Quality, like pornography, is somewhat difficult to define, but is usually recognizable when it is seen. Quality in health care has become a major focus of attention in recent years for several reasons, including the (1) burgeoning costs of health care, (2) emergence of data on the frequency of errors in this field, and (3) the need to reconcile the first 2 issues with increasing public clamor for the elimination of errors. As is often the case, the public and the media express outrage at the revelations of data on errors, but are quite unwilling to invest public funds into precautionary systems to minimize the opportunities for errors. Unfortunately, the avaricious medicolegal climate in the United States has almost certainly been further stimulated by the open and frank publication of data on medical errors. Against this background of media, legal, and medical pressures to reduce errors, federal and state governmental regulatory agencies have ratcheted up the volume and complexity of their regulations, which in turn have further increased the cost of providing health care.

It would be both misleading and overly simplistic to equate quality with freedom from errors, but there is broad consensus that effective systems to make significant reductions in errors must be an integral part of any part of health care that is recognized as having a high level of quality.

Perhaps because of the data management issues intrinsic to laboratory medicine, clinical laboratorians have been innovators and trend setters in the establishment of parameters for the measurement of quality, especially in laboratory services. In some respects, it has taken several decades for the rest of medicine to catch up. While clinical pathology in general has been very active in setting up systems to evaluate the frequency and nature of laboratory errors, transfusion medicine has been in the forefront because of the well-established potential for immediate and clinically catastrophic consequences of errors in the transfusion of blood. For probably 60 years, medical students have been taught about ABO hemolytic reactions, and for many decades, most hospitals have established sample identification criteria that are often more strict for transfusion medicine blood samples than for those destined to be tested in other clinical laboratories. This strategy was perhaps one of the earliest forms of laboratory error prevention based on unfortunate clinical experience and very simple root-cause analysis. This early analysis obviously indicated that blood sample identification errors were causative in a significant number of transfusion errors leading to severe adverse outcomes.

This article outlines the philosophical underpinnings and experience with a program that evolved over many years at our institution. We readily admit that this is not a prescription for how a program can best be developed at any other institution, because the mission, size, and resources (including the experience level and attitudes of personnel) are all unique to each institution and are constantly changing and evolving. We merely describe what has worked at our institution so that others may glean ideas that can be modified and applied in their own settings. How each institution goes about setting up their quality programs will depend on their existing programs, the resources that are available to them, and those issues that internal or external audits or inspections indicate are deficiencies in their existing programs. There is no “one size fits all” approach that can be simplistically applied. We believe that developing a supportive attitude among top management and implementation of an adequate quality and educational infrastructure are more important than the choice of a particular error management system, of which there are several available.

ERRORS

Errors can be defined in a number of ways, but most would agree that an error involves a failure (deviation) in the performance of a standard operating procedure (SOP). Looked at from a patient's perspective, an error in the transfusion of blood can occur in a variety of situations; some of which are listed here.

  1. An error is made in identifying the patient, who then receives the wrong product.

  2. A patient receives the wrong blood product and has an adverse reaction, for example, hemolytic reaction. Investigation of the reaction uncovers the transfusion of the wrong blood product.

  3. A patient almost receives the wrong product, but the error is caught at the bedside (so-called near miss).

  4. An error in crossmatching is caught before the blood is released from the laboratory (another near miss).

  5. An error is made with an intralaboratory sample mix-up and is caught prior to release of the blood.

  6. An error is made by the venipuncture technicians when identifying a blood sample.

  7. An error is made by the venipuncture technicians when identifying the patient.

  8. An error is made by the ordering physician who orders blood for a patient whom he/she did not intend to transfuse.

This list is by no means complete and only refers to one aspect of the lengthy, circuitous process involved in collection, processing, labeling, storage, testing, ordering, and administration of blood products. This process is intended to function so that safe, appropriately tested, stored, and matched blood products are delivered to the correct patient at the time and place intended by the ordering physician. When one considers the hundreds of different SOPs and the dozens of individual technologists, physicians, and nurses involved in carrying out the multitude of steps in the process, it is astonishing that even more errors are not made.

