Full text of DOI:10.1111/j.1553-2712.2000.tb00474.x

1272
CONVERSION DISORDER
Glick et al. • SUSPECTED CONVERSION DISORDERS
Suspected Conversion Disorder: Foreseeable Risks
and Avoidable Errors
THOMAS H. GLICK, MD, TOM P. WORKMAN, MD, SLAVA V. GAUFBERG, MD
Abstract. The authors reviewed the occurrence in and attempting to prevent such occurrences. Pro-
their emergency department of cases of serious neu- posed interventions focus on education regarding the
rologic problems initially thought to be conversion difficulty of diagnosis, patient-based risk factors, and
disorders or similar psychogenic conditions. Their physician-based attitudes and thought processes. The
aim is to indicate the significance of this issue for authors also include suggestions for systemic ‘‘safety
emergency physicians (EPs) because of its contribu- nets’’ that will help to ensure quality of care, such as
tion to the incidence of medical errors. Although there appropriate imaging and consultation. Review of
are no national statistics, the authors estimate by ex- texts and journals readily accessible to EPs revealed
trapolation that thousands of such cases probably little attention to this subject. Key words: conversion;
have occurred and large numbers may still occur each hysteria; emergency; error; neurology; psychiatry.
year in the United States, sometimes resulting in pa- ACADEMIC EMERGENCY MEDICINE 2000; 7:
tient injury. They have identified ways of anticipating 1272–1277
MERGENCY physicians (EPs) frequently
confront the problem of acute neurologic
C
ASE REPORTS
E
symptoms in a patient who has no convincing, ob-
jective neurologic findings. This may arouse the
suspicion of a ‘‘psychogenic’’ disorder. The specific
nature of the presumed ‘‘nonorganic’’ problem is
frequently not distinctly formulated, but is often
diagnosed as a conversion disorder (so-called ‘‘hys-
terical conversion’’). In this paper we highlight the
occurrence of diagnostic mistakes, illustrate pit-
falls, and offer suggestions for processes that phy-
sicians and nurses in emergency medicine (EM)
can implement to minimize such errors.
We were stimulated to study this problem by a
series of cases in which excellent caregivers oper-
ating under routine circumstances made diagnos-
tic mistakes or delays that, in retrospect, might
seem extraordinary. Such cases highlight the dif-
ficult challenges encountered in the emergency
care of complicated and unfamiliar patients. We
have pursued this investigation from the point of
view of error detection and prevention with the
purpose of quality improvement and prioritization
of training objectives and content.
Study Setting and Case Finding. We studied
patients presenting to our emergency department
(ED) who came to the attention of the neurology
consultation staff over a six-year period (1994
through 1999) because of suspicion of a conversion
disorder or similar psychogenic explanation of po-
tentially serious neurologic symptoms. Identifica-
tion of cases with misdiagnosis or delayed diag-
nosis of a neurologic syndrome resulted from both
retrospective review of consultation records and
prospective surveillance by means of educating
physicians about our concern over potential occur-
rences of this type. No additional cases that might
have been subsequently evaluated elsewhere were
communicated to us. We were not able to identify
accurately the denominator of all cases in which
conversion or other psychogenic causes of neuro-
logic symptoms were suspected, nor were we able
to specify reliably the number of true conversion
cases during this time period. The correct neuro-
logic diagnosis was verified by definitive findings
on neurologic workup, including imaging. Patients
with actual or suspected pseudoseizure were ex-
cluded from this analysis because they represent a
distinct problem that has received much attention.
We have focused on the recent past in order to
place diagnostic decision making well within the
era of modern imaging and within the time that
we have had access to a nearby resource for emer-
gency magnetic resonance imaging (MRI). The ED
of this small, urban, community-oriented teaching
hospital in the period of this study averaged more
From the Department of Neurology (THG) and the Division of
Emergency Medicine (TPW, SVG), Harvard Medical School,
and the Division of Neurology (THG) and the Department of
Emergency Medicine, The Cambridge Health Alliance (TPW,
SVG), Cambridge, MA.
Received June 7, 2000; revision received July 5, 2000; accepted
July 29, 2000.
