The most important papers published tend to start with a deceptively simple question. When David Rosenhan sat down to write "On Being Sane in Insane Places" in 1972, his simple question was: "If sanity and insanity exist, how shall we know them?" [1]. In order to investigate, he planted a number of perfectly sane patients in mental hospitals to see if staff would notice. The shocking results of his experiment sparked a fierce debate that continues to this day.

"If sanity and insanity exist, how shall we know them? The question is neither capricious nor itself insane. However much we may be personally convinced that we can tell the normal from the abnormal, the evidence is simply not compelling." Rosenhan quoted the example of conflicting evidence from psychiatrists in murder trials, but since this is a blog and I have an unhealthy appetite for T.V., I'll refer instead to the common plot device of the sane man who infiltrates a mental hospital, only to find himself struggling to convince the staff that he's actually sane when he tries to leave.

Presumably bored with television, Rosenhan decided to do something really cool - he decided to play out the scenario in real life. His experiment involved "getting normal people (that is, people who do not have, and have never suffered, symptoms of serious psychiatric disorders) admitted to psychiatric hospitals and then determining whether they were discovered to be sane and, if so, how. If the sanity of such pseudopatients were always detected, there would be prima facie evidence that a sane individual can be distinguished from the insane context in which he is found."

But what if the sane people weren't detected?

- - - - - - - - - - - - -

Rosenhan recruited eight, sane volunteers - five men and three women - to secretly gain admission to twelve hospitals using false names and occupations. The institutions selected represented a variety of settings across five states on the East and West coasts of the U.S.A.; some old and shabby, some shiny and new; some with many staff, some with few; and with a variety of different sources of funding - private, public, university. No hospital staff were in on the ruse.

The admissions worked as follows. The pseudopatient would call the hospital for an appointment, and then present to the admissions office complaining that they had been hearing voices. "Asked what the voices said, he replied that they were often unclear, but as far as he could tell they said 'empty,' 'hollow,' and 'thud.' The voices were unfamiliar and were of the same sex as the pseudopatient. Such symptoms are alleged to arise from painful concerns about the perceived meaninglessness of one’s life. It is as if the hallucinating person were saying, 'My life is empty and hollow.' The choice of these symptoms was also determined by the absence of a single report of existential psychoses in the literature."

Aside from faking their name, occupation and initial symptoms, no other alterations of the pseudopatients history or circumstances would be made. The patients would talk about their own history and personal events as they had happened, and describe real relationships with friends, colleagues and family. Real upsets and real satisfactions were described. A point to bear in mind is that, as Rosenhal highlights, none of their histories or current behaviors were indicative of any mental health issues. As soon as the pseudopatients were admitted to the psychiatric ward, they behaved completely normally. None of them really expected to be admitted so easily, and all expected to be quickly exposed and embarrassed.

They needn't have worried.

- - - - - - - - - - - - -

So what happened with the admissions? Well the patients all passed the entry exam, and all bar one were admitted with schizophrenia . When eventually discharged, each was given a diagnosis of "schizophrenia in remission". As Rosenhan explains, "if the pseudopatient was to be discharged, he must naturally be 'in remission'; but he was not sane, nor, in the institution’s view, had he ever been sane." Not only did the institutions fail to spot that the patients were fake, but Rosenhan suggested that they even failed to diagnose sanity.

This apparent failure had nothing to do with the quality of the hospitals, as several were considered excellent, and the length of hospitalization ranged from 7 to an incredible 52 days (averaging 19), so the staff had plenty of time to make an assessment. During all that time, the pseudopatients behaved perfectly sanely.

The fake patients weren't picked out by staff, but they didn't remain undetected - the other patients cottoned on very quickly. Rosenhan notes that "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"

Rosenhan, believed that the psychiatrists were falling into the trap of what statisticians call the "type 2 error". Just as owning a dog sometimes makes you more away of other dogs around you, doctors are often more inclined to call a healthy person sick, and psychiatrists are more inclined to diagnose illness.

Could this bias be reversed? In order to find out, Rosenhan recruited a research and teaching hospital who were briefed on his findings, but "doubted that such an error could occur in their hospital". The staff were informed that at some time during the following three months, one of more fake patients would attempt to be admitted. In that period, 193 patients were admitted, and off those: 41 were suspected by at least one member of staff; 23 were suspected by at least one psychiatrist, and 19 were suspected by one psychiatrist and one other staff member.

In fact, there were no pseudopatients at all.

Were the 19 highly-suspect candidates sane, or were the psychiatrists erring in their caution to avoid making type 2 errors? There's no way we can tell, but as Rosenhan pointed out, "any diagnostic process that lends itself too readily to massive errors of this sort cannot be a very reliable one."

- - - - - - - - - - - - -

During their stay, one thing became apparent to Rosenhan - once the patients had been labeled schizophrenic, there appeared to be little they could do to overcome that level. In fact, Rosenhan and his volunteers suggested that normal behaviours were either ignored, or severely misinterpreted.

The pseudopatients all took notes publicly, which surprisingly raised no suspicion. In fact, in a startling piece of irony, this was recorded "patient engaged in writing behavior" by the nurses in one case. Regular writing was seen as a pathological behaviour.

In fact many behaviours were misinterpreted that way. A patient pacing the corridors was described as "nervous", before pointing out that he was simply bored by the sterile environment. Often a patient would become upset at some witting or unwitting mistreatment by an attendant, but nurses would rarely ask about the circumstances behind the incident, instead assuming that the upset came simply from from the illness. "One psychiatrist pointed to a group of patients who were sitting outside the cafeteria entrance half an hour before lunchtime. To a group of young residents he indicated that such behavior was characteristic of the oral-acquisitive nature of the syndrome. It seemed not to occur to him that there were very few things to anticipate in a psychiatric hospital besides eating."

