Development and evaluation of data entry templates based on the entity-attribute-value model for clinical decision support of pressure ulcer wound management

PurposesThe purpose of this study was to develop and evaluate the functionality of structured data entry templates using the entity-attribute-value (EAV) model for clinical decision support of pressure ulcer wound management.MethodsA data set for wound assessment of pressure ulcers that has commonly been recommended by clinical practice guidelines was identified, and then the EAV models on each data were developed. Structured data entry templates and a database were developed based on these EAV models. These were integrated with a knowledge engine into the clinical decision support system (CDSS) to provide patient-specific recommendations on pressure ulcer wound management. The functionality of the EAV model and structured data entry templates for the CDSS was evaluated heuristically by five nurse experts using clinical scenarios.ResultsThe data set containing 13 entities was identified and EAV models of these entities were created. Cardinalities and data types of attributes were defined to represent the models in more detail. Terms used in the EAV models were mapped to SNOMED CT concepts. Six data entry templates and the relational database with ten tables were developed. Five nurses successfully entered all data in the scenarios except one data element and retrieved expected recommendations successfully from the clinical decision support system when all data were entered correctly.ConclusionsThe clinical data models and structured data entry templates developed in this study were useful in supporting clinical decision making on pressure ulcer wound management.

► We identified 13 data entities for pressure ulcer wound management.
► Thirteen entity-attribute-value models were created to represent these entities.
► Six structured data entry templates and the relational database based on these models were developed and integrated with the CDSS to provide patient-specific recommendations.
► Five nurses successfully entered all data and retrieved expected recommendations in six scenarios except one.
► We conclude that the data models and structured data entry templates were useful in supporting decision making.