DSM data are often used to build statistics on the types of patients to whom mental health services are offered. These statistics are normally based on broad categories that cluster finer subcategories for specific diagnoses. These broad categories can be built in a number of ways. The purpose of this study is to determine the effect of the choices made in forming broad DSM categories, by observing the resulting differences in clinical health statistics. Results based on 3496 clinical cases show that DSM-III-(R) classifications can be presented in different formats, leading to different outcomes. To avoid wrong conclusions based on one format, DSM statistics should always be presented in varying formats.