Site Development Manual: Chapter 13

Exhibit 13-5: CONTINUOUS QUALITY IMPROVEMENT IN HEALTH CARE - A NEW APPLICATION OF AN OLD IDEA

by
Marc E. Babitz, M.D.

What is quality in health care? As health care professionals we believe that we can recognize quality in our health care delivery system when we see it, but can we define it? More important, how can we assure it, or manage it, or, best of all, improve it?

In this article I would like to suggest answers to these questions. First, I will share a definition of quality and compare the significant differences between traditional quality assurance and continuous quality improvement. I will then discuss and describe the application of continuous quality improvement methods in health care settings. In preparing this review article, I have drawn upon the works of several individuals (including Shewart, Deming and Juran) who initially developed these techniques for traditional business settings. In addition, I have borrowed from the writings of health care professionals (including Berwick, Benson, Kritchevsky, Simmons, Laffel and Blumenthal) who have applied these techniques to health care. Lastly, I must also give credit to Drs. Fried and Wallace who have provided much of my direct education on this topic. This article is intended to offer the reader enough information to actually begin implementing this process in their own health care setting.

Kritchevsky and Simmons have stated that "quality is like beauty: it has a positive connotation but denotes nothing measurable. For quality to be managed, it must be defined in terms of specific measurable attributes." This concept of specific and measurable attributes is a cornerstone for the development of a quality improvement program.

Continuous quality improvement (CQI) is actually a cog from a bigger management tool called Total Quality Management (TQM). While it isn't necessary to fully discuss TQM (whose concepts were developed by Shewart, Deming and Juran) we need to realize that this approach has been highly successful, and that at least 2 TQM principles are important to our discussion of CQI.

Does CQI and TOM work? I hope to convince you about the value and benefit of CQI in the health care setting; however, to appreciate the success of TOM, one need look no further than the success of Japanese automakers and electronics manufacturers over the past 10 - 20 years as compared with similar American producers. The means by which Sony and Honda became synonymous with quality and successfully penetrated world markets was in large measure due to their adoption of these management principles and tools.

To better understand CQI, we need to briefly discuss the TOM concepts of "customer" and the "role of management." Unfortunately the term customer is poorly accepted by health care professionals because of our desire to focus discussions about quality health care around the term "patient." The TQM approach suggests that all organizations, whether small (i.e. a solo provider's office) or large (i.e. a multi-specialty, multi-disciplinary practice), have customers. Customers can be described as internal and external. The external customer is the traditional recipient of our services, which for health care systems would be our patients. The internal customer represents those individuals and organizations upon whom we rely to provide the materials, services or other support necessary for our health care system to function. TQM teaches us the need to be customer focused; to continually assess the adequacy of our relationship with our customers and to make adjustments to our organization (i.e. health care system) which can improve those relationships. The concept of customer focus goes far beyond the traditional issue of concern with cost, to include concerns such as timeliness, accessibility, satisfaction, reliability and needs for training or education.

The traditional management structure (typical of American businesses prior to 1980) can be described by the pyramid on the left. In this structure the highest level executives were considered, borrowing a term from Dr. Wallace, to be the "Source of All Wisdom" (SOAW) regarding the operation of their organization. Since they knew what was best for their organization, the information flow in such a structure was primarily unidirectional (from the top down), as depicted by the arrow. While this type of structure may have been appropriate when we were dealing with smaller, less complex organizations in limited markets (e.g., the solo, general practitioner of yesteryear in a cash-for-service practice); today's managers and leaders are often faced with extremely complex organizations which operate in a variety of markets (e.g., the modern health care organization dealing with issues such as narrowly specialized staff, the demands of advanced technology, multiple payors of care and numerous systems of review and reporting). Even today's solo practitioner (admittedly an endangered species) could, in many respects, qualify as a complex organization in which one individual is unable to be an expert in all areas of the operation. For reasons such as these, TOM has developed a management approach best pictured as the inverted pyramid on the right. In this modern organization, we recognize that our front-line personnel (e.g., provider, receptionist, billing clerk) are likely to have the greatest expertise about the operation of their part of our organization. We recognize the need for a two-way flow of information between front-line personnel and managers, as depicted by the arrow. And, finally, we recognize that a critical part of management now must include the functions of coach, or facilitator, or supporter, or enabler, in addition to the traditional skills of organizational and financial management and leadership.

Now that we have offered a glimpse of some applicable TQM principles, let us return to our focus on CQI. To illustrate some key benefits of this approach it is helpful to contrast the CQI approach with the traditional Quality Assurance (QA) approach (see Figure 2). These points have been brought out by individuals such as Berwick, Kritchevsky, Simmons and Fried.

