reliability + validity

• Inter-rater reliability


  Diagnosis is reliable when different clinicians reach the same diagnosis for the same individual.
• Test-retest reliability


  When the same clinician reaches the same diagnosis for the same individual on two occasions
• Reliability means consistency
• Validity concerns whether clinicians are assessing what they are intending to assess, whether it is accurate
• Factors that contribute to the low validity in diagnosis of SZ:

  • complexity and overlap of symptoms
  • subjective nature of diagnosis
  • lack of clear biological tests
• Ward et al. (1962) identified three types of factors which affect the reliability of diagnosis:

  1. Client/patient factors
  2. Clinician factors
  3. Classification factors
• Client/patient factors


  • Clients may not be able to talk about their mental state clearly - may feel ashamed or leave out information
  • The person (a relative) speaking on the client's behalf may have a vested interest in the diagnosis being made and minimise or maximise the severity
  • Atypical presentations mean that not all clients present with the textbook symptoms and they may not exactly fit the DSM criteria
• Clinician factors


  • How well the clinical interview is conducted depends on the rapport with the client and how the semi-structured nature of the interview unfolds
  • How well the clinician is trained and where they were trained
  • What approach in psychology they follow
  • How much they rely on observation or tests
• Classification factors


  • Differences between DSM and ICD (the WHO system) descriptors of the illness and which is used by the clinician
  • The use of DSM criteria for billing purposes in private medical bills may lead to a bias in use
• Sane in Insane Places

  • A psychological experiment conducted by Rosenhan in 1973
  • he explored the validity of psychiatric diagnoses, particularly in distinguishing between sane and insane individuals
• Cheniaux et al. (2009) had two psychiatrists assess the same 100 clients.

  • One psychiatrist diagnosed 26/100 patients according to DSM model and 44/100 according to ICD model
  • The other psychiatrist diagnosed 13/100 during the DSM model and 24/100 using the ICD model
• Mason et al. (1997)

  • followed 99 patients with schizophrenia over 13 years to see if the validity of the main classification systems worked
  • found that if they used the symptoms that were present over six months, there was good validity
• Jansson and Parnas (2007)

  • reviewed 92 studies that applied different definitions of schizophrenia to the same patient samples
  • found that both ICD-10 and DSM-5 had low accuracy with diagnosis
• Cheniaux et al. (2009) has proven that schizophrenia is much more likely to be diagnosed using the ICD model, thereby calling into question the validity of the diagnosis systems
• What classification model does the UK use for diagnosing schizophrenia?

  ICD Model (2 or more negative symptoms present)