Chapter 26. Approaches to Quality Improvement in Anesthesia Care

1. In the United States there is significant public dissatisfaction with the level of safety associated with medical care. This issue has been taken up by many legislative and governmental regulatory bodies at both the federal and state level.

2. The demand for greater patient safety (ie, error reduction) has created multiple new levels of regulation and demands for accountability that go well beyond the traditional.

3. The traditional approach of practitioner accountability is giving way to approaches that additionally emphasize systems redesign and group-managed processes.

4. The demands for accountability require that individual practitioners, groups of practitioners, hospitals, and entire health systems implement methods that allow documentation of outcome and process.

5. The reliance of review systems such as case conference and focus on the individual patient is not adequate for the level of error remediation demanded. New approaches that allow examination of aggregate performance across a hospital or health system need to be developed and implemented. This is particularly challenging for perioperative anesthesia care, considering the already low level of adverse outcome.

6. For many practitioners, reimbursement will be linked to both outcome and the demonstrated compliance with process variables such as perioperative antibiotic administration. Systems that allow documentation at the individual patient level need to be developed or reimbursement and accreditation may be compromised.

7. The level of sophistication with respect to outcome and process evaluation in the intensive care unit has undergone much greater development than that for the operating room. A significant body of work exists to guide the improvement of care and to decrease the error rate in this area of anesthesia practice.

8. Safety and evaluation of outcome for acute pain therapy lags. Part of this may be related to the fact that decisions about acute pain therapy are often split between multiple groups. In addition, there is little definition of what adequate levels of acute pain relief are, how to accurately measure adequacy of pain relief, or how to monitor to prevent the complications of opioid-based pain therapy.

Since the publication of the Institute of Medicine monograph To Err Is Human in 1999, there has been a dramatic increase in the public's concern about the quality of patient care and the determination of the extent with which errors occur during the provision of that care.1 This concept of quality includes ensuring that the care is both beneficial and cost-effective and has placed unprecedented demands on all health care providers to prove the safety and value of the care they deliver.

Until recently, the autonomy of the individual practitioner to provide care in a manner in which they deemed best was paramount. In the hospital setting, the quality of care was ensured by preemployment credentialing with casual periodic renewal and episodic case conferences with the primary reliance on the individual professionalism of each practitioner. This no longer suffices. Groups, regardless of specialty, are being asked to identify potential sources of error both within their practices and the entire health care system. It is now assumed by hospital chief executives and medical officers that groups will establish procedures and protocols to minimize risk and to collaborate with other professional groups within a health system. For example, it is expected that anesthesiologists will collaborate with surgeons and nurses to cross-check and confirm correctness at certain critical points in a patient's care (eg, time-out before incision or checking transplant organ compatibility). As part of the continued support of a group, they are expected to document the approaches being used to evaluate and improve the quality and safety of the care provided.

These concerns about patient safety are being driven at many levels (Table 26-1). In addition to traditional accrediting agencies, hospital boards of directors and credentialing committees, state boards of medicine, legislatures, the federal government including Congress, and even the courts are increasingly involved. Patient safety and quality improvement is an area that is rapidly changing, the terminology often cryptic and the sources of information outside that usually consulted by anesthesiologists. The key goal of this chapter is to demystify the processes related to quality improvement and discuss several toolsets that are available to provide quality assurance including the approach in use at the University of Pennsylvania Health System. Questions to be addressed include: How does one determine the quality of care provided by a group of anesthesiologists? How does one identify errors in care? What is meant by a critical incident? How does one establish a process to improve the quality and safety of care provided? How should this process be documented?

A variety of terms are used to define aspects of patient safety, risk reduction, and quality of care. Some of these terms are duplicative. For a number of terms the meaning has changed over time. The following are terms and their definitions used by the Institute of Medicine (IOM) and the Joint Commission in their recent standards and reporting documentation.

Health Care Quality

In Crossing the Quality Chasm, the Institute of Medicine (IOM) provided a 6-part definition of health care quality that many view as the emerging standard. According to the IOM, health care should be safe, avoiding injuries to patients from the care that is intended to help them; effective, providing services based on scientific knowledge to all who could benefit and refraining from providing services to those not likely to benefit (avoiding underuse and overuse, respectively); patient centered, providing care that is respectful of and responsive to individual patient preferences, needs, and values and ensuring that patient values guide all clinical decisions; timely, reducing waits and sometimes harmful delays for both those who receive and those who give care; efficient, avoiding waste, including waste of equipment, supplies, ideas, and energy; and equitable, providing care that does not vary in quality because of personal characteristics such as gender, ethnicity, geographic location, and socioeconomic status.2

Patient Safety Culture

Lucien Leape, at the Inaugural Gala of the Leape Institute in 2008, said, "A just culture and transparent culture, where no one is punished for making or reporting errors, while at the same time everyone is held accountable. A culture of mindfulness and vigilance where every individual feels personally responsible for practicing safely and for reducing hazards. A culture where everyone feels valued and respected, but processes are standardized and simplified, teamwork trumps autonomy, and safety takes precedence over productivity" (used with permission).

Medical Error

Medical error has been defined as the failure of a planned action to be completed as intended (failure of execution) or use of a wrong plan to achieve an aim (error of planning).3 This is clearly a very broad definition. With regard to unintended acts, it is important to recognize that activities once considered appropriate care may now be considered a medical error and vice versa.

Sentinel Events

As defined by the Joint Commission, a sentinel event is "an unexpected occurrence involving death or serious physical or psychological injury, or the risk thereof. Serious injury specifically includes loss of limb or function. The phrase "risk thereof" includes any process variation for which a recurrence would carry a significant change of a serious event outcome." The Joint Commission notes the term sentinel reflects an event that requires immediate investigation and response. At a leadership level, the Joint Commission requires an integrated patient safety program that includes in part (1) A definition of the types of occurrences to be addressed, (2) a mechanism to ensure that all relevant areas of the health care organization participate and that their activities are integrated, (3) a procedure of immediate response including care of the affected patient, containment of the risk to others, and preservation of information for future review, (4) clear system for internal and external reporting, (5) a defined mechanism for responding to various types of occurrences, and (6) reporting at least yearly on occurrence of errors and efforts to improve patient safety both proactively and in response to actual occurrences. Thus reporting, categorizing, and review of serious events are required and mechanisms need to be established to accomplish this goal (Table 26-2).

In addition to the Joint Commission, other organizations have begun to require the documentation and reporting the equivalent of the Joint Commission sentinel event. Currently in the United States (2010), about 39 states require some form of reporting. Although the specific events and terminology may vary, there are common concepts that tend to follow the Joint Commission descriptions and more recently National Quality Forum Definitions. For example, New Hampshire's Adverse Event Reporting law (HB 592) was passed in 2009 and took effect January 1, 2010. The law requires hospitals to report certain adverse events to the NH Department of Health and Human Services (DHHS) by filing the following via a secure e-mail to DHHS: (1) An Initial Report, due within 15 days of the discovery of an adverse event, and (2) the hospital's Root Cause Analysis (RCA) and Corrective Action Plan (CAP) or a no corrective action taken statement, due within 60 days of the discovery of an adverse event.

