[Patient opinion surveys in the hospital: revalidation and improvement of a questionnaire].

Measures to improve quality of care in hospitals should best be based upon the results of patient surveys. The present paper reports in the validation of a questionnaire originally developed in the US in 658 patients from two Swiss hospitals in a tertiary care setting. Questionnaires were sent out two weeks after release from hospital, more than 75 percent of questionnaires were returned. Based upon the experience that feedback of many single items rarely results in distinct actions to improve quality of care a factor analysis followed by an analysis of reliability was undertaken to validate the instrument and to reduce the amount of information, to help hospital officials to deal with the data, and finally to reduce the number of items in the questionnaire. Factor analysis revealed a 3-factor-solution based upon 25 out of 37 deficit-oriented questions, explaining 34.4 % of total variance. The factors can be described as: 1. Patient information (9 items; explained variance: 13.0 %); 2. Patient involvement and shared decision making (11 items; explained variance: 11.2 %); 3. Nursing quality (5 items; explained variance: 10.2 %). Cronbach's alpha values of these scales are 0.75, 0.72, and 0.69, respectively. As further analyses show, reducing the item pool does not lead to the loss of relevant information. Single wards show substantial differences in scores based on these factors (e.g. between 10 and 27 percent deficit in factor 2). The sensitivity of the factors in describing specific profiles of single wards will help to target measures to improve quality of care to those wards where improvement is most utterly needed.