Content uploaded by Allen Tien
Author content
All content in this area was uploaded by Allen Tien on Dec 22, 2014
Content may be subject to copyright.
Distributions of Hallucinations in the Population
Allen Y. Tien
From: Department of Mental Hygiene
The Johns Hopkins University School of Hygiene and Public Health
Baltimore, Maryland 21205
301-955-3910
ACKNOWLEDGEMENTS
The Epidemiologic Catchment Area Program is a series of
five epidemiologic research studies performed by independent
research teams in collaboration with staff of the division of
Biometry and Epidemiology (DBE) of the National Institute of
Mental Health (NIMH). During the period of data collection,
the ECA Program was supported by cooperative agreements. The
NIMH Principal Collaborators were Darrel A. Regier, Ben Z.
Locke, and William W. Eaton (10/1/78 - 10/1/83)/Jack Burke
(10/1/83 - 3/1/87); The NIMH Project Officers were Carl A.
Taube and William Huber. The Principal Investigators and Co-
Investigators from the five sites were: Yale University, UO1
MH 34224-- Jerome K. Myers, Myrna M. Weissman, and Gary L.
Tischler; Johns Hopkins University, UO1 MH 33870-- Morton
Kramer, Ernest Gruenberg, and Sam Shapiro; Washington
University, St. Louis, UO1 MH 33883-- Lee N. Robins and John
Helzer; Duke University, UO1 MH 35386-- Dan Blazer and Linda
George; University of California, Los Angeles, UO1 MH 35865--
Marvin Karno, Richard L. Hough, Javier I. Escobar, M. Audrey
Burnam, and Dianne Timbers. This research was supported by
grants from the National Institute of Mental Health (MH41908,
MH14592, MH33870) and the Academic Data Processing Center of
the JHU School of Hygiene and Public Health, and by NIMH
Contract #87M020298301D.
Introduction
Although not specific to any single disorder,
hallucinations are an important type of psychiatric symptom.
The occurrence of visual, auditory, olfactory, or somatic
sensations without external stimuli suggests a significant
mental disturbance (1). In clinical populations
hallucinations are common and are associated with many factors
which can affect brain function, but less in known about
hallucinations in the community. Existing estimates of the
prevalence of hallucinations in non-clinical samples suggest
rates around 10% (2,3,4).
The largest of these studies, the Sidgewick Study, was
published in 1894, and is of particular interest because of
its size and methology. However, the Sidgewick (and other
previous studies) only assessed hallucinations cross-
sectionally in time. The Epidemiologic Catchment Area Program
(ECA), a collaborative multi-site study carried out between
1980 and 1984, provides data at two points in time on adult
community residents (5). This design allows estimates of
incidence as well as prevalence. Questions which this data
can address include: What proportion of the general
population experiences hallucinations? What are the age and
gender incidence distributions? Do different sensory
modalities of hallucinations differ in distribution? Another
question is whether the distribution of hallucinations has
changed over the past century. Comparison of results from the
Sidgewick Study to the ECA Program provides information on
this last question.
METHODS
Data
Data are from two sources, the Sidgewick Study carried
out by the Society for Psychical Research, and the NIMH
Epidemiologic Catchment Area Program (ECA) (2,6). The
Sidgewick Study was carried out between 1889 to 1892, and
studied 17,000 adults (at least 21 years old) primarily in
England, but also in Russia and in Brazil. A standard
interview schedule was created and used by the study. About
90% of the Sidgewick interviewers had a professional level of
education. Many were psychologists or physicians. No
information on their interview training was available.
Between 1980 and 1984, the ECA Program studied adults 18
years and older in New Haven, Baltimore, Durham, St. Louis,
and Los Angeles (7). To measure occurrence of psychiatric
conditions, subjects were assessed twice using the NIMH
Diagnostic Interview Schedule (DIS; 8). The DIS was based on
the DSM-III (9). At baseline 18,572 neighborhood residents
were assessed. At follow-up roughly one year later data was
obtained on 15,258 of the baseline participants.
The ECA interviewers were not clinicians, but nearly all
had at least some college education. At the New Haven site,
interviewers received a total of 24 hours of training, while
at the other four sites interviewers received an average of 53
hours of training. At all sites training consisted of
lectures, videotapes, live demonstrations, homework exercises,
mock interviews, and interactive practice sessions (10).
Confidentiality was explicitly provided in both studies. The
DIS did not enquire about gustatory hallucinations, and the
Sidgewick Study did not inquire about gustatory or olfactory
hallucinations. Both studies made efforts to exclude
sleep-related hallucinations.
In the ECA, assessment of hallucinations began with a
symptom question. For example:
“Have you ever had the experience of seeing something or
someone that others who were present could not see--that is
had a vision when you were completely awake?”
In the Sidgewick Study, the structure of the interview
was similar to the DIS approach. A screening question
(“Schedule A”) inquired as to the occurrence of any
hallucinations. The wording of Schedule A resembles wording
of the DIS items.
