gms | German Medical Science

Background: Clinical practice guidelines contain important diagnostic and therapeutic information extracted from trials, reviews and meta-analyses. Their lengthy textual form hinders easy access to the information at the point of care. Computerized Decision Support Systems (CDSSs) can improve the implementation of guidelines, by providing patient-specific point-of-care information, based on patient data extracted from the electronic health record and on rules (in the form of 'if... then ..), extracted from the guidelines.

Aim: To collect existing methodologies for guideline formalization into computer algorithms for electronic decision support and to apply these methods to guidelines for cardiovascular prevention .

Methods: We conducted a literature search in Pubmed and Web of Science for relevant studies. Retrieved methods were applied to the field of pharmacological cardiovascular prevention.

Results: We retrieved 20 relevant publications, discussing different methodologies for extracting information: consensus methods, document mark-up and ontological annotation, all with their specific advantages and disadvantages. Formalization, clarification and disambiguation are multidisciplinary processes requiring cooperation between medical domain experts and knowledge engineers. The process results in computer-interpretable and more transparent formulation of the original guidelines, which is in it self a form of validation. This is illustrated by applying the methods to the Belgian Consensus Conference Report on cardiovascular prevention.

Conclusions: Multidisciplinary human formalization of guidelines is a necessary prerequisite for building CDSSs. The field of cardiovascular prevention, because of its rich and sophisticated evidence base, is particularly apt to apply multi-faceted guideline formalization methods.