Treatment of schizophrenia with antipsychotic drugs is frequently sub-optimal. One reason for this may be heterogeneity between patients with schizophrenia. The objectives of this study were to identify patient, disease and treatment attributes that are important for physicians in choosing an antipsychotic drug, and to identify empirically subgroups of patients who may respond differentially to antipsychotic drugs. The survey was conducted by structured interview of 744 randomly-selected psychiatrists in four European countries who recruited 3996 patients with schizophrenia. Information on 39 variables was collected. Multiple component analysis was used to identify dimensions that explained the variance between patients. Three axes, accounting for 99% of the variance, were associated with disease severity (64%), socioeconomic status (27%) and patient autonomy (8%). These dimensions discriminated between six discrete patient subgroups, identified using ascending hierarchical classification analysis. The six subgroups differed regarding educational level, illness severity, autonomy, symptom presentation, addictive behaviors, comorbidities and cardiometabolic risk factors. Subgroup 1 patients had moderately severe physician-rated disease and addictive behaviours (23.2%); Subgroup 2 patients were well-integrated and autonomous with mild to moderate disease (6.7%); Subgroup 3 patients were less well-integrated with mild to moderate disease, living alone (11.2%); Subgroup 4 patients were women with low education levels (5.4%), Subgroup 5 patients were young men with severe disease (36.8%); and Subgroup 6 patients were poorly-integrated with moderately severe disease, needing caregiver support (16.7%). The presence of these subgroups, which require confirmation and extension regarding potentially identifiable biological markers, may help individualizing treatment in patients with schizophrenia.