Saravanan Kuppuswamy, MD

• Assistant Professor, Department of Internal Medicine
• University of Missouri
• Columbia, Missouri

It is estimated that there are 10 cholesterol test principle generic 40mg lipitor with mastercard, 000 or more health- Copyright National Academy of Sciences cholesterol levels measurement buy discount lipitor 5mg on-line. There is however cholesterol test scotland discount lipitor 10 mg free shipping, much variability in the accuracy and completeness of health information found on the Web (Biermann et al questran cholesterol medication generic lipitor 10mg without prescription. The effect of these trends on health care will be a fundamental transformation in the ways services are organized and delivered and clinicians and patients interact cholesterol hormones buy lipitor 5 mg with visa. Increasingly cholesterol levels for life insurance lipitor 10 mg without a prescription, they are also bringing information to their physicians to obtain help in interpreting or judging its value for themselves cholesterol medication causes memory loss discount lipitor 5mg mastercard. To better understand how information technology can contribute to improving quality high cholesterol chart australia generic 5 mg lipitor otc, the Committee on the Quality of Health Care in America held a workshop in September 1999 at which participants identified five key areas in which information technology could contribute to an improved health care delivery system: Access to the medical knowledge-base. Through use of the Web, it should be possible to help both providers and consumers gain better access to clinical evidence. Embedding knowledge in tools and training clinicians to use those tools to augment their own skills and experience can facilitate the consistent application of the expanding science base to patient care. The automation of patient-specific clinical information is essential for many types of computer-aided decision support systems. Automation of clinical data offers the potential to improve coordination of care across clinicians and settings, which is critical to the effective management of chronic conditions. Information technology can contribute to a reduction in errors by standardizing and automating certain decisions and by aiding in the identification of possible errors, such as potential adverse drug interactions, before they occur. Information technology can change the way individuals receive care and interact with their clinicians. Similarly, patients would be able to go online and obtain test results, inform their clinicians about how they are doing, send pictures and data, participate in interactive care management services, receive after-care instructions, and participate in support groups. Appropriately structured e-mail communication between patient Copyright National Academy of Sciences. In recent years, some applications have become commonplace, such as online searching for health information by patients and providers. Others, such as remote and virtual surgery and simulations of surgical procedures, are in early stages of development. Although opportunities to improve access, quality, and service abound, the health care industry has been slow to invest in information technology. In 1996, the industry spent only $543 per worker on information technology, compared, for example, with $12, 666 spent by securities brokers, and ranked 38th out of 53 industries surveyed (U. In a recent survey of 30 health plans, it was found that all had established Web sites to allow patients to obtain certain types of information and interact with the organization. But none had automated entire service functions, such as online medical management, which would require significant changes in business strategy, involve many employees and/or partners, and entail sizable capital investments. There are many technical, organizational, behavioral, and public policy challenges to greater use of information technology. Over the long run, however, organizational challenges may play the greatest role in constraining the adoption of various types of Internet applications. The diverse and highly decentralized structure of the health care industry, as discussed above, makes the business models for new applications complex and difficult, resulting in slow adoption of even highly successful pioneering applications. Efforts to introduce new applications also encounter resistance from health care professionals for a variety of reasons, including uncertainties about how such applications will alter relationships among and between clinicians, patients, and health care organizations (National Research Council, 2000). Fee-for-service payment, the most common method of payment for physicians, does not compensate clinicians for time spent on e-mail communication. State-based professional licensing requirements and restrictions on practice have stymied widespread use of