Reliability and validity of a self-administered questionnaire of patient health care behavior and satisfaction

The reliability of a self-administered questionnaire of health care behavior and patient satisfaction, and also the validity of measuring patient satisfaction by a single question, were investigated on 754 outpatients and 204 inpatients. In order to assess the reliability and the validity, the questionnaire was administrated twice. Subjects for reliability tests were 194 outpatients and 91 inpatients and validity tests were performed with 280 outpatients and 84 inpatients. To assess the validity of patient satisfaction, supplementary questions about satisfaction (outpatients 22 items, inpatients 14 items) were included in the questionnaire. Reliability was assessed using kappa coefficient and proportion of agreement. To assess validity, principal component analysis was performed, and the correlation between principal-components and the score of patient satisfaction was analyzed. With the exception of patient understanding of the doctor's explanation and patient satisfaction, kappa coefficient and proportion of agreement were high (kappa: 0.4 approximately 0.82), indicating that question contents were objective and easily understood. By making the question categories simple, a higher kappa coefficient and proportion of agreement in patient understanding about doctor's explanation and patient satisfaction were achieved. This suggests that in questionnaire, similar question category must be made simple or similar categories should be put together in analysis. Spearman's correlation coefficient between the question about patient satisfaction and the first principal component (which summarizes all of the questions about patient satisfaction) was high (outpatients: 0.63, inpatients: 0.67). Therefore the validity of measuring general information about patient satisfaction by using a single question is suggested.