t the ASA 2001 Annual Meeting, the Board of Directors of the Anesthesia Patient Safety Foundation (APSF) issued the following statement:

The APSF endorses and advocates the use of automated record keeping in the perioperative period and the subsequent retrieval and analysis of the data to improve patient safety¹

This brief statement is in accord with national efforts that recognize the pressing need to develop an information technology infrastructure to support the delivery of health care.² Despite this endorsement by APSF and related national initiatives, the role of information technology in anesthesiology practice remains quite limited. Indeed, at the present time, the majority of anesthesia practitioners in the United States render good quality care without utilizing automated record keeping. Nevertheless this author predicts that within 10 years, information systems for documenting the anesthesia care process are likely to be a standard part of anesthesiology practice. Personal predictions aside, there is little doubt that by 2050, information technology will be completely integrated into medical practice and will have a profound impact on the quality and cost of health care.

Why is it that APSF has decided to promote the use of automated record keeping in anesthesiology practice? The rationale behind this decision recognizes the limitations of current efforts to improve patient safety and outcome. In response to legislation, ethical considerations and mandates by the Joint Commission on Accreditation of Healthcare Organizations, most anesthesiology departments have an ongoing quality assessment/improvement (QI) process based upon voluntary reporting of recognized outcome indicators. Cases are discussed in a departmental forum designed to protect the content of the discussion so that reporting is encouraged and discussion uninhibited. Although this process is valuable for evaluating the impact of clinical practice on patient care, the methodology suffers from several inherent limitations. Such limitations follow:

Outcomes are under-reported: Voluntary reporting requires individuals who will be asked to explain their role in an undesired outcome to report the event. This approach undoubtedly leads to some degree of under-reporting.

Data to support the discussion are often unavailable: Even if you are able to capture all of the desired indicators, when the case is discussed, important data required to truly understand what might have led to the outcome are often forgotten or unavailable.

The “apprentice” approach to training can limit the perspective on practice patterns: We are all victims of our training to some degree and tend to approach discussions of acceptable practice patterns from the perspective of our prior experience. Departments with staff that have trained in different institutions may benefit from varied perspectives on practice patterns. Nevertheless discussions of practice patterns tend to fall within the constraints of local standards.

You only find the problems that you look for: By focusing on a specific list of outcome indicators, by definition, you may not examine all of the outcomes of interest. Furthermore the focus of QI efforts solely on indicators of untoward events prevents us from looking to optimize outcomes. Consider a patient who undergoes an uneventful anesthetic for laparotomy but ends up admitted to the intensive care unit on postoperative day one for respiratory distress. At the present time, we cannot determine whether this patient suffered a complication that might have been prevented by better anesthesia care or not. Was intraoperative attention to respiratory status optimal? Was postoperative analgesia as effective as possible?

Events that occur in the postoperative period are difficult to identify: Attention to outcome measures in the postoperative period is a major limitation of the anesthesia QI process. It is difficult, if not impossible, for most anesthesiology departments to identify outcomes that might relate to intraoperative management after the patient leaves the postanesthesia care unit.

The QI process typically does not cross departmental lines: Despite the fact that multiple departments (anesthesiology, surgery, nursing, respiratory care, pathology, etc.) impact the perioperative care process, it is rare for all of the relevant departments to participate in the QI effort.

The low incidence of undesired outcomes prevents individual departments from identifying the scope of a problem: Individual departments are unlikely to have a large number of significant bad outcomes in a given year. Outcomes discussed as part of the QI process are not typically entered into a database so that they can be retrieved and analyzed when similar events occur again. When viewed from a departmental level, a particular outcome may not appear as a major problem. When viewed from a national level, however, the scope of the problem may be quite significant. The recent recognition that surgery in the prone position is associated with visual loss is an example of such an outcome.

