Objective To evaluate the reliability and reproducibility of the thoracolumbar osteoporotic vertebral compression fracture (OVCF) severity score system (TLOFS),so as to provide basis for clinical decision-making.Methods A total of 320 cases of thoracolumbar OVCF hospitalized between January 2014 and June 2015 were scored by TLOFS.There were 107 male and 213 female patients,aged 52-90 years (mean,63.78 years).All cases presented with a chief complaint of low back pain.Thirty-nine cases (TLOFS ≤3 in 21 cases,TLOFS =4 in 17 cases) were treated non-operatively,while 259 cases (TLOFS =4 in 23 cases,TLOFS ≥ 5 in 259 cases) were treated surgically.Pain was evaluated by the visual analogue scale (VAS).Reliability,reproducibility and content validity of the TLOFS were analyzed.Results Cronbach' sα coefficients of all TLOFS subgroups including evaluation of morphological changes,MRI,bone mineral density,clinical situation and therapy recommendation and total score evaluation were within the range of higher to high reliability (0.76-0.94).Kappa coefficients of all TLOFS subgroups were within the range of higher to high reproducibility (0.84-0.95).Mean content validity of all TLOFS subgroups was 0.916.VAS improved from preoperative(7.8 ± 2.1) to (1.3 ±0.7) at the final follow-up.Rates of correct diagnosis,sensitivity and specificity were 95.7％,87.6％ and 96.5％ respectively.Conclusion TLOFS exhibits good reliability and reproducibility,easy operation,and accurate and comprehensive evaluation,which is effective in guiding the clinical decision making in treatment of thoracolumbar OVCF.