The ‘Rosenhan experiment’ is a well known experiment examining the validity of psychiatric diagnosis. It was published in 1975 by David Rosenhan in a paper entitled ‘On being sane in insane places’
The study consisted of two parts. The first involved ‘pseudopatients’ – people who had never had symptoms of serious mental disorder – who, as part of the study, briefly reported auditory hallucinations in order to gain admission to psychiatric hospitals across the United States.
After admission, the pseudopatients no longer reported hallucinations and behaved as they ‘normally’ would. Despite this many were confined as inpatients for substantial periods of time and all were discharged with the diagnosis of a psychiatric disorder.
For the second part of the experiment staff at a teaching hospital, whose staff had learned of Rosenhan’s above results, were informed that one or more pseudopatients would attempt to be admitted to their hospital over an ensuing three month period. Many patients were subsequently identified as likely pseudopatients but in fact no pseudopatient had been sent.
‘On being sane…’ also examines, though the experience of the pseudopatients, the patient experience of psychiatric inpatient wards. This part of the paper is discussed often only in passing.
Rosenhan’s conclusion was stark: A psychiatric diagnosis is more a function of the situation in which the observer finds a patient and reveals little about a patient themselves.
“It is clear that we cannot distinguish the sane from the insane in psychiatric hospitals“
Despite being over thirty years old the Rosenhan experiment remains well known and is often cited. Accounts of the experiment are widespread on the internet, but critiques are rarer and many people accept the study’s conclusions at face value.
This was an audacious experiment and the subsequent paper had an extremely good title, but was Rosenhan justified in his conclusion? Anthony Clare, amongst others, wrote that Rosenhan was ‘theorising in the absence of sufficient data’. But if Rosenhan was correct then his experiment remains extremely important; as if diagnoses are in ‘the mind of the observer’ and do not reflect a quality inherent a patient, they are of little use.
If you wish to read the original paper it can be found here.
Spitzer’s 1975 critique is: Spitzer, Robert L More on pseudoscience in science and the case for psychiatric diagnosis Arch Gen Psychiatry Vol 33 April 1976
Davis’s critique here. Davis, Douglas A. On being detectably sane in insane places: Base rates and psychodiagnosis. Journal of Abnormal Psychology, Vol 85(4), Aug 1976, 416-422
Clare’s ‘Psychiatry in dissent’ is available in preview here.
Circumstances of diagnosis and the detecting of sanity.
In the experiment eight pseudopatients presented at psychiatric hospitals complaining of hearing a voice. Asked what the voices said, they replied that the voices were often unclear, but as far as they could tell, said “empty,” “hollow,” and “thud.” Beyond alleging this symptom, and falsifying their names and vocations, no other falsehoods were told. Upon admission to the ward the pseudopatients are reported to have ceased to claim symptoms and behaved as they ‘normally’ would.
Length of hospitalization was an average of 19 days during which time no pseudopatients were identified as fraudulent. All pseudopatients except one (diagnosed with bipolar disorder) were discharged with a diagnosis of ‘schizophrenia in remission’. In light of this Rosenhan regards there to have been ‘uniform failure to recognise sanity’. Rosenhan refused to identify the hospitals used on the grounds of his concern for confidentiality. This is laudable in some respects, but it makes it impossible for anyone at the hospitals in question to corroborate or refute this account of how the pseudopatients acted or were perceived.
It is a difficulty that Rosenhan seeks to answer whether patients can be identified as ‘sane’ or ‘insane’, whilst psychiatrists, whose practice he wishes to scrutinize, do not make such distinctions in their practice but instead aim to identify and treat what they view as psychiatric disorders. This objection aside, and working within this terminology, in his 1975 critique Spitzer identifies three possible meanings for ‘detecting of sanity’.
- Recognition, when he is first seen, that the pseudopatient is feigning insanity as he attempts to gain admission to the hospital. This would be detecting sanity in a sane person simulating insanity.
- Recognition, after having observed him acting normally during his hospitalization, that the pseudopatient was initially feigning insanity. This would be detecting that the currently sane person never was insane.