DETECTION OF ERRORS

It has been long established that there are limitations to the ability of humans to perform defined, simple tasks (however well trained and knowledgeable they are) repeatedly without committing a human error.1,2 Data from the Division of Transfusion Medicine at Mayo Clinic (Donor Center and Transfusion Service, Rochester, Minn) indicate that this limit is about 1 in 10 000 performances at best and may be considerably more frequent in situations in which personnel are tired, overworked, stressed, distracted, or harassed; are poorly or incompletely trained; or are incompetent for any other reason. Because of the certainty of human errors, it is critical that there be systems in place to identify all errors occurring at any stage of each process involving 1 or many SOPs. If all errors cannot be analyzed, cataloged, and characterized in a consistent fashion, logic-based solutions cannot be implemented and instead, a reactive, episodic, and unplanned series of temporary approaches will be hurriedly put in place whenever a particular error or group of errors happens to attract attention, either because they get all the way to a patient who is harmed or are feared likely to do so. There are several published and well-tried systems designed to help institutions proactively collect data on errors in ways that facilitate the regular, consistent analysis and characterization of these errors so that trends may be followed in a methodical fashion.2–4 In our experience, the American Association of Blood Banks' Quality Systems Essentials have been of great value. This well-organized system provides an excellent starting point upon which to add locally applicable modifications. Another excellent event-reporting system is the Medical Event Reporting System for Transfusion Medicine developed by Kaplan et al.5 This system is particularly good for ensuring complete reporting of errors and consistent application of rules of analysis of those data. It is self-evident that implementation of such systems and strict application of their tenets greatly facilitates accreditation and inspection processes, which are often designed to determine the degree of compliance with these very same error management systems. Even more helpful has been the insistence of the Food and Drug Administration (FDA) that blood collection centers (freestanding or hospital based) operate in an environment that applies current Good Manufacturing Practices (cGMPs).

While the extensive rules and regulations applied by FDA are at times onerous, cGMPs, if consistently adhered to, force institutions to thoroughly plan, document, and validate each and every SOP and process. Errors can then be much more closely visualized against a background of comprehensively and clearly written SOPs so one can determine just what went wrong. That is the first essential step in the analytical process. For the collection and documentation of errors to be dependable, there must be clear delineation of responsibilities for performance of tasks. For obvious reasons, there should be a degree of separation of quality-specific responsibilities so that the individual who is the supervisor of a particular laboratory is not the sole person who counts, characterizes, and analyzes the errors occurring in that laboratory. No one should oversee the quality indicators and direct appropriate responses based on errors in their own personal work. Objectivity requires that the analysis of the data be directed by someone who has less vested interest in the operational performance of that area. However, this separation cannot be so great that the operations supervisor is totally excluded from the process of analysis of the errors in his or her area or in the development of plans for corrective action. Supervisors and technologists in a particular work area are usually the most highly knowledgeable about SOPs and systems in that area and are essential to the planning and validation of corrective actions. By getting these individuals deeply involved in root-cause analysis, corrective options, and planning for such, one makes them part of the solution rather than part of the problem. However, they should not be in a position to minimize or deflect appropriate attention from a problem in their area. By including supervisors and technologists as members of a team evaluating the problem, one can ensure objectivity and accountability in the process while also painlessly furthering their education concerning the real value of good data, thorough objective analysis, and the need for understanding of the big picture rather than just their own more parochial concerns. This approach also helps to inculcate the concept that quality management is the responsibility of each and every individual, rather than just the realm of an elite few quality management personnel. In other words, balance between the operations and the quality assurance units is critical to a successful error detection and management system (Figure 1).

Figure 1.

When operations and the quality assurance (QA) unit are in balance, a healthy tension for error management is created (shaded boxes). When operations is responsible for all aspects of error management, the objectivity and regulatory review provided by the quality assurance unit is lost (vertically hatched box). When the quality assurance unit is responsible for all aspects of error management, the concept of quality management as the responsibility of each and every individual is lost (horizontally hatched box)

Figure 1.

