Marlot Kuiper

234 Connective Routines Secondly, ‘the artefact’ does not exist, because artefacts are constantly changing. In the literature, scholars often speak of ‘the artefact’ that models the routine, as objects that once they are created, are out there to model performances (Bredillet, Tywoniak& Tootoonchy, 2018; Cohen et al., 1996; Essén, 2008). The episodic observations allowed me to identify artefacts as dynamic rather than static entities. Throughout the fieldwork, not only were artefacts added, but also, they were amended. For instance by indicating responsibilities or including items in the software module to allow for continuation only when boxes are ticked. Artefacts are thus identified as items that can influence the course of routines, while at the same time they themselves are evolving. A temporal dimension is therefore important when considering artefacts as models for routines. 7.6.2 Creating workable artefacts in performance-oriented con- texts is political Artefacts are introduced to create a routine that increases performance. At first sight, creating and amending artefacts seems quite ‘pragmatic’. Actants are busy creating ‘workable artefacts’. The findings show however, that this seemingly pragmatic exercise actually is highly political. Two elements seem of extra importance here: who has the agency to amend artefacts, and how are they amended. Firstly, it matters who amends artefacts. Following D’Adderio (2011, p. 198-199) who stated that artefacts are “evolving as a consequence of their appropriation by certain agencies in specific contexts”, I delved into the issue of ‘appropriation by certain agencies’. In this study, ‘certain agencies’ mostly referred to the departmental managers (without medical background), often in collaboration with frontrunning professionals that were very enthusiastic about the checklist. Most of the professionals who work with the checklist on a daily basis were not the ones who determined the appropriateness of the checklist. The question on ‘workable artefacts’ therefore particularly became one of: how to design artefacts in such a way that they result in compliant behaviour? Secondly, it matters how artefacts are amended. Clearly indicating responsibilities was seen as one of the solutions to improve compliance, for instance by making changes in the software system to trace back actions to individual professionals. It has been claimed that increased control can be achieved by embedding a rule (in this case a checklist) in a machine or software, since this tends to make them more persistent and durable (D’Adderio 2008, Pentland & Feldman, 2008). The findings of this study indeed point towards reinforced feelings of