Inter-observer agreement for quality measures applied to online health information.

Many quality criteria have been developed to rate the quality of online health information. However, few instruments have been validated for inter-observer reliability. Therefore, we assessed the degree to which two raters agree upon the presence or absence of information based on 22 popularly cited quality criteria on a sample of 21 complementary and alternative medicine websites. Our preliminary analysis showed a poor inter-rater agreement on 10 out of the 22 quality criteria. Therefore, we created operational definitions for each of the criteria, decreased the allowed choices and defined a location to look for the information. As a result 15 out of the 22 quality criteria had a kappa >0.6. We conclude that even with precise definitions some commonly used quality criteria to assess the quality of health information online cannot be reliably assessed. However, inter-rater agreement can be improved by providing precise operational definitions.