When operations and the quality assurance (QA) unit are in balance, a healthy tension for error management is created (shaded boxes). When operations is responsible for all aspects of error management, the objectivity and regulatory review provided by the quality assurance unit is lost (vertically hatched box). When the quality assurance unit is responsible for all aspects of error management, the concept of quality management as the responsibility of each and every individual is lost (horizontally hatched box)

ANALYSIS OF ERRORS

Again, there are multiple schemata described for data analysis. A particularly effective one is included in the Medical Event Reporting System for Transfusion Medicine developed by Kaplan et al.4 This system uses the concept of analysis of near-miss events as a particularly rich vein of information. Like other successful schemata, it stresses the importance of a nonpunitive reporting system. For collection of complete information on errors, individuals must not only understand why they should be alert to even minor errors and report them, but also that such reporting of their own and others' mistakes will not lead to retribution. Nothing will dry up the fund of information on errors as quickly as the fear of retribution. This nonpunitive philosophical approach must be very clearly and widely understood throughout the workforce and must be emphasized and reemphasized (preferably publicly) by the top officers of the organization, such as the medical director and administrator. The goal should be to create a culture that rewards the behaviors desired, that is, reporting near misses and real errors as potential opportunities for improvements. The system should not be designed to identify “culprits” for punishment when an adverse event occurs.

A great variety of different sortings of the error data can be performed, from simple to very detailed, depending on what computer system support one has available. Data might be stratified according to type of error (clinical, performance, interpretive, etc) by work unit, by time of day or night, by experience level of personnel, etc. The particular approach used will depend on the circumstances of the errors and the leads provided by initial analyses. Naturally, comprehensive data entry is necessary if the stratifications are to be meaningful. Similarly important is the consistent application of the predetermined rules for characterization of each error. In our setting, a team of quality technologists collectively and simultaneously perform the error data characterization and cataloging so that they maintain consistency of interpretation. We have also established a Quality Council, which consists of the medical director as chair, the administrator, all transfusion medicine staff physicians, and the Quality Team members. In addition, members of our Education Resource Team are included. This council acts as an oversight body to establish policies, review error data and other quality indicators, and review inspections and competency testing results. These same data are also discussed at a monthly quality meeting for all supervisors, who then disseminate the information to all technologists at their regular work area meetings. In our experience, this regular open discussion of errors and system failures is a vital part of the continuous education process within the Division of Transfusion Medicine. This multiteam approach, with overlapping membership for wide dissemination of data, discussion points, and ideas for change, creates an atmosphere in which all personnel feel free to comment and in which all members are encouraged to participate in planning for corrective action (Figure 2).

Figure 2.

A multiteam approach for analysis of errors. QA indicates quality assurance; FDA, Food and Drug Administration

Figure 2.

A multiteam approach for analysis of errors. QA indicates quality assurance; FDA, Food and Drug Administration

MANAGEMENT OF ERRORS

We reiterate the need for a nonpunitive approach only because of its critical role in establishing an atmosphere of open and frank discussion of errors. As mentioned, our management of quality has involved the establishment of clearly delineated sets of responsibilities for quality-specific activities and the foundation of an administrative infrastructure to achieve the following goals: (a) involve personnel at all levels; (b) establish clear lines of communication regarding quality issues; (c) provide oversight and accountability; and (d) provide guidance, policies, direction, and documentation.

The cornerstone of our program is that all errors are to be reported, documented, and characterized. These steps are followed by data analysis by the Quality Assurance Team, including the medical director and administrator. This analysis includes examination of trends and root-cause evaluation. These deliberations often result in 1 or more of the following: (a) collection of more data; (b) performance of focused audits; (c) restratification of data; (d) corrective action planning; (e) validation of corrective action; (f) implementation of corrective action; (g) reevaluation of results of corrective action; and (h) application of lessons learned to other aspects of operations.

Corrective action planning is often quite laborious, because one has to restrain one's natural inclination to apply the quick fix in order to get the satisfaction of seeing the problem drop from the radar screen. However, for corrective action to be most effective, the root-cause analysis has to be performed very thoroughly. Since most errors are the result of system failures, these systems must be carefully evaluated not only under the high-power microscope of the local work area review, but also under the low-power objective looking at all systems functioning both throughout the division of transfusion medicine and, in some cases, as they apply to the institution as a whole. For example, a transfusion to the wrong patient may result from an error in blood sample identification simply due to inadequate institutional policies and practices regarding patient wristband information and application. This is not to say that one should always be looking for a scapegoat outside one's own sphere of responsibilities, but rather that the big picture must always be kept in mind when dealing with errors, however local the problem may initially seem.

