Cliff Prior from the National Schizophrenia Fellowship is absolutely correct to separate danger from diagnosis ('Personality disorder debate', letters, 25 March).

However, the conceptual problems of personality disorder are no different from those applying to all psychiatric diagnosis of functional disorders, including schizophrenia.

The latter has poor predictive validity about either prognosis or dangerousness.

Moreover, disjunctive category-two patients with different symptom profiles can receive the same label.

It is true that psychiatrists trained in the same diagnostic method (say ICD-10 or DSM-IV) are more likely to agree on the diagnosis, but reliability is not validity.

Functional diagnoses ipso facto have no hard biological signs, have unknown or contested aetiology and are constructed from professional descriptions of the patients' behaviour and verbalisations alone (symptoms).

Thankfully more and more mental health practitioners are abandoning the bio-medical obsession with point diagnoses, which are scientifically weak and socially stigmatising, and are looking instead at how to ameliorate specific presenting problems and improve the quality of life of their patients.

Professor David Pilgrim

Consultant clinical psychologist

Communicare trust