Marlot Kuiper

69 Research Perspective: Professional routines performances in a for themdesirable way. Still, when routine participants actually start working with the artefact, the performances are not necessarily what the implementers had in mind (ibid). Earlier in this chapter, I described how various pressures force professional organisations like hospitals to adapt their ways of working, for example to increase (multidisciplinary) collaboration, to decrease costs, and to improve patient safety. From a routine dynamics perspective, it is thus argued that new patterns of action not automatically come about after implementing a checklist (artefact). Taking on routines as a unit of analysis provides valuable opportunities to go beyond the so called ‘implementation problem’. It allows for examining actual patterns of action that emerge – not just the proposed patterns reflected in artefacts. 3.2.2 Routine Interactions Since the recognition of internal routines dynamics, scholars have attempted to unravel internal routine dynamics to analyse how routines evolve over time. Though the basic idea that routines occur in ‘bundles’ has been recognized for many years (e.g. Nelson and Winter, 1982). This idea indicates the need to consider the multiplicity of routines. However, “we have studied stability and change in individual routines, but there has been less focus on how routines affect one another and how they work together to support stability and change” (Feldman et al., 2016,p. 509, emphasis added). To move beyond a consideration of ‘isolated’ routines, I explicitly conceptualize routines as practices that are shaped by interactions with other routines through their continuous performance (Schatzki, 2011). To understand such generative ‘bundles of routines’ I thus need to analytically trace the connections between the routines. It is therefore necessary to consider how different actors are involved in the performance of various routines, so I can identify what roles the different actors play in creating, maintaining or modifying routine connections (Nicolini, 2013). I therefore expand the analytical framework that was initially developed by Feldman and Pentland (2003), by including an analytical focus on routine interactions (figure 1, 2). In chapter 4 I will consider the methodological implications of this approach. I consider a focus on routine interactions especially relevant for a study in a high-risk environment like surgical care for two main reasons. First of all, very little is known about the interaction of routines, especially those in high-complex professional domains. Most studies conducted on checklist use - that are, as argued earlier, conducted from a health care management perspective - analysed 3