psychiatric diagnosis can be in clinical practice.
The research uses much stricter principles. They may test for
inter-rater reliability, i.e. they check to make sure the physicans are
all diagnosis roughly the same way. This is how the science works. If
they can reliably make a diagnosis which is the same as the quality of
the diagnosis in research then there's a reasonable expectation the
patient will follow the negative prognsosis. This reasonable chance is
the validity of psychiatric diagnosis. Or the other way round. Can't
remember. Bentall explains it better.
In practice I think psychiatrists don't diagnose the same way. I think
many of them don't use the strict criteria and often use methods they've
developed or learned to make a diagnosis.
Sometimes they'll refer to the differential diagnosis information in
psychiatric textbooks. This is the information which is used to strictly
define differences between different disorders. This is necessary
because, for example, schizophrenics and manic depressives can exhibit
the same set of symptoms at different times or concurrently. These two
types are in practice often difficult to separate because some people
display symptoms which someone might expect the other type to display.
Their type isn't schizoaffective. This is a different thing because of
the way it is diagnosed,. It isn't inbetween schizophrenia and bipolar
except in the outcomes. The diagnosis doesn't type the person as
inbetween schizophrenia and bipolar. In fact clinical physicians are so
poor at diagnosing this disorder that the APA want to remove it in DSM-V
even though studies show it exists as a different outcome to bipolar and
This problem of reliability of diagnosis in clinical practice is
important because it means patients could get the wrong medication or
treatment package. They may also have to face stigma when they don't
have to. There's also the unexplored effect of expectations and
labelling. Frankly this is beyond my capability to explain but
essentially a label can cause better or worse performance. The Rosenthal
experiment showed children who teachers are told are "late bloomers"
(they have had average performance but are expected to excel in the
coming year) at random perform better on teacher assessments and
standardised IQ tests. It has not been tested whether the opposite is
true, e..g. iif a label of schizophrenia causes worse outcomes, because
it may be an impossible experiment and it would be highly unethical.
Patients want to know the right label and they want the right treatment.
If doctors can't get it right and if the research community can't design
diagnostic systems which actually work in clinical practice then what
patients get is chaos and the placebo effect as the most effective
treatment in modern mental healthcare.