The aim of this study was to adapt the instrument 'Good Nursing Care Scale for Patients' to Swedish conditions as a measure of patients' satisfaction, as well as estimating its reliability and validity. Following a pilot test, discussions in the author group, testing for readability among patients and judgement of content validity by a panel of experts, the final version was reduced to 72 items focusing on good caring. The refined instrument was assessed for internal consistency in 447 surgical in-patients, for 2 week test-retest reliability in 100 patients and subjected to orthogonal principal components factor analysis with varimax rotation, followed by second-order factor analysis. The internal consistency item-item correlation coefficient ranged from 0.15 to 0.91, correlation between each item and the total scale was >or=0.30 for 70 items, Cronbach's alpha coefficient for the final scale was 0.79 and test-retest reliability was 0.75. An orthogonal principal components factor analysis with varimax rotation was conducted on the final 71 items and the 15 first-order factors with eigenvalues >or=1 explained 66% of the total variance. A second-order factor analysis of these 15 factors as items resulted in a seven-factor solution. The total variance explained by the seven factors was 79%. Cronbach's alpha coefficient for the seven factors ranged between 0.32 and 0.95. The instrument seems reliable and valid to assess the patients' satisfaction with what happened during their hospital stay. To confirm the factor structure and improve factor consistency additional development and testing is suggested.