An example of this type of situation in our institution was that of the introduction in the mid-1980s of the same-day surgery concept. This practice change has often resulted in the patient arriving in the operating room at about the same time as their blood sample for antibody screening and crossmatching arrives in the laboratory. Appropriate institutional planning should have included discussions with all downstream service areas, such as the laboratories, radiology, etc, about the possible consequences of having patients admitted on the day of surgery. Corrective action for this particular problem in our division originally involved practice agreements with relevant surgical specialties to ensure that samples were obtained in a timely fashion or else surgery was scheduled for later in the day. This approach was initially successful.6,7 However, recent reappearance of the problem required a different approach. We recently implemented a presurgical sample program for surgery. This program, along with computer-assisted crossmatching, allows eligible patients to have their compatibility sample collected well in advance of their scheduled surgery. Implementation of these process changes has the potential to eliminate a large portion of the same-day surgery problems related to late arrival of testing samples in the laboratories. It is too early to judge whether this new practice will adequately address the problem. Yet another example of a problem that originated upstream from Transfusion Medicine was the increasing rate of blood orders for the wrong patient, which our error program detected a few years ago. By using control charts, we monitored this problem and timed its occurrence to coincide with a change in institutional practice, which had previously required both handwritten and stamped patient identifiers (name, unique number) on all blood requests. Removal of the requirement for handwritten information was associated with the increase in errors and its reinstatement with the return from 6/10 000 administrations to previous levels of 1/10 000 administrations (S. B. Moore, M. L. Foss, unpublished data, June 2003).

CREATING A QUALITY CULTURE FOR ERROR PREVENTION AND MANAGEMENT

For any management system to be effective, it must have a strategic framework and an administrative infrastructure. These are integral parts of the quality plan required under cGMP regulations. We believe that the most critical element in this strategic framework is the establishment of a quality culture. Accomplishing this goal is a lot more difficult than merely articulating a mission statement and declaring that the organization is strongly supportive of quality. It requires that, starting with top management, there be high levels of understanding of quality concepts and cGMPs, a demonstrated commitment to those concepts, a planning process to ensure their implementation, top-level realistic understanding of the resources necessary, and the determination to acquire such resources. We believe that this process of establishing a quality culture must be clearly seen to be driven by top management personnel who “walk the talk,” and it must provide for an extensive and ongoing educational effort to inculcate both knowledge and commitment throughout the workforce.

In essence, for the establishment of a quality culture throughout any health care institution there has to be a firmly held conviction that each individual employee is privileged to participate in the care of sick people, however far removed from direct patient contact the employee may be (eg, laboratories). To help ensure the long-term, consistent motivation necessary to maintain this culture, each employee must be trained to consider the importance of the care of every patient to be equivalent to that they would ardently wish for if that patient were their own child, parent, or spouse. The collective, committed, and cohesive willpower and inventiveness of virtually the entire work force at all levels of an organization will provide an enormous resource for accomplishing change and improvement. If they each consider what would be needed institutionally to ensure that their child or mother got the best care, then the direction of the entire entity will become more patient-oriented.

To instill this dedication to excellence, we have initiated a Quality School, whereby modules are developed for a wide variety of workforce categories. When individuals successfully complete their appropriate modules, they are then, and only then, authorized to perform the appropriate functions. Other measures that help inculcate the quality culture include the authors' meeting with each new employee within a few weeks of commencing work to discuss the fundamentals of our philosophy on quality, errors, and reporting as they relate to the overriding importance of providing only the best service to all patients. In addition, each morning there is a working/teaching conference for transfusion medicine technologists and residents, at which a transfusion medicine staff physician reviews challenging clinical cases of the prior 24 hours, options for providing service for ongoing cases, and any errors or practical problems already detected in their management. All serologic red blood cell workups seen in the reference/crossmatch laboratory (other than negative screens) and complicated crossmatches are also discussed and signed out at this daily conference. This work/education conference fosters real-time discussion of relevant current transfusion problems and recently detected errors in service, and encourages discussion of their management. Because of the enormous educational value of this conference, we frequently have medical students and medical residents from anesthesiology, hematology, and pediatrics in attendance. Their attendance and involvement in discussions not only helps foster good working relationships between our transfusion medicine physicians/technologists and the clinical services, but also educates the clinicians regarding issues such as patient and sample identification, informed consent, and transfusion medicine–based treatment options, for example, therapeutic apheresis and intraoperative salvage. It also provides an excellent opportunity to discuss and emphasize the need for greater clarity in communications between clinicians and the laboratories. Greater understanding of all of these issues helps promote attitudes that facilitate our overall efforts toward error prevention, detection, and management. The Table lists examples of this and other educational initiatives that foster the development and maintenance of a quality culture.