“Have you ever, when believing yourself to be completely
awake, had a vivid impression of seeing or being touched by a
living being or inanimate object, or of hearing a voice; which
impression, so far as you could discover, was not due to any
external physical cause?”
In both studies, positive responses to these screening
questions led to a set of probe questions which inquired in
more detail about the nature of the experiences. Here there
were more differences. The Sidgewick Study was interested in
the content of the hallucinations because the study was
motivated by an attempt to prove the existence of telepathic
phenomena. For example, visions of dying relatives were
considered important for this question. Therefor, the probe
questions (“Schedule B”) were designed to detail the
“particulars” of the hallucination and.
The ECA had no specific hypotheses and was motivated by
questions on the occurrence of mental disorders; thus the
probe questions attempted to determine whether reported
hallucinations were due to alcohol or other drug use, any
medical or physical disorder, and whether they caused distress
or interfered with life. Each positive response to a DIS
symptom question was followed by structured probe questions.
The probe questions were designed to assess the severity and
consequences of the symptom, and to exclude the possibility of
an organic cause (i.e., use of drugs, medications, or medical
conditions). There are 5 levels of response coding for
hallucinations in the DIS. The first level is no reported
occurrence of hallucinations. Level 2 indicated the
occurrence of hallucinations but no distress or impairment of
function, and no reported association with alcohol or drug use
of medical or physical problems. Level 3 indicates the
subject believes the hallucination to be due to alcohol or
drug use, and level 4 to be due to a medical or physical
cause. Level 5 indicates the occurrence of a hallucination t
due to drugs or medical problem and causing distress or
deterioration of function. The age distributions of level 2,
level 5, and level 2 and level 5 combined were examined.
Raw age distribution data was not available from the
Sidgewick Study but the paper reported estimates for the
occurrence of hallucinations across ten year age groups. The
age of occurrence was dependent on the recall of the subjects,
and all subjects were over the age of 21 but occurrence as far
back as the age of ten was included. Finally, the information
was presented in terms of events of hallucination rather than
individuals with hallucinations. However, since the overall
ratio of subjects reporting hallucinations to the number of
hallucinations was close to 1, the age distribution of
occurrence of hallucinations must be similar to what would
have been estimated as the age distribution of people with the
occurrence of hallucinations.
The Sidgewick approach to estimate the incidence
distribution has weaknesses. Use of retrospective data to
assess age-incidence hallucinations is problematic because
people have difficulty remembering whether they had an
experience and what age they were when it happened. The
structure of the ECA data permits a stronger approach which
lessens the effects of these problems. By excluding subjects
reporting hallucinations at the initial interview, subsequent
followup reports of hallucinations can be considered to
represent new occurrence. This approach also excludes
subjects with a tendency for false positive responses shown by
positive symptom reports at baseline. Furthermore, when
subjects repeated the DIS there was a global decrease in
frequency of reported symptoms suggesting that subjects were
more conservative in reporting psychopathology (11).
To estimate the age distribution of onset of
hallucinations in the ECA data, a Generalized Additive
Interactive Model (GAIM) program was used (12). The GAIM
approach is an extension of general linear models without the
restriction of using only linear relationships to fit data.
This permits estimating a non-linear age curve which is likely
to better reveal underlying age variations of occurrence.
The DIS approach differentiates benign hallucinations
from those with attributed negative consequences. The age
distributions of these different levels were examined
separately for males and females. However, as the
consequences of experiencing a hallucination may be distinct
from the causes of the hallucination, these levels were
combined together in comparing age distributions by gender.
RESULTS
Sidgewick
Of 17,000 subjects assessed, 2,272 supplied affirmative
responses. 353 of these were considered false positive due to
errors in interviewing such as recording of dreams. A number
of subjects were excluded based on further criteria, shown in
Table 1. After corrections, the sample size was 15,316.
There were 7,717 men and 7,599 women.
Table 2 shows the lifetime prevalence of hallucinations
in both studies by gender. Data from both baseline and
follow-up interviews in the ECA are shown. There are
consistently higher rates for women than for men. The rates
in the ECA are lower than the Sidgewick Study and are lower at
follow-up than at baseline interview. An adjustment could be
made with the Sidgewick data as they provided information on
the proportion of visual hallucinations either occurring
immediately after awakening or while lying in bed awake
(423/1112). These would probably be considered to be sleep-
related and not true hallucinations by modern criteria. For
auditory hallucinations the proportion was 169/494. For
tactile hallucination 79 of 179 occured while the subject was
in bed. An adjusted value for overall prevalence of
hallucinations in the Sidgewick Study is included in Table 2.
Figure 1 shows the estimated age distribution for
occurrence of hallucinations in three sensory modalities in
the Sidgewick data. The rate for visual hallucinations is
much higher than auditory or tactile. Both visual and
auditory hallucinations showed a peak between ages 20-29.
ECA
Figure 2 shows the age distributions for each type of
hallucination by DIS level. There appears to be a higher
proportion of auditory hallucinations associated with distress
or deterioration of function than for the other sensory
modalities. There do not appear to be substantial differences
in the shapes of the distributions for level 2 versus level 5
responses. Somatic hallucinations appear to decline later in
life in contrast to other hallucinations.
