Article written as a part of the Masters program in Public Health at Uppsala University, Fall 2016
Abstract
Clinical Decision Support Systems (CDSS) can be used to predict patient outcomes, identify likely diagnoses and allocate healthcare resources. The literature investigating the ethical aspects of rule-based CDSS has identified issues relating to standards of care, the appropriate use and users of such systems, and their impact on patient-provider relationships. This paper extends this discussion to machine-learning based systems with particular attention given to their effects on distributive justice. Frameworks for examining these issues based on consequentialist, procedural and deontologoical normative theories of justice are proposed and discussed.