The state's objectives with the new reporting are to improve safety by encouraging systematic, open, and transparent reporting of serious adverse events and to improve knowledge and understanding of why serious adverse events occur and how they can be prevented by identifying common underlying causes. HB 592 identifies 28 reportable events in the categories of surgical, product or device related, patient protection, care management, environmental, and criminal. (National Quality Forum: Serious Reportable Events in Healthcare Update 2010, A Consensus Report; www.qualityforum.org). The events are significant and result in either death or a permanent change in function or ability. Not all states have comparable reporting requirements to New Hampshire. However, most states publish an annual report of the number of reportable events that allow a certain level of comparison on this issue.

In some states such as New Jersey, specifically delineated events that might be associated with perioperative anesthetic care include hemolytic reactions from administration of ABO incompatible blood or blood products, electric shock, burns, injuries from the use of restraints, malfunction of ventilators or infusion pumps, surgery on the wrong body part, wrong patient, and wrong surgical procedure. They also include intraoperative or postoperative (within 12 hours) coma, death, or other serious preventable adverse event for any American Society of Anesthesiologists (ASA) class 1 inpatient or any same-day surgery patient regardless of ASA class. Other states such as Pennsylvania are less prescriptive. The Pennsylvania statute, Medical Care Availability and Reduction of Error (MCARE) Act (Act 13 of 2002), does not use the "sentinel event" terminology but rather defines a "serious event" as an event, occurrence, or situation involving the clinical care of a patient in a medical facility that results in death or compromises patient safety and results in an unanticipated injury required the delivery of additional health care services to the patient. The act further defines the term incident as an event that could have produced unexpected injury to a patient in a way similar to that defined under serious event. Several investigators have begun using state databases of reported events to determine the safety of different practices such as office-based anesthesia care.

Near Miss

This is defined as a "process variation" that did not affect the outcome but for which recurrence carries a significant chance of a serious adverse outcome. These events are not subject to review by the Joint Commission but need to be appropriately reviewed and changes made to decrease the risk of the event happening again.

Hazardous Conditions

These are defined as circumstances (not including the disease or condition for which the patient is being treated) that significantly increase the likelihood of a serious adverse outcome. A wet floor, for example, presents a hazardous condition that might increase the risk of a patient fall. The Joint Commission has now made prevention of patient falls a national patient safety goal, and therefore identification of both those at risk and hazardous conditions in those at risk is a national priority.

Sentinel Event Alerts

This is a mechanism introduced by the Joint Commission to bring attention to situations that might lead to sentinel events. Among the many are a number that appear potentially to involve the practice of anesthesia or intensive care, including high concentrations of potassium chloride, wrong surgery (wrong side, wrong patient, wrong procedure, etc), restraint deaths, high-alert medications (ie, anticoagulants, hypoglycemics, sedative/analgesics, paralytics, cardio- and vasoactive medications), operative and postoperative complications, infusion pump malfunction, look-alike sound-alike drugs, medical gas mixups, needles and sharps injuries, dangerous abbreviations, ventilator-related events, delays in treatment, nosocomial infections, surgical fires, anesthesia awareness, and patient-controlled analgesia. Once an alert occurs, it is important that it is disseminated throughout the organization.

The specialty of anesthesiology has recently addressed one of these sentinel alerts more formally. On October 6, 2004, the Joint Commission issued an alert, Preventing, and Managing the Impact of, Anesthesia Awareness. The alert went on to state:

To overcome the limitations of current methods to detect anesthesia awareness, new methods are being developed that are less affected by the drugs typically used during general anesthesia. These devices measure brain activity rather than physiological responses. These electroencephalography (EEG) devices (also called level-of-consciousness, sedation-level and anesthesia-depth monitors) include the Bispectral Index® (BIS), spectral edge frequency (SEF) and median frequency (MF) monitors. These devices may have a role in preventing and detecting anesthesia awareness in patients with the highest risk, thereby ameliorating the impact of anesthesia awareness. A body of evidence has not yet accumulated to definitely define the role of these devices in detecting and preventing anesthesia awareness; the Joint Commission expects additional studies on these subjects to emerge. In its review of the Bispectral Index (BIS) ® monitor, the Food and Drug Administration determined that "Use of BIS monitoring to help guide anesthetic administration may be associated with the reduction of the incidence of awareness with recall in adults during general anesthesia and sedation."

The Joint Commission recommended that health care organizations that perform procedures under general anesthesia develop and implement an anesthesia awareness policy. Because many practitioners were concerned that a particular monitoring device was being endorsed, they developed a sample departmental policy on intraoperative awareness that is available to members. It is important to recognize that resources are available to help address some of these alerts.

A broad definition for reportable events such as those found for sentinel events or serious events presented earlier may be difficult for practitioners to use on a day-to-day basis to decide what should or should not be reported. Without detracting from the need to report at a broad level, defining specific reporting indicators provides operational definitions that speeds this decision making. In 2009, the ASA established the Anesthesia Quality Institute (AQI). Multiple data sources are now flowing into AQI's central repository (Fig. 26-1). The mission of the AQI is to develop and maintain an ongoing registry of case data that helps anesthesiologists assess and improve patient care. The eventual goal of the AQI is to provide a resource for anesthesiologists to obtain patient safety and quality management data and to meet regulatory requirements designed to improve patient care. A key component of the AQI is the National Anesthesia Clinical Outcomes Registry (NACOR), which includes a group of key quality indicators developed by the Committee on Performance and Outcome Measures of the ASA. (Table 26-3). There are important process indicators also, such as on-time antibiotic administration or time-out documentation.

Reporting of adverse events needs to be encouraged. The aviation industry has used self-reporting for a number of years to detect situations at risk for serious events (for an overview, see Helmreich and Merritt.4 This reporting is nonpunitive, whereas failure to report may compromise one's employment position. An environment in which reporting is encouraged can only be created if the act of reporting does not create negative repercussions. Anonymous reporting may be helpful but compromises the ability to follow up and determine events at the level of detail that might be required. There has been debate about whether or not voluntary reporting is adequate, and a number of institutions have reported their experiences with scanning of anesthesia records for vital sign excursions beyond predefined boundaries. Others have used chart review by sampling a portion of the anesthetics performed. Given the relative rarity of many of the most serious anesthesia-related incidents and events, sampled chart reviews may miss critical cases. Chart scanning for acceptable vital sign boundaries might generate more false positives than can be reasonably reviewed, depending on the patient population of a particular institution. "Buy-in" for voluntary reporting provides the ability to identify the uncommon but serious events. (As mentioned, in some states, there is also mandatory reporting of medical errors.) Voluntary reporting need not be limited to members of an anesthesia department. We have found that observations by others, such as nurses, surgical house staff, and so on, can be valuable additions to the quality improvement process.

