Objective: The aim of this study was to evaluate the impact of a centralised referral mechanism (CRM) upon the number and type of "inappropriate referrals" made to a countywide physical activity referral scheme. Design: Case study. Method: Phase 1: Hierarchical Content Analysis of 458 "inappropriate referrals" made to a countywide scheme over a 28-month period. Phase 2: Chi-square analysis to compare patients removed on the basis of psychosocial or medical reasons by reason for original referral and gender. One-way ANOVA to compare both groups by age. Results: Reasons for referral were classified into first, second and third order themes. Two third level themes emerged from the data: medical contraindicator and psychosocial reasons. Significantly more people were removed on the basis of psychosocial factors rather than medical reasons (x[superscript 2] = 77.17; p less than 0.01 ). Further those referred for overweight/obesity (x[superscript 2] = 49.31; p less than 0.01) or for musculoskeletal conditions (x[superscript 2] = 40.96; p less than 0.01) were more likely to be removed on the basis of psychosocial factors; whilst those referred for a cardiac condition were more likely to be removed for a medical reason (x[superscript 2] = 11.76; p less than 0.01). Women were significantly (x[superscript 2] = 10.78; p less than 0.01) more likely to be removed for psychosocial reasons. Patients removed from the scheme for medical reasons were also significantly (F = 13.00; p less than 0.01 ) older than those in the psychosocial group. Conclusions: This study provides support for the introduction of a CRM within a countywide physical activity referral scheme. The need for existing referral guidelines to be extended to include holistic recommendations with regard to both the medical conditions that are appropriate for referral and also the stage of readiness of the patient to take-up the referral is recommended. (Contains 2 tables and 1 figure.)