Politics and Patient Safety Don't Mix: Understanding the Failure of Large-Scale Software Procurement for Healthcare Systems

President Obama has recently announced an additional $50 billion to support the development of healthcare informatics and electronic patient records systems. Public attention has, therefore, focused on ensuring that such investments do not suffer from the failures that have afflicted previous large-scale software procurements. This paper analyses a number of recent failures that have affected applications used by the Veterans Health Administration (VHA). The following sections describe the technical causes for these problems; tracing them back into the software engineering practices and project management techniques that have been used within the Department of Veterans Affairs (VA). However, the key argument in the paper is that these causes have been obscured by political arguments. In 2005, the VA centralized responsibility for IT infrastructure under a Chief Information Officer. Before this, each center had considerable autonomy in managing their IT budget and in organizing software development projects. We argue that attitudes to recent software failures are determined by individual perspectives on the political and organizational consequences of this reorganization. Some stakeholders have argued that recent software failures are due to the introduction of standardized development practices and centralized management techniques that ignore the end-user. They maintain that the changes of 2005 have obscured the clinical focus that is seen as essential for the successful development of healthcare informatics. In contrast, other agencies have argued that previous failures, which threatened patient safety, were often under-reported in the decentralized approach. They argue that standardization is essential for progress towards an integrated healthcare system and for ensuring accountability when failures do occur. We argue that these two perspectives have become so entrenched that they underestimate the complexity of software engineering and obscure the additional pressures that political uncertainty places on the technical development of healthcare information systems. It is concluded that by mixing politics and patient safety, there is a danger that we will waste the opportunities provided by new investments in healthcare informatics.