Figure 2 shows the gender by age distributions.
Discussion
Conclusions
There appears to have been a change in the relative
proportions of different sensory modalities, with fewer visual
experiences now.
Study biases might also have accounted for the high occurrence
of visual hallucinations in the Sidgewick Study
Overall occurrence of hallucinations appears little changed.
This depends greatly however on the measurement method
The results support the concept of schizophrenia beginning
later in life, consistent with the change in DSM-III-R on age
criteria.
The results are consistent with age-gender differences in
onset of schizophrenia.
The results support different etiologies for different
modalities of hallucinations.
Questions
Has there been a shift in diathesis - stress factors or simply
a change in the expression of these? Perhaps religious factors
or socialization made seeing visions more common in the 19th
century than now.
Is a change in environment partly responsible for differences
in visual hallucinations? That is, what is the effect of
television on the occurrence of visual hallucinations? Could
be tested perhaps in countries with different exposures.
Table 1. Comparison of 1889 to 1980 Data on the Prevalence of
Hallucinations in the General Population
1889
Visual
Male
Female
Total 8.4%
Auditory
Male
Female
Total 3.3%
Tactile
Male
Female
Total 0.9%
Total affirmative answers to the probe were 9.9%, 7.8% male,
12.0% female
Primacy of hallucinations and delusions in psychiatric
phenomenology in modern psychiatry. Interpretation of
hallucinations as evidence for telepathy in Sidgewixk study.
Concept of hallucinations in the bicameral mind theory, and
implications for perspective on schizophrenia, positive and
1. Asaad G and Shapiro B. Hallucinations: Theoretical and
Clinical Overview. Am J Psychiatry 1986, 1088-1097
2. Sidgewick H, Johnson A, Myers FWH, et al. Report on the
Census of Hallucinations. Proceedings of the Society for
Psychical Research, 1894, 34:25-394
3. West DJ. A mass-observation questionaire on hallucinations. J
of the Society for Psychical Research 1948, 34:187-196
4. Posey TB and Losch ME. Auditory hallucinations of hearing
voices in 375 normal subjects. Imagination, Cognition, and
Personality 1983, 3:99-113
5. Eaton WW and Kessler LG, Eds (1985). Epidemiologic Field
Methods in Psychiatry: The NIMH Epidemiologic Catchment Area
Program. Academic Press, New York
6. Eaton WW and Kessler LG, Eds (1985). Epidemiologic Field
Methods in Psychiatry: The NIMH Epidemiologic Catchment Area
Program. Academic Press, New York
7. Eaton WW and Kessler LG, Eds (1985). Epidemiologic Field
Methods in Psychiatry: The NIMH Epidemiologic Catchment Area
Program. Academic Press, New York
8. Robins LN, Helzer JE, Croughan J, and Ratcliff KA. National
Institute of Mental Health Diagnostic Interview Schedule: Its
history, characteristics, and validity. Arch Gen Psychiatry
1981; 38:381-389
negative symptoms, brain anatomy and function. Paper on
verbal hallucinations theories.
Other recent studies on hallucinations in the community.
Distinction between community and clinical samples, context of
behavior for psychopathology. How subject to false positive
and false negative errors are the DIS items. Clinical
Reappraisal study (Eaton and Nestadt paper).
Differentiating sleep related hallucinations from other types,
noted also in the Sidgewick study.
The relationship of schizophrenia to schizophrenia spectrum
and to predicted prevalence of schizotaxia (Meehl). The
historical stability of the findings of these two studies in
terms of prevalence of hallucinations and implication of
stability for rates of schizophrenia.
The link of occupations and psychotic symptoms in Muntaner's
paper. Noisome occupations in particular. Also relationships
to alcohol and drug use, and aging. Education effects.
Although hallucinations alone are not equivalent to
schizophrenia, current models of schizophrenia are based on
concepts of interacting diathesis and stress (Meehl), and it
is possible that hallucinations are also the product of
underlying vulnerabilities and overt events, representing
overlapping of mechanisms. Although substantial difficulties
face separate measurement of each of these constructs in the
population, examination of distributions of hallucinations
might provide clues as to the operation of these factors.
9. American Psychiatric Association: Diagnostic and Statistical
Manual of Mental Disorders, Third Edition. Washington, DC,
American Psychiatric Association, 1980
10. Munson ML, Orvaschel H, Skinner EA, et al. Interviewers:
Characteristics, Training, and Field Work. In Eaton WW and
Kessler LG (Eds.) (1985). Epidemiologic Field Methods in
Psychiatry: The NIMH Epidemiologic Catchment Area Program. New
York: Academic Press
11. Eaton WW, Kramer M, Anthony JC, et al. The incidence of
specific DIS/DSM-III mental disorders: data from the NIMH
Epidemiologic Catchment Area Program. Acta Psychiatr Scand 1989,
79:163-178
12. Hastie T and Tibsharani R. Generalized Additive Models,
Chapman and Hall, London, 1990
