmental
durations show that most phobias are persistent. They begin
phobias, as storms and water. However, phobias of storms (PP
while the women are still children and persist into adulthood.
0.1%) and water (PP 0.3%) occurred rarely. Situational pho-
bias were evident in 1.9% of the women, with driving phobia
3.4. Comorbidity of SPs
being the most common (PP 0.9%), followed by phobias of
lifts and enclosures (PP 0.5%) and ?ights (PP 0.4%). The
Comorbidity is illustrated in Table 3. Comorbidity was de-
‘‘other’’ category contains phobias of noise, costumes, and
?ned as the occurrence of disorders at a lifetime prevalence
exams and occurred hardly at all. The 12-month prevalence
level. The lifetime prevalence of the disorders were: 28.3% for
and the lifetime prevalence showed a similar pattern of occur-
anxiety disorders, 13.7% for affective disorders, 3.2% for soma-
rence. This suggests that SPs are rather stable since prevalence
toform disorders, 2.2% for substance related disorders, 4% for
rates do not change much over the different time frames.
eating disorders, and 9.8% for disorders usually occurring in
childhood. An odds ratio (OR) larger than 1 indicates a positive
3.3. Age of onset of SPs
relationship between SP and other mental disorders. As the con-
?dence intervals (CIs) show, however, not all of these relation-
Table 2 shows the age of onset (median and mean) and the
ships are reliable. Furthermore, the N of some cells were so
duration of SPs. The number of participants is sometimes
small that the CIs are very wide. Thus, only ORs that are signif-
higher in the aggregated groups than in the speci?c groups
icant with Ns bigger than 4 are interpreted and printed in bold in
Prevalence rates in percent and N for speci?c fears and speci?c phobias
Life time prev. phobias
12 month prev. phobias
Point prev. phobias
(N ¼ 326)
(N ¼ 264)
(N ¼ 243)
Environment (except heights)
Blood, injuries and shots
E.S. Becker et al. / European Psychiatry 22 (2007) 69e74
with substance use disorders and childhood disorders, but the
Mean age of onset and mean duration of speci?c phobias
wide CIs have to be taken into account, and therefore the signif-
icant associations have to be interpreted with caution. Height
phobia, the other environmental phobia, is only associated
with anxiety disorders, showing a different pattern from the
other environmental disorders. Phobia of blood and injury is re-
lated to anxiety disorders, too, and also to disorders usually oc-
curring in childhood. Physical phobias are also related to the
anxiety disorders. The situational phobias show signi?cant asso-
ciations with affective disorders, disorders usually occurring in
childhood, and probably also with substance use disorders.
3.5. SPs and their effect on quality of life
Table 4 shows indicators of the severity of SP and of the de-
gree of impairment in everyday life. All ratings were provided
by interviewers. As the results show, there were almost no dif-
ferences concerning the severity ratings between the different
SPs. On average, all disorders displayed moderate severity
(values slightly above 4 on a scale ranging from 0 to 8) with
comparable standard deviations, and an analysis of variance
revealed no differences between them (F(6/307)
n.s.). The same was true for the impairment caused by the
¼ 0.56, n.s.). However, differences emerged
regarding the ratings of distress in everyday life (F(6/
¼ 2.47, p ¼ 0.02), with phobias of heights being the least
distressing and ‘‘other’’ SPs the most. To aid interpretation,
the current ratings may be compared to those of a group of
women who had other mental disorders than SP at the time
of the interview (N
¼ 155; all other diagnoses included).
These women showed similar ratings, with mean severity of
4.00 (SD 1.38), mean impairment 3.41 (SD 1.81), and mean
Table 3. Animal phobias are signi?cantly associated with all
distress 4.20 (SD 1.74).
other disorder groups except eating disorders. Interestingly,
the eating disorders show no relation to any speci?c phobia. En-
vironmental phobias (of water and storm) might be associated
Speci?c phobias are a common problem for young women.
The prevalence rates of SPs in our sample (12.4%) are consis-
tent with several previous studies (e.g. 10.2% ), 14.9%
, and 10.1% ) with others report slightly higher rates
(e.g. 22.7% ), 25.9% , or 20.2% ). Comparison of
prevalence rates reported here to those of other studies is dif-
?cult as studies differ with regard to the classi?cation methods
used. Age, gender, and a varying number of feared objects or
situations included in the survey also vary. Our ?ndings show
animal phobias (LP 5.0%) to be the most prevalent, followed
by blood phobia (LP 2.4%) and phobia of heights (LP 1.9%).
This ?nding is consistent with those of Agras et al.  and
Curtis et al.  who also found animal fears to be the most
prevalent. Taken together, situation phobias were prevalent
among 2.6% of the young women, with fear of driving being
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
the most common.