Peer Review

Peer review has been a part of anesthesia practice for over 70 years. One of the earliest documented uses of peer review for anesthesia-related mortality was by Ruth in 1935 when he helped establish the first anesthesia study commission to analyze perioperative deaths.5 The commission relied on voluntary submission of cases and determined the cause of death by majority vote. The peer review process actually approaches the analysis of care in a much more systematic way today, but clearly there is a long tradition in the field of trying to identify those cases in which the actions of the anesthesiologist contribute to mortality and morbidity. The magnitude of iatrogenic morbidity and mortality was not quantified across the spectrum of health care until the Harvard Practice Study, published in 1991.6 This seminal study used a structured peer review methodology to estimate that iatrogenic failures occur in approximately 4% of all hospitalizations. Specific categories of surgical adverse events have been identified in subsequent review of over 14 700 hospitalizations in Colorado and Utah.7 The nature of surgical adverse events was categorized by type of injury and by preventability (Table 26-4).

It is also common today for many of these categories of complications to be reviewed in a quality assurance (QA) process by multiple individuals who were not directly responsible for the care of the patient. For example, it may include a QA officer in the anesthesia department but may also include individuals in surgery, nursing, and pharmacy.

Mortality and Morbidity Conferences (Table 26-5)

The "M&M" conference has a long history of reviewing negative outcomes in medicine. The goal of this traditional conference is to learn how to prevent future patients from suffering similar harm and thus incrementally improve care. However, frank discussion of error is often limited in M&M conferences. Also, the actual review practices fail to support deep learning regarding systemic vulnerabilities; indeed, because M&M conferences do not explicitly require medical errors to be reviewed, errors are rarely addressed. One prospective investigation of 4 US academic hospitals found that a resident vigilantly attending weekly internal medicine M&M conferences for an entire year would discuss errors only once. The surgical version of the M&M conference was better with error discussion. However, although surgeons discussed adverse events associated with error 77% of the time, individual provider error was the focus of the discussion and cited as causative of the negative outcome in 8 of 10 conference discussions.8 Surgical conference discussions rarely identified structural defects, resource constraints, team communication, or other system problems. Further limiting its usefulness, the M&M conference is reactive by nature and highly subject to hindsight bias. This is the basis for most clinical outcome reviews, focusing solely on medical providers and their decision making.9 In their report, "Nine Steps to Move Forward From Error," in medicine, human factors experts Woods and Cook challenged the medical community to resist the temptation to simplify the complexities practitioners face when reviewing accidents post hoc. Premature closure by blaming the closest clinician hides the deeper patterns and multiple contributors associated with failure and ultimately leads to naive "solutions" that are weak or even counterproductive.10 The IOM has also cautioned against blaming an individual and recommending training as the sole outcome of case review.1 Although the culture within medicine is to learn from failure, the M&M conference does not typically achieve this aim.

The single-case approach slights trends and risks that can only be detected by grouping events and performing various types of data analyses. Although this type of conference still has a place as a tool to improve patient safety, it has been less than optimal, and other approaches need to be implemented as well.

One approach to using the morbidity and mortality conferences more effectively is to conduct them in an interdisciplinary manner. At the University of Pennsylvania, cases are periodically presented to a joint group of surgeons and anesthesiologists. There is a focus on cases in which more effective teamwork or communication could have led to better outcomes. These conferences have proven to be a very effective method for obtaining more widespread buy-in to quality assurance processes. Senior faculty members and the respective departmental leaders have been very willing to participate. The concept has expanded from joint conferences with general surgery and anesthesia to conferences with other surgical specialties and even with our colleagues in perioperative nursing. At Dartmouth-Hitchcock Medical Center the Department of Anesthesiology has used a variant of the M&M conference called Quality and Patient Safety Conference. It is multidisciplinary and presents all adverse events reviewed by the departmental Quality Assurance and Quality improvement Committee. The method of review identifies contributing factors and priority corrective actions. Hundreds of problems have been identified and improvements implemented using this approach over the 10 years it has existed.

Root Cause Analysis

Root cause analysis is a more structured process than traditional peer review for identifying the causal or contributing factors that lead to major morbidity or mortality (see www.patientsafety.gov/rca.html). A root cause analysis must include the following:

1. Determination of human and other factors

2. Determination of related processes and systems

3. Analysis of underlying cause and effect systems through a series of why? questions

4. Identification of risks and their potential contributions

5. Determination of potential improvement in processes or systems

The approach demands that bad outcomes not be attributed to the first error discovered, but rather it is important to review all of the potential sources of error or systems problems that led to the adverse event. Analysis depends on a robust model of accident causation. Reason's model is widely respected in high hazardous industries including health care.11 This model recognized that a combination of triggering factors, unsafe acts by individuals, latent conditions, and faulty defenses, allow a minor incident to progress until harm results (Fig. 26-2).

A systems approach allows one to discover all of the potential areas that can lead to an adverse event. We routinely use an approach that systematically investigates for contributory factors.12,13

Reason14 popularized the concept of latent conditions, that is, the "accidents waiting to happen," in a system or process. For example, an anesthesiologist may give the wrong medication, but contributing factors may be containers with similar size and lettering, low light levels in an operating room, a distressingly high level of music, and a malfunctioning inspiratory valve whose correction has created distraction. Typically these factors do not all occur at once, and the patient receives the correct medication. Focusing only on the last error (the giving of the wrong medication) will not correct the underlying cause of the error, nor will it decrease the probability of avoiding a similar error in the future.

Failure Modes and Effects Analysis

Failure modes and effects analysis (FMEA) is a systematic method of evaluating a process to identify where and how it may fail. Whereas root cause analysis assesses cause after an event has occurred, the goal of FMEA is to anticipate risks, alter processes, and thus avoid adverse events. According to the Institute for Healthcare Improvement, FMEA includes review of the following:

• Steps in the process
• Failure modes (what could go wrong?)
• Failure causes (why would the failure happen?)
• Failure effects (what would be the consequences of each failure?)

Tools have been developed to perform a FMEA (http://www.patientsafety.gov/safetytopics.html). Each failure mode gets a numeric score that quantifies (1) likelihood that the failure will occur, (2) likelihood that the failure will be detected, and (3) the amount of harm or damage the failure mode may cause to a person or to equipment. The product of these 3 scores is the risk priority number (RPN) for that failure mode. The sum of the RPNs for the failure modes is the overall RPN for the process.

