Seite - 215 - in Applied Interdisciplinary Theory in Health Informatics - Knowledge Base for Practitioners

As noted in section 3, the development or application of theory in health informatics is predominantly implicit rather than explicit. Readers (and editorsof theory textbooks!) are usually left to infer the theoretical contribution. The STARE-HI guideline for reporting of evaluation studies in health informatics [10] includes “theoretical background” as a section in Methods – this is very far from routinely followed. We certainly would not want to see our field adopt an extreme position where theory is idolized and academic papers become weighed down with ponderous and pretentious intellectual displays, as has been reported in the field of management [29], but health informatics generally appears to be at the opposite end of that spectrum and needs a nudge towards a stronger theoretical approach. 7. A research agenda for theoretical health informatics Finally, we consider theoretical topics for future research that have been identified in the literature and some that we propose based on our learning in preparing this textbook. From Table 4, weoffer some specific areas for consideration: Theory of CPOE implementation Theory of sociological design of EHRs Theory of computational diagnostic support Theory of clinical communication patterns Theory of healthcare protocol adoption Theory of systemic evaluation Theory of personally controlled electronic health records. The converging paradigms of precision medicine, Learning Health Systems and implementation science seem to offer a particularly fruitful ground for theory development given the central role of informatics in each of these fields [14; 52; 59]. A key aspect of this convergence is clinical decision support, which has long been an important area of study in health informatics [27; 42], though the robustness of its evaluation still has a way to go in terms of scientific measurement practice [53]. The need for sound theoretical foundations for this work has been recognised [23;24] but in some quarters seems to be perceived as merely a technical implementation challenge. We argue that this is a prime area where we should expect to see emerging theory. In addition to these specific topics, there is the general lack of replication studies in health informatics [19]. Without such a culture of replication studies, our field will be dominated by single-case evaluations that do not lend themselves to broader theoretical generalisation. Theories may not lend themselves to the same form of replication of findings but there is still a need for validation. Theoretical approaches often take a triangulation approach or utilise member validation methods. We have already noted in section 4 that criteria need to be developed for selecting relevant theory and assessing theory quality. As suggested in [7], there should be transparent reporting of the criteria used to select theories in research studies. This implies the need for a comprehensive list of criteria that are used to choose a theory. Such an approach in health informatics wouldencourage reflective thinking, explication and generalisability. The theory criteria would include the identification of key constructs; informing data collection; enhancing conceptual clarity; clarifying terminologyand hypothesising relationships [7]. This would be a useful contribution to the field and could inform the revision of existing reporting guidelines [10]. P.J.Scott etal. /ReflectingandLooking to theFuture 215