DSM data are often used to build statistics on the types of patients to who
m mental health services are offered. These statistics are normally based o
n broad categories that cluster finer subcategories for specific diagnoses.
These broad categories can be built in a number of ways. The purpose of th
is study is to determine the effect of the choices made in forming broad DS
M categories, by observing the resulting differences in clinical health sta
tistics. Results based on 3496 clinical cases show that DSM-III-(R) classif
ications can be presented in different formats, leading to different outcom
es. To avoid wrong conclusions based on one format, DSM statistics should a
lways be presented in varying formats.