Electronic Charting and Process Control
High-Level Description of General Strategy for Process Monitoring & Process Control:
Our goal is to leverage electronic documentation systems to improve anesthesia care by providing the right information to the right person at the time when it will be most useful. Borrowing terminology from other industries, we call this approach 'real-time process monitoring & process control'. The workflow is illustrated in a very generic way in the flow diagram to the right:
At Massachusetts General Hospital, process monitoring and control happen electronically, using software written by the Department of Anesthesia, Critical Care and Pain Medicine (DACCPM). 'Process exception' is a generic term indicating that the monitored process is not progressing according to the process model defined by the user. Process models against which actual process execution can be compared are the key feature. While the process model itself can often be very simple to state ? e.g., "preoperative antibiotic management should be documented prior to incision" or "the patient should go only to their scheduled operating room," the underlying logic and database queries that support process monitoring can be quite complex.
When a process exception is detected, the system generates an alert (the 'Exception!' flag in the figure). These alerts are 'pushed' to relevant stakeholders to provide the information as soon as it is known, with the goal of providing information when it is most beneficial. In other words, the system ideally provides a reminder just before it is needed. The two modalities we have currently used to 'push' alerts are pop-up windows in the anesthesia application, and text pages generated through the hospital paging system
The central notion that drives our implementation of process monitoring and control comes from Deming: when the typical clinician cannot reliably achieve the desired level of performance using the tools, methods and systems provided, then the onus is on the organization to provide the means (through better tools, materials, methods and systems) to improve performance.
Anesthesia Information Management Systems for Process Monitoring & Process Control
The Massachusetts General Hospital Department of Anesthesia, Critical Care and Pain Medicine installed its AIMS in 2002 to improve the quality of documentation and facilitate billing. It was discovered that simply installing the system was not sufficient to improve billing efficiency. Hence the AIMS group developed the anesthesia billing alert system (ABAS). Using the ABAS, the time to correct documentation errors and the residual number of errors that were never corrected could both be reduced by an order of magnitude (essentially to zero). Although not explicitly described in the publication, the ABAS simultaneously monitors adherence to dozens of process models for each case as it is being performed.
In 2006, the DACCPM began participating in 'pay-for-performance' contracts. Performance on potential metrics such as temperature management, perioperative antibiotic administration, perioperative oxygen therapy, and deep vein thrombus prophylaxis could be gauged by reference to clinical documentation. Hence, successful participation depends in part on documentation that is complete and timely.
Accordingly, we set out to test the notion ? and successfully demonstrated ? that meaningful improvement in documentation could be achieved with a single alert during the case, without reminders.
Statistically and functionally significant improvements in documentation occurred in the first 5-day period of alerting. The fraction of charts initially lacking documentation also fell abruptly. This indicates that people are learning to avoid the alert by adopting timely & complete documentation. Finally, the effect lasted for two months after the alerts ended, only diminishing concomitant with the arrival of a new resident cohort.
Real-time Location Systems
Beginning in 2002, Massachusetts General Hospital has experimented with an "indoor positioning system" (IPS). In a proof-of-concept pilot, we tested a system for automated patient location monitoring and management. Real-time IPS data provides patient locations that are compared to an expected-process model derived from the hospital scheduling system. The system automatically flags wrong-location events as soon as they occur and sends messages via the hospital paging system. The architecture is illustrated below.
The system detected all mock "wrong location" events during testing, and all actual locations were correctly assigned within 0.50 ± 0.28 minutes (mean ± SD). All "wrong-OR" events were correctly annunciated via the paging function. This experiment demonstrates that current technology can automatically collect sufficient data to remotely monitor patient flow through a hospital, provide decision support based on pre-defined rules and automatically notify stakeholders of errors. During the testing, we detected an actual wrong-patient, wrong-location event.
Recently, Mass General's OR installation of the IPS has facilitated just-in-time location of unique equipment needed to manage medical emergencies. We have also used the system to track patient wait times in the Mass General Same Day Surgery Unit.
The IPS is part of a family of 'location-aware' technologies that includes bar-coding, passive RFID, active RFID and ultrasound-based tracking, collectively known as auto-ID technologies. We recognized that no single auto-ID technology would satisfy the requirements of all healthcare applications. We also realized the true power of auto-ID technology as a quality and safety tool could only be harnessed by integration of the data from all deployed auto-ID systems, scheduling systems and clinical information systems. Accordingly, one member of the team developed the critical but heretofore missing middleware.
This new middleware imports location data from any auto-ID technology. It also interfaces with hospital information systems. This allows the software to provide 'state-awareness'. For example, when a tracked patient and multiple tracked pieces of equipment are in the same place, and the scheduling system shows that a procedure is underway, the software can update the state of the tracked equipment from "available" to "in use".
Data Integration & Team-Shared Awareness
To complement the automatically generated alerts, it is important to create a shared awareness among the care team of the most important patient data. We believed that the OR environment had enough digital data that such a shared awareness could be created by thoughtful integration of information already available in isolated systems. We also specified that this integration should occur with no requirement for user input or interaction with the system. The resulting visually integrated OR system is in use in the CIMIT/Mass General OR of the Future and is now being propagated to other ORs.
The integrated OR data display resides directly adjacent to the surgical display. The surgical display provides the entire OR team with an overview of the surgical situation. The OR data integration system is designed to provide a similar overview for key patient perioperative data.
The integrated information display collects information from devices and information systems present within the operative suite, including the physiologic monitor, surgical equipment, the Anesthesia Information Management System and Nursing Perioperative Record. External information sources including the Partners Enterprise Allergy Repository, the OR Dynamic scheduling system and the IPS are also inputs. Staff and patient locations can be imported from the IPS.