- Recognition, during hospitalization, that the pseudopatient, though initially appearing to be ‘insane’ was no longer showing signs of psychiatric disturbance.
Only the first two involve identifying a pseudopatient as a fraud and Spitzer feels that it is these that Rosenhan implies are all that are relevant to the central research question. He disagrees, writing that when the third definition of detecting of sanity is considered Rosenhan’s conclusions cannot be sustained.
This assertion hinges on Rosenhan’s report that all the pseudopatients were diagnosed as being ‘in remission’, that is recognised as being, currently, without signs of mental disorder or ‘sane’. By this view the data as reported by Rosenhan contradicts Rosenhan’s own conclusion. Spitzer also writes that ‘schizophrenia in remission’ was a diagnosis rarely used by psychiatrists at the time of the experiment, and as such this indicates that the diagnoses given were a function of the patients’ behaviours and not simply of the environment in which they were made.
Should a psychiatrist be able to able to detect that a patient is a fraud? That is, should a psychiatrist be able to detect that, after observing a patient acting normally, that they were initially feigning insanity? Rosenhan reports that this possibility was considered by the pseudopatients’ fellow patients but by no clinical staff:
“It was quite common for the patients to “detect” the pseudopatient’s sanity. During the first three hospitalizations, when accurate counts were kept, 35 of a total of 118 patients on the admissions ward voiced their suspicions, some vigorously. “You’re not crazy. You’re a journalist, or a professor (referring to the continual note-taking). You’re checking up on the hospital.” …. The fact that the patients often recognized normality when staff did not raises important questions.”
Rosenhan reports that the psychiatrists did not spend much time with the pseudopatients. Other patients of course had ample time to formulate their own theories. Whilst the medical staff’s lack of engagement with the pseudopatients is regrettable, it does point towards poor clinical skills rather than an indictment of psychiatric classification. Clare again:
“Rosenhan and those many critics of psychiatry who have greeted his paper with enthusiasm seem in fact to be saying that, since the doctors did not appear to have the faintest idea as to what constitutes the operational concept of ‘schizophrenia’ and yet applied it with haste to people showing virtually no signs or symptoms whatsoever, the whole diagnostic approach should be scrapped!”
Rosenhan later wrote that he considered the patients apparent insight over that of the psychiatrists as due to the ‘experimenter effect’ or ‘expectation bias’. The professionals expected to see a patient with a mental illness, so they looked for reasons to believe it, and eventually they convinced themselves that the pseudopatients were actually suffering from schizophrenia.
People do sometimes simulate mental illness for their own ends and this is a genuine diagnostic problem. It is a situation not unique to psychiatry and how easily a disorder psychiatric or otherwise can be feigned tells us little about the worth of the psychiatric classification system. Kety has something to say on this.
“If I were to drink a quart of blood and, concealing what I had done, come to the emergency room of any hospital vomiting blood, the behavior of the staff would be quite predictable. If they labeled and treated me as having a bleeding peptic ulcer, I doubt that I could argue convincingly that medical science does not know how to diagnose that condition”
Clare makes a similar point using the example that the signs and symptoms of diabetes exist independently of whether they are correctly elicited or not.
Rosenhan does consider in his paper that that a mental illness is a life sentence:
“A broken leg is something one recovers from, but mental illness allegedly endures forever”
If a disorder was known to be always chronic and unremitting, it would illogical not to question the original diagnosis if the patient was later found to be asymptomatic and it is at this that Rosenhan is presumably driving. If the pseudopatients ‘recovered’ from an incurable illness whilst under the gaze of their psychiatrists and this did not alter the diagnosis then this would be an example, just as Rosenhan says, of the hospital environment influencing diagnostic decision making. But in stating that mental illness is something that endures forever Rosenhan is taking a very selective view of the wide range of presentations all of which come under the umbrella of ‘schizophrenia’. Schizophrenia has acute subtypes from which full recovery is possible and can also relapse and remit.