The first point has to do with standards. Traditional quality assurance has relied on the establishment of standards against which performance (usually that of a provider) is to be measured. These standards are often reflective of the recommendations of national, narrowly focused individuals or organizations (both private and governmental). In addition, these standards mayor may not be known to the involved parties prior to the actual performance review. The review determines whether the standard has been met (allegedly a sign of good quality) or has not been met (allegedly a sign of poor quality). Unfortunately, this type of judgment has little utility for most organizations. First, it is highly unlikely that these standards are truly applicable to every possible health care setting, considering the variances in resources and problems being faced. Second, the mere fact that an organization has met some arbitrary standard has little meaning in terms of quality until we know whether that organization is able or desirous of performing at an even higher level. CQI suggests (as Benson and others have noted) that standards should be established by the organization after input from appropriate staff. Based upon the results of the review for which the standard was set, the organization should determine their own level of satisfaction (i.e. appropriateness) with the standard. From this point, the organization can begin to set goals and objectives (specific and measurable) for quality improvement and pursue them within the limits of their resources and interest. In summary, the first point represents the choice between meeting arbitrary (outside) standards versus a plan for the continual improvement of current performance.

The second point has to do with the focal point for quality assessment. Traditional QA has focused on the performance of providers (and, in a limited way, support staff). In traditional QA we randomly audit charts to assess provider compliance with our "standard" in the hopes of ferreting out the non-compliant source of poor quality in our organization. If such an individual is discovered, they will be subjected to a variety of interventions, from benign (e.g., educational) to malignant (e.g., professional banishment). Kritchevsky and Simmons have pointed out the limited utility of this negative approach to quality management. CQI teaches us (as will be described later in this article) that the performance of systems has a much greater impact on measurable quality than the performance of individuals.

The third point has to do with the negativism of traditional QA as described above, versus the positive approach toward personnel which we learn from TQM. Traditional QA makes a negative assumption; namely, that the assurance of quality is dependent upon the uncovering of poor performers (i.e. our statistical outliers). In other words, one (or more) of us must be the cause (scapegoat) for a particular problem in our organization. The CQI approach makes the positive assumption that the vast majority of our staff want to do a good job (i.e.; do the right thing) in the delivery of health care services. The CQI approach has demonstrated that the major barriers to quality improvement in organizations are not the people, but rather the systems in which they work. We can have a much greater impact on quality in our organization if we support, coach, and educate our staff on the improvement of systems than by making negative assumptions about their performance or motivation.

To further influence the reader positively about the potential benefit of the CQI approach, I will share with you 2 examples which help illustrate the 3 points just made. The first is about Dr. Semmelweiss and puerperal sepsis, and the second is about a hospital ICU and central venous line infections.

As a medical student in the early 19th century, Dr. Semmelweiss made a troubling observation. He noted that the puerperal sepsis (a postpartum infection, nearly always fatal) rate for women delivered in his hospital was significantly greater than the rate for women delivered at home by midwives. Since the implication of this rather radical observation was unsuitable to the hospital physicians, Dr. Semmelweiss was later dismissed from that hospital. Some years later, around 1843 (which was well before the establishment of the germ theory of disease), Dr. Semmelweiss became chief of staff of his own hospital in Vienna. Again, he made a similar observation; namely, that women admitted to his lying-in (birthing) hospital who were attended to by physicians had a puerperal sepsis rate of approximately 27%, whereas the women who were delivered by midwives at home had a much lower death rate. Dr. Semmelweiss also made another observation about the delivery system in his hospital which differed from the system employed by the midwives. He noted that physicians were often performing other duties prior to attending deliveries. These duties included working in the anatomy and pathology laboratories, as well as attending other sick patients. He also noted that the midwives had a practice of always washing their hands prior to attending a delivery. If Dr. Semmelweiss had been a practitioner of traditional QA techniques, he would have undoubtedly conducted an audit of all the cases of puerperal sepsis to discover which physicians had the highest rates. He then would have instituted some form of disciplinary (or corrective) action toward those "outliers." Such a review would have had a minimal impact on the overall rate of puerperal sepsis for the patients of the remaining physicians. Instead, Dr. Semmelweiss chose a CQI approach which consisted of a system change. He ordered all physicians to wash their hands prior to attending a birth. I suspect that you know the rest of the story. In fact, the puerperal sepsis rate at this hospital dropped dramatically and approached the rate experienced by the midwives.

