Many business owners pay a portion or all of their employees' insurance premiums. As the cost of providing health care benefits increases, they must raise the prices of their products or services and thus lose their competitive advantage. Some small businesses have even gone bankrupt.

Governments' Perspective

All levels of government are under a lot of pressure from different stakeholders. Many people feel that it is government's responsibility to provide health care coverage for all the uninsured patients or to reduce the financial burden of health benefits on business owners. The business owners' assumption is that this government action will allow business owners to reduce their costs and prices to make U.S. business more competitive in the world economy.

The overall impact of the different stakeholders each making decisions for their own best interest is that the current health care system is fragmented with no one accountable to provide integrated total care for the patients. To improve care, Lean, Six Sigma, and the Balance Score Card methodologies have been applied to healthcare but they have not created breakthrough results for the overall health care system. Lean, Six Sigma, and Business Process Re-engineering (BPR) have resulted in process improvements. However, these local improvements have not translated into significant reductions in cost or improvement in the health area. The cost of care is continually rising and overall stakeholder satisfaction with the health care services is very low.

Defining the Goal of the Health Care System

A clear goal and a vision of the future are prerequisites for any system to embark upon the Process of Ongoing Improvement (POOGI). One also needs a view of the system itself. In healthcare, the goal is to increase the percentage stock of the healthy population. Several factors7 are necessary to achieve this goal. The key factors are preventive care and velocity of cure rate as indicated in the system model provided in Fig. 31-1. The model is based upon a system dynamic model with rates of flow. The rate of identifying, treating, and preventing diseases will have a significant impact on the stock of health of our population. All strategies can be developed and stakeholder's interests can be aligned with the goal of the system.

Figures 31-1 and 31-2 give us a summary view of the medical practice system.

The model of the health care system in Fig. 31-1 points out that the improvements in healthcare must be done at the system level. Figure 31-1 shows that inputs affecting the disease rate (moving people from the healthy population state to the diseased population state) include the amount of preventative care, the environment of the populations, the genetics of the individuals, the lifestyle and psychosocial behavior of the individuals. The health care system provides inputs such as financial capability, system capability for quality/reliable care, system capability for rapid response to patient needs, and access to care to patients measured by the cure rate (defined as moving patients from the diseased population to the healthy population). The lower the cure rate, the higher the death rate. The model shows the significance of system capability to respond rapidly to patient needs and to provide quality and reliability of the disease management system in improving the cure rate. The goal of the health system as shown in Fig. 31-2 is to transform a patient from a diseased state to a healthy state as fast as possible. The system must be integrated with all of its parts functioning together and well to provide high quality, reliable treatment without unnecessary delays and to exceed patient expectations. The diagram further shows that all systems are made up of processes and subprocesses that can be broken down to the task or subtask levels of each stakeholder. The system must never be bogged down in the details of individual processes or lose sight of the main goal; that is, to serve the patient efficiently. The goal of aligned value chains is to satisfy patients' wants and needs FIGURE 31-1 Model of a health care system.

FIGURE 31-2 Process flow model of a health care system.

Improving Quality and Quantity of Patient Flow through Health Systems

Looking inside the systems pictured in Figs. 31-1 and 31-2, TOC's basic premise is that even the most complex systems have one key constraint or weakest link. For a POOGI, the Five Focusing Steps8 (5FS) (Goldratt, 1990b, 7) are useful in identifying and managing the constraint within the system and improving flow. The 5FS are: 1. Identify the system's constraint.

2. Decide how to exploit the system's constraint.

3. Subordinate everything else to the above decision.

4. Elevate the system's constraint.

5. If in the previous steps a constraint has been broken, go back to Step 1, but do not let inertia cause a system constraint.

Elaborating on the 5FS

The 5FS process assumes that there is a clear goal or vision of the system's performance: the goal of making more money (Throughput) now and in the future in a for-profit business. Further, since a constraint is the weakest link in a system, it determines the Throughput of the entire system. One strategy to increasing Throughput is to be much better than your competitors are in meeting the customers' needs.

The strategy of a small or large health care system or even the entire value chain should be to develop a decisive competitive edge (DCE) in providing a high quality, reliable delivery and health care system. The goals of each of the subsystems have to be consistent with the overall health care system in the larger context. Example: If the specialty hospital has cardiac surgery as the financial model, it should work within the umbrella of the epidemiology or disease management model. This hospital must work to improve the velocity of flow of diseased patients through the health care system but at the same time invest in research to prevent the cardiac disease; not encourage people to get sick so that the hospital can continuously make profits. The strategies of government, insurers, hospitals, private practitioners, and businesses must also be aligned to provide high quality, reliable health care on both the prevention and curative fronts to the patients.

