The Hospital Anxiety and Depression Scale (HADS) is widely used as a tool for assessing psychological distress in patients and non-clinical groups. Previous studies have demonstrated conflicting results regarding the factor structure of the questionnaire for different groups of patients, and the general population. This study investigated the factor structure of the HADS in a large heterogeneous cancer population of 1474 patients. It also sought to investigate emerging evidence that the HADS conforms to the tripartite model of anxiety and depression (Clark & Watson, 1993), and to test the proposal that detection rates for clinical cases of anxiety and depression could be enhanced by partialling out the effects of higher order factors from the HADS (Dunbar et al., 2000). The results demonstrated a two-factor structure corresponding to the Anxiety and Depression subscales of the questionnaire. The factor structure remained stable for different subgroups of the sample, for males and females, as well as for different age groups, and a subgroup of metastatic cancer patients. The two factors were highly correlated (r =.52) and subsequent secondary factor analyses demonstrated a single higher order factor corresponding to psychological distress or negative affectivity. We concluded that the HADS comprises two factors corresponding to anhedonia and autonomic anxiety, which share a common variance with a primary factor namely psychological distress, and that the subscales of the HADS, rather than the residual scores (e.g. Dunbar et al., 2000) were more effective at detecting clinical cases of anxiety and depression.