Early diagnosis and referral of oral cancer is essential. Successful implementation of clinical guidelines must include current practitioners and students. Objective: To evaluate the diagnostic accuracy of students at oral cancer screening and to assess the effectiveness of clinical referral guidelines. Study Design: Fifth year dental students were randomly allocated to either control (n=19) or experimental groups (n=18). Both received the customary training in oral diagnosis. The experimental group underwent a 2 hour workshop where the guidelines for the referral of suspicious lesions were discussed. Three months later, a set of 51 clinical cases including benign, malignant, and precancerous conditions/lesions were used to assess the screening ability of each subject. Results: All 37 students entered the study. Sensitivity (control group) ranged from 16.7% to 66.7%; the experimental group scored from 16.7% to 83.3%. Fifty percent of the experimental students reached sensitivity values ?62.5% (p=0.01). Diagnostic specificity (control group) spanned from 80% to 93.3% (median=50%); amongst experimental group it ranged from 82.2% to 97.8% (median=92.8%); (p=0.003). Concordance -control group- was X=82.5 (SD=3.2), and X=88.2 (SD=4.3) for the experimental, (p>0.001). Cohen's kappa test was poor (K<0.40) for the controls and moderate for the experimental group. The experimental group referred more oral cancers urgently (p=0.002) and left less unreferred cancers (0.04). This group also referred more precancerous lesions/conditions urgently (p=0.02). Conclusions: The implementation of a clinical referral guideline at undergraduate level has proved valuable, under experimental conditions, to significantly increase diagnostic abilities of the examiners and thus to improve screening for oral cancer. © Medicina Oral S. L.