Recognizing the inherent limitations of the typical QI process, ASA promotes efforts to assess outcome that go beyond the departmental level. These efforts have identified outcome problems that would have been unlikely to be detected by individual departments. The association of spinal anesthesia with sudden cardiac arrest identified by analysis of the Closed Claims Project database is a case in point.3 The Closed Claims Project hints at the promise of information technology to enhance patient care. The Closed Claims project now consists of 6,448 claims that have been systematically entered into a database. Any ASA member can request a search of the database <www.asaclosedclaims.org>. Under the direction of the ASA Committee on Professional Liability and investigators at the University of Washington in Seattle, the Closed Claims Project has spawned other national database efforts. Both the Pediatric Perioperative Cardiac Arrest (POCA) <depts.washington.edu/asaccp/POCA/index.shtml> and the Postoperative Visual Loss (POVL) <depts.washington.edu/asaccp/eye/index.shtml> registries provide a mechanism for accumulating data from multiple departments to gain insight into a problem that cannot be analyzed on the departmental level alone. Although the value of these national database projects is indisputable, they are limited by many of the same factors as the departmental continuous quality improvement process. Most importantly, they only provide a sampling of the events that are reported, not a representative sample of the outcomes that actually occur.

Recognizing the limitations of voluntary reporting methods, there are international efforts to bring scientific rigor to outcomes assessment. Examples include the Multicenter Study of Perioperative Ischemia <www.mcspi.org> and the Outcomes Research Institute <www.or.org>. These efforts are typically funded through research grants and extramural support rather than being integral to the health care process.

How will the limitations of our current quality improvement efforts be addressed by information technology in the future? The first steps are already under way. A fundamental first step is to establish standards for storing data so that information can be shared between institutions. The Data Dictionary Task Force (DDTF) was formed by APSF to address that issue for anesthesiology <www.apsf.org/ddtf>. Merely agreeing on data standards is only an enabling step; however, it will not solve the challenge of analyzing outcomes.

So here is the vision for 2050. A completely electronic medical record will be maintained for all patients. Each patient will likely have a “key” of some sort for gaining access to his or her medical record. The actual data stored on the record will exist in many locations, but network technology will link the data together and provide access to the data required to support care. At the point of care, the medical record will not consist of multiple data items from each department involved with the patient, but will offer an integrated picture of the care process. This integrated picture will foster care processes throughout health care institutions that optimize outcome.

In this future model, public health practitioners will become the powerbrokers of health care delivery as they will have the tools to utilize the information infrastructure to benefit the greatest number of patients. The data on the medical record will be stored in a standard format so that information from multiple patients and multiple institutions can be linked and studied. The ability to readily analyze population data will be the foundation of an evidence-based system for providing health care. Such a system will allow health care to move beyond the limitations imposed by apprentice training. Best practices will emerge based upon outcomes measures that are not limited to untoward events but rather focused on optimizing the outcome of the care delivery process. Outcomes will be viewed through the lenses of economics and quality of life and will offer a better understanding of the relationship between cost and quality. Patients will have a much clearer understanding of the expected outcomes of care and the level of individual risk, and medico-legal costs will be a fraction of the current level, if not disappear completely. Patients will have full access to the information underlying their health care decisions, and “standards of care” will not be articulated based upon individual opinion but objective measures that link practice to outcome.

The challenge to the health care system today is to embrace the promise of information technology and identify the obstacles that must be overcome to realize the benefit of this technology. We have already begun to standardize the data collection process. Information technology that will capture the care process in an electronic medical record is proliferating, although it will be many years before it is universally adopted. Along the way, we will address the challenges of privacy and funding for this infrastructure and we will develop methods for storing and analyzing the immensely large data sets that will result from this process.

It is unclear how information technology will influence anesthesiology practice in 2050, but what is clear is that some form of information infrastructure will exist that improves outcomes and controls costs. Without information technology, we simply cannot meet the health care challenges of the future.

References:

1. APSF Endorses Use of Automated Record Keepers. APSF Newsletter. 2001-2002; 16(4). <www.apsf.org/newsletter/2001/winter/02ARK.htm>.

2. Crossing the Quality Chasm: A New Health System for the 21st Century. Washington, DC: National Academies Press; 2001.

3. Caplan RA, Ward RJ, Posner K, Cheney FW: Unexpected cardiac arrest during spinal anesthesia: A closed claims analysis of predisposing factors. Anesthesiology. 1988; 68:5-11.

Jeffrey M. Feldman, M.D., is Associate Professor of Clinical Anesthesia, University of Pennsylvania School of Medicine, Children’s Hospital of Philadelphia, Philadelphia, Pennsylvania. He is President of the Society for Technology in Anesthesia.