Designing Digital Diagnostics
Abstract Since the 1990s Western clinical radiology has been confronted with a fundamental media-induced change - the so-called analogue-digital migration. Film-based diagnostics and archiving of radiological images are transformed into digital interfaces and infrastructures. Networked software applications, namely picture archiving and communication systems (PACS), provide a new basis for processing and displaying image data. The design and implementation of PACS and their (user) interfaces challenged, amongst others, the search for data standards for digital diagnostics. The data format DICOM (Digital Imaging and Communication) was developed to provide the technological basis for encoding image data. Simultaneously, DICOM determines how patients’ bodies are rendered machine-readable and how radiologists are able to gain software-based insights. A main function of DICOM metadata is encoding and continuously actualising patient identification for technological and human actors. A misidentification of image data and specific patient could lead to fatal errors in the furthe+r treatment process. Accordingly, metadata themselves meander between being invisible to the human user and being essential and hence necessarily visible information for diagnostics. Shifting between normativity and fluidity, DICOM metadata enables new practices of radiological diagnostics, which literally bear vital consequences for patients and, on another level, for the profession of radiology. The paper analyses inherent politics and tensions of metadata from a media theoretical point of view by employing the case of the DICOM standard. Based on subject-specific discourses, data models as well as an in-depth examination of exemplary DICOM metadata it shows how (meta)data politics redefine diagnostic infrastructures and routines as well as gain impact on epistemic and aesthetic practices at the turn of the analogue-digital migration.