Address for correspondence and reprints: Thomas H. Glick
MD, 1493 Cambridge Street, Cambridge, MA 02139. Fax: 617-
665-1671; e-mail: thomas࿝glick@hms.harvard.edu

ACADEMIC EMERGENCY MEDICINE • November 2000, Volume 7, Number 11
1273
than 22,000 visits per year. The separate psychi- tendon reflexes were 1 plus, plantar responses
atric ED averaged more than 5,000 visits per year. flexor, and no sensory deficit was recognized, her
symptoms were assumed to be psychogenic. She
was admitted to the hospital. Two days later an
Definitions. Our definition of conversion disor-
MRI revealed a thoracic epidural abscess com-
der conforms to the criteria of the Diagnostic and
pressing the cord.
Statistical Manual of Mental Disorders, 4th Edi-
tion (DSM-IV),1 which can be paraphrased, in part,
Patient 3. A 43-year-old male prisoner was
as: One or more symptoms or deficits affecting vol-
brought in from jail (just prior to a court appear-
untary motor or sensory function that suggest a
ance) because of complaint of right-sided weak-
neurologic or other general medical condition; as-
ness, numbness, and pain in his neck and back. He
sociation with psychological factors, and lacking
was found to be very hard to examine because of
feigned symptoms; not fully explainable by appro-
his severe pain and apparent inconsistencies in his
priate medical investigation; significant distress or
motor and sensory examination. Conversion dis-
impairment of function, not limited to pain or sex-
order or malingering was suspected and discharge
ual dysfunction.
to jail with prescription of analgesics was planned,
but the patient angrily insisted on a more thor-
‘‘Conversion’’ in practice often serves as a sur-
rogate for a broader variety of nonmalingering
ough workup, since he had ‘‘a history of being mis-
presentations in which the neurologic deficits are
diagnosed’’ for prior problems. With this implicit
believed to be ‘‘unreal.’’ We have used traditional
threat of litigation, the consulting neurologist was
dualistic brain (‘‘organic’’) vs mind (‘‘psychogenic’’)
called, who detected a substratum of weakness be-
terms for convenience, recognizing the need for
neath the variable effort. An MRI demonstrated a
new ways of expressing the difference between
cervical disc herniation with compression of the
presently identifiable and nonidentifiable nervous
cervical cord.
system mechanisms.
In order to assess the educational resources
available to EPs and general physicians, we re-
Patient 4. A hostile, known sociopathic, 39-year-
old male intravenous drug abuser, with a history
viewed more than 30 leading textbooks and man-
of incarceration for major violence, presented late
uals of EM, medicine, family practice, psychiatry,
one night with complaints of abdominal pain, low
and neurology, and searched the journal literature
back pain, and bilateral leg numbness. After a lim-
to identify the guidance provided to nonneurolo-
ited exam, complicated by frightening threats
gists.
made by the patient, the EP suspected that the
problem was psychogenic, either conversion disor-
der or possibly malingering and drug seeking.
However, the EP later reconsidered his degree of
diagnositic certainty and called the neurologist,
who suggested further evaluation, which revealed
pathologic reflexes and urinary retention. Armed
with these ‘‘hard findings,’’ an MRI now appeared
justified and showed spinal cord compression at
Cases. In the six-year period 1994 through 1999,
we identified six cases in which major, presump-
tively neurologic symptoms were initially sus-
pected to be due to conversion disorder (or a su-
perficially similar psychogenic cause), but in which
serious organic pathology was finally diagnosed.
Patient 1. A 66-year-old woman with schizophre- cord level T 9–10, due to Staphylococcus aureus
nia presented with inability to void and complaint abscess.
of ‘‘my legs hurt, I can’t move them.’’ She was re-
Patient 5. A 24-year-old woman in the seventh
month of pregnancy presented to the ED because
of one week of increasing low back pain, numbness
in the legs and the saddle area, and lower limb
weakness. An orthopedic resident, called in con-
sultation, diagnosed a conversion disorder. The
consulting obstetrician harbored doubts, however,
and somewhat apologetically summoned the neu-
rologist. A cauda equina syndrome was diagnosed
and documented by MRI to be due to a herniated
lumbar disc.
garded as a verbally rambling, anxious, and un-
reliable historian with no focal neurologic deficits
and no bona fide neurologic problem. She was ad-
mitted and treated for a urinary tract infection,
but her neurologic complaints continued to be min-
imized as functional. The cause for her leg symp-
toms was not correctly diagnosed until two days
later, when an MRI showed spinal cord compres-
sion due to a compression fracture with a thoracic
disc herniation.