The mean age of onset for SPs was 7.9. Lindal and Stefans-
son  found a slightly later onset of 9.6 years in their sam-
ple. In our study, animal phobias started early in life
(5.5 years). Similarly, O
¨ st  found the mean age of onset
Fig. 1. Cumulative age for speci?c phobia subtypes.
to be 6.9, while Starcevic and Bogojevic  found
E.S. Becker et al. / European Psychiatry 22 (2007) 69e74
Lifetime comorbidity of speci?c phobias with other mental disorders: N, Odds ratios, and 95%-CIs
N, OR, 95%CI
The reference group for determining the OR was ‘‘no speci?c phobia’’.
a somewhat later onset at 9.7 years. In the Starcevic and Bogo-
eating disorders, and animal phobias were related to all other
jevic study , blood phobias started quite late with
mental disorders. Otherwise, all subgroups showed a different
21.1 years, whereas O
¨ st  found a much earlier age of onset
pattern of comorbidity. Overall, phobias start early in life, and
for blood phobia, 8.8, which is comparable to our ?nding of
pose a risk for developing a second mental disorder. This, and
7.0 years. In contrast, situational phobias started later in life
the rather long duration, make the need for treatment of SP in
around 15e17 years. Consistent with this, Starcevic and Bogo-
childhood and adolescence obvious.
jevics  found a mean of 23.6 years.
In addition, this study investigated how strongly women
In terms of age of onset, our data show that there is wide
were impaired by their speci?c phobia. Speci?c phobias are
variation in speci?c situational phobias, with fear of elevators
often considered less impairing than other disorders, since
starting much earlier than fear of ?ying, and fear of driving be-
the feared object or situation is circumscribed and therefore
ing the latest phobia. The fear of infections appears to begin
its avoidance is much easier than, for example, in social pho-
much later than the phobias of other physical illnesses and
bia or agoraphobia . Regarding ratings of severity, impair-
may be related to an early start of hypochondria rather than
ment, and associated emotional stress, however, SPs did not
speci?c fears. Overall, phobias concerning situations or phys-
differ from other mental disorders. Moreover, there were no
ical fears appeared to be acquired in young adulthood rather
differences between the SP subtypes regarding impairment,
than during childhood. Phobias appear relatively stable and
and only small differences regarding distress. Impairment in
have a long duration. Taken together, this study found some-
everyday life seems to be as high as for other mental disorders
what earlier onset of SPs than other surveys. But since the
in this young sample, thus there is no hint that SPs are
data are retrospective (as in most studies), the exact age of on-
a smaller problem. Overall, the majority of the disorders in
set is hard to determine.
this young sample is not extremely impairing.
This study also assessed speci?c relations between SP sub-
Some limitations regarding the generalizability of this
types and other mental disorders. Almost all subtypes were
study should be noted. First, only women were included in
associated with other anxiety disorders, except the environmen-
this survey. Women usually show higher rates of SP . Sec-
tal phobias. These phobias are probably related to substance use
ond, our sample consisted of young women. Therefore, no
disorders and disorders occurring in childhood, but these rela-
conclusions can be drawn for males or for an older sample. Al-
tions have to be interpreted with care because of the low number
though phobias start early in life, this study showed that some
of cases. Interestingly, no speci?c phobia subtype was related to
start much later than others (e.g., situational phobias), and
therefore some SPs might be underrepresented. Furthermore,
all participants lived in Dresden, Germany. The ECA study
Impairment ratings: means and standard deviations
showed that prevalence rates vary across cities . Therefore,
?ndings from one site might not be representative of ?ndings
at another. As the socio-demographic data showed, more
highly educated women participated in our study, but data
Heights (N ¼ 39)
from the of?cial census show that this is rather typical for
Environment (N ¼ 15)
Dresden, being a city with a large university.
Blood, injuries, injections
The data suggest that SPs are a heterogeneous group. Age of
N ¼ 50)
onset varies widely for different subgroups, as do patterns of
comorbidity. Even the phobias within one subgroup (e.g., natural/
Others (N ¼ 13)
environment or situational phobias) de?ned by DSM-VI differ
E.S. Becker et al. / European Psychiatry 22 (2007) 69e74
regarding age of onset. Age of onset might in some cases be related
 Boyd JH, Rae DS, Thompson JW, Burns BJ, Bourdon K, Locke BZ, et al.
to vulnerable times during normal development. It can be assumed
Phobia: Prevalence and risk factors. Soc Psychiatry Psychiatr Epidemiol
that the likelihood of developing a phobia is related to the exposure
 Brown TA, DiNardo PA, Barlow DH. Anxiety Disorders Interview
to an object or situation that children fear while growing up, thus
Schedule for DSM-IV (ADIS-IV). Albany: Graywind Publications;
following a prepared pattern. Furthermore, some phobias are
clearly related to experience, like fear of driving. Since most par-
 Curtis GC, Magee WJ, Eaton WW, Wittchen HU. Speci?c fears and pho-
ticipants fear driving as a driver and not as a passenger, this phobia
bias. Br J Psychiatry 1998;173:212e7.
