OBJECTIVE To establish the reliability and validity of the translated version of the Safety Attitudes Questionnaire (SAQ) by evaluating its psychometric properties and to determine possible differences among nurses and physicians regarding safety attitudes. DESIGN A cross-sectional study utilising the Albanian version of the SAQ and a demographic questionnaire. SETTING Four regional hospitals in Albania. PARTICIPANTS 341 healthcare providers, including 132 nurses and 209 doctors. MAIN OUTCOME MEASURES The translation, construct validity and internal validity of the SAQ. The SAQ includes six scales and 30 items. RESULTS A total of 341 valid questionnaires were returned, for a response rate of 70%. The confirmatory factor analysis and its goodness-of-fit indices (standardised root mean square residual 0.075, root mean square error of approximation 0.044 and comparative fit index 0.97) showed good model fit. The Cronbach's α values for each of the scales of the SAQ ranged from 0.64 to 0.82. The percentage of hospital healthcare workers who had a positive attitude was 60.3% for the teamwork climate, 57.2% for the safety climate, 58.4% for job satisfaction, 37.4% for stress recognition, 59.3% for the perception of management and 49.5% for working conditions. Intercorrelations showed that the subscales had moderate-to-high correlations with one another. Nurses were more hesitant to admit and report errors; only 55% of physicians and 44% of nurses endorsed this statement (χ(2)=4.9, p=0.02). Moreover, nurses received lower scores on team work compared with doctors (N 45.7 vs D 52.3, p=0.01). Doctors denied the effects of stress and fatigue on their performance (N 46.7 vs D 39.5, p<0.01), neglecting the workload. CONCLUSIONS The SAQ is a useful tool for evaluating safety attitudes in Albanian hospitals. In light of the health workforce's poor recognition of stress, establishing patient safety programmes should be a priority among policymakers in Albania.