Fuad Abujarad

Symbolic synthesis of masking fault-tolerant distributed programs

We focus on automated addition of masking fault-tolerance to existing fault-intolerant distributed programs. Intuitively, a program is masking fault-tolerant, if it satisfies its safety and liveness specifications in the absence and presence of faults. Masking fault-tolerance is highly desirable in distributed programs, as the structure of such programs are fairly complex and they are often subject to various types of faults. However, the problem of synthesizing masking fault-tolerant distributed programs from their fault-intolerant version is NP-complete in the size of the program’s state space, setting the practicality of the synthesis problem in doubt. In this paper, we show that in spite of the high worst-case complexity, synthesizing moderate-sized masking distributed programs is feasible in practice. In particular, we present and implement a BDD-based synthesis heuristic for adding masking fault-tolerance to existing fault-intolerant distributed programs automatically. Our experiments validate the efficiency and effectiveness of our algorithm in the sense that synthesis is possible in reasonable amount of time and memory. We also identify several bottlenecks in synthesis of distributed programs depending upon the structure of the program at hand. We conclude that unlike verification, in program synthesis, the most challenging barrier is not the state explosion problem by itself, but the time complexity of the decision procedures.

Multicore Constraint-Based Automated Stabilization

Given the non-determinism and race conditions in distributed programs, the ability to provide assurance about them is crucial. Our work focuses on incremental synthesis where we modify a distributed programs to add self-stabilization. We concentrate on reducing the time complexity of such synthesis using parallelism. We apply these techniques in the context of constraint satisfaction. In particular, incremental synthesis of self-stabilizing programs requires adding recovery actions to satisfy the constraint that are true in the legitimate states. We consider two approaches to speedup the synthesis algorithm: first, the use of the multiple constraints that have to be satisfied during synthesis; second, the use of the distributed nature of the programs being synthesized. We show that our approaches provide significant reductions in the synthesis time.

Automated constraint-based addition of nonmasking and stabilizing fault-tolerance☆

We focus on the constraint-based automated addition of nonmasking and stabilizing fault-tolerance to hierarchical programs. We specify legitimate states of the program in terms of constraints that should be satisfied in those states. To deal with faults that may violate these constraints, we add recovery actions while ensuring interference freedom among the recovery actions added for satisfying different constraints. Since the constraint-based manual design of fault-tolerance is well known, we expect our approach to have a significant benefit in automating the addition of fault-tolerance. We illustrate our algorithm with four case studies: stabilizing mutual exclusion, stabilizing diffusing computation, a data dissemination problem in sensor networks, and tree maintenance. With experimental results, we show that the complexity of our algorithm is reasonable and that it can be reduced using the structure of the hierarchical systems. We also reduced the time complexity of the synthesis using parallelism. We consider two approaches to speedup the synthesis algorithm: first, the use of the multiple constraints that have to be satisfied during synthesis; second, the use of the distributed nature of the programs being synthesized. We show that our approaches provide significant reduction in the synthesis time. To our knowledge, this is the first instance where automated synthesis has been successfully used in synthesizing programs that are correct under fairness assumptions. Moreover, in three of the case studies considered in this paper, the structure of the recovery paths is too complex to permit existing heuristic-based approaches for adding recovery.

Patient-Centered Decision Support: Formative Usability Evaluation of Integrated Clinical Decision Support With a Patient Decision Aid for Minor Head Injury in the Emergency Department

Background The Canadian Computed Tomography (CT) Head Rule, a clinical decision rule designed to safely reduce imaging in minor head injury, has been rigorously validated and implemented, and yet expected decreases in CT were unsuccessful. Recent work has identified empathic care as a key component in decreasing CT overuse. Health information technology can hinder the clinician-patient relationship. Patient-centered decision tools to support the clinician-patient relationship are needed to promote evidence-based decisions. Objective Our objective is to formatively evaluate an electronic tool that not only helps clinicians at the bedside to determine the need for CT use based on the Canadian CT Head Rule but also promotes evidence-based conversations between patients and clinicians regarding patient-specific risk and patients’ specific concerns. Methods User-centered design with practice-based and participatory decision aid development was used to design, develop, and evaluate patient-centered decision support regarding CT use in minor head injury in the emergency department. User experience and user interface (UX/UI) development involved successive iterations with incremental refinement in 4 phases: (1) initial prototype development, (2) usability assessment, (3) field testing, and (4) beta testing. This qualitative approach involved input from patients, emergency care clinicians, health services researchers, designers, and clinical informaticists at every stage. Results The Concussion or Brain Bleed app is the product of 16 successive iterative revisions in accordance with UX/UI industry design standards. This useful and usable final product integrates clinical decision support with a patient decision aid. It promotes shared use by emergency clinicians and patients at the point of care within the emergency department context. This tablet computer app facilitates evidence-based conversations regarding CT in minor head injury. It is adaptable to individual clinician practice styles. The resultant tool includes a patient injury evaluator based on the Canadian CT Head Rule and provides patient specific risks using pictographs with natural frequencies and cues for discussion about patient concerns. Conclusions This tool was designed to align evidence-based practices about CT in minor head injury patients. It establishes trust, empowers active participation, and addresses patient concerns and uncertainty about their condition. We hypothesize that, when implemented, the Concussion or Brain Bleed app will support—not hinder—the clinician-patient relationship, safely reduce CT use, and improve the patient experience of care.

