There are two aspects of diagnosis that are important in psychology.  First, is the validity of diagnosis.  If the diagnosis is valid, it should lead to successful treatment. Then there is the reliability of a diagnosis. A diagnosis is reliable when more than one psychiatrist diagnoses a patient the same way using the same diagnostic criteria. There is much research that shows that the diagnosis of psychological problems is not highly reliable, but the research is problematic. If a diagnosis is reliable, this does not mean that it is valid; however, if a diagnosis is not reliable it cannot be valid.

Rosenhan carried out a classic study on the validity of diagnosis.  Rosenhan had a group of students play the role of confederates, going into different psychiatric hospitals and claiming that they heard a voice saying, “One, two, three, thud.”  Most of the students were admitted to hospital.  Even though they immediately stopped saying that they had symptoms, they were kept in hospital.  When eventually released, they were diagnosed with schizophrenia in remission.  The study claimed to show that the doctors did not make a valid diagnosis – and argued that the validity of diagnosis is problematic at best.

However, this study has been challenged. First, the study was based on deception.  One of the difficulties of making a valid diagnosis is that psychiatrists are dependent on self-reported symptoms.  In this case, the “patients” lied about their symptoms.  As patients cannot be tested for auditory hallucinations, it is not possible to verify their symptoms.  This trust between a doctor and a patient is key to a valid diagnosis.  Secondly, disorders are “clusters of symptoms” and many symptoms are representative of several different disorders.  This also influences the validity of diagnosis.  In this case, the study is old and based on the DSM III.  At that time, auditory hallucinations were considered to be a sign of schizophrenia.  Today, no single symptom would be used to make a diagnosis.  A problem, however, is that there are no blood tests that can be given to test for the disorder, although it is possible that in the future genetic testing may make diagnosis both more valid and reliable.

Finally, the behaviour of the hospital staff was described by the pseudo-patients.  They wrote that the nurses took notes about their “pathological behaviour”, but the actual notes did not show this to be true. The study suffers from researcher bias, where those that were being observed assumed the intentions of the nurses on staff.  Based on the symptoms that were presented at the time, under the DSM III, the diagnosis of “schizophrenia in remission” seemed appropriate.  Although the study led to an improvement in the way that hospitals admit mental health patients, it does not tell us much about the validity of diagnosis.

One study that looked at the reliability of diagnosis was carried out by Lipton and Simon. The researchers randomly selected 89 patients with schizophrenia in a New York hospital and then had a team of psychologists diagnose them to see if their diagnosis would match the original diagnosis.  They found that only 16 of the patients received a diagnosis of schizophrenia.  A limitation of this study is that it was done using the DSM III.  Today we have the DSM 5, so it is possible that diagnosis would be more reliable.  More importantly, the patients had already received treatment.  The treatment is then a confounding variable in making a reliable diagnosis.  It could be that the reliable diagnoses were made for patients that had shown the least improvement.

Research that investigates the reliability of diagnosis shows that reliability for the diagnosis of some disorders is stronger than others.  Lobbestael et al used the DSM IV to diagnose 151 participants, consisting of both patients and non-patients.  All participants took part in a clinical interview that was audio taped. A second psychiatrist watched the original interview but did not know the diagnosis that was made by the first psychiatrist. The results showed that the reliability of diagnosis for personality disorders was around 85%, whereas for depression it was 70%.  These results are higher rates of reliability than are seen in older studies which used the DSM III.  This could mean that improvements in the diagnostic guides have increased the reliability of diagnosis.

A strength of this study was that only audio tapes were used, so non-verbal behaviour or the appearance of the patient did not affect the diagnosis process.  However, this can also be seen as a limitation. This is not how a psychiatrist would actually diagnose a patient. This means that the experiment was highly artificial and lacks ecological validity.

Testing the validity of diagnosis is difficult because it depends on the success of treatment.  And it is difficult to rule out the placebo effect in treatment, meaning that we cannot guarantee that the treatment was the sole reason for improvement.  A way to counter this would be to use researcher triangulation in diagnosis. When more psychologists reach the same diagnosis, this increases the potential for a valid diagnosis. In addition, reliability may be increasing as psychologists have a better understanding of disorders and the DSM has been updated to reflect this.

Words: 862