Software Engineering research is one of the enablers in societal grand challenges such as Health Care, and Software Engineering researchers need to be involved to attain long-term workable solutions. For example, current efforts to make health data available electronically on a national scale could benefit from research in software architectures and software engineering economics to avoid locking into systems with inherently limited capabilities and evolvability. This project will conduct a case study of health record systems, in collaboration with the School of Medicine, leading to architectural specifications that will be the basis for prototypes and a test bed for interoperable Health Care applications. This will enable software researchers to understand the costs and risks of architectural design decisions in building the Health Care systems of the future. Demonstrations and evaluation will be built on the Eclipse-based IBM Jazz platform. The project will explore what role that the large body of Software Engineering research can play in this societal grand challenge.
The problem this project addressed is that the current state of knowledge in the fields of in software and systems engineering, while vital, is inadequate to meet the needs of or societal-scale information systems. Examples of such systems include air transportation control, national-scale smart grid electrical power systems, and national--scale health information systems that could monitor disease outbreaks and biological attacks, enable precision healthcare for individuals, and inform research on the safety and effectiveness of clinical, pharmacological, and other health-related practices and interventions. The traditional approach of software and systems engineering is based on rigorous centralized control of system design, implementation, deployment, and operation to ensure compliance with pre-determined system requirements. Yet at societal scale, centralized control of system design and operation is neither possible or desirable. For example, no single entity does, can, or should control the thousands of health care facilities from which clinical data might be drawn for privacy-preserving analytics at national scale. An analogy is helpful. If the kinds of systems addressed by traditional methods are like buildings, which can be designed by traditional engineering methods, then societal-scale systems are like cities. They are planned, and their infrastructures, services, incentive systems, and governance mechanisms generally are engineered, but the city as a whole evolves organically. A city, supported by engineered systems, as a whole admits planning and support but not control. We are now entering an era of societal-scale, information- and computation-driven systems : from smart cities to smart grids to learning healthcare systems. In this context, researchers in the field of computer and information science and systems engineering face two broad challenges: how to better design key system components that can be engineered in traditional ways, and how to better plan and shape those aspects that are inherently not centrally controlled so that what emerges will nevertheless meet societal needs and expectations. This project pursued both of these questions, albeit in different ways. It developed and evaluated a novel approach to the problem of software production through software synthesis from formal specifications. Such approaches will, the PI believes, be vitally important to the future production of software that is both affordable and dependable, especially for key architectural components of the cyber-infrastructures that will be needed to support societal-scale information systems. These new methods were developed and tested within several sub-projects, one of which developed and tested a laboratory-scale information architecture and infrastructure system for connecting clinical sources of health information to patient-selected destinations. In addition to these technical activities, the PI has developed extensive interactions with the health informatics research community through both the Institute of Medicine and the Office of the National Coordinator of the Department of Health and Human Services. Societal-scale systems cannot be built by technologists alone. New communities of researchers spanning traditional disciplinary boundaries will be needed. The PI, with the support of the National Science Foundation and this NSF Eager grant has launched an initiative that promises both technical advances and the building of research communities of the future to enable ongoing exploration of the scientific and engineering challenges that must be addressed if we are to succeed in building societal-scale information systems, whether in healthcare, energy, education, or any other field of vital interest to our nation.
