Reliability and Validity

Reliability

• For classification system to be reliable, different clinicians using same system should arrive at same diagnosis

• Reliability - level of agreement on the diagnosis by different psychiatrists across time and cultures; stability of diagnosis over time given no change in symptoms

AO3

• Diagnosis of schizophrenia is difficult as the practitioner has no physical signs but only symptoms (what the patient reports) to make a decision on

• Jakobsen et al. (2005)

  • Tested the reliability of the ICD-10 classification system in diagnosing schizophrenia.

  • 100 Danish patients with a history of psychosis assessed

  • Concordance rate of 98% obtained

  • High reliability of clinical diagnosis of schizophrenia

• Comorbidity - suffering from two or more mental disorders

  • e.g. schizophrenia + depression

  • More difficult to confidently diagnose schizophrenia

  • Overlap of symptoms e.g. low motivation - does it reflect schizophrenia, depression, or both?

• Gender bias

  • Loring and Powell (1988) found that some behaviour was regarded as psychotic in males but not in females

Validity

• Validity: the extent to which schizophrenia is a unique syndrome with characteristics, signs and symptoms

• For classification system to be valid should be meaningful and classify a real pattern of symptoms, which result from a real underlying cause

AO3

• Schizophrenia - single disorder? No such thing as a ‘normal’ schizophrenic exhibiting the usual symptoms

• Problems with validity of diagnosis classification - unsuitable treatment may be administered, sometimes on an involuntary basis - practical and ethical issues.

• Are we really testing what we think we are testing?

  • USA - 20% of psychiatric patients classes as having schizophrenia in 1930s → 80% in the 1950s

  • London - rate remained at 20%

  • Neither group had a valid definition of schizophrenia

• Neuropsychologist Michael Foster Green - neurocognitive deficits in basic functions such as memory, attention, central executive and problem solving skills may combine to have an outcome which we are labelling “schizophrenia” as if it were the cause when in fact it is simply an umbrella term for a set of effects

• Predictive validity

  • If diagnosis leads to successful treatment - diagnosis valid

  • Some schizophrenics successfully treated, others not

  • Heather (1976) - only 50% chance of predicting what treatment a patient will receive based on diagnosis, suggesting not valid

• Aetiological validity

  • For a diagnosis to be valid, all patients diagnosed should have same cause. Not the case with schizophrenia - biological, physical, both

• Rosenham (1973)