Patient 2. An 82-year-old woman with a history
Patient 6. A 36-year-old woman with a history of
of depression was evaluated because of ‘‘not being
able to get up’’ and because her legs were numb remote psychiatric problems and substance abuse
and painful. Because she ‘‘moved’’ all limbs, deep was evaluated for weakness in her left arm, first

1274
CONVERSION DISORDER
Glick et al. • SUSPECTED CONVERSION DISORDERS
noted while lying on her side in bed. She was ob- ter’s 1965 report from a selective, neurologic refer-
served to drool and to have a possible mild left fa- ral institute, focus on chronic neurologic conditions
cial droop, and she was believed to be undercon- and apply only to the emergency situation in a gen-
cerned. She had just learned that she was eral hospital insofar as the mind-set and diagnos-
pregnant. As she confided this information, she tic process of the clinician are concerned.6–10
‘‘brightened.’’ She did not move her left arm at all
Two of our six cases (patients 5 and 6), would
when requested, but was seen by the triage nurse have been candidates for the diagnosis of true con-
to push back her hair with her left arm. The triage version disorder, were it not for the incomplete
nurse diagnosed ‘‘r/o conversion reaction’’ because neurologic workup. Conversion disorder was less
of seeming inconsistencies in her deficits and be- clearly an appropriate differential diagnosis in pa-
cause of her psychic conflict. The ED medical res- tients 3 and 4, since the putative psychogenicity
ident and the psychiatry resident followed suit was muddied by concerns about malingering and
(based in part on a nondermatomal sensory loss, drug seeking, respectively. Nonetheless, in sus-
judged to be nonanatomic.) The patient would not pecting a psychologic explanation of the motor and
answer some questions, posed by someone stand- sensory deficits, the physicians initially viewed
ing to her left. Sent home for follow-up the next these cases as if they were conversion. Even in the
day, she returned and was admitted to psychiatry. two elder women with background mental illness,
An expedited head computed tomogram (CT) was who would have been unlikely candidates for con-
requested, which revealed a large right frontopa- version disorder because of age and alternative
rietal hematoma, cause undetermined. The neu- psychiatric diagnoses, the tenor of thinking was
rosurgeon elected nonoperative management, and that the neurologic symptoms ‘‘weren’t real.’’ In
the patient improved substantially over time.
contemporary EM and general medical practice,
this is the way that we sometimes see the term
‘‘(hysterical) conversion’’ used—as a surrogate
term for psychogenic, ‘‘functional,’’ ‘‘unreal’’ neuro-
logic problems that are often assumed to be trig-
gered by a psychological event, including stress.
Thus, from a practical standpoint, we are not con-
cerned in the ED with the accuracy of the psychi-
atric terminology, but with the management of
cases falling under the popular rubric of suspected
conversion, broadly construed.
With patients 1 and 2, the delay in diagnosis
and intervention may possibly have worsened the
outcome, and certainly there is a serious potential
for so doing in instances of spinal cord compres-
sion. It is axiomatic that time is of the essence in
diagnosis and treatment of incipient paraplegia or
quadriplegia because of the risk of permanent dis-
ability, thus placing a premium on an early and
high index of suspicion. Thus, from the standpoint
Review of Didactic Sources. Our extensive re-
view of the journal and textbook literature that
EPs might be likely to consult revealed remarkably
little attention to issues of misdiagnosis of conver-
sion. The neurology text by Adams and colleagues
provides quite an extensive discussion of several
facets of conversion, but, as with other sources,
only very limited, specific guidance for avoiding er-
rors is offered.2 Generalizations about the need for
appropriate medical workup are standard fare.
However, the literature does not convey the great
difficulty of diagnosis for EPs and what processes
will contribute to effective ED management of such
presentations.
DISCUSSION
We identified three prior studies that addressed of process, delay in diagnosis should be considered
actual or suspected conversion disorder from an tantamount to misdiagnosis, as should emphasis
ED perspective. Dula and DeNaples drew on ED on vague or marginal diagnoses that distract from
records of patients who were established to have the central problem. In addition, in these cases, as
true conversion disorder.3 More pertinent to our well as with patients 5 and 6, a measure of pre-
paper was a study, published in the psychiatric lit- served function (especially in the absence of ‘‘hard
erature by Fishbain and Goldberg, showing the op- findings’’) appeared to negate the reality of the
posite side of the conversion coin: More than 4,000 neurologic symptoms or signs, although most such
patients seen in 1982 for new psychiatric symp- problems develop by degrees.