Back to the Bedside: Developing a Bedside Aid for Concussion and Brain Injury Decisions in the Emergency Department.

Context: Current information-rich electronic health record (EHR) interfaces require large, high-resolution screens running on desktop computers. This interface compromises the provider’s already limited time at the bedside by physically separating the patient from the doctor. The case study presented here describes a patient-centered clinical decision support (CDS) design process that aims to bring the physician back to the bedside by integrating a patient decision aid with CDS for shared use by the patient and provider on a touchscreen tablet computer for deciding whether or not to obtain a CT scan for minor head injury in the emergency department, a clinical scenario that could benefit from CDS but has failed previous implementation attempts. Case Description: This case study follows the user-centered design (UCD) approach to build a bedside aid that is useful and usable, and that promotes shared decision-making between patients and their providers using a tablet computer at the bedside. The patient-centered decision support design process focuses on the prototype build using agile software development, but also describes the following: (1) the requirement gathering phase including triangulated qualitative research (focus groups and cognitive task analysis) to understand current challenges, (2) features for patient education, the physician, and shared decision-making, (3) system architecture and technical requirements, and (4) future plans for formative usability testing and field testing. Lessons Learned: We share specific lessons learned and general recommendations from critical insights gained in the patient-centered decision support design process about early stakeholder engagement, EHR integration, external expert feedback, challenges to two users on a single device, project management, and accessibility. Conclusions: Successful implementation of this tool will require seamless integration into the provider’s workflow. This protocol can create an effective interface for shared decision-making and safe resource reduction at the bedside in the austere and dynamic clinical environment of the ED and is generalizable for these purposes in other clinical environments as well.

Rap Backs: Continuous Workforce Monitoring to Improve Patient Safety in Long-Term Care

The Michigan Workforce Background Check (MWBC) system is a Web-based application that centralizes the screening process for job applicants at long-term care facilities by integrating the checking of registries and databases, and by providing secure communication between the system and state agencies. A key feature of the system is the rap back, which is a process whereby fingerprint images submitted by healthcare workers are maintained in a database and compared against arrest fingerprints to track subsequent criminal activity. Rap backs eliminate the need for subsequent fingerprinting to update criminal record information. A user-centered design process was used to create the interaction design during the development process, resulting in very few post-release adjustments and saved time and financial resources. Patient safety is improved due to this rap back process because unfit workers are immediately identified and removed from long-term care facilities.

Improving Patient Safety through User-Centered Healthcare Background Check System Design

The Michigan Workforce Background Check (MWBC) system is a cost-effective, usable web application for conducting employability screenings for healthcare workers in long-term care settings. This system is an exemplar for demonstrating how a user-centered design process facilitated the flow of critical information, improved applicant data accuracy, and improved patient safety by eliminating access by known abusers and criminals to vulnerable adults in long-term care settings. The redesigned system 1) increases the efficiency in the background check process, 2) collects more accurate and comprehensive applicant data, and 3) ultimately improves patient safety by quickly identifying unfit workers.

Real-world user-centered design: the michigan workforce background check system

The Michigan Workforce Background Check system demonstrates how an iterative user-centered design (UCD) process enhances organizationallevel communication practices and efficiency. Well-designed information communication technology is an essential component of effective public health management. Usability and accessibility testing informed subsequent design and development. The iterative improvement in the background check application demonstrates that UCD should be a component of public health management projects in particular, and online project development in general.

Constraint Based Automated Synthesis of Nonmasking and Stabilizing Fault-Tolerance

We focus on constraint-based automated addition of nonmasking and stabilizing fault-tolerance to hierarchical programs. We specify legitimate states of the program in terms of constraints that should be satisfied in those states. To deal with faults that may violate these constraints, we add recovery actions while ensuring interference freedom among the recovery actions added for satisfying different constraints. Since the constraint-based{\em manual} design of fault-tolerance is well-known to be applicable in the manual design of nonmasking fault-tolerance, we expect our approach to have a significant benefit in automation of fault-tolerant programs. We illustrate our algorithms with three case studies:stabilizing mutual exclusion, stabilizing diffusing computation, and a data dissemination problem in sensor networks. With experimental results,we show that the complexity of synthesis is reasonable and that it can be reduced using the {\em structure} of the hierarchical systems. To our knowledge, this is the first instance where automated synthesis has been successfully used in synthesizing programs that are correct under fairness assumptions. Moreover, in two of the case studies considered in this paper, the structure of the recovery paths is too complex to permit existing heuristic based approaches for adding recovery.

Weakest Invariant Generation for Automated Addition of Fault-Tolerance

One of the important tasks in evolving a fault-intolerant program into a fault-tolerant one is to identify the legitimate states (its invariant) from where the original program satisfies its specification. This allows us to ensure that the fault-tolerant program recovers to these legitimate states from where it satisfies its specification. It is desired that the invariant be the weakest possible so as to provide maximal options to the algorithm for adding fault-tolerance. Requiring the designer to specify such weak invariant increases the burden on the designer. In this paper, we present a novel approach for automating the generation of the weakest invariant from the program actions and specifications. Our algorithm is efficient and manages the state explosion problem with the use of BDDs. We demonstrate our approach through two case studies and we show that generating such invariants is extremely fast. For example, for a mutual exclusion program with a state space of size 10^8^4 states, it took 0.9 of a second.