The second story involves an American hospital that was concerned about the rate of infection for central venous lines in their ICU patients. With a traditional QA approach, this hospital would probably have audited the charts of patients experiencing these infections to determine which physicians and/or nurses had the most patients with infections. Once these "outliers" were identified, they would probably have been subjected to some type of corrective intervention. Fortunately, this hospital had already adopted the CQI approach to problem solving and quality management. They carefully studied all aspects of this issue which included seeking input from front-line staff about possible system problems which might contribute to the infection rate. The end result of the process was that the hospital discovered that the infection rate for one particular brand (out of 3 or 4 total) of intracath was associated with much higher infection rates than the others. Without bothering to investigate the causal etiology for this observation, the hospital merely discontinued stocking this particular brand of intracath. Again, I suspect that you know the rest of the story. In fact, the overall infection rate for central venous lines in this hospital's ICU dropped significantly.

If I have, by now, convinced you that this approach has merit, or at least maintained your interest, we need to turn our discussion to the concept of a system. Much of the material for this portion of my article comes from a 1991 article by Dr. Kritchevsky and Dr. Simmons in the Journal of the American Medical Association. These authors defined a system as "a sequence of actions by, and interactions between, functional units that bring about the manufacture of a product or the delivery of a service." They further point out that systems have both explicit and implicit definitions. Explicit definitions of a system are those written rules, policies, procedures or protocols which describe how the system is to function. Implicit definitions of a system are not formalized or written and are based primarily on the social and cultural values and expectations which might impact on a particular system. The other important point made by these authors is that "no matter how brilliantly a system is designed, there will be a certain percentage of bad outcomes or defects." In fact, when we monitor systems we can expect to see variation about a norm (or median) value (see Figure 3). The fact that there is variation, and that the norm is not a "perfect" value, demonstrates that a system cannot always account for the innumerable potential mistakes which might occur during these complex interactions.

There are actually two different kinds of variation for which we can monitor. The first is variation due to systemic problems as described above. This type of variation is generally constant over time, with a fairly consistent norm and degree of variance from that norm. Again, this pattern of variation (Figure 3) reflects explicit and implicit interactions present in the system. On the other hand, we may find variations due to extra-systemic problems.

Extra-systemic problems produce different amounts of variation which tend to cluster by person, place, and time (see the change in Figure 4 as compared with Figure 3). An unfortunate example of extra-systemic variation occurred in a hospital which regularly monitored the death rates in their critical care units. When a substantial rise in this rate was noted, the hospital looked for some problem unrelated to the basic critical care system. In this case, they discovered a person who was hastening patients to their demise by giving them IV injections of KCl (potassium chloride). On the other hand, we are sometimes able to recognize extra-systemic variation of a positive nature. In that case, we can determine the cause and try to incorporate that activity into our regular system with the goal of improving our overall performance.

Juran writes about two ways for organizations to achieve gains in quality. The first is through careful monitoring and maintenance of current systems, and the second is through the implementation of system changes. System maintenance involves making sure that the desired rules and procedures are actually being followed, and that extra-systemic problems are either prevented or quickly dealt with. System changes offer the most fruitful area for substantially improving the quality we deliver to our customers as our preceding examples show. System changes are particularly effective because they improve the level of quality for everyone rather than limiting improvements to those affected by extra-systemic problems (i.e. by the outliers). Examples of system changes include enhanced training, technical innovations or procedural improvements.

Before implementing a CQI process in your center you need to understand the broad scope of activities to which this can be applied, and you need to understand the four basic tasks of the CQI process.

CQI can be applied to any, and every, aspect of your health center's operation. To help health professionals think beyond the performance of providers in seeking ways to improve quality, Dr. Dale Benson has developed the "Ten Parameters of Ambulatory Care" (see Table I). Each of these parameters offers opportunities for quality improvement activities. I would encourage health professionals to use this list (along with other quality of care materials by Dr. Benson) to help them move beyond the assessment of providers and staff as the only basis for a quality management program. Table II (from Benson) basically indicates how a parameter is broken down into its component parts which are more specific and measurable. Again, the term standard, as used in this context, refers to locally developed and agreed upon standards which mayor may not be consistent with those of another national organization.

Let us now turn to the four basic tasks of the CQI process: first, being able to separate externally caused problems from systemic problems; second, monitoring the system to make sure that no new extra systemic problems occur; third, studying the system to identify the source of problems and implementing appropriate changes; and, fourth, evaluating the efficacy of changes to the system. To perform these tasks effectively requires the use of new management and assessment tools. A discussion of the tools for CQI is well beyond the scope of this article; however, I hope that the reader's interest in this topic will cause them to seek further education and training in this area. On the other hand, anyone can begin to implement CQI with just the information presented here.

Steps one and two can easily be considered together since they involve the establishment of a mechanism to determine the current level of performance (i.e. the level of quality) of a particular system and the normal variance expected within the system. Once this is established, and norms are recognized, monitoring of this performance will document the present level of quality and identify the appearance of extra-systemic events (either positive or negative), as in Figure 4, which can be dealt with.