toms included three cases (with many striking sim-
With patient 3 the patient’s attitude and behav-
ilarities to ours) of serious neurologic problems ior, seemingly driven by ulterior motives, but ac-
misdiagnosed as conversion.4 Jones and Barklage tually reinforcing the need for definitive workup,
reported two cases of misdiagnosed brain tumor, prevented a potentially bad outcome. A nonneuro-
one of which involved the ED, and the authors em- logic specialist ‘‘pulled the chestnuts from the fire’’
phasized the hazards of the diagnosis of conversion with patient 5 by refusing to accept a diagnosis of
disorder.5 The classic studies of misdiagnosis of conversion, and an EP did likewise with patient 4
conversion disorder, stimulated especially by Sla- by critically reflecting and making a last-minute,

ACADEMIC EMERGENCY MEDICINE • November 2000, Volume 7, Number 11
1275
apologetic call to the neurologist. This resulted in 4. The findings are inconsistent, effort is hesitant,
redirecting the care process to the right conclusion. resistance is yielding (‘‘give-away’’ weakness), or
Not uncommonly, the management of such cases results are unreliable, as with patients 3 and 6.
balances on a razor’s edge of critical reappraisal vs (This is a major trap! There are many physiologic
a variety of pressures to act without further reflec- and pathologic reasons for variable, inconsistent
tion or consultation. One of these pressures is a findings, including the examiner’s inconsistent in-
reluctance to seek help from consultants who put. Apparent improvement with reassurance does
might sometimes communicate ill-disguised dis- not prove psychogenicity, as neurologic outputs are
dain for the gullibility of the EP or primary phy- very dependent on the input mode.)
sician. With patient 6, the patient’s large intrace- 5. The patient’s ‘‘inappropriate’’ affect or behavior
rebral hematoma probably narrowly missed a (histrionic, flat, unconcerned) reveals the diagno-
critical threshold for a very bad outcome.
sis. (In fact, seemingly inappropriate affect might
come from overt cerebral pathology, as with patient
The Role of Imaging. The major barriers to cor- 6. In addition, there is enormous individual vari-
rect diagnosis in cases of suspected conversion dis- ation and lack of specificity).11
order have always been cognitive and attitudinal, 6. ‘‘Psychologic explanations can validate a diag-
as much as technical. Nonetheless, advances in nosis of conversion.’’ (As with patient 6, psychologic
noninvasive imaging have shifted the burden of di- explanations can be very seductive to the clinician.
agnostic management. The prominence of spinal Neurologic deficits can masquerade as psychologic
cases emphasizes how difficult the assessment of problems. In this case there was inattention to the
reported problems with stance, gait, falls, and back left side, clinically associated with a right brain
pain can be, and how much MRI helps when it is lesion.)
available.
As with patient 6, emergency physicians and
For EDs without access to MRI, the threat of psychiatric physicians and nurses, triage nurses,
spinal cord compression poses the difficult di- and even emergency medical technicians may be
lemma of how best to manage an ambiguous case complicit in skewed interpretation of the present-
when the options are extensive scanning by CT, if ing history and behavior. Biased perceptions may
available, myelography with its discomforts and then lead to selective observation, examination,
risks, and the hazards of simple clinical observa- and test ordering, all contributing to a vicious
tion. The widespread use of CT scanning, which circle of reinforcing errors. The corrective attitude
suffices to rule out most cerebral mass lesions and is one of critical thinking and tenacious open-
subarachnoid hemorrhage, has undoubtedly mindedness by every participant at every step.
served to reduce the misdiagnosis of symptoms and Even patients themselves and family members and
signs referable to the brain. The absence of easy friends sometimes unwittingly join this interplay
access to MRI, in patients with symptoms poten- of unwarranted assumptions, attributions, and im-
tially referable to cord compression, places a bur- pressions. Observations that bias caregivers are of-
den on EPs. With further improvements in access ten repeated from one observer to the next, result-
to sophisticated imaging for emergency patients, ing in a chain of presumption in favor of a
EPs will increasingly be able to collaborate with psychogenic cause and against a neurologic expla-
consultants to rule in or out, on the spot, the most nation.
critical neurologic syndromes in cases of suspected
conversion and other so-called ‘‘functional’’ disor- Patient-based Factors in Misdiagnosis. Based
ders.