Step 1: Identify the System's Constraint

In a complex health care system, there is one constraint that most influences the patients' flow. It is usually the most expensive resource: human, machine, or physical space. For example, in a small practice the constraint is usually the physician, dentist, chiropractor, or veterinary surgeon. In a larger system, it could be operating rooms, recovery rooms, emergency rooms, or CT/MRI machines. Ideally, the physician or surgeon performing the services (without whom the patients cannot flow) should be the constraint. However, due to the high investment in operating rooms, government regulations, and nursing or anesthesiology shortages, the constraint might be in one of these areas. The first challenge in the 5FS is to identify the constraint. To that end, a high-level value stream map (VSM) can help clarify the key obstruction or constraint to the flow of patients and information. The current or as-is VSM is provided in Figs. 31-3a and b and shows the flow of the patient, the value-added time (41 minutes), the wait time (52 minutes), the value-added quotient of 44 percent (the value-added time/total time in system), and the constraint (doctor) location in the system.

FIGURE 31-3 High level and lower value stream maps of health care systems.

Once we understand the relationship of the constraint to Throughput and the achievement of the system's goal, we can develop protection (a buffer for the constraint), and policies and procedures for the constraint and supporting staff to maximize the system Throughput.

In Fig. 31-4, the VSM can be drawn showing the complete value chain of health care businesses for a patient treatment. The flow of the patient through the supply chain will be dependent upon the dentist or lab when viewed from the larger system view. Similarly, the constraint to the flow of the patient in the hospital might be the capacity of the imaging department, the blood lab, or the recovery room nurses.

FIGURE 31-4 Value stream map of a series of businesses providing a patient treatment.

Step 2: Decide How to Exploit the System's Constraint

Once we have identified the constraint, we can determine how to make the constraint both effective and efficient. In our case, this is done by determining the best use of the doctor's time measured as Throughput/doctor time unit. Here we begin to see how combining TOC with other tools can be effective.

Example: Let us examine my practice-a dental surgeons' practice. We identify the dental surgeon as the constraint (the scarcest and most expensive resource). We next exploit the doctor's (dental surgeon's) time by using the Lean tools to effectively and efficiently use the doctor's time. We can implement the total kit concept, which ensures that all the documents, medical clearance, lab results, and imaging information are available to the doctor prior to seeing the patient. The Lean tools such as standard workflow and 5 S are useful in organizing our work-place to ensure everything is in its assigned place and is visually available for the doctor. Total preventive maintenance and mistake proofing ensure that the doctor's time is never wasted.

Six Sigma measures are implemented to ensure that processes are capable of achieving the desired results. DMAIC methodology is used to ensure the control of the system that protects doctor time utilization. DFSS (Design for Six Sigma) methodology can be helpful in redesigning certain processes where system capability is too low and new services have to be started to stay competitive. An example might be the use of patient focus groups to develop new services through quality function deployment (QFD). QFD methods can be utilized as in Fig. 31-5 to identify the needs of potential patients, referring doctors, and other healthcare stakeholders. Once the needs have been identified, the functional requirements and design parameters can be determined. See Fig. 31-6. This process links new services development with a company's strategic goals.

Improving the effectiveness of the constraint is a very important part of exploitation of the constraint. The product-mix decision identifies the services that we must process through the constraint to best achieve the system goal. We have to look at the goal and the supporting strategy to determine if this is the best action to increase Throughput per constraint (doctor) time unit. In Fig. 31-7a, a pie chart of the current distribution of doctor time is provided. The chart indicates the total revenues collected and approximate time for specific procedures based on time blocks in the schedule. Notice that a large amount of time is directed to facial trauma surgery (the least profitable service), while little time is devoted to wisdom tooth extraction and dental implants (the most profitable use of doctor time). In Fig. 31-7b, crowns and bridgework provide Throughput of $400 ($1000 $200 variable cost o 2 doctor hours = $400) and fillings and veneer services provide Throughput of $400 per doctor hour, while extractions/RTC provide $350 per hour and implants provide $250 per hour. Clearly, the surgeon should focus more on the crowns and bridgework and fillings and veneer services and less on the implants.

Data mining and an understanding of TA will help determine the services and patients that should be sought to provide consistency with the organization goal. Throughput/constraint unit time or doctor unit time (DU) is the key factor used in pie chart in Fig. 31-7. Instead of taking each procedure and each patient that is highly variable, we aggregated the data over time. If the total collections from trauma services divided by the doctors' time utilization is significantly lower than the T/DU from rendering services in other areas, the focus must be directed to those services with the higher Throughput.

FIGURE 31-5 Quality function deployment matrix sometimes called the house of quality.

FIGURE 31-6 Design process.

FIGURE 31-7 Current and ideal product mix.