For example, FMEA can be used to look at blood bank processes even if no adverse event occurred. It can also be used by groups of clinicians to test or simulate a process before it is incorporated into routine clinical care. This may lead to improvements in the process. Paradoxically, a proposed "improvement in care" may be found to increase the RPN rather than decrease it and therefore should not be implemented. The RPN can also be used to determine if implemented process changes actually resulted in lower potential risk.

Data Acquisition

Observer Reporting

Acquiring information about patient outcomes, adverse events, and incidents that might have led to adverse events is the first step for any quality review process. Event reporting by the caregivers or observers has been contrasted to chart review. Manual chart review can be labor intensive, and the yield may be low for the infrequent events of most interest to anesthesiologists. Chart review may be complementary to incident reporting and may in fact identify different populations of patients.15 Reporting by those involved needs to be encouraged. Ease of reporting,16 comfort providing reports, and the feeling that reporting may lead to useful evaluation and change may contribute to higher reporting rates. A process that leads to retribution for error will decrease reporting.

Whether or not caregiver incident reporting produces adequate response rates is a topic for debate. Sanborn et al suggested that in his institution there was a low level of compliance with voluntary reporting of defined intraoperative incidents.17 In contrast, Lagasse described a program claiming a near 100% response rate with a nonthreatening quality assurance system.18,19 Katz and Lagasse reported on incidents captured after 37,924 anesthetics by self-reporting, chart reviews, and an incident report process.20 They found that self- reporting for events resulting in disabling patient injury was very high. They found no significant benefit to chart review compared with self reporting.

Automated Data Acquisition

Over the past 15 years there has been a marked acceleration of the implementation of hospital information systems and continuing refinement of intraoperative data record keepers or anesthesia information management systems (AIMS). The AIMS have the ability to capture and identify multiple outcomes of interest and intraoperative events (eg, blood pressure changes, drugs). Data from individual patients can be aggregated and linked to predefined clinical and/or resource utilization outcomes. Although there is concern about the accuracy of the information obtained from the AIMS related to the potential "smoothing" of intraoperative data, studies in this area have reported a higher degree of accuracy than data obtained from a traditional "handwritten" record or self-reporting.21,22 Of greatest concern in the potential use of these large databases is the potential for "data mining." It is important to recognize that the associations that might be found from these types of analyses are just associations, and when a large number of variables are available there is a good statistical probability that these associations will be found. It is therefore important to recognize that these are hypothesis generating, rather than hypothesis confirming or disproving. Nonetheless, the increasing implementation of AIMS should lead to an improved ability to identify morbidity and mortality and to reconstruct the potential factors that led to the complications.

Computer-Assisted Data Acquisition

Computers can assist in the acquisition of incident and event information. In 2002, the University of Pennsylvania Medical Center went online with a computerized report entry system. Embedded within the system are the quality improvement indicators that we had been following for anesthesia-related events. The system is accessible from any computer within the medical center's domain that includes computers in every operating room, in strategic locations outside the operating room, postanesthesia care units, intensive care units, and all faculty offices. Prior to this system going online, we would typically receive about 50 to 100 patient-related event reports per year. Most of these came from anesthesia faculty and residents. The year after the computer system went online, 250 events were logged in with at least half coming from nonanesthesia observers. Currently we are receiving about 300 reports per year of events or incidents that might be related to the activities of anesthesiologists. The system has greatly improved our capture of potential issues. In addition there is faster access to these reports than was possible when a paper-based system was in use.

Data Verification

The medical record is the ultimate source of data verification. With respect to serious events, discussions with those involved before memory fades, review of the anesthesia record, and other associated records are critical. We have found that single reports may be inaccurate or the reporter not aware of the entire clinical situation. Thus reports are a start, and depending on the nature of the event, more investigation may be required. In some situations we undertake these reviews internally. When there appears to be significant multidisciplinary issues, we have asked our medical center risk management group to conduct a formal root cause analysis.

Database Entry

The data obtained often contain patient identifiers and information of a sensitive nature. Clearly basic precautions to avoid theft of this data or its publication need to be taken. Password protection of computer files using strong passwords and possibly encryption appear necessary. Paper records should be kept in locked offices or filing cabinets. We do not store any of our patient safety data on portable computers. We store the data on medical center servers that are routinely backed up and maintained behind a firewall. Patient records are not removed from the medical center because records of this nature have been stolen from automobiles; briefcases have been mislaid in restaurants, and so on. We attempt to maintain all of our patient safety/quality improvement data within the peer review system to decrease the risk of discoverability during any medical legal processes.

Consideration must also be given to the personnel entrusted with database entry. Diversion of records from their intended purposes by clerks and others has occurred. We also make sure that the records are not left out on desks when those doing data entry leave the area.

Review of Data

Moving beyond the Assessment of a Single Case

The single case unless intensively studied may provide little information. Aggregating the data will provide information about trends and special risks depending on the frequency of event occurrence and the time over which the data is accumulated. Particularly for rare events such as anesthesia-related deaths, occurrence and cause may not at all be evident without some sort of data accumulation. However, the key first step is to review the incident to determine its relevance to the anesthetic care provided.

Database Management and Incident Analysis Programs

Systematic data review can be done reasonably well using a computerized spreadsheet. Data entry and simple analysis such as frequency, changes over time, or events based on practitioner, type of surgery, or other variables of interest can be relatively easily obtained. The ASA has made available a program from Quality Assurance Research called QA/PDX. Members can apply and then download it from the ASA Web site (for information, see www.asahq.org/qmdaform.htm). According to the information provided by the developers, the program is a relational database that can be used on virtually any computer that uses a Microsoft operating system, including DOS legacy machines. The program has a defined data entry format as well as preconfigured reports.

Computer-Assisted Reporting Programs

Many large centers have developed their own reporting systems. However, a number of reporting systems can be purchased. The advantages of using a computer-based (as opposed to a paper-based) system for the entire institution (in contrast to a single department) have been mentioned earlier. Although we have not had personal experience with any of these programs, each has been installed and recommended in general way by a current user. These include Risk MonitorPro; WEBagent by Peminic; Quantros Occurrence report management system; International Developers Healthcare Management System, which includes modules for incident and complaint management, quality, and peer review management, as well as other process management programs.