on these six illustrative cases and our earlier ex-
periences with such problems, we postulate a num-
Cognitive and Attitudinal Factors of the Phy- ber of ‘‘patient factors’’ that may predispose phy-
sician and Other Caregivers. Our cases illus- sicians to medical errors, whether these are
trate many of the pitfalls of diagnosing conversion significant delays or outright misdiagnosis. The
disorders and other psychogenic explanations of patient is not to be faulted for these characteris-
neurologic symptoms:
tics, but they are real risk factors and may interact
1. A symptom or other aspect of the history or with biases and understandable reactions to past
exam is ‘‘too bizarre—it can’t be real.’’ (But it may experiences on the part of the clinician. These con-
be! Some neurologic events are even stranger than ditions need to be recognized as ‘‘red flags,’’ just as
psychiatric ones.)
with medical background disorders, such as HIV
2. ‘‘I’ve never seen anything like it’’—it’s ‘‘beyond positivity, or with altered physiologic states, such
all of my experience.’’ (Your experience may be lim- as pregnancy. It is critical that physicians remain
ited).
3. The findings are ‘‘nonanatomic.’’ (But is your diagnosis of neurologic or other medical illness in
anatomic knowledge foolproof?) a patient who has these characteristics (even if
aware that they may be biased against making a

1276
CONVERSION DISORDER
Glick et al. • SUSPECTED CONVERSION DISORDERS
that patient is equally or more likely to harbor an incidence of mistakes is the rule, not the excep-
tion.13,14 Fine doctors in fine institutions make er-
organic problem):
rors, and these are unimaginable only in retro-
spect. The following general guidelines for
approaching suspected conversion disorder (and
other unspecified ‘‘psychogenic’’ conditions) may be
helpful:
1. history of past or concurrent psychiatric illness
or encounters
2. history of substance abuse or drug-seeking be-
havior
3. history of somatization
1. Recognize the great difficulty of accurately di-
agnosing a psychologic cause for potentially seri-
ous neurologic symptoms. The EP who considers a
diagnosis of conversion should have a very low
threshold for seeking sophisticated imaging or
neurologic input if there is any doubt.
2. Even when the patient is suggestible and elab-
orates symptoms, look for an ‘‘organic nidus,’’ or
minor, true neurologic deficit upon which the pa-
tient elaborates.
4. current or temporally related emotional crisis
5. high level of stress
6. cultural or language differences
7. abnormal or unexpected affect
8. histrionic behavior
9. sociopathic or hostile behavior
10. known or potential medical conditions, such as
diabetes, that might mask or confound detection or
interpretation of symptoms or signs
3. As shown by the potential for worsening of out-
come in two of our cases, and the close brush with
dangerous misdiagnosis in three others, the stakes
can be very high. Virtual certainty in ruling out
serious and treatable neurologic disease should be
the goal. The downside of a workup is minimal,
especially when there is access to noninvasive im-
aging.
4. Clinicians need to be reflective and self-aware
of their level of understanding, degree of experi-
ence with such cases, and ability to interpret con-
founding factors, such as severe pain (as with our
patient 3). They must also remain self-aware re-
garding their own biases regarding off-putting pa-
tients.
5. Contrary to some texts, it is not easy to inter-
pret the array of clues and ‘‘tricks’’ that are often
touted to unmask pseudoneurologic signs. Miscon-
struing the results of such maneuvers (‘‘the patient
showed inconsistent effort’’) may compound the
risk of diagnostic error.
6. Respecting the difficulty of such cases, clini-
cians need to avoid premature closure once they
find a plausible explanation. Careful attention to
process includes a thorough history and exam that
is independent of prior assumptions; articulation
of a full differential diagnosis and worst-case sce-
narios; critical appraisal of any postulated diag-
nosis of conversion or other psychogenic attribu-
tion; and a readiness to seek assistance when there
is any doubt. We have adopted the tactic of warn-
ing ourselves, whenever we are tempted to make
the diagnosis of conversion disorder, ‘‘I am taking
the chance of making a terrible error. Do I really
have compelling evidence for this diagnosis?’’ We
then make doubly sure that a mistake is not in the
offing.
