Objectives. To assess discrepancies between perception of Belgian rheumatologists on eligibility of RA patients for anti-TNF treatment and Belgian reimbursement criteria and to compare Belgian with Dutch criteria and UK guidelines. Methods. Consecutive MTX-experienced patients with active RA were recruited from 25 private and academic rheumatology practices. Discrepancies between eligibility for anti-TNF treatment according to the rheumatologist and fulfillment of Belgian reimbursement criteria [HAQ ⩾25%, tender joint count (TJC) and swollen joint count (SJC) ⩾8; ⩾1 erosion; failure of ⩾2 DMARDs including MTX; no tuberculosis] were recorded. Reasons for failing the Belgian criteria and results of applying Dutch reimbursement criteria and UK guidelines on the dataset were analysed. Results. Of 492 patients, rheumatologists considered 135 (27.4%) as eligible, whereas Belgian criteria were fulfilled for only 34 (6.9%). Positive predictive value (PPV) of rheumatologists’ perception on eligibility for fulfillment of Belgian criteria was 22.9%, whereas negative predictive value (NPV) was 99.1%. The 104 patients (21.1%) considered eligible despite criteria not being fulfilled had significantly greater TJCs, SJCs, disease activity score (DAS28) and Rheumatoid Arthritis Disease Activity Index scores than the 385 patients (78.2%) in the no-discrepancy group. Number of swollen joints, HAQ and erosions mainly accounted for discrepancies. Of 492 patients, 263 (53.4%) qualified for Dutch criteria and 41 (8.3%) for UK guidelines. PPV of Belgian rheumatologists’ perception was 72.6% for fulfilling Dutch criteria (NPV 49.6%) and 23.4% for UK guidelines (NPV 96.7%). Conclusions. Rheumatologists consider more RA patients eligible for anti-TNF treatment than would be reimbursed according to Belgian criteria. Dutch guidelines, based on DAS28, match more closely eligibility according to Belgian rheumatologists.