| Year: 2006 | 
Download Free Adobe Acrobat Reader |
| LHNCBC-2006-029 |
| Text Features and Readability: Expert Evaluation of Consumer Health Text |
| Rosemblat G, Logan R, Tse T, Graham L |
| Mednet 2006: 11th World Congress on Internet in Medicine the Society for Internet in Medicine. October 14-16, 2006. Toronto, Canada. |
| Background: Previous research suggests that consumers frequently have difficulty understanding written health information. Objectives: This exploratory study investigated the influence of linguistic and stylistic features on the readability of consumer health texts. Specifically, the research goals were (1) to examine the importance of previously identified predictors of general readability in the consumer health domain, based on expert judgment, and (2) to characterize patterns associated with expert ratings of readability across the various predictors. Methods: Health communication experts (N = 4) reviewed a sample of 22 consumer health texts on two common health topics, asthma and weight management. Each expert independently rated the contribution of 15 specific features on the readability of all documents in the sample. Results: Simultaneous multiple regression found that a 15-variable model significantly predicted readability for a general audience. However, only two variables, Vocabulary and Main Point significantly predicted general audience readability. A factor analysis of all ratings for the 15 features across the 22 documents revealed three clusters of features representing expert perceptual orientations: (1) discourse-level features, (2) sentential-level features, and (3) semantic features (Vocabulary and Main Point). Conclusions: The preliminary results suggest that developing consumer health-specific readability tools may require both modifying existing general measures, such as including health-related vocabulary, as well as adding new predictive features, such as ability to detect the "take-home message." Future work includes verification of this expert evaluation by consumers. Document URL: http://www.mednetcongress.org/fullpapers/MEDNET-192_RosemblatGracielaA_e.pdf |
| PDF |