As for the non-existent impostor experiment it is surprising that it was agreed to by the teaching hospital in question. The poor reliability of psychiatric diagnoses means that the design of the experiment could only produce an outcome where actual patients were incorrectly identified as pseudopatients.
Conditions on the ward
All of the pseudopatients took extensive notes. Rosenhan makes much of this writing being “seen as an aspect of their pathological behaviour” on the grounds of the nursing entry that read “engages in writing behaviour”. Spitzer argues that was routine for nursing staff to frequently and intentionally comment on non-pathological activities in which a patient engages to enable other staff members to have knowledge of how the patient spends his time. As such, a comment about note taking is therefore inevitable and unremarkable. He is struck by what he sees as Rosenhan’s actual failure to provide data demonstrating where normal hospital experiences were categorized as pathological.
Rosenhan’s account of the conditions on the psychiatric wards is, for me, the most interesting part of the paper. The staff and patients were strictly segregated, the professional staff and especially the psychiatrists being rarely seen and having little patient contact.
“Staff and patients are strictly segregated. Staff have their own living space, including their dining facilities, bathrooms, and assembly places. The glassed quarters that contain the professional staff, which the pseudopatients came to call “the cage,” sit out on every dayroom. The staff emerge primarily for care-taking purposes – to give medication, to conduct therapy or group meeting, to instruct or reprimand a patient. Otherwise, staff keep to themselves, almost as if the disorder that afflicts their charges is somehow catching.”
This description bears resemblance to modern UK psychiatric wards. Psychiatrists spend little time with the patients in their care and nurses are occupied for a great deal of their time sitting in a locked room doing paperwork. The healthcare staff members with the most patient contact are the least qualified. This is far from ideal, and a target for improvement, but it should be noted that within healthcare this distance between staff and patients is not restricted to psychiatric wards and the pressures on staff due to the number of patients in their care means that a more desirable personal service is something with which the NHS struggles in all its domains.
Rosenhan’s description of the depersonalising effect of a long stay on the wards is also powerful. Despite their commitment to the experiment in which they are taking part, their wish to resist the powerlessness they experience leads several of them to jeopardise the study.
“The patient is deprived of many of his legal rights by dint of his psychiatric commitment. He is shorn of credibility by virtue of his psychiatric label. His freedom of movement is restricted. He cannot initiate contact with the staff, but may only respond to such overtures as they make. Personal privacy is minimal. Patient quarters and possessions can be entered and examined by any staff member, for whatever reason. His personal history and anguish is available to any staff member (often including the “grey lady” and “candy striper” volunteer) who chooses to read his folder, regardless of their therapeutic relationship to him. His personal hygiene and waste evacuation are often monitored. The water closets have no doors.”
Attendants were reported to deliver verbal and occasional physical abuse to patients, something that can in no way be justified. Rosenhan’s report of this leads to an interesting inconsistency. Despite initial descriptions of abusive staff behaviour, in his conclusion Rosenhan describes the staff as overwhelmingly ‘committed and … uncommonly intelligent’. Spitzer considers that this is because of Rosenhan does not wish to direct attention toward shortcomings of the staff, rather wishing to concentrate on diagnostic labels.
Validity of diagnosis.
There are two issues here. Where the psychiatrists who met his pseudopatients wrong to make a diagnosis of schizophrenia within the DSM II diagnostic framework? And are psychiatric diagnoses of use or should they be replaced by an alternative?
The ease with which the pseudopatients gained admission on the basis of what are reported to be mild symptoms was remarked upon by Anthony Clare in Psychiatry in Dissent.
“It is a matter of some interest that a solitary complaint of a hallucinatory voice in the absence of any other unusual experience or personal discomfort should actually persuade certain American hospitals to open their doors. Such is the current demand for a psychiatric bed within the National Health Service and the prevailing emphasis on treating patients outside hospitals and in the community that the average admitting doctor in Britain is likely to find himself under strict instructions to avoid admitting any patient who can see, speak, and do all of these things without bothering himself or others to an significant extent. On suspects that, in Britain, Professor Rosenhan might well be advised to go home like a good man, get a decent night’s rest and come back again in the morning.”