Feedback

Obtaining and analyzing data is a QA function. However, quality improvement requires mechanisms to feed back this information to the involved practitioners and to other stakeholders. Because quality improvement involves system and behavior change, this part of the patient safety process is the most difficult and critically dependent on feedback mechanisms. In teaching departments, the normal turnover of residents and in many cases junior faculty means that the "institutional" memory is short. The intervals required to review or reteach this information may be difficult to determine, as is maintaining an up-to-date list of "what needs to be relearned." Change management in complex sociotechnical systems like health care is not simply common sense. One will quickly find that the major barriers to designing and implementing behavioral change are cultural, and with regard to systems and process, it involves managing the uncertainty found in health care. Interestingly, at both Dartmouth Hitchcock and University of Pennsylvania, we have very similar experiences with managing change to improve quality and patient safety. At both institutions, many reacted to new requirements for universal precautions during central line insertion by viewing them as an impediment rather than a process to be embraced and reliably practiced. We find communication to be critical and that a variety of approaches are needed to reach most of those who need to receive such specific patient safety messages. We regularly use conferences, e-mail both in the form of a periodic "newsletter" and individual e-mail, individual conversation, and policy development. We find that developing policies that can alter processes at the level of the operating nurse or those who prepare the operating room between cases can often effect change faster then working through physicians. We have also found that residents are more receptive than faculty to implementing change rapidly. The need for repetition and follow-up suggests that in larger departments, resources are needed to be devoted to the education component of the patient safety, quality improvement process. Several examples may be helpful. Intraoperative cardiac arrest related to potassium was not finally eliminated until concentrated potassium solutions were removed from the operating room. We achieved virtual 100% compliance with operating room central line placement guidelines when a sterile gown, drape, and towel pack specifically for central line placement was provided with every request for a central line kit. In addition, the circulating nurse and scrub nurse have become actively involved as a double check to ensure compliance.

One of the first reports in which anesthesia-related patient care data was aggregated was by Beecher and Todd, who studied anesthetic death in 10 institutions and published their work in 1954.23 Their study included 599,548 anesthetics. The cause of mortality was determined at the local institution by consensus of a surgeon and the chief anesthetist of the institution. Each death was characterized as having one primary cause and may also have had multiple secondary causes. This approach allowed a more thorough analysis of the causes of mortality beyond a primary one.

Dripps and colleagues at the University of Pennsylvania surveyed their experience during the 10-year period from 1947 to 1957.24 They noted 1285 operative deaths (death within 30 days) in approximately 120,000 anesthetics, for a gross mortality rate of 1.1%. This definition includes late deaths, as opposed to many studies that focus on the intraoperative period or the first 48 postoperative hours. After review of the hospital records, they determined if anesthesia was definitely or possibly contributory to death. This approach allowed the authors to identify systematic issues that could be addressed.

Clifton and Hotten reported on 162 deaths associated with anesthesia in 205,640 operations performed in the Royal Prince Alfred Hospital in Sydney, Australia, between 1952 and 1962.25 One cause of postoperative mortality was respiratory insufficiency. These authors argued that many of these complications would have been prevented by the use of a recovery unit. The potential safety advantage of a postanesthesia care unit (PACU) was a general theme in these reports from the 1960s, and this quality assurance process led to the routine availability of a PACU.

Numerous other groups have examined the incidence and etiology of perioperative morbidity and mortality over the years. For example, Tiret and colleagues carried out a prospective survey of complications associated with anesthesia in France from 1978 to 1982 in a representative sample of 198,103 anesthetics chosen at random from hospitals throughout the country in a study under the direction of the French Ministry of Health.26 The investigators evaluated either deaths or coma within 24 hours of surgery. The French survey confirmed previous findings that major complications occurred more frequently in older patients, those undergoing emergent operations, and those with more extensive comorbidity, as measured by the ASA Physical Status Classification.

One of the most important findings of the survey was that postanesthesia respiratory depression was the largest cause of death and coma that was totally attributable to anesthesia. Almost all of the patients with respiratory depression leading to a major complication had received narcotics and muscle relaxants that had not been reversed. They also reported a high incidence of "anaphylactoid shock." The authors contended that this was primarily due to Althesin and succinylcholine. Importantly, there was no category of drug overdose, which may have been a more appropriate label for some of these cases.

The need to look into perioperative morbidity and mortality on a national level was recognized in England. The pioneering work of Lunn and others led to the development of the Confidential Enquiry into Perioperative Deaths (CEPOD), which assessed nearly a million cases of anesthesia during a 1-year period in 1987 in 3 large regions of the United Kingdom.27,28 Unique to this study was the establishment of "crown privilege" by the government to allow total confidentiality: "The Secretary of State is satisfied that the disclosure of documents about individual cases prepared for the Enquiry into Perioperative Deaths would be against the public interest and would undermine the whole basis of a confidential study. Therefore, the data/information sent to the confidential Enquiry intro perioperative deaths is protected from subpoena."

Deaths within 30 days of surgery were included in the study. Anesthesia was considered the sole cause of death in only 3 individuals, for a rate of 1 in 185,000 cases, and anesthesia was contributory in 410 deaths, for a rate of 7 in 10,000. An important aspect of the CEPOD study was that it established both anesthesia- and surgical-related factors that contributed to mortality. Of the 410 perioperative deaths, there were 9 cases of aspiration or vomit and 18 cases of cardiac arrest.

There remains some debate as to whether anesthetic mortality has actually decreased or, if it has not, what factors (such as increasingly complex procedures or extension of surgical care into the extremes of age) may influence the data. Lagasse reviewed perioperative deaths in the 1990s and concluded that the anesthesia-related mortality rate had remained stable at approximately 1 death per 13,000 procedures, with wide variation making it impossible to detect trends in anesthesia safety. (See Chapter 25 for an extensive discussion of anesthesia risk and mortality.)29

Although some uncertainty remains, these types of publications add greatly to understanding the cause of rare outcomes and provide individual clinicians with areas to focus their attention.

Studies similar to the CEPOD study have not been performed in United States, most likely because of the legal system. Therefore, potential causes (and treatments) of perioperative mortality had to be obtained from other sources. This led the Professional Liability Committee of the ASA to conduct a nationwide survey of closed insurance claims for major anesthetic mishaps, which has resulted in a series of publications over the past several decades. The Closed Claims Study allowed analysis of very rare events and potential treatments.

