Modeling and Analysis of Clinical Care for Health Information Technology Improvement

Electronic health records (EHRs) and health information technology (IT) applications have great potential to reduce healthcare costs while increasing efficiency and quality. However, there is resistance to EHR adoption due to high startup costs, unpredictable benefits, and disruption to the workflow of clinical care. Two fundamental steps to realize the full potential of health IT are to identify the optimal role of computing in clinical care and to design systems to improve care in predictable ways.

The project developed new techniques for model-based analysis and design of health IT systems using MATH, a method that captures workflow improvements and connects them to health IT software design. MATHflow is the modeling tool that supports concurrent engineering of workflow and health IT, and represents clinician workflow tasks and the information resources required to support them. The project also aimed to demonstrate how these techniques can be applied to design systems that are highly usable and predictably beneficial to healthcare in outpatient clinics treating multiple sclerosis (MS) and chronic pain, and for an emergency department (ED).

The specific aims of the project were as follows:

• Apply ethnographic research to discover the way care is actually done with existing information resources at each clinic. 
• Model research findings as baseline workflows using the new modeling techniques of MATHflow. 
• Analyze how care should be improved methodically in measurable ways with new health IT design concepts. 
• Develop software prototypes of the most promising designs to evaluate usability, workflow impact, and technical feasibility. 
• Refine MATH, a model-based design method, based on experience with each successive clinic. 

The project modeled and analyzed clinical care workflows, networks, and decision making at three sites: The Veterans Affairs Puget Sound Multiple Sclerosis Regional Program, the University of Washington Medicine Center for Pain Relief, and the Baylor University Medical Center ED. Researchers captured and documented baseline clinical workflows using MATH, and then analyzed how health IT could improve workflow in measurable ways using the MATHflow modeling tool. Five health IT system prototypes were modeled, designed, and evaluated to address the needs of: 1) case management for MS, 2) procedure ordering for pain care, 3) patient data collection and analysis for pain care, 4) room turn-around measurement for ED care, and 5) ED decision support for admissions.

Study results found that that new workflow modeling tools such as MATH and MATHflow can explicitly represent both the clinical workflow and the needed improvements in a single, integrated model. These new techniques for model-based analysis and design have promise to reduce unnecessary health IT variation and correct the gap between health IT and the needs of clinical care. Closer adherence to optimal procedures and workflows can “build in” care quality to health IT systems.