Executable Medical Best Practice Guidance Systems
In 1999, the challenge of preventable medical errors was first raised in a landmark study by the American Medical Association: To Err is Human. This study reported that “at least 44,000 people, and perhaps as many as 98,000 people, die in hospitals each year as a result of medical errors that could have been prevented.” In 2013 Journal of Patient Safety reported that more than 400,000 people die every year because of preventable medical errors. The negative economic impact was estimated at a colossal $1 trillion per year in a US Senate hearing in 2014.
A tenfold increase of preventable medical errors in over a decade tells us that the clinical practice needs a fundamental change. Preventable medical errors are caused by unintended/unjustified deviation from medical best practices. In spite of medical community’s continued effort, the complexity of medicine and the clinical workflow have overwhelmed the current manual process of healthcare. Think of it this way, giving any map that you need, can you always drive flawlessly in any big city at a moment’s notice? Medical knowledge is way more complex than information presented in maps.
A paradigm shift is needed. Like how GPS navigation transformed how we drive, we need Executable Medical Best Practice Guidance System to transform the practice of medicine.
Preventable medical errors are not a medical knowledge problem. It is a medical cyber-physical system challenge in the form of medical information and workflow management challenge, where the medical devices, doctors, nurses, and technicians have to work together flawlessly in real-time. Currently, we are focusing on two projects: i) cardiac arrest resuscitation, where there is no room for errors; 2) sepsis. The CDC estimates that sepsis accounts for as many as half of all hospital deaths, more than cardiovascular disease. What has made sepsis particularly difficult to treat and resolve has been the complex interaction of the infection and the body’s immune and defense response. There are multiple interactions between laboratory values, clinical data and appropriate treatment regimens that need a constant attention for as long as 24 hours.
Funded by NSF, we are working with hospitals to develop medical best practice guidance systems.
- Scientific Impact: Changing the representation of medical knowledge from informal natural language texts to executable formal representations in the form of networked organ automaton and best practice automaton. The precise and repeatable nature of formal representation removes errors from subjective interpretations and memorization problems.
- Clinical Impact: Executable Best-practice Medical Guidance systems will fundamentally transform the implementation of medical best practices, with the rapid and consistent timing of medical interventions, prevent unintended deviation from standardized medical treatment guidelines, improved team situation awareness and precise record keeping.
- Cardiac Arrest Resuscitation Guidance Systems in cooperation with Dr. Karen White, Director of ICU at Carle Foundation Hospital. This system has approved for phase 1 of clinical trials at Carle’s ICU. We are now working with Dr. Priti Jani of University of Chicago Pediatric ICU to extend it for pediatric resuscitation.
- Sepsis Best Practice Guidance System for adults in cooperation with Dr. Karen White, Director of ICU at Carle Foundation Hospital. This system has entered the late stage of design review. Also, we are working with Dr. Richard Pearl and Dr. Jonathan Gehlbach of OSF Illinois Children Hospital to extend the system for pediatric sepsis.
- Heart Transplant Perioperative Guidance System with Dr. Jai Raman at Oregon Health and Science University. This system is still in the early R&D stage.
- Courses development for our Nation’s first integrated medical and engineering school: Carle–UIUC Medical School. We are working with Dean King Li. MD, to develop interactive teaching and training based on medical best practice for complex diseases.
Low Complexity Real-Time Virtual Computer
National Academy of Science’s Committee on Certifiably Dependable Systems wrote, “One key to achieving dependability at a reasonable cost is a serious and sustained commitment to simplicity, including simplicity of critical functions and simplicity in system interactions. This commitment is often the mark of true expertise. There is no alternative to simplicity. Advances in technology or development methods will not make simplicity redundant; on the contrary, they will give it greater leverage”.
Historically, a landmark in the distributed systems is the invention of the atomic transaction technology, which allows software engineers to program a distributed system if each transaction were run alone, even though transactions are running concurrently and sharing distributed data. The reduction of the complexity of distributed transaction processing is truly remarkable.
Funded in part by ONR, NSF, Lockheed Martin and Rockwell Collins, we are well on our way to the creation of a low complexity real-time virtual computer, which allows software engineers to program a networked real-time multicore computer system as if it were a large single core computer.
Scientific Merit: The foundation for real-time computing assumes that 1) the worst-case execution time when a task runs alone is the same when it is running concurrently with other tasks, and 2) distributed nodes have consistent views and perform consistent actions every period. The first assumption was violated when a multicore computer is used as is. The second assumption is violated by the fact that each node has its own clock. Clocks skews can be bounded but not eliminated. Letting the tasks at each node direct driven by their local clocks creates the potential of distributed race conditions that will lead to inconsistency in distributed views and actions from time to time. Yet it is impractical/inefficient/unreliable to use a global physical clock to drive distributed computation. We are creating the architecture and protocols that solve these touch challenges elegantly and efficiently.
