Seite - 78 - in Applied Interdisciplinary Theory in Health Informatics - Knowledge Base for Practitioners

Each of these models has associated principles extracted from the literature on distributed cognition [10]. Table 1 shows how some of these are applied. These distributed cognition principles have been shown to help analysts gain further insight into complex sociotechnical systems compared to contextual design alone [14]. These principles should not be seen as a comprehensive and prescriptive check list of distributed cognition features, but as a set of sensitizing concepts to enrich what can be seen and described – some are likely to be relevant in a given context and some will not be. This approach has been applied in different contexts by our group and others (e.g. [15, 16]). DiCoT-CL is an extension of DiCoT, which adds ‘concentric layers’ to each of the models to encourage the analyst to think about micro, meso and macro layers of distributed cognition [17]. DiCoT-CL has been applied to evaluate the design and use of a blood glucose meter on a ward [17] and to investigate the safety around infusion practices on a ward [18]. In both cases we treat the micro as the layer that is closest to the interactions to do with the device, procedure or technology under study, e.g. by the bedside; the meso is the layer out from this which might include different professionals as a team, e.g. at the scale of the ward; and macro is the layer that might be as broad as the hospital or above, e.g. national guidance. These different layers have been effective in showing that determinants for success and failure in a system might not be proximate to technology use but might be further away in space and time. For example, the configuration of infusion pump alarms in the hospital layer had downstream consequences for staff and patients in the micro layer [18]. Figure 1. DiCoT-CL framework has three concentric layers of the sociotechnical system, where each layer is divided into five areas that reflect the themes of the different DiCoT models, i.e. information flow, artefact, physical, social and evolutionary models (adapted from [17]). Reproduced with permission. D.Furniss etal. /DistributedCognition:UnderstandingComplexSociotechnical Informatics78