A Quality Assurance Approach to Healthcare: Implications for Information Systems

Abstract

Despite decades of research, healthcare information systems have been characterised by cost over-runs, poor specifications and lack of user uptake. A new approach is required which provides organisations with a reason to invest in this type of software. W Edwards Deming argues that quality is not an entity but derives from using feedback, iteratively to seek improvement to processes, in order to increase productivity and to make better use of resources. The authors propose that supporting this form of quality assurance (QA) using information systems (IS) has the potential to deliver a return on investment. An object-oriented analysis, where healthcare is viewed as the delivery of interdependent processes to which Deming’s form of QA is applied, results in a class model of data types that has some useful characteristics. It is able to store data about medical and nonmedical events; to save descriptions of procedures and to represent the QA process itself. With software based on the model, organisations will have a memory of previous attempts at making improvements as well as data about feedback from patients and staff to drive future change. A critical research in information systems (CRIS) analysis of this model proposes a number of criticisms deriving from theories about rationality; concepts of technology; politics and hidden agendas, as well as the social consequences of technology. The view that QA is a standardised, ongoing conversation about the important characteristics of a process pre-empts many of these counter arguments. The CRIS critique also highlights the need to ensure that development is in harmony with the needs of the many stakeholders in healthcare IS. These concepts lead to new directions in healthcare IS research. The class model needs to be tested against clinical and non-clinical use-cases for its viability not only as support for QA but also as an electronic patient record. A standard terminology is required for processes and for how objects from the model should be used to represent them. The model predicts that user interfaces will have to collect more detailed data than hitherto. Also use of the software should be tested in controlled trials to demonstrate whether the required improvements in quality not only benefit the patient but also the organisations managing their care.