Emergency physicians are very good at pattern recognition. Whether the skill leads us to our chosen profession or the profession hones that skill is unclear. Certainly we are more practiced at it. The average emergency physician will hear between 2,000 and 3,000 chief complaints this year, and at the end of a three- to four-hour sojourn, will have a diagnosis, disposition, and treatment plan for each of them. We have become very adept at putting together the puzzle pieces of the patient's subjective complaints, physical findings, and lab and radiology data to form a recognizable entity. We become rapid synthesizers of data. No other medical specialty has that experience!
Richard Bukata, MD, the chairman of ACEP's Benchmarking Task Force and an EMN columnist, likes to say that emergency physicians should be pure, roving intellects. He observes that physicians frequently are bogged down with busy work, clerical tasks, and data-chasing. In the perfect ED, the physician would be free of most of those tasks to roam the department, interacting with patients, supervising staff, and keeping busy with cerebral endeavors, using pattern recognition skills to the utmost in an unfettered environment. Can information technology take us to such a utopian practice environment?
As I visit emergency departments around the country, speaking and educating physicians about quality improvement, I hear an overwhelming sense of frustration with existing information technology in emergency medicine. It is useful to realize that emergency practitioners are not the only ones feeling disappointed by unrealized operational gains though information technology. According to Todd Warden, MD, an emergency physician and president of Emergenuity, a health information management firm, American businesses went on a historic technology binge between 1998 and 2000, spending $6.5 billion.
CEOs and CFOs then felt bewildered and burned by money lost. The industry secret is out: Most applications purchased over the past decade have been disappointing. They have missed projections on return on investment, ease of use, and ability to simplify processes. At times, they have made situations worse by adding more levels of complexity. Like companies, hospitals have been caught in a cycle of trying to get software to fix the software. Dr. Warden notes, “A better answer lies in a new approach of integrating and coordinating existing technology at the process level and putting the power of that coordination in the hands of those closest to the clinical issue.” (Business Integration, February 2005.)
BPM and EDA
One reason to be optimistic about ED information technology lies in a basic change in health care information management. Emergency health care has complex transactions occurring simultaneously on many different proprietary systems, and early systems used financial software to manage clinical data! Business process management (BPM) technology is at the forefront of an emerging trend toward event-driven architecture (EDA) that holds great promise for the challenges of today's clinical information management in the emergency department.
BPM technology offers a lightweight, highly flexible way to turn existing applications and systems into real-time, event-driven infrastructure. BPM technology generally can be installed in a third of the time and at a fifth of the cost of traditional methods because it effectively knits together and integrates existing systems. Purchasing new applications is not necessary, and the emphasis is on letting clinical leaders use the data being created to drive processes in real time. In the best designs, it can be monitored and altered by managers without IT intervention. (DM Review, Feb. 24, 2004; Event-Driven Architecture, www.webmethods.com. Accessed Oct. 13, 2005.)
BPM provides built-in integration capability that spans the heterogeneous environment of today's complex emergency department, manages multiple processes spanning physical geographies, ensures enterprise level performance, supports up to hundreds of simultaneous users with a configurable run-time architecture, and streamlines organizational efficiency.
Event-driven architecture uses unidirectional messaging to communicate among two or more application processes. (Computerworld, May 12, 2003; “ED Dashboard Indicators: Process Improvement in Real Time.” Welch SJ, presented at Urgent Matters Conference, Las Vegas, Oct. 27, 2005.) The communication is initiated by an event trigger that typically corresponds to some occurrence, receipt of an order, for example. In medicine, this would be used for ED events such as an abnormal lab value or completion of a test. Such data accrual in real time can be used in real-time process improvement. EDA uses event triggers, which prompt processes to occur or change in real time.
All of this is occurring at various institutions around the country. In particular, some homegrown tracking systems have used BPM in ED information systems to integrate all relevant data systems at an institution. These systems are informed by frontline physicians, and are beginning to show the benefits that can be reaped with such sophisticated technology.
EDA also is being used to employ real-time data triggers. At our institution, we developed dashboard indicators, which take the individual's tracking system data and combine it for an aggregate view of how overall ED operations are performing in real time. All islands of ED relevant information have been integrated and the passive transfer of data occurs automatically. (Figure 1.)
Process improvements can be enacted in real time as backlogs and delays occur. These dashboards indicate when there are operational problems at any of the seven processes or operations a patient may pass through during an ED visit: triage, door-to-physician, lab, radiology, nursing intervention, discharge, or admission. The universally recognized red, yellow, and green traffic light indicates the status of each of those operations for the aggregate of patients in the department in real time. (Figure 2.)
These dashboards alert ED staff when a particular process or operation is becoming inefficient. For each operation, we track backlogs (patients lining up for that operation) and delays (when time expectations are not met).
We currently are setting up VHA operational benchmarks, and are working on a theoretical model to predict exactly when an operation is becoming inefficient. Armed with this type of information, we can adjust our processes to remedy the backlog or delay in a specific operation and do so in real time as operational problems occur. Some examples of real-time process improvement:
▪ When the dashboards show a queue in triage, we set up a secondary mobile triage to get patients quickly into the system.
▪ When we have a long line of patients waiting for plain x-rays, or the simple x-ray is taking more than an hour to turn around, we institute an alternative procedure where in-house radiology technicians move patients upstairs and shoot some of the films up there until we are caught up. This is dubbed secondary radiology.
▪ When the dashboards show a queue of discharges or that it is taking more than 30 minutes from discharge order to actual discharge, we create an ad hoc discharge team.
▪ The point is that process improvements can be performed in real time with this dashboard technology. (“ED Dashboard Indicators: Process Improvement in Real Time.” Welch SJ, presented at Urgent Matters Conference, Las Vegas, Oct. 27, 2005.) The old paradigm of looking retrospectively at QI data for process improvement has just been turned on its head. We also are incorporating sophisticated cueing and trigger functions to alert caregivers to clinical data that are critical or operational problems that are imminent. By keeping the development of these systems close to the clinicians, many problems with functionality have been avoided.
Todd Taylor, MD, has written a comprehensive article that is helpful to the emergency physician looking to advance IT in his department. (Emerg Med Clin NA 2004;22:241.) He provides a blueprint for the analysis and decision-making necessary for implementing new information technology in the ED. IT planning for the ED is complex and relatively new to emergency medicine despite the fact that other industries have been doing it for many years. It has been estimated that less than 15 percent of EDs have a comprehensive information system in place. ED automation is challenging and often fraught with pitfalls. Active physician and nurse involvement is essential in the process if the new system is to be accepted at the user level.
Dr. Taylor makes some interesting observations regarding return-on-investment expectations for information technology in the emergency department: EDIS and IT computers rarely save time and money. What they do is allow physicians to do things previously not possible. Often these tasks take more time and money so when considering return on investment, it is staff efficiency, ability to gather data for management, patient safety, improved data distribution and access, and work flow automation that are the real returns on investment.
Information technology can support the varied tasks of the emergency physician. Though the past decade has been rife with false starts and unfulfilled promises, frontline practitioners ought not despair. Recent breakthroughs in event-driven architecture and BPM are in the wings to make good on the bargain. They have given rise to such innovations as ED dashboards and real-time process improvement. The day is not far off when sophisticated information technology tools will support our work, enabling us to be that pure, roving intellect busy with pattern recognition and cerebral endeavors.