Disrupted routines: Team learning and new technology implementation in hospitals

Administrative Science Quarterly,  Dec, 2001  by Amy C. Edmondson,  Richard M. Bohmer,  Gary P. Pisano

• E-mail
• Print
• Link

In developing a process model for how to implement new technological routines, we have emphasized the role of implementation leaders and built on the observation that technological frames shape the way technology is used (Orlikowski, 1993). Our data shed light on the process through which such frames arise and are communicated in an organization through the efforts of implementation teams. We found that team leaders play a critical role in communicating and reinforcing a particular technological frame, which affects how others think about a new technology and the nature of the challenge it presents. This in turn may give rise to self-reinforcing processes in which use of the same technology process is alternatively seen as drudgery and pain or as opportunity and privilege.

The process model that emerged from these data is, on the one hand, mundane: (1) carefully select a team, (2) practice and communicate, (3) work to encourage communication while experimenting with new behaviors in trials, and (4) take time to reflect collectively on how trials are going so that appropriate changes can be made. This process has much in common with long-standing descriptions of the learning process (e.g., Kolb, 1984) and the quality improvement process (e.g., Hackman and Wageman, 1995). On the other hand, although individual learners have been shown to follow such iterative practices instinctively (Schon, 1983), teams are less likely to do so. Organizational and group factors often conspire to preclude interpersonal learning (Argyris, 1982) and team learning (Edmondson, 1999), especially when teams are multidisciplinary (Dougherty, 1992). Moreover, these simple practices were seen as radical in the context in which we found them. Encouraging low-status OR team members to speak up and challenge high-status surgeons went against the grain of the cultural and structural context of cardiac surgery. This context and its traditions are neither arbitrary nor irresponsibly harsh but, instead, reflect a well-established process that functions effectively. Surgeons have years of specialized training, are medically and legally responsible for patients' care, and conventional surgical technology allows them the highest quality, most direct access to data on a patient's well-being in the OR. The kind of top-down, one-way communication that was problematic in learning MICS can be essential to saving lives in critical moments during conventional cardiac surgery.

Our process model attempts to explain how new routines were implemented in this particular context, and it suggests steps for designing an implementation effort. A range of theories of organizational learning describe adaptive processes that occur naturally, generally not in optimal ways, such as trial and error, selection and retention, and diminishing openness to alternatives (Levitt and March, 1988). In contrast, we propose an iterative learning cycle that must be actively managed by local leaders; in that sense, it is a teleological process model, in which the implementation team acts in ways that are purposeful and adaptive (Van de Ven, 1992). Our findings thus plant the seeds of theory that is as much normative as descriptive (Argyris, 1996).