The third step truly focuses on quality improvement. If we have determined that we are not satisfied with the baseline information (either the norm or the degree of variation) obtained in steps 1 and 2, then we need to find opportunities for improvement within that system. The organization needs to explicitly understand the functioning of the system under study in order to identify all the potential factors which might contribute to a less than desirable level of quality. This type of analysis can only be successful when front-line staff is intimately involved in the process. Once the analysis is fairly complete, the study group should propose one or more modifications to the system (just as Semmelweiss proposed hand washing) which potentially could improve the level of quality. These proposals need to, be discussed and reviewed by management prior to implementation to determine the acceptability of their fiscal and personnel impact; however, management's prime role should be to encourage staff input and support suggested improvements whenever possible. The final stage of this step is to actually implement one or more changes to the system (e.g., washing hands, or not stocking a certain brand of intracath). This implementation process should be done in an explicit fashion. In other words, the new system should be clearly described in policies, procedures or protocols; and, it should be thoroughly discussed and explained to all personnel involved in the system.

The fourth task is to evaluate the efficacy of the changes which were made to the system. This is relatively easy since we already established our monitoring process for this system in steps 1 and 2. As we continue to monitor the results of the system (e.g., the incidence of puerperal sepsis, or the incidence of central line infections) we will be able to determine whether the changes have, in fact, improved our level of quality. An important point needs to be emphasized for this fourth step. As Kritchevsky and Simmons have stated, "large improvements following a system change are not to be expected; the ideal is continual improvement. The goal is improvement through a continuous series of incremental improvements rather than through great leaps."

Before offering some concluding remarks, I must make one final point regarding the CQI approach. As an, organization develops their monitoring systems as a prelude to implementing changes, we must decide exactly what we wish to evaluate. There are essentially 2 types of evaluation; process and outcome. Process evaluation measures whether the proposed change has actually been implemented as planned (e.g., Semmelweiss could have measured whether physicians were, or were not, washing their hands; and, our hospital could have measured whether or not any or our undesirable intracaths were still being used). On the other hand, outcome evaluation measures whether the implemented changes actually have had a positive outcome for our customers, in other words, resulted in an improvement in their health status. As you have surmised, this is what Semmelweiss did when he determined that the sepsis rate in his hospital dramatically dropped; and, what the hospital did when they determined that the central line infection rate had decreased. Clearly, outcome measurements are potentially more difficult to measure and are not always appropriate; however, we should strive to utilize them whenever possible since they more accurately document improvements in quality of care.

In conclusion, I will again borrow from the writing of Kritchevsky and Simmons. These authors have noted 3 unique obstacles (special considerations) to the implementation of CQI in health care settings, and they offer 3 fundamental lessons to be learned from CQI.

The first obstacle they describe relates to management structure. Industry usually deals with a single management structure, which may be true of ambulatory health care settings. However, when CQI is applied to hospitals (or larger ambulatory care organizations) there may be more than one management structure to deal with (e.g., the hospital administration, the medical staff and the nursing staff). The point is that all members of an organization must work together if CQI is to be successfully implemented.

The second obstacle relates to the role of management in industry. Managers in industry are not the front-line workers; however, in health care we often find individuals (especially providers) who function in both roles. Sometimes these conflicting roles for key staff may impede the CQI process. The point is to recognize this potential for conflict and assure that adequate front-line input (not skewed by a manager's role) is sought.

The third obstacle relates to our past experiences with Quality Assurance programs, as discussed near the beginning of this article. Since these were often punitive in nature, staff may be more interested in hiding problems as opposed to recognizing them as opportunities for system improvement. The CQI process requires open, honest and non-punitive analyses of the problems inherent in our systems before we can suggest means for improvement.

Finally, Kritchevsky and Simmons offer these three fundamental lessons to be learned from CQI.

1. "Quality means specific, measurable attributes. Without the ability to define quality unambiguously, quality management is impotent."
2. "Quality is an inherent feature of any system. Once providers of health care realize that better quality can be gained through system changes, the opportunities for improvement increase enormously because providers begin to consider what changes could be made so that bad outcomes would be prevented."
3. "Those who control the system control the level of quality. Quality is determined through the rules and procedures; therefore, those who set out the rules and procedures are ultimately responsible for the level of quality. By refocusing attention on the system of care, CQI provides a conceptual framework for the continuous improvement of health."

As health care professionals we are dedicated to providing our patients the highest quality of care available within the limitations of our environment. Continuous Quality Improvement offers us the philosophy, the process and the tools to reach this goal effectively and efficiently. I hope that you will give it a try.