Classification 23differences between diagnostic groups should be more immediately rele-
vant to aetiology, associated problems, treatment response, or prognosis.
It is possible to have satisfactory diagnostic categories but an unsatisfac-
tory overall classification. This is true when too many cases fail to meet
the criteria for any category, or have to be fitted into ‘atypical’ or ‘miscel-
laneous’ categories. An ideal classification is as valid and as comprehensive
as possible, but these two aims sometimes pull in opposite directions.
Phenomenology above all
The classification of psychopathology at all ages has increasingly focused
on the presenting features of each disorder rather than on the supposed
aetiology or pathogenesis. When disorders are defined in this way, it is
possible to study aetiology and pathogenesis with an open mind. Diagnos-
tic categories based on pathogenesis, such as ‘minimal brain damage’ or
‘reactive psychosis’, have generally impeded rather than facilitated clinical
and research progress. Although most child and adolescent psychiatric
disorders are currently defined on the basis of phenomenology alone, a
few disorders such as ‘reactive attachment disorder’ and ‘post-traumatic
stress disorder’ are defined both in terms of phenomenology and the
presumed cause.
Dimensions or categories?
Many sorts of psychopathology seem to be extreme values on a continuum
that extends into the normal range, with many children and adolescents
exhibiting lesser degrees of the same features. Imposing a cut-off between
normality and abnormality is sometimes an arbitrary but convenient way
of converting a dimension into a category. To take an example from
general medicine, blood pressure is continuously distributed, with progres-
sively higher blood pressures leading to progressively higher rates of stroke
and heart disease. Keeping blood pressure as a dimension retains more in-
formation than imposing a cut-off that fairly arbitrarily divides individuals
into those with normal blood pressure and those with ‘hypertension’ –
and yet a simple dichotomy has the potential advantage of feeding directly
into a straightforward action plan for busy practitioners: treat those who
are hypertensive and leave individuals with normal blood pressure alone.
On the other hand, a single dichotomy based on a single cut-off may
oversimplify the options. For example, it may be a good idea to titrate
the intervention more flexibly, for example, advising simple dietary or
lifestyle changes at borderline blood pressures and adding in medication at
higher values. Because similar issues apply in mental health, the architects
of DSM 5 and ICD 11 are wrestling with how to create hybrid schemes that
will retain the simplicity of diagnostic categories alongside the flexibility of
dimensional scores.
While dichotomising dimensions is sometimes arbitrary, there are in-
stances where individuals with extreme values are genuinely a case apart.
There are three possible indications of discontinuities between normal and
extreme values. First, the distribution may be bimodal, for example, with a