Background: GPs have referred patients for psychological treatment under the Better Outcomes in Mental Health Care, Access to Allied Psychological Services (ATAPS) Program since 2003. It is not known how GPs might select patients for referral. We explored which characteristics identified ATAPS patients compared to usual GP patients. Methods: The study was conducted in GP Access, a Division of General Practice (Newcastle and Lower Hunter) in NSW, Australia. It was a case-control design with 63 cases (ATAPS patients), and 64 controls (GP patients never referred to ATAPS). Unadjusted and sequentially adjusted logistic regressions were used to identify independent predictors of being an ATAPS case based on official referral guidelines: ICD-10 diagnosis of depression or anxiety and scores on the K-10 (psychological distress) and DASS-21 (psychological symptoms). A multivariable logistic regression was also used to determine the best minimum set of predictor variables. Results: Eight-three per cent of ATAPS cases had anxiety or depression. In unadjusted models, any mood disorder, OR 7.68 (95% CI: 3.47, 17.01), any anxiety disorder, OR 2.88 (95% CI: 1.37, 6.05), higher K-10 score, OR 1.06 (95% CI: 1.04, 1.14) and higher DASS-21 score, OR 1.06 (95% CI: 1.03, 1.09) were associated with being an ATAPS case. Any mood disorder, any anxiety disorder, K-10 score and DASS-21 scores remained significant in most adjusted analyses and all models showed change when adjusted for mental disability and physical disability. Three variables predicted being an ATAPS case in the multivariable regression: greater mental disability, lesser physical disability and greater number of substances misused. Conclusion: Cases had higher levels of mental disability and greater substance misuse, but lower levels of physical disability. This may reflect GP referral decision making and have implications for policy development.