And many people have been critical of the way the pseudopatients were diagnosed with schizophrenia on the basis of hallucinations – a single symptom and not even essential for the diagnosis. Anthony Clare again:
“…the doctors did not appear to have the faintest idea as to what constitutes the operational concept of ‘schizophrenia’ and yet applied it with haste to people showing virtually no signs or symptoms whatsoever…”
Spitzer remarks that the doctors should have been wary of making a diagnosis of schizophrenia in a previously unknown patient presenting without any history of insidious onset. However he is more lenient toward the pseudopatients’ psychiatrists, writing that, given the information available, schizophrenia was the most reasonable diagnosis. Davis and Weiner agree, respectively arguing from statistical and attribution theory standpoints that schizophrenia was the most likely diagnosis. Rosenhan himself presents no differential diagnosis.
Hunter takes exception to Rosenhan’s assertion that the pseudopatients acted ‘normally’ in the hospital:
“The pseudopatients did not behave normally in the hospital. Had their behaviour been normal, they would have talked to the nurses’ station and said “Look, I am a normal person who tried to see if I could get into the hospital by behaving in a crazy way or saying crazy things. It worked and I was admitted to the hospital but now I would like to be discharged from the hospital”.
We in fact learn very little about the diagnostic process beyond the initial presentations of the pseudopatients. It should be noted that the pseudopatients would likely not have been, unlike Rosenhan’s assertion, admitted on the basis of their hallucinations solely. Their presentation to hospital and request for admission may also have carried diagnostic weight as it suggested much greater distress. However, whatever the fine detail, throughout their stay, the pseudopatients do not appear to have been assessed in detail.
The poor diagnostic skills and apparent lack of curiosity of the psychiatrists that the pseudopatients met is not an indictment of the classification per se, rather its application. The Rosenhan paper offers no insight as to why psychiatric classification had developed into the shape that he found it in 1973.
The purpose of a disease classification system is that it allows healthcare professionals to:
- Communicate with each other about the subject of their concern
- Avoid unacceptable variations in diagnostic practice
- Predict their outcome disorders and suggest a treatment.
- Conduct research
Amongst others Richard Bentall has made a career out of pointing out that psychiatric diagnosis is neither particularly valid nor reliable. However in Spitzer’s view the historical precedent is that classification in medicine has always been preceded by clinicians using imperfect systems. These have then improved on the basis of clinical and research experience. The clinician is forced to do the best he/she can until something better comes along.
In contrast to psychiatric disorders, the diagnosing of physical medical conditions is often portrayed as being solid and dependable. This does not bear close inspection, as many medical conditions are at least as vaguely described as psychiatric disorders. Although it is true that by-and-large a physical illnesses diagnosis rests on biological ‘facts’, the accompanying negative impact on person is the most important factor and this is highly subjective. For instance we all have bacteria in the back of our throats, but do not consider ourselves to have an infection.
In light of his experiment, rather than the syndromal classification system, Rosenhan would favour a classification system based on behaviours:
“It seems more useful … to limit our discussions to behaviours, the stimuli that provoke them, and their correlates”
Yet despite this early on in the paper he writes that “Anxiety and depression exist”, suggesting he favours an ad-hoc classification system at least.
“It is clear that we cannot distinguish the sane from the insane in psychiatric hospitals. The hospital itself imposes a special environment in which the meaning of behavior can easily be misunderstood. The consequences to patients hospitalized in such an environment – the powerlessness, depersonalization, segregation, mortification, and self-labeling – seem undoubtedly counter-therapeutic.”
There are sufficient objections to the design of Rosenhan’s experiment – not least that his study consisted of only eight subjects- to doubt whether he is justified in writing his initial sentence. Rosenhan’s observational study of conditions on psychiatric wards – to which the rest of the above paragraph alludes – still has relevance today and remains a note of caution for anyone who works in mental health.
Mind changers 27 July 2009 Radio 4: The pseudopatient study