Broader Impact: A notable invention of our team in this area is the invention of Physically Asynchronous Logically Synchronous (PALS) architecture. Rockwell Collins Inc demonstrated in their lab that using PALS, the verification time using model checking time for a dual redundant flight control system has reduced from 35 hours to less than 30 seconds, winning the 2009 David Lubkowski Memorial Award for the Advancement of Digital Avionics from American Institute of Aeronautics and Astronautics (AIAA).
Running on top of PALS middleware, engineers can design, verify and run a networked (single core) real-time control system as if it were a single core computer at the fastest possible speed permitted by the platform.
Our team has since invented Single Core Equivalence technology that allows engineers to use each core in a multicore computer as if it were a single core computer. This work has already gained support from academics, industry, and governments across US, Europe, and Asia. And we are in the process of creating the Virtual Single Core computer that will allow engineers to use a set of cores as if it were a higher schedulability single core computer.
Our goal is the create the Real-Time Virtual Computer that allows engineers to use a networked multicore control system as if it were a single core computer. The success of this project will impact the development of real-time systems ranging from avionics, automobiles, IoT to smart cities.
Lui Sha graduated with Ph.D. from CMU in 1985. He worked at the Software Engineering Institute from1986 to1998. He joined UIUC in 1998. Currently, he is Donald B. Gillies Chair Professor of Computer Science, the University of Illinois at Urbana-Champaign.
He led the research on real-time computing theory, which was cited as a major accomplishment in the selected accomplishment section of the 1992 National Academy of Science’s report, “A Broader Agenda for Computer Science and Engineering” (P.193). He led a comprehensive revision of IEEE standards on real-time computing to support the application of this work, which has since become the best practice in real-time computing systems. Later, he led the development of Complexity Reduction and Control architectures for dependable real-time systems, including Simplex architecture and Physically Asynchronous Logically Synchronous architecture.
He is a widely cited author in real-time and embedded computing community. His work has impacted many large-scale high technology programs including GPS, Space Station, and Mars Pathfinder. Now it is widely used in system real-time constraints such as airplanes, robots, cars, ships, trains, medical devices, power generation plants and manufacturing plants.
In recent years, his team is developing the technologies for secure and certifiable multi-core avionics with aviation community and computational pathophysiology models for medical best practice guidance (“GPS”) systems with the medical community.
His Research impacts on National high technology projects include:
- Global Positioning Satellite: Contributions to the worldwide navigation. “The navigation payload software for the next block of Global Positioning System upgrade recently completed testing. … This design would have been difficult or impossible prior to the development of rate monotonic theory”, L. Doyle, and J. Elzey “Successful Use of Rate Monotonic Theory on A Formidable Real-Time System, technical report, p.1, ITT, Aerospace Communication Division, 1993.
- International Space Station: “Through the development of Rate Monotonic Scheduling, we now have a system that will allow [Space Station] Freedom’s computers to budget their time, to choose between a variety of tasks, and decide not only which one to do first but how much time to spend in the process”, Aaron Cohen, Deputy Administrator of NASA, October 1992 (p. 3), Charting The Future: Challenges and Promises Ahead of Space Exploration.
- Mars Pathfinder: “When was the last time you saw a room of people cheer a group of computer science theorists for their significant practical contribution to advancing human knowledge? 🙂 It was quite a moment. … For the record, the paper was L. Sha, R. Rajkumar, and J. P. Lehoczky. Priority Inheritance Protocols: An Approach to Real-Time Synchronization. In IEEE Transactions on Computers, vol. 39, pp. 1175-1185, Sep. 1990.” reported by Dr. Michael Jones in http://catless.ncl.ac.uk/Risks/ 19.49.html
Honors and Awards
- He is a co-recipient of IEEE Simon Ramo Medal, “for technical leadership and contributions to fundamental theory, practice, and standardization for engineering of real-time systems”, 2016.
- He has been appointed by NASA Administrator Bolden to NASA Advisory Council’s Aeronautics Committee, providing review and advice on NASA’s research programs, 2015 to 2017.
- He was a member of National Academy of Science’s Committee on Certifiably Dependable Software, 2005 to 2007
- He was a member of National Foundation of Science’s CPS Initiative planning group that formulated and launched the CPS program.
- He is a co-recipient of David Lubkowski Award “for the Advancement of Digital Avionics, 2009.”
- He is a fellow of the ACM “for contributions to real-time systems”, 2005.
- He was an IEEE Distinguished Visitor, 2005 – 2007.
- He was awarded for Outstanding Technical Contributions and Leadership in Real-Time Systems, IEEE Technical Committee on Real-Time Systems, Dec. 2001.
- He was Chair of IEEE Real-Time Systems Technical Committee from 1999-2000.
- He is a Fellow of the IEEE, “for technical leadership and research contributions which enabled the transformation of real-time computing practice from an ad hoc process to an engineering process based on analytic methods.” 1998
- Teachers Ranked as Excellent by Their Students, UIUC, 1999 and 2000
- GE Scholar, the Academy for Excellence in Engineering Education, UIUC, 1999.
- For young researchers: Elements of Successful Research and How to Write Research Papers
- Ph.D. and Postdoc positions
- Part-time jobs in the lab: software development and/or hardware device interfaces