Issues of reliability
Reliability refers to the consistency of a measuring instrument, such as a questionnaire or scale, to assess, for example the severity of their schizophrenic symptoms.
They can be measured in terms of whether two independent assessors give similar diagnoses (Inter-rater reliability) or whether tests used to deliver these diagnoses are consistent over time (test-retest reliability).
Inter-rater reliability
The DSM-III was specifically designed to provide a much more reliable system for classifying psychiatric disorders.
(Carson, 1991) claimed that it has fixed the problem of inter-rater reliability once and for all.
(-) Little evidence that DSM is routinely used with high reliability by mental health clinicians.
(-) (Whaley, 2001)– inter-rater reliability correlations in the diagnosis of schizophrenia as low as +.11.
(-) (Rosenhan, 1973)– when normal people presented themselves to psychiatric hospitals in the US claiming they heard an unfamiliar voice in their head saying the words ’empty’, ‘hollow’ and ‘thud’, they were all diagnosed with Schizophrenia and admitted. Throughout their stay none of the staff recognised that they were actually normal.
(-) Unreliable symptoms- (Mojtabi and Nicholson, 1995)– When 50 senior psychiatrists in the US were asked to differentiate between ‘bizarre’ and ‘non-bizarre’ delusions, there was an inter-reliability correlation of only around +.40, forcing the researchers to conclude that even this central diagnostic requirement lacks reliability.
(-) (Cheniaux et al., 2009)– although inter-rater reliability was above +.50 for both classificatory systems, S was more frequently diagnosed according to ICD-10 than DSM-IV.
Test-retest reliability
Cognitive screening tests are important in the diagnosis of S as they measure the degree of neuropsychological impairment.
(Wilks et al., 2003) administered two alternate forms of the test to schizophrenic patients over intervals varying from 1-134 days. The test-retest reliability was high as +.84.
(Prescott et al., 1986)– analysed test-retest reliability of several measures of attention and information processing in 14 chronic schizophrenics. Performance on these measures was stable over a 6-month period.
Issues of validity
Validity refers to the extent that a diagnosis represents something that is real and distinct from other disorders and the extent that a classification system such as the ICD or DSM measures what it claims to measure.
Diagnosis cannot be valid if it is not reliable.
Symptoms- (Ellason and Ross, 1995)– people with dissociative identity (DID) actually have more schizophrenic symptoms than people diagnosed as being schizophrenic.
Ethinicity may lead to misdiagnosis- (Harrison et al., 1997)– the incidence rate was eight times higher for African-Caribbean groups than for white groups. Explained as a result of poor housing, higher rates of unemployment and social isolation. Misdiagnosis may result from factors such as, cultural differences in language and mannerisms and difficulties in relating between black patients and white clinicians.
Comorbidity
Refers to the extent that two or more conditions co-occur. (Buckley et al., 2009) estimate that comorbid depression occurs in 50% of patients, and 47% of patients also have a lifetime diagnosis of comorbid substance abuse.
This creates difficulties in the diagnosis of a disorder and also in deciding what treatment to advise.
(+) (Weber et al., 2009)– examined nearly 6 million hospital discharge records to calculate comorbidity rates. Psychiatric and behaviour related diagnoses accounted for 45% of comorbidity. However, also found that many patients with a primary diagnosis of S were also diagnosed with medical problems, including hypothyroidism, asthma, hypertension and type 2 diabetes. Concluding- patients tend to receive a lower standard of medical care, which in turn adversely affects their prognosis.
(Kessler et al., 1994)– the rate for attempted suicide rose from 1% for those with S alone to 40% for those with at least one lifetime comorbid mood disorder.
Positive or negative symptoms?
(Klosterkotter et al., 1994) found that positive symptoms were more useful for diagnosis than were negative symptoms.
Prognosis
The prognosis for patients varies with about 20% recovering their previous level of functioning, 10% achieving significant and lasting improvement, and about 30% showing some improvement with intermittent relapses. They therefore have little predictive validity – some people never appear to recover, but many do.