Page - 90 - in Applied Interdisciplinary Theory in Health Informatics - Knowledge Base for Practitioners

1.3. The contested nature of Actor-Network Theory and its limitations ANT is constantly evolving as it gets interpreted and re-interpreted by different scholars, which makes it somewhat hard to define its nature at any point in time [7,10]. It also has several limitations, and these have been extensively debated in the literature. The notion of non-human actors and their ability to possess agency is a particularly hot topic, with some doubting the contribution of this notion.[13-16] Whilst it is beyond the scope of this chapter to discuss all of ANTs shortcomings in detail, the most pertinent ones in relation to health technology implementation are discussed below. Most importantly, it has been argued that ANT is not much of a theory at all as it lacks predictive power.[17,18] Predictive power is the ability of a theory to prospectively predict a phenomenon under investigation. In relation to health technology, this may for instance include postulations about how certain design features can lead to certain workarounds of users. According to Wacker, for a theory to be “good” it needs to have internal consistency and empirical riskiness - these are areas that ANT does not fare very well in.[18] Internal consistency refers to a theory providing logical and adequate explanations of reality. However, despite providing a vocabulary to describe social phenomena, ANT lacks the ability to explain and integrate the relationships between various human and non-human actors. As a result, ANT accounts can describe how clinical users and technology are related but may leave the reader questioning the actual contribution of applying the lens. Empirical riskiness encompasses the need for a “good” theory to be both risky and testable – but ANT cannot really be tested and lacks specificity.[18] Its terminology (see Box 2) is extremely loosely defined and its networks are potentially limitless, which can result in a lack of focus.[19,20] Consequently, if a theory cannot be subject to prospective tests, it may have limited usefulness. Other criticisms have included the following: [10,21] • ANT’s perceived inappropriate equal treatment of both human (e.g. clinical users) and non-human actors (e.g. technology, paper); • the inability of the human observer/researcher to be fully agnostic (as postulated by ANT); and • the lack of attention to the role of macro structures (e.g. economic, political environments) in influencing micro contexts (e.g. work practices, usability). Nevertheless, ANT can help to facilitate interpretations of the researcher and provides a helpful vocabulary that can be used to explore a view of a world consisting of networks. This view of the world has several advantages when exploring the implementation of health IT. As such, ANT may be most appropriately viewed as a tool for theory development or a methodological approach in evaluating technology implementations. 2. Using Actor-Network Theory to evaluate health information technology ANT has been employed by several medical sociologists to explore how artifacts and technologies can shape social processes in healthcare settings. In doing so, it has been applied rather pragmatically as a lens to examine specific aspects of technology K.Cresswell /UsingActor-NetworkTheory toStudyHealth InformationTechnology Interventions90