Page - 183 - in Applied Interdisciplinary Theory in Health Informatics - Knowledge Base for Practitioners

monitoring). Transfer of knowledge on PAC procedures and related practice updateswas observed to occur informally during other work, leaving the intervention vulnerable to staff attrition (as no formal process existed to ensure that this knowledge was transferred to new appointees). 2.3. Limitations of NPT Before concluding with an exercise to help readers apply NPT to health informatics developments we will discuss some of its limitations, the first of whichconcerns lack of sensitivity to wider contextual factors beyond the immediate site of implementation. For example, Clarke et al. (2013) used NPT to evaluate implementation of a training programme for carers of stroke patients, within a cluster randomized controlled trial (RCT) [13]. The multi-site nature of this trial meant that variations in implementation context were present at the local level (e.g. service, resources, divisions of labour), in addition to regional and national policy changes (with differences in local responses to such changes providing further sources of complexity). In their evaluation, Clarke et al. noted that while NPT had been useful for identifying mechanisms and processes that inhibited implementation of the training programme, it did not capture the impact of these wider contextual factors [13]. At a national level, recruitment to the cluster RCT began shortly after the launch of a new National Stroke Strategy in 2008[13]. In addition, many sites experienced competing demands on MDT members’, patients’, and care givers’ time and resources from other service development initiatives [13]. All the hospital services involved were working towards the goal of stroke survivors spending all or part of their stay on a stroke unit, while most were also planning or introducing thrombolysis services. In addition, many sites were introducing early supported discharge schemes or reorganization of existing services, which required changes in staff locations and roles. While the impact of these factors may have been visible indirectly through their impact on other kinds of process identified through NPT (e.g. Resource allocation), Clarke et al. found that theory did not account fully for their role in the implementation context. These observations indicate both the vulnerability of service developments (including health informatics innovations) to organizational turbulence, and how building relationships and processes that are resistant to such turbulence is essential in complex healthcare settings. Theyalso indicate the importance of attending to contextual factors that shape implementation processes, a concern that has driven ongoing development of the theory [14]. In addition, authors such as Johnson et al. (2017) have sought to address these limitations in their application of the theory, by presenting adapted models that link the constructs with wider organizing structures and social norms (e.g. policies, public expectations of services, political contexts – see Figure 2). Elsewhere, in a systematic review of NPT use in feasibility studies and process evaluations, May et al. (2018) noted a number of additional criticisms from researchers: that NPT constructs overlapped; that the technical vocabulary of the theory was difficult; and that as a result coding qualitative data was difficult [1]. May et al. noted that problems of this nature seemed less evident when researchers used a more inductive approach to qualitative data analysis than they did when authors employed a framework approach[1]. M.BracherandC.R.May / ImplementingandEmbeddingHealth InformaticsSystems 183