There is another factor besides T/DU that is excessive: cost utilization of other resources. Some of the trauma cases require excessive paperwork, legal documentation, and court appearances by practice administrators or by the doctors to be paid for the services. This is a simplified version of activity-based costing, called CUT (cost utilization) in aggregate. In health care, the nursing staff, specialized billing or coding staff is an expensive resource. The cost utilization of these resources in addition to constraint resource time utilization can help with correct decision making about whether to perform or not to perform certain procedures. We also might have to make a decision about whether to perform certain procedures or refer the patient to someone else. We take into account our TA equation NP = T OE. If Throughput is the function of effective and efficient use of doctor time, and OE are all the salaries, utilities, cost of inventory, etc., we must take into account the large cost of administrative work required to do certain procedures. The increase in OE can offset the gains in Throughput.

This concept could raise many questions, but with payments capped on many procedures, for-profit health care organizations cannot survive without taking these things into account.

In the hospital context, with the operating room being the constraint, the different services such as oral and maxillofacial surgery, orthopedic surgery, neurosurgery, general surgery, plastic surgery, otolaryngology, urology, cardiothoracic, and gastroenterology should be evaluated based upon Throughput generated divided by the allocated time to the specialty services. Due to variability in patient demands in these services, most of the time these services either do not use their allocated time completely or they need additional time. Any percentage of time that was not utilized after allocation to a service or practice should be accounted as the time given to a specific service. Scheduling footprints (histories) can be developed with priority to the service that yields higher Throughput9 per unit of operating room time allocated and the service that has greater utilization of its block time.

Example: A community hospital has 10 operating rooms, which are the constraint in the system. The hospital is losing money and it has to improve its net profit (NP); otherwise, it will face closure. When we do the data analysis, we find that General Surgery has a time block of two operating rooms for two days. The Throughput or dollars collected from General Surgery is far lower than Throughput of Neurosurgery for the equivalent time block. The priority will be given to Neurosurgery. If General Surgery only utilizes 60 percent of its time and the demand for Neurosurgical patients is high, the hospital might take away General Surgery's unutilized time and give to Neurosurgery. If the hospital makes 75 to 80 percent utilization of a particular threshold, the hospital could then open the block times when another service is using less than its threshold levels. The hospital could also negotiate with staff, nursing, and administrative staff to open operating rooms for longer hours including Saturdays and Sundays. The goal should be to have flexible capacity to respond to patients' needs and wants.

In examining the data in Fig. 31-7, one is cautioned to examine related factors like customer service and patient's total comprehensive needs, which must be taken into account. We cannot always look at just one procedure in isolation of total patient care. That is why it is important to take the data for each procedure and view segments of population rather than dividing the total population of patients by each procedure. It is equally important not to be guided only by the details of this analysis by looking at just Throughput per unit of doctor time because it can result in partial care and dumping of patients on other practitioners, which can have serious negative effects. Example: A practitioner selects the higher reimbursement procedures over the low reimbursement procedures and sends those procedures to other specialists. A maxillofacial surgeon in private practice can refuse to treat facial trauma patients and send them to plastic surgeons or otolaryngologists or vice versa. This might not be consistent with customer service and reputation goals.10 FIGURE 31-8 Scheduling the doctor's time based on buffers and BM.

Now having some idea of the type of exploitive steps that might be taken by physicians and hospitals, we move on to examine what it means to "subordinate."

Step 3: Subordinate Everything Else to the Above Decision

TOC offers the following methods to subordinate to the constraint: DBR, CCPM, and BM. In scheduling the patient with the doctor, the schedule procedure should be setup such that the doctor's time is fully utilized. See Fig. 31-8. Once a time in the doctor's schedule has been identified, the patient is given an appointment (arrival) time such that he or she should arrive at the office with ample time to sign-in, show insurance card, fill out forms, be shown to the examining room, and be prepared for the doctor's arrival. On average, the patient should have a short wait prepped in the examining room prior to the doctor's arrival. This short wait is provided such that Murphy may occasionally strike, but the doctor performing his or her procedure is not delayed. Both the appointment time schedule and the checkout schedule are derived from the doctor's schedule. All resources in the process should have ample capacity to respond to unexpected events (Murphy) and should do everything possible to keep the doctor on schedule. This extra or protective (or sprint) capacity of all supporting resources is available in case it is needed. This is subordination to the constraint. The buffer in Fig. 31-8 is the time it takes to get the patient to reach the doctor. There is much variability in patient's arrival time, the skill sets of multiple staff members interacting with the patients, patient personalities, their medical conditions, the mental conditions of patients and staff on the particular day of interaction, etc. This interaction of variability among various factors results in delays or queuing in front of workstations. TOC provides techniques and tools of managing this interacting variability using buffers and BM reports. The buffers are placed strategically to protect the constraint resource-the doctor time. An experienced staff member takes the role of flow or buffer manager. He or she has two goals on a daily basis-to ensure that the doctor's time is efficiently utilized and the patient is not in the system longer than he or she expected. If the doctor has scheduled short procedures every 30 minutes, then we have 30 minutes to get the patient from arrival time at the receptionist to the doctor. We have buffer time of 30 minutes with 10 minutes of green zone, 10 minutes of yellow zone, and 10 minutes of red zone. If the patient arrives 15 minutes late, we must expedite this patient by doubling up the resources or doing several tasks parallel to ensure that the patient reaches the doctor in 15 minutes when he or she is done with the first procedure. Protection of doctor time is the priority. Similarly, the checkout or discharge is also important so that the patient is not waiting in the system after the doctor care is completed. The buffer reports tell us the trends where we have delays. If we have check-in workstation delays, we provide staff training to identify and eliminate the delay; we then implement the Lean systems (5 S, mistake proofing, setup reduction, kitting, etc.) and re-evaluate. If we work on changing patient behavior to come on time by reminding them by phone, e-mails, text messaging, or penalizing late arrivals, then after a while, we could start reducing the buffer time when we have control over the internal variability.