Magnitude of the Problem. The occurrence of
one serious, acute case per year of misdiagnosis of
conversion or ‘‘near-miss’’ in this small hospital
during a six-year period should alert EPs, primary
physicians, and specialists alike to a vulnerable
spot in the provision of emergency care. Fishbain
and Goldberg4 found three cases of suspected con-
version disorder, or an incidence of 0.07% of all
cases of new symptoms in their psychiatric ED,
lower than the 0.23% and 1% incidences of conver-
sion in other studies they cite. However, all three
of their cases of suspected conversion represented
misdiagnosis of serious neurologic disorders. Nei-
ther their data nor ours are sufficient to establish
generalizable incidences of correctly or incorrectly
suspected conversion disorder in EDs. However,
assuming that our care might be typical, our ex-
perience suggests that the vulnerability to mis-
takes may be widespread.
To frame the possible magnitude of the prob-
lem, we can extrapolate from the number of actual
instances in our hospital and that of Fishbain and
Goldberg to the more than 6,000 hospitals in the
United States.12 If we were to apply even half of
our average occurrence of one problem case per
year or their single-year total of three cases (both
from EDs of modest size), we might predict be-
tween 3,000 and 18,000 cases per year. Even if the
results of this extrapolation were excessive (espe-
cially since the cases of Fishbain and Goldberg
were from the pre-MRI era), it is likely that the
annual number of such occurrences in the United
States has been, and may still be, in four figures.
With attention to this problem, the true incidence
can be reduced dramatically before prospective
studies definitively quantitate it.
Avoidance of Error. The misdiagnoses illus-
trated here might seem like inexplicable outliers
of a reasonable standard of care. However, we
C
ONCLUSIONS
Misdiagnosis of serious neurologic conditions as
know from the ‘‘errors’’ literature that a significant conversion disorder (or other psychogenic prob-

ACADEMIC EMERGENCY MEDICINE • November 2000, Volume 7, Number 11
1277
3. Dula DJ, DeNaples L. Emergency department presentation
of patients with conversion disorder. Acad Emerg Med. 1995;
2:120–3.
4. Fishbain DA, Goldberg M. The misdiagnosis of conversion
disorder in a psychiatric emergency service. Gen Hosp Psychi-
atry. 1991; 13:177–81.
5. Jones JB, Barklage NE. Conversion disorder: camouflage
for brain lesions in two cases. Arch Intern Med. 1990; 150:
1343–45.
6. Slater E. Diagnosis of ‘‘hysteria.’’ Br Med J. 1965; 1:1395–
9.
7. Crimlisk HL, Bhatia K, Cope H, David A, Marsden CD, Ron
MA. Slater revisited: 6 year follow up study of patients with
medically unexplained motor symptoms. Br Med J. 1998; 316:
582–6.
8. Marsden CD. Hysteria—a neurologist’s view. Psychol Med.
1986; 16:277–88.
9. Mace CJ, Trimble MR. Ten-year prognosis of conversion dis-
order. Br J Psychiatry. 1996; 169:282–8.
10. Tissenbaum MJ, Harter HM, Friedman AP. Organic neu-
rologic syndromes diagnosed as functional disorders. JAMA.
1951; 147:1519–21.
lems) represents a significant challenge in the field
of medical errors, although the frequency in an in-
dividual ED may be low. Legal vulnerability aside,
the current focus in the United States and else-
where on reducing errors and iatrogenic injury re-
inforces the need for the education of physicians
and nurses in the difficult area of suspected psy-
chogenicity of acute neurologic symptoms. Critical
thinking and an understanding of both physician-
based and patient-based risk factors can contribute
to correct diagnosis, optimal care, and patient
safety. In addition to educational interventions, a
quality improvement approach would require reas-
sessment of the diagnostic ‘‘safety net’’ to deter-
mine whether procedures for supervision, review,
testing, and consultation are sufficient to prevent
mistakes.
11. Lazare A. Current concepts in psychiatry: conversion
symptoms. N Engl J Med. 1981; 305:745–8.
12. American Hospital Association. Hospital statistics (1997,
table 110). Chicago: AHA, 1999.
References
13. Leape L. Error in medicine. JAMA. 1994; 272:1851–7.
14. Kohn LT, Corrigan JM, Donaldson MS (eds). To Err Is Hu-
man: Building a Safer Health System. Report of the Institute
of Medicine. Washington, DC: National Academy Press, 1999.
1. Diagnostic and Statistical Manual of Mental Disorders, 4th
Ed. Washington, DC: American Psychiatric Association, 1994.
2. Adams RD, Victor M, Ropper AH, Principles of Neurology,
6th ed. New York: McGraw-Hill, 1997.
ⅷ