Examples of Educational Initiatives for Error Management

Examples of Educational Initiatives for Error Management
Examples of Educational Initiatives for Error Management

SUMMARY

For errors to be prevented, there must be an effective system for consistently sorting, characterizing, and cataloging errors in a timely fashion. There must be separation of responsibilities for some aspects of error data collection/processing and operations, but key operations personnel must be involved in the root-cause analysis and corrective-action planning, validation, and implementation. To facilitate these activities, a quality plan must be in effect to provide not only an essential administrative infrastructure, but also a mechanism for setting policies and strategic planning.

For the management of errors, there has to exist a quality culture that makes it plain that quality is not the prerogative of a few designated quality technologists or a quality team, but is the responsibility of each employee. Logically, there must be a nonpunitive approach to errors to foster reporting and open, frank discussions necessary for root-cause analysis and planning of corrective action. There must be a widespread enthusiasm for grasping the opportunities for improvement provided by the detection and appropriate analysis of error data. Finally, and perhaps most importantly, there must be strong leadership from top management with concomitant determination on its part to provide the resources necessary to establish the desired quality within all segments of the operations; that is, we need to put our money where our mouth is. The leaders must be so convinced of the financial value to the institution of quality that they can entice the ultimate financial decision makers to provide resources to do what is ethically and morally appropriate in terms of quality care for patients.

References

References
Taswell
,
H. F.
and
C. L.
Sonnenberg
.
Error analysis: types of error in the blood bank.
In: Smit Sibinga CTh, Taswell HF, eds. Quality Assurance in Blood Banking and Its Clinical Impact. Hingham, Mass: Martinus Nijhoff Publishers; 1984:227–237.
Motschman
,
T. L.
,
P. J.
Santrach
, and
S. B.
Moore
.
Error/incident management and its practical applications.
In: Duckett JB, Woods LL, Santrach PJ, eds. Quality in Action. Bethesda, Md: American Association of Blood Banks; 1996
.
Motschman
,
T. L.
and
S. B.
Moore
.
Error detection and reduction in blood banking.
Clin Lab Med
1996
.
16
:
961
973
.
Kaplan
,
H. S.
,
J. B.
Battles
,
T. W.
Van der Schaaf
,
C. E.
Shea
, and
S. Q.
Mercer
.
Identification and classification of the causes of events in transfusion medicine.
Transfusion
1998
.
38
:
1071
1081
.
Kaplan
,
H. S.
,
J. L.
Callum
,
B.
Rabin Fastman
, and
L. L.
Merkley
.
The Medical Event Reporting System for Transfusion Medicine: will it help to get the right blood to the right patient?
Transfus Med Rev
2002
.
16
:
86
102
.
Moore
,
S. B.
,
R. K.
Reisner
,
T. J.
Losasso
, and
S. K.
Brockman
.
Morning admission to the hospital for surgery the same day: a practical problem for the blood bank.
Transfusion
1987
.
27
:
359
361
.
Moore
,
S. B.
,
R. K.
Reisner
, and
K. P.
Offord
.
Morning admission for a same-day surgical procedure: resolution of a blood bank problem.
Mayo Clin Proc
1989
.
64
:
406
408
.

Author notes

Reprints: S. Breanndan Moore, MD, Division of Transfusion Medicine, Mayo Clinic, 200 First St, SW, Rochester, MN 55905 (moore.breanndan@mayo.edu)