Evaluating Model-Driven Development for large-scale EHRs through the openEHR approach

•The separation of clinical and technical concerns seem to be rather aspirational.•End-user control does not apply for large-scale implementations.•Modelling the domain entail modelling clinical work practice.

PurposeIn healthcare, the openEHR standard is a promising Model-Driven Development (MDD) approach for electronic healthcare records. This paper aims to identify key socio-technical challenges when the openEHR approach is put to use in Norwegian hospitals. More specifically, key fundamental assumptions are investigated empirically. These assumptions promise a clear separation of technical and domain concerns, users being in control of the modelling process, and widespread user commitment. Finally, these assumptions promise an easy way to model and map complex organizations.MethodsThis longitudinal case study is based on an interpretive approach, whereby data were gathered through 440 h of participant observation, 22 semi-structured interviews and extensive document studies over 4 years.ResultsThe separation of clinical and technical concerns seemed to be aspirational, because both designing the technical system and modelling the domain required technical and clinical competence. Hence developers and clinicians found themselves working together in both arenas. User control and user commitment seemed not to apply in large-scale projects, as modelling the domain turned out to be too complicated and hence to appeal only to especially interested users worldwide, not the local end-users. Modelling proved to be a complex standardization process that shaped both the actual modelling and healthcare practice itself.ConclusionA broad assemblage of contributors seems to be needed for developing an archetype-based system, in which roles, responsibilities and contributions cannot be clearly defined and delimited. The way MDD occurs has implications for medical practice per se in the form of the need to standardize practices to ensure that medical concepts are uniform across practices.