Two Key Studies of Validity & Reliability of Diagnosis: Rosenhan et al. (1973); Nicholls et a. (2000) (DP IB Psychology)

Key Study One: Rosenhan (1973)

Rosenhan: a man with a plan…

Key study one (validity of diagnosis): Rosenhan (1973)

Aim:  

• To investigate the validity of mental illness diagnosis

• To investigate the consequences of the ‘sticky label’ of a mental illness diagnosis

Participants: 

• The study used naive participants from the following:

• The staff and patients from 12 mental hospitals from across the USA

• The hospitals varied in terms of age, location, staff-patient ratios, expertise

Observers: 

• Rosenhan recruited eight confederates who comprised his sample of pseudopatients who infiltrated the mental hospitals and made covert observations of the hospital staff and patients

• The pseudopatients consisted of 3 females and 5 males with Rosenhan himself assuming a pseudopatient role as well

• The pseudopatients were from a range of different backgrounds and none of them had a mental illness

• The pseudopatients were told to use fake names and occupations when they presented themselves for diagnosis

Procedure:

• The confederates recruited by Rosenhan (known as ‘pseudopatients’ as they would be faking their symptoms) were instructed to present themselves at one of the 12 hospitals selected by Rosenhan

• Upon getting an appointment with a doctor they were told to report the following symptoms: I have been hearing a same-sex voice in my head which repeats the words ‘empty’, ‘hollow’ and ‘thud’

• The pseudopatients were told to behave normally during the consultation and not to fake any other symptoms of mental illness

• All but one of the pseudopatients were admitted to hospital with a diagnosis of schizophrenia (one of them was admitted with a diagnosis of bi-polar disorder)

• Once the pseudopatients had been admitted to hospital Rosenhan’s instructions were that they were to never mention their (fake) symptoms again, to behave normally and to persuade the hospital to release them as soon as possible

• Rosenhan also told the pseudopatients to keep notes of what they observed during their time in hospital relating to both staff and patients

• The pseudopatients were told not to take any drugs administered to them by hospital staff but to dispose of them discreetly

• The dependent variable was the number of days spent in hospital before release

• The overarching method of this research is a covert participant observation

Results: 

• The notes made by the pseudopatients while in hospital detailed the everyday interactions between staff and patients

• Interactions between staff and patients was sparse, with staff often ignoring patients, dismissing their requests (e.g. asking when visiting hours were), making little eye contact with the patients

• Normal behaviours were often interpreted by staff as aspects of mental illness e.g. three pseudopatients were told that their writing was evidence of pathological behaviour, labelling this is ‘writing behaviour’ rather than simply ‘writing’

• One one occasion a psychiatrist pointed to a group of patients queuing for lunch and labelled this behaviour as ‘oral-acquisitive syndrome’ rather than simply accepting that they were just queuing up for lunch

• None of the staff suspected that the pseudopatients were fake, however 35 out of 118 patients approached the pseudopatients and voiced their suspicions that the pseudopatients were not actual patients (some of the patients thought that the pseudopatients might be undercover journalists)

• The pseudopatients spent from 7 to 52 days in hospital (mean=19 days)

• All but one of the pseudopatients were released from hospital with a diagnosis of ‘schizophrenia in remission’ 

Conclusion: 

• There are questions to be asked re: the validity of mental illness diagnosis as the doctors should not have diagnosed any of the pseudopatients with schizophrenia or bi-polar disorder as their (fake) symptoms do not align with either of these diagnoses

• Once someone has been diagnosed with a mental illness this becomes a ‘sticky label’ through which all subsequent behaviours are viewed and judged

• Patients hospitalised with a mental illness experience depersonalisation due to the indifferent, sometimes hostile treatment at the hands of hospital staff

Evaluation of Rosenhan (1973)

Strengths

• The use of research in the field via covert observational methods means that the observed participants are unlikely to have succumbed to the observer effect, making the findings high in ecological validity

• This was a controversial, ground-breaking study which provoked important discussion about how people suffering from mental disorders are treated by institutions 

Limitations

• The study does raise some ethical concerns: the staff and patients of the hospitals were deceived; the hospital participants could not give their informed consent or be given the right to withdraw plus their privacy was compromised

• A sample of only 8 pseudopatients is not enough from which to draw strong and meaningful conclusions plus there is the possibility that the pseudopatients might have succumbed to confirmation bias in reporting their observations

One of Rosenhan’s key findings was that mental hospitals rob people of their individuality.

Key Study Two: Nicholls et al. (2000)

Aim: To evaluate the reliability of diagnostic classification systems for eating disorders when applied to children and young adolescents.

Participants: 81 children aged 7-16 who had been selected via random sampling from a population of 226 child patients attending a clinic specialising in eating disorders.

Procedure: 

• Each child was assessed by one of six clinicians 

• The clinicians were asked to use either the DSM-IV, the ICD 10 or the Great Ormond Street Hospital (GOSH) diagnostic manual in to form their diagnosis of each child

• Each clinician gave their diagnosis as to which specific eating disorder the child was suffering from, using one of the three diagnostic manuals cited in the above bullet point

• Two clinicians assessed each child (each clinician having used a different diagnostic manual to the other) without knowing about each other’s diagnosis i.e. they were blind to the pre-existing diagnosis

Results:

• Inter-rater reliability values were calculated for each of the three diagnostic manuals used to come to reach the diagnosis

• The higher the inter-rater value is, the more reliable the diagnosis is

• The results per diagnostic manual were as follows:

• GOSH: 0.879

• DSM-IV: 0.636

• ICD 10: 0.357

• The GOSH definitions included anorexia and bulimia nervosa, food avoidance emotional disorder, selective eating and pervasive refusal to eat amongst their classification of eating disorders

• GOSH criteria bad been specifically developed to classify child and adolescent eating disorders: they were more reliable than the DSM IV and ICD 10 criteria, which showed little consistency, especially the ICD 10, which had the lowest inter-rater reliability of all the classification systems

• The DSM-IV and the ICD 10 focused too much on body shape and weight which are invalid criteria when diagnosing eating disorders in children

Conclusion: The DSM and ICD are not suitable classification systems  for the diagnosis of eating disorders in children; a clinician working diagnosing children with eating disorders requires tailor-made criteria such as those supplied by GOSH.

Evaluation of Nicholls et al. (2000)

Strengths

• The study’s use of blind clinicians (who did not know the diagnosis given by their counterpart) increases the validity of the findings as it helps to eliminate bias from the assessments provided

• The findings are vital in that they pinpoint flaws in the more traditional classification systems and highlight how children with eating disorders should be diagnosed

Limitations

• A sample of 81 children from the UK is small and unrepresentative of the wider population, making the results difficult to generalise

• The research only highlights how children with eating disorders should be diagnosed, it does not account for other disorders which may also require a separate and specific classification system