In Fig. 31-9a, four patients are already scheduled by CCPM. The most heavily used resource in the system is black, the strategic resource: the doctor. These networks provide the basis for scheduling the doctor's time throughout the day. If the doctor is the black resource, he or she cannot be with four patients at the same time so some shifting of the networks based on the black resource must be performed. In Fig. 31-9b, the doctor time from each network is shown. Notice that the doctor is fully utilized most of the time.

In Fig. 31-9a, the black resource is the doctor and he or she is supported by resources shown in other colors. The usual scenario in health systems is multitasking. The doctors and other resources are jumping back and forth to different patients without completing a single patient. This results in delays for everyone. We believe that the solution better than the DBR explained previously is CCPM. Each patient is unique and multiple providers or support staff have to work on them to get the Throughput. Multiple patients enter our system (practice) and several staff members work on these patients simultaneously. The system is prone to multi-tasking and unnecessary delays. CCPM for multiple projects with short durations can be utilized effectively to flow the patients rapidly. The Critical Constraint Resource is shown as black in Fig. 31-9a. As we can see, the black resource is overlapping in all four patients. This is overly optimistic scheduling that will result in delays, and the patients will be upset. Figure 31-9b is the first attempt to start scheduling the patient by staggering the schedule based upon black constraint resource or doctor time. Usually after three to four patients, a buffer is kept to absorb variability and Murphy which accumulated across patients. The buffers can be dynamically designed based upon customer input. If the patients start complaining within 30 minutes or 15 minutes of wait time, the psychological management of queues could be implemented. Usually for different procedures, the customers have different tolerance levels for waiting.11 FIGURE 31-9 Scheduling the doctor's time based on patient critical chain networks.

This mapping of the networks for all procedures can be performed manually and then the networks shifted around to fully utilize the doctor's time, but software programs are being developed with patient care mapped out as a project. Multiple patients or patients with different needs and wants flow through our systems. We schedule a finite time that it will take to ensure that the patient exits the system within the promised time and quality of outcome. The project must start with an understanding of the patient's expectations in addition to the medical diagnostic test results. The necessary conditions like finances (insurance, Medicare, etc.), time available and required, and patient's existing medical condition are identified prior to starting diagnostic tests and treatment plans. After these initial steps, the best treatment designs or plans are chosen based upon the evidence-based medicine. Part of the treatment plan must take into account the patient's inability to understand these complex concepts about their own care. Increasing the patient's understanding or comprehension about the solutions to his or her problems will be an important step in the execution of the project for patient care so that we get full compliance from the patients.

Step 4: Elevate the System's Constraint

Elevate the constraint when we need to increase system capacity or make significant investments to offload the constraint time. To understand our investment options for elevating the constraint, we need to understand something about the TA terms to be used. We defer these now, but examine TA in more detail later in the chapter. For discussion of the Exploit step, we only need to understand the following accounting terms. Definition of these terms is woven into our discussion of the Exploit decision.

First, we must look at the impact on NP and return on investment (ROI) in making the elevate decision.

T = PriceTotally Variable Cost

NP = T OE (Net Profit = Throughput12 Operating Expense13)

ROI = NP/I (Return on Investment = Net Profit/Investment)

Keeping this in mind, we can ensure that all of our investments in elevating the constraint will result in increases in Throughput greater than increases in OE and the ROI greater than cost of capital.

Step 5: If in the Previous Steps a Constraint Has Been Broken, Go Back to Step 1, But Do Not Let Inertia Cause a System Constraint

Sometimes the environment changes or in implementing Step 4 Elevate, the constraint moves. In these cases, one should go back to Step 1 Identify. Changes in reimbursement or regulations from insurance companies or Medicare can cause changes in product mix, for example.

This Five Focusing Step process (5FS) is one of the TOC POOGIs.

Thinking Processes14 for Identifying Root Cause of Physical Constraints to the Flow of Patients