can not start earlier than at around 18 years (the legal age for driv-
 Essau CA, Conradt J, Petermann F. Frequency, comorbidity, and psycho-
social impairment of speci?c phobia in adolescents. J Clin Child Psychol
ing in Germany). Interestingly, subgroups of phobias seem to have
their own patterns of comorbidity, also. Physical illness phobias
 Fredrikson M, Annas P, Fischer H, Wik G. Gender and age differences in
and blood phobias, for instance, seem not to be related to affective
the prevalence of speci?c fears and phobias. Behav Res Ther
disorders, in contrast to situational phobias are. Thus, age of onset
and patterns of comorbidity show that SPs may not be a uniform
 Keller A. Die Klassi?kation psychischer Sto¨rungen nach DSM-IV mit
Hilfe eines strukturierten Interviews (F-DIPS). Unpublished dissertation,
group, and where subgroups are not homogeneous, doubts about
University of Heidelberg; 2000.
the validity of the groups are raised.
 Kendler KS, Neale MC, Kessler RC, Heath AC, Eaves LJ. The genetic
epidemiology of phobias in women. Arch Gen Psychiatry 1992;
 Lindal E, Stefansson JG. The lifetime prevalence of anxiety disorders in
Iceland as estimated by the US National Institute of mental health diag-
These data show that SPs are common, impairing, and phenom-
nostic interview schedule. Acta Psychiatr Scand 1993;88:29e34.
enologically heterogeneous among young women in the commu-
 Magee WJ, Eaton WW, Wittchen HU, McGonagle KA, Kessler RC. Ag-
oraphobia, simple phobia, and social phobia in the national comorbidity
nity. In addition, these data carry messages for prevention and
survey. Arch Gen Psychiatry 1996;53:159e68.
classi?cation of SPs. First, the data suggest that age of onset
 Margraf J, Schneider S, Soeder U, Neumer S, Becker ES. F-DIPS: Diag-
may vary widely by phobia subtype and, as such, efforts aimed
nostisches Interview bei Psychischen Sto¨rungen (Forschungsversion).
at primary prevention need to take into account these speci?c dif-
Unpublished manuscript; 1996.
ferences. Second, the results indicate that the use of subtypes em-
 Marks IM. Fears, phobias, and rituals. New York: Oxford University
ployed in the current diagnostic classi?cation system may obscure
 McConaughy SH, Achenbach TM. Comorbidity of empirically based
phenomenological differences within and across subtypes as they
syndromes in matched general population and clinical samples. Child
occur naturally in the community. Future research should look
Psychol Psychiatry 1994;35(6):1141e57.
closely at the differences and their signi?cance for treatment.
¨ st LG. Age of onset in different phobias. J Abnorm Psychol
 Regier DA, Rae DS, Narrow WE, Kaelber CT, Schatzberg AF. Preva-
lence of anxiety disorders and their comorbidity with mood and addictive
disorders. Br J Psychiatry 1998;173:24e8.
The research was supported by grant DLR 01EG9410, Min-
 Robins LN, Helzer JE, Weissman MM, Orvaschel H, Gruenberg E,
Burke JD, et al. Lifetime prevalence of speci?c psychiatric disorders in
istry of Science, Research and Education. We would like to
three sites. Arch Gen Psychiatry 1984;41:949e58.
thank the many people who helped with this study.
 Schneider S, Margraf J. DIPS: Diagnostisches Interview bei psychischen
Sto¨rungen. Berlin: Springer; 2006.
 Smoller JW, Lunetta KL, Robins J. Implications of comorbidity and
ascertainment bias for identifying disease genes. Am J Med Genetics
(Neuropsychiatric Genetics) 2000;96:817e22.
 Agras S, Sylvester D, Oliveau D. The epidemiology of common fears and
 Starcevic V, Bogojevic G. Comorbidity of panic disorders with agorapho-
phobia. Compr Psychiatry 1969;10(2):151e6.
bia and speci?c phobia: Relationship with the subtypes of speci?c
 Antony MM, Brown TA, Barlow DH. Heterogenity among speci?c pho-
phobia. Compr Psychiatry 1997;38(6):315e20.
bia types in DSM-IV. Behav Res Ther 1997;35(12):1089e100.
 Wittchen HU, Nelson CB, Lachner G. Prevalence of mental disorders and
 American Psychiatric Association. Diagnostic and statistical manual of
psychosocial impairments in adolescents and young adults. Psychol Med
mental disorders. 4th ed. Washington, DC: APA; 1994.