This apparent interpretative bias even seemed to extend to behaviour that occurred before the patient was admitted. You would expect that personal history would affect the initial diagnosis, but in fact, Rosenhan's work suggested that the opposite occurred - the psychiatrists reinterpreted the pseudopatients' histories according to the diagnosis. In one example, the patient said that had had been close to his Mother, but more distant from his father until adulthood, when he became closer to his father his relationship with his mother cooled. He had a close, warm relationship with his wife, with the odd argument. The children had rarely been spanked (this was 1972, folks).

There was nothing abnormal about this history, but the case summary stated that: "[he] manifests a long history of considerable ambivalence in close relationships, which begins in early childhood. Affective stability is absent. His attempts to control emotionality with his wife and children are punctuated by angry outbursts".

- - - - - - - - - - - - -

"There is by now a host of evidence that attitudes toward the mentally ill are characterized by fear, hostility, aloofness, suspicion, and dread. The mentally ill are society’s lepers." Rosenhan made this assertion without backing it up particularly well, but what was undeniable was that such attitudes seemed to be common among professionals that the pseudopatients interacted with. Rosenhan described the relationship of staff with patients as "exquisitely ambivalent", and the team's examples of staff-patient interaction are particularly shocking.

Immediately apparent to the pseudopatients was the segregation that occurred. Staff and patients were kept strictly apart, with separate staff facilities in glassed quarters overlooking the dayrooms, which the pseudopatients came to call "the cage". The staff came out when necessary, but otherwise kept to themselves. The segregation was so severe in four hospitals that "when an attempt was made to measure the degree to which staff and patients mingle, it was necessary to use 'time out of the staff cage' as the operational measure." On average, attendants spent only 11.3% outside of their cages, rarely talking or interacting with patients. It proved impossible to gather data on the amount of "mingling time" for nurses, because the amount off time they spent out of the case was too brief. Psychiatrists were rarely seen at all.

As part of the experiment, pseudo-patients attempted to interact with the staff, by asking a simple, appropriate and courteous question. "By far, their most common response consisted of either a brief response to the question, offered while they were 'on the move' and with head averted, or no response at all. The encounter frequently took the following bizarre form: (pseudopatient) 'Pardon me, Dr. X. Could you tell me when I am eligible for grounds privileges?' (physician) 'Good morning, Dave. How are you today? (Moves off without waiting for a response.)'" The mean average time of daily contact that psychiatrists, psychologists, residents and physicians combined had with the patients was just 6.8 minutes.

Patients were utterly powerless and virtually dehumanized. The volunteers had the sense that they were invisible, or at least unworthy of account. On the ward, attendants delivered verbal and occasionally serious physical abuse to patients in the presence of others (the pseudopatients) who were writing it all down. Abusive behavior, on the other hand, terminated quite abruptly when other staff members were known to be coming. Staff are credible witnesses. Patients are not. This quickly become very troubling for the pseudopatients - who remember were perfectly sane individuals. They found themselves caught up fighting the process of dehumanization, latching on to any way of becoming a "person" again, much as Jack Nicholson would go on to do in his Oscar-winning portrayal
of R.P. McMurphy a few years later.

How many sane people were locked up in mental institutions in the 1970s and before? And how many apparently insane people would have behaved normally if removed from this damaging environment? Rosenhal's sensational conclusion was that "we cannot distinguish the sane from the insane in psychiatric hospitals."

- - - - - - - - - -

Rosenhal's research didn't just expose an error in scientific methodology, it uncovered a national scandal. The publication of his research in Science in 1973 create a storm of controversy that raged for several years, and there were some very serious criticisms of the work.

Undoubtedly, Rosenhal's analysis was stronger in the areas where he had gathered hard data - particularly his criticisms of staff conduct. Where his arguments were weaker were in his suggestions that a patient couldn't overthrow the initial diagnosis - that "labeling" a patient too readily could be seriously detrimental. There was plenty of anecdote and innuendo suggesting this conclusion - particularly from Rosenhan's presentation of the facts - but no really hard data.

A major bone of contention surrounded the deception that occurred during initial diagnosis. As psychiatrist Robert Spitzer observed: "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 labelled and treated me as having a peptic ulcer, I doubt I could argue convincingly that medical science does not know how to diagnose that condition." In other words, given that all doctors rely to some extent on information from the patient to reach conclusions, it was unfair to give the professionals false information and then criticize their diagnosis. Additionally, it was pointed out that every patient was correctly diagnosed as exhibiting no pathological symptoms when discharged.

All papers are flawed, and criticisms like these were certainly valid, but they didn't really get to the guts of the problem. The initial diagnosis was less important to Rosenhal's work than the treatment and evaluation of the patient once diagnosed, and the concerns raised skirted around the fact that the diagnosis could be so easily skewed in experiments.

Ultimately though, the value of the paper was the sheer ferocity with which it challenged the field of psychiatry - and in that respect it should perhaps be judged as a great piece of journalism rather than a fine example of scientific method. The article, and the questions and fierce debate that it created, would go on to have a dramatic impact on psychiatry in the years to come, inspiring new diagnostic texts such as the "Diagnostic and Statistical Manual of Mental Disorders", and highlighting the need to improve conditions in hospitals. While the scientific merits of this article are perhaps debatable, the impact was enormous, and has benefited medicine ever since.

[1] Rosenhan, D.L. (1973). Rosenhan, D.L. (1973). On being sane in insane places. Science, 179, 70, pp. 250-8.. Science, 179(70), 250-258.