Page - 6 - in Applied Interdisciplinary Theory in Health Informatics - Knowledge Base for Practitioners

before dissemination. So, each predictive principle should be subjected to rigorous evaluation studies in relevant settings, to test if it really does contribute to making systems better. Only if it passes these tests should it be accepted as a predictive principle for our discipline, to be applied in systems development and taught as part of accredited educational programmes for newcomers [10]. The details of designing and carrying out such theory-based evaluations are beyond the scope of this chapter but are summarized in a chapter introducing the concept of evidence-based health informatics [10] and detailed in a textbook on evaluation methods [11]. These evaluation principles need to be understood not only by academics developing and testing new principles but also by system designers and developers applying new principles, so that they can confidently carry out a critical appraisal of the studies that have been conducted to test the principle before applying it. Thus, the central idea discussed in this book of using theories or predictive principles more widely is closely linked with the idea of developing and testing theories central to evidence based health informatics. Figure 2 below illustrates the suggested process, from identifying a theory relevant to HI, deriving a predictive principle from it then incorporating the principle into system development and testing if this improves the system, for example by making the system more usable, accurate or effective. Figure 2. How to identify and use a useful health informatics principle An example of this kind of evaluation is the study we carried out to test the applicability of Fogg’s theory of credible website design, based on the design of eCommerce websites, to inform health-related decisions [12]. We designed two version of a website about organ donation with near identical content and usability, one of which followed all Fogg’s credibility recommendations while the other site lacked all of these. We then recruited over 800 students via email and randomized them to experience either the credible website or the less credible version. After 4 weeks, we asked participants to join the NHS Organ Transplant Register. Surprisingly, an identical proportion of about 38% joined the register in each group, demonstrating that credible website design had no role in taking this decision [12]. J.C.Wyatt /TheNeed forTheory to InformClinical InformationSystems6