| Summary: | Thesis (S.M.)--Harvard-MIT Division of Health Sciences and Technology, 2004. === Includes bibliographical references (p. 40-43). === Background: Local adaptation of guidelines may increase compliance with guidelines that have been developed at a national level and are often not used in practice because of contextual factors. We have developed a representation scheme known as HieroGLIF, that facilitates a two-step approach to development and implementation of guidelines. In the first step, professional medical societies create a setting-independent guideline. In the second step, the guideline is adapted to local settings. The scheme represents the setting-independent guideline knowledge in a hierarchical structure for which we use axiomatic design theory to guide a top-down design. This representation extends the Guideline Interchange Format (GLIF), a frame-based ontology that has been previously developed for representation of guidelines. Methods: We encoded conventional national guidelines in the setting-independent format which were adapted to local clinical settings of their practice by primary care physicians (PCPs). We conducted a qualitative analysis of the type of changes that were made by the local adaptors. For each of the guidelines two patients' scenarios were created for which two sets of guideline recommendations were generated, one from the adapted hierarchical guideline and one from the conventional, non-hierarchical guideline. We evaluated in a randomized controlled trial the potential impact of the local adaptation methodology on adherence to guidelines. For each recommendation, 70 PCPs responded to a questionnaire that inquired if PCPs would follow the recommendations and their ratings for several relevant attributes. We also analyzed the data to look for attributes of recommendations that are important for their acceptance by PCPs. Results: In 8 out of === (cont.) recommendations the changes made were significant. The most common types of changes were additions of practical information (6/19) and deletions of unnecessary information (6/19). The PCPs significantly preferred the adapted version for one recommendation in which addition of practical information and deletions of unnecessary information were done. The response for changes in clinical content in another recommendation resulted in rejection of the adapted version. The most important attributes for acceptance of recommendations found were agreement with the clinical content and then in descending order--fitness to the practice environment, availability of practical information and succinctness. Conclusions: We conclude that the potential role of local adaptation is mainly for adding relevant practical information and deletions of unnecessary information. This combination of changes led to significant preference of adapted guidelines' version. On the contrary, changes to clinical content led to rejection of an adapted version recommendation and should be done cautiously. In addition, efforts should be made to make clinical recommendations fit the practice environment, contain practical information and be succinct. We consider the real impact of local adaptation in this study lower that expected. We also believe that the study sharpens the issues of the optimal level of adaptation in specific and the real impact of local adaptation in general. === by Yaron Denekamp. === S.M.
|
|---|
