Archetype based Electronic Health Record

Today`s health care professionals can access an ever increasing amount of patient-related information and clinical knowledge, one of the most-promising applications being the Electronic Health Record (EHR). While several clinical and economic benefits are expected by the EHR, it may also lead to information overload, as the relevant information item that an EHR user searches for may be hidden in vast amounts of information within the lifelong EHR. Therefore, we see an urgent need for supporting EHR users in selectively retrieving information that is relevant in their respective search context. Besides document meta-data, archetype-based dual-model EHR architectures propose a corresponding support, but it is unclear whether they fulfil this promise. Overall, it is unclear what the information needs of health care professionals are, and how to support them adequately by dual-model EHR architectures. The objectives of this project are therefore 1. to identify the information needs of health care professionals when accessing the EHR, considering the respective search context, 2. to develop concepts to fulfil these information needs by combining document meta-data and dual-model EHR architectures, 3. to evaluate the developed concepts in a trial implementation. The project is innovative in the sense that information needs will be analysed (and not only information access patterns), and that we will propose concepts how to combine meta-data-based and archetype-based searching to optimize precision and recall of EHR queries. The concepts will be based on international standards such as IHE XDS to provide a future-proof, vendor independent solution. For the archetype-based search, we will take into account EN/ISO 13606 and HL7 CDA, both important international EHR standards. As a proof-of-concept we will focus on information needs in the treatment of patients suffering from diabetes. This focus was chosen as diabetes represents a chronic disease in which much information from different sources accumulates within the EHR over time, thus making the identification of the relevant information a challenging task. The information needs analysis will be based on interviews and observation of clinical EHR users and the analysis of clinical guidelines. The resulting information needs patterns will then form the basis for the concept of a generic document crawler that will support EHR users in finding information efficiently. The trial implementation of the document crawler will be evaluated by assessing precision and recall of information retrieval within the EHR, combined with user satisfaction surveys and usability evaluations. The outcome of the project will be a better knowledge of information needs of clinical EHR users - the initial scope of diabetes treatment can easily be expanded in follow-up projects - and a concept and a trial implementation of efficient EHR information retrieval techniques. We expect that our results will also influence the further development of parts of international standards such as IHE XDS.

The electronic health record that is currently being introduced in Austria is discussed controversally. Health care professionals often fear that the electronic health record will contain too much information, thus making retrieval of specific patient information, such as the history or allergies of a patient, too cumbersome or even impossible for them. In the framework of the EHR Arche project we developed a concept and a prototype to support health care professionals in searching information in an electronic health record. In an evaluation study with physicians we could see that our approach increased the timeliness and quality of information retrieval in the electronic health record. In this project we focused on patients suffering from Diabetes Mellitus, as this chronic disease normally leads to the generation of a large number of documents to be stored in an electronic health record. We started the project with analysing the specific information needs of Diabetes specialists by conducting observations of clinical encounters, interviews with Diabetes specialists, and analyses of available documentation systems. We found that information needs depend on the clinical context; for example, during a rountine check, other information is needed than in the case of an emergency encounter. The identified and formalized information needs were then mapped to search queries, which were implemented in a prototype based on international standards. In the future, this prototype could be integrated in the Tyrolean electronic health record system. In a controlled evaluation study we compared the search options available nowadays with the new search queries offered by our prototpye. We found that using our prototype, relevant information about a patient such as his allergies, previous surgeries, or recent medication can be found much quicker and more successfully. All eight health care professionals that participated in the evaluation study appreciated the new search options. Our EHR Arche project offers some insight on how to design the upcoming electronic health record system in Austria, in order to support information retrieval and to avoid information overload of health care professionals. This will help to improve the quality of health care by electronic health records in the future.