One example involved unexpected cardiac arrest during spinal anesthesia, which was observed in 14 healthy patients from the initial 900 claims.30 The cases were analyzed to identify patterns of management that may have led to the event. Two patterns were identified: oversedation leading to respiratory insufficiency and inappropriate resuscitation of high spinal sympathetic blockade. In another example, Caplan et al reviewed the closed claims study for respiratory events.31 They identified inadequate ventilation, esophageal intubation, and difficult tracheal intubation as the primary causes of respiratory events. Most of the outcomes were deemed to be preventable with better monitoring. Much of this work formed part of the basis for the ASA Difficult Airway Guidelines and algorithm.32 More recently, they have evaluated complications that occur in remote locations.33 Investigators in England have also used this approach to review claims from the National Health Services.34

Based on concern about quality of care within the US Veterans Administration, Congress mandated the development of the National VA Surgical Risk Study (NVASRS) in 1986. Between October 1991 and December 31, 1993, the NVASRS was conducted in 44 VA Medical Centers and developed a risk-adjustment model for predicting 30-day outcome. Based on the NVASRS, the National VA Surgical Quality Improvement Program (NSQIP) was established in January 1994.35,36 NSQIP collected data on 40 preoperative clinical risk factors (eg, diabetes and heart disease), 20 categories of 30-day postoperative morbidity (eg, venous thrombosis, wound infections, and pneumonia), and 30-day postoperative mortality on patients having major operations under general, spinal, or epidural anesthesia. This VA-led initiative has published numerous articles demonstrating the factors associated with poor outcome and medical errors.37 As of 2002, the 30-day postoperative mortality after major surgery in the VA had decreased by 27% and the 30-day morbidity by 45%. By 2004, the private sector piloted and then the American College of Surgeons implemented a NSQIP program. Currently, more than 211 independent sites participate and contribute to the shared data and costs of data management and administration. The previously described ASA Anesthesia Quality Institute, although still in its early stages of development, is expected to accelerate identification of factors associated with poor outcomes specific to anesthesia practice.

As part of some of the national patient safety initiatives, a critical component has been the involvement of the patient and family. For example, the Joint Commission and the Institute for Healthcare Improvement (IHI) have both advocated informing the patients to ask their physicians and other health care providers to wash their hands. This may seem like a simple request, but multiple observations have suggested that hand washing is not a routine activity between patient encounters. Similarly, the Surgical Care Improvement Project (SCIP) has developed patient "tip sheets" to advise patients about appropriate questions and best practices that should be adopted at the individual hospital. For example, patients are informed to ask about the use of protocols to maintain perioperative β-blockade in those patients already taking the medication and other protocols to prevent deep vein thromboses. It is the authors' belief that this is becoming common practice and that clinicians should be prepared for an educated public who questions their practices.

See Table 26-6.

Over the past several years, a series of national patient safety goals have been advanced by The Joint Commission based on sentinel events and other sources of medical errors that occurred in the past. Table 26-6 represents a list of those currently being assessed during unannounced site inspections. The evidence-based goals are updated by The Joint Commission annually (Table 26-7), and links to the most current National Patient Safety Goals (NPSGs) and related literature are available at their Web site (http://www.jointcommission.org/patientsafety/nationalpatientsafetygoals/). Because the operating room is a high-risk health care setting, adoption of systems to reduce this risk is deemed critical and adherence to this collection of patient safety best practices is scrutinized.

On a national level, the ability to disseminate best practices has taken the form of standards and guidelines produced by national associations and the federal government. Practice policies or guidelines are the summation by clinicians of the available evidence about the benefits and risks of a treatment plan. Guidelines are a method of codifying recommendations regarding the use of a given technology or practice. Several types of recommendations fall into the general category of a practice parameter. A standard implies that a therapy or practice should be performed on patients with a particular condition. Standards are only approved if an assessment of the probabilities and utilities of the group indicate that the decision to choose the treatment or a strategy would be virtually unanimous. If a particular therapy or strategy is considered a standard, it is cost-effective for those for whom it is recommended. Standards are intended to be applied rigidly. The ASA first adopted Standards for Intraoperative Monitoring in 1986, which were developed from safety guidelines adopted at the Harvard hospital system. Guidelines are intended to be more flexible than standards, but they should be followed in most cases. Depending on the patient, setting, and other factors, guidelines can and should be tailored to fit individual needs. Like standards, guidelines should be cost-effective. The ASA has developed a series of guidelines that they term practice parameters or practice guidelines on a number of issues such as pulmonary artery catheter use and blood transfusions.38-40 Similarly, the American Heart Association/American College of Cardiology has established Guidelines for Perioperative Cardiovascular Evaluation for Noncardiac Surgery.41 Local practices can then be benchmarked against these national norms and local performance improvement initiatives can be developed. One example is the development of β-blocker protocols for patients who present to the hospital currently taking β-blockers. As more high-quality studies of best practices become available, practice guidelines will be used to disseminate these practices.

In the absence of strong evidence to support a given practice, expert opinion can help define best practice. For example, there is little high-quality randomized data to support the decision to perform a routine laboratory test before surgery. After reviewing the data, the ASA Task Force on Preoperative Evaluation chose to develop a practice advisory.42 Practice advisory uses expert opinions and survey of practice results to help inform clinicians.

Most of the discussion has focused on outcome measurements, death, or myocardial infarction, for example. However, because anesthesia-related complications are rare, observing outcome changes requires fairly large sample size and may be hard to separate from other factors such as underlying disease or the surgical intervention. For this reason, quality improvement methods typically implement and measure process-based indicators such as antibiotic administration, β-blocker administration, or other processes that have been demonstrated to lead to improved outcome.43,44 The key issue here is the relationship between the process being measured and outcome. Unfortunately, high compliance with process measures alone may not guarantee an acceptable outcome. Thus the previously discussed patient safety and quality improvement activities cannot be abandoned. Of concern is the use of process variables to institute "pay for performance" financial incentive paradigms and whether or not the processes mandated will in fact be correlated with improved patient care.

As summarized by Wu et al, in the United States, intensive care units (ICUs) account for about 10% of inpatient acute care beds.45 ICU mortality has been estimated to be between 8% and 10%, accounting for about 400,000 to 500,000 deaths each year.46 Errors and adverse events appear to be common in ICUs and may be on the order of 1.7 errors per patient per day.47 Consequently many believe that ICUs present an ideal area to study the effects of patient safety initiatives on outcome.48 The National Quality Forum has recently evaluated for endorsement a composite measure of ICU length of stay coupled with mortality. It is important to recognize that shorter ICU lengths of stay are beneficial if and only if they are not associated with increased mortality or readmission rates.

Communication errors may contribute to as many as 67% of adverse events in the intensive care setting.49 Pronovost and colleagues have described several tools, "the morning briefing" and the concept of "daily goals," that have improved certain measured outcomes.49,50 The morning briefing approach highlights those patients with the most critical issues; it helps determine the order of rounds and provides focus for the day's activity. The morning briefing form asks 3 questions: "What happened overnight that I need to be aware of? Where should I begin rounds? What are your concerns regarding potential problems for today?" The authors note that this approach provides an immediate overview of the totality of care requirements in the ICU and allows a more rational allocation of resources. They also note anecdotally enhanced teamwork, better identification of defects, and an improved admissions/discharge process.

The "daily goals" is a second form. It is "low tech" but has decreased ICU length of stay in the studied ICU from a mean of 2.2 to 1.1 days and allowed an additional 670 patients to be cared for during the year.50 According to the authors, this approach allows explicit delineation of the goals for each patient and improved communication with nurses, house staff, and families.

The frequency of ICU-based incidents and their deleterious effect on patient outcome has led to considerable effort to define relevant variables that alter outcome. Berenholtz et al described outcome variables such as ICU mortality rate, average days on mechanical ventilation, suboptimal management of pain; process variables such as appropriate use of blood transfusion, appropriate peptic ulcer disease prophylaxis; access variables such as delayed admissions and rate of delayed discharges; and complication variables such as rate of unplanned ICU readmission and the rate of resistant infections as a starting point to determine which factors seemed the most relevant to improving patient outcomes in the ICU.51 There is increasing awareness that postoperative pulmonary complications may be an important determinant of outcome.

Others have described various processes for data collection of adverse events, including collaborative, multi-institutional Web-based reporting systems.46,52 Approaches to encourage reporting within institutions include the trigger tool described by Resar et al, which is a mix of process and outcome variables that led to further review.53 Triggers in this series include chest tube insertion, positive blood cultures, death, renal failure, and a number of others. The selection was empirical. However, the authors determined a prevalence of 11.3 adverse events per 100 ICU days. Schuerer et al describe an error reporting system involving the use of preprepared cards listing various reporting indicators.54 They noted that reporting increased with this system compared with the previous online system. This was attributed mostly to ease of reporting, although the attention paid to reporting including inservice sessions may have been a factor.

Other types of data acquisition besides reporting have been attempted. Beckmann et al compared incident reporting to a medical card review.55 Although the incident reporting and medical chart review often found similar events, medical chart review seemed to detect infections, pain management problems, and myocardial infarction that were missed by the incident reporting system. The authors felt that the 2 systems were complementary, but that the medical chart review required a greater investment of resources.

Adverse events also occur while patients are being moved from other areas of the hospital to the ICU. Gillman et al, for example, determined that during a 6-month period, 22% of 290 patients transferred from the emergency department to the ICU experienced some type of adverse event during the transfer process. These ranged from incorrect identification bands to hypertension, hypotension, hypothermia, and transport equipment problems.56 Thus any program developed for ICU-related patient care quality needs also to account for patients coming into and leaving the unit.

Despite great efforts by a number of institutions, monitoring for patient safety and improving outcomes in the ICU setting is still in the developmental stage. For efficacy, Pronovost and Holzmueller note that quality improvement and patient safety within the ICU setting is multidisciplinary.57 They emphasize the need to develop a culture that emphasizes these aspects of patient care. They note that developing work processes that reduce complexity, standardize, and automate may be most helpful for reducing error.

Quality and safety standards with respect to acute pain appear have been especially difficult to develop. Part of the problem relates to the fragmentation of acute pain care between surgeons, nurses, and anesthesiologists that occurs in most institutions. Part is also related to philosophical differences toward the treatment of acute pain and the lack of a quantitative measure outside of patient observation. Practice guidelines from the ASA emphasize that a single approach to pain management may not be appropriate for all types of patients and that multimodality approaches tailored to the patient's clinical conditions and therapeutic requirements have the greatest probability of producing adequate pain relief with the least risk of undesirable side effects.58 The Joint Commission has published specific pain management standards that may help increase the resources devoted to this often ignored aspect of patient care.59 The requirement that patient pain needs to have appropriate assessment and management provides the basic underpinning of determining the efficacy and safety of pain control efforts. There are 2 problems. The first is that many institutions have not yet implemented appropriate pain assessment and control programs. The American Pain Society guidelines, now close to 11 years old, provide a very precise approach to instituting an effective approach to acute pain care.60 The second problem is the focus of this section: Specifically, once an acute pain care plan is in place, how does one go about monitoring for safety and efficacy?

Garnerin et al discuss a single event related to a pump error and the process of root cause analysis related to that error.61 Karlsten et al found that over a 3-year period the assessment of pain according to the set protocols improved with repeated staff training, meetings, and audits.62 Bardiau et al reported improved pain relief as indicated by visual analog scale (VAS) scores with the implementation of a multimodal pain therapy approach, routine VAS measurement, and nurse training.63 Sartain and Barry reported improved pain control with the institution of an acute pain service.64 Miaskowski et al described the results of a prospective multisite study involving 5837 patients of whom 49% were cared for by an anesthesia-based pain service.65 Patients cared for by the acute pain service had significantly lower pain scores; they also had a lower level of complications. Meissner et al described an ongoing benchmarking protocol for quality assessment.66 This approach used a specialized pain nurse who interviewed patients on the surgical wards for quality of pain relief. Benchmarks included pain at rest, pain with ambulation, and maximal pain intensity since surgery. Specific side effects such as nausea, vomiting, and sedation were also determined. Other potential quality indicators might include the documentation of VAS, the frequency of VAS documentation, the appropriateness of pain medication for the expected level of pain, efficiency of transition for one pain modality to another, and complications.67 The large number of patients being treated for pain allows the construction of control charts based on type of surgery, pain therapy modality, and physical condition of the patient. This might allow determination of the success of pain control efforts and better determination of which approaches are the most effective in a particular patient population.

Postoperative sedation/analgesia care has been associated with preventable deaths, leading to calls for increased monitoring in the postsurgical care setting. Multiple studies have shown that failure to rescue due to sedative/analgesic-induced respiratory depression is relatively common given the millions of patients having surgery annually. Although there is debate as to the effectiveness of rescue teams and methods for deploying them, recent focus on early detection supports expanded monitoring for unexpected respiratory depression.68-72

At Dartmouth Hitchcock, surveillance monitoring of all postsurgical patients was associated with significant reductions of rescue activations and the need for ICU transfer when compared with units without this monitoring practice.

Many physicians are concerned about the confidentiality and discoverability of discussions of patient outcome. Most states provide protection for discussions held within the purview of peer review committees. The degree of protection may vary but tends to protect the members of the committee, those who testify to the committee, and the written work product of these committees. Original patient records are of course discoverable even if taken into the "protected" environment of the peer review committee. Discussions outside the committee environment may be discoverable in legal proceedings. In general, case conferences and mortality and morbidity conferences avoid many of these issues by avoiding the use of specific patient identifiers.

Changing Patient Expectations

Patient expectations are changing. Many want greater involvement in the care process. Some demand input into the exact drugs used and evidence significant dissatisfaction with the care provided when their desires are not met. Many also want a near risk-free experience and have little tolerance for what many of us view as relatively minor events such as small degrees of hematoma or bruising, intravenous infiltration, delays in start of surgery, and so on. "Service recovery," measures taken to satisfy the patient or their family when care expectations are not met, has long been used to compensate patients for dental injury. It is now being expanded at our institution to reimburse the copay or uninsured portions of postoperative consultations related to certain aspects of anesthesia care such as ophthalmology consults in the case of eye irritation, plastic surgery consultation for intravenous infiltration, or otorhinolaryngologic consultation for voice changes or persistent sore throat. Patients are also more willing to "complain" to hospital leadership when their expectations are not met.

Changing Payer Expectations

Payers for medical care expect better outcomes and are willing to pay for these better outcomes through the use of incentives (ie, pay-for-performance, or P4P). There are 2 different philosophical approaches to P4P: (1) focusing on the outcomes themselves or (2) on care processes that have been associated with the improved outcomes. An example of an outcome-oriented program is the NSQIP. As described previously, the American College of Surgeons has recently adapted NSQIP (ACS-NSQIP) to the private sector and allowed private hospitals to send a standard set of data to a national center and obtain observed versus expected mortality.

The process variable approach is best exemplified by the Surgical Care Improvement Project (SCIP). SCIP represents a partnership that was formed in 2003 and included representatives from the American Hospital Association, the American College of Surgeons, the ASA, the Association of Perioperative Registered Nurses, the Joint Commission on Accreditation of Healthcare Organizations, the IHI, the Department of Veterans Affairs (VA), the Agency for Healthcare Research and Quality (AHRQ), the Centers for Medicare and Medicaid Services (CMS), and the Centers for Disease Control and Prevention (CDC) (see www.medqic.org/scip). In 2005, the group set the goal of reducing surgical complications by 25% by the year 2010. Four major areas of surgical complications were identified: prevention of surgical site infections, perioperative myocardial infarction, postoperative pneumonia, and venous thromboembolism (pulmonary embolism and deep vein thrombosis).

The partnership has used expert panels that included members from the steering committee and from more than 20 additional organizations. The goals are to identify best practices supported by the literature or preferably identified as class I recommendations in guidelines developed by the appropriate societies. For example, the recommendations for perioperative β-blockade were restricted to class I recommendations by the American College of Cardiology/American Heart Association Guidelines for Perioperative Cardiovascular Evaluation. In fact, the Focused Update on Perioperative Beta Blockade was published in part to ensure that the SCIP measures would be consistent with the most current evidence. The measures for venous thromboembolism are based on recommendations from the American College of Chest Physicians. The most well-described measures are those related to the prevention of surgical site infections. Both timing and appropriateness of perioperative antibiotics were incorporated from the previously commissioned Surgical Infection Project (SIP). Additional measures such as perioperative normothermia in colon surgery and perioperative glucose control in cardiac surgery were also part of this "module." At this time, the SCIP measures are voluntary. Hospitals can report on any one of the 4 major areas, but they must report on all of the measures in that area. The IHI has also included the SCIP project in their national agenda. However, the antibiotic measures of SCIP that were incorporated in SIP are part of the new P4R, or pay-for-reporting, initiative incorporated into the Medicare reimbursement policy. Many private insurers and states are also incorporating this measure, and eventually the overall performance on this measure may affect payment (pay-for-performance, often called "P4P") rather than just reporting. Although the US Congress has clearly identified P4P as an important approach to controlling health care costs and improving quality, there are several investigators who are concerned that better performance on process variables (ie, antibiotic timing) may not lead to improved outcome. Werner and Bradlow found that hospital performance measures only predict small differences in hospital risk-adjusted mortality rates.73 In a recent study of hospitals in the Premier Inc Perspective Database reporting SCIP performance, adherence measured through a global all-or-none composite infection-prevention score was associated with a lower probability of developing a postoperative infection. However, adherence reported on individual SCIP measures, which is the only form in which performance is publicly reported, was not associated with a significantly lower probability of infection. Importantly, when performance measurement was evaluated over time (2004-2006), hospital process performance improved and was associated with better patient and quality outcomes. For acute myocardial infarction, performance improvements were associated with declines in mortality rates, lengths of stay, and readmission rates. As part of health care reform, an increasing number of measures are being developed for physicians through the American Medical Association Physician Consortium for Performance Improvement. The Consortium is composed of more than 100 national medical specialties and state medical societies; the Council of Medical Specialty Societies; American Board of Medical Specialties and its member boards; experts in methodology and data collection; the Agency for Healthcare Research and Quality; and Centers for Medicare & Medicaid Services. It is committed to the development, testing, and maintenance of evidence-based clinical performance measures and measurement resources for physicians. The measures are then proposed to the National Quality Forum (NQF) and will become those measured included in the P4P measures for CMS.

Changing Hospital Expectations

Medical centers increasingly are expecting members of their clinical staff to actively participate in programs to improve patient outcomes and safety. The active participation of anesthesiologists in the patient safety time-out process and in the proper identification of organ transplant blood types are just 2 of many. Requests to participate in multidisciplinary process planning groups are increasing. There is an increasing concern with patient satisfaction, and clinician groups are being increasingly asked to review processes that produce less than optimal outcomes, significant patient dissatisfaction, or have higher costs. This will only increase, and a good understanding of the sources of patient dissatisfaction and poorer outcomes within the context of a specific practice will become increasingly necessary.

A robust quality improvement system is essential to ensure that we provide our patients with the highest quality of care. The history of the specialty of anesthesiology has been marked by a focus on patient safety and assessment of outcomes. Anesthesiology has been lauded as the specialty that has focused on achieving a Six Sigma approach to quality. Numerous national projects have attempted to identify factors that contribute to perioperative morbidity and mortality and proposed systems to reduce complications through the publications of guidelines. These guidelines are now being converted into P4P initiatives to ensure that all practitioners practice according to the best evidence. Despite these national efforts, individual providers and groups of providers should develop their own systems to both identify the causes of complications and disseminate practices to reduce these same complications. As an example, participation in the Anesthesia Quality Institute will allow anesthesiologists to benchmark their performance versus peers in an effort to continue to improve.

• Miaskowski C, Crews J, Ready LB, Paul SM, Ginsberg B. Anesthesia-based pain services improve the quality of postoperative pain management. Pain. 1999;80:23-29.
• O'Neil AC, Petersen LA, Cook EF, Bates DW, Lee TH, Brennan TA. Physician reporting compared with medical-record review to identify adverse medical events. Ann Intern Med. 1993;119:370-376.
• Quality improvement guidelines for the treatment of acute pain and cancer pain. American Pain Society Quality of Care Committee. JAMA. 1995;274:1874-1880.
• Reason JT. Human Error. Cambridge, UK: Cambridge University Press; 1990.
• Rubin HR, Pronovost P, Diette GB. The advantages and disadvantages of process-based measures of health care quality. Int J Qual Health Care. 2001;13:469-474.
• Sanborn KV, Castro J, Kuroda M, Thys DM. Detection of intraoperative incidents by electronic scanning of computerized anesthesia records. Comparison with voluntary reporting [see comments]. Anesthesiology. 1996;85:977-987.
• Wu AW, Pronovost P, Morlock L. ICU incident reporting systems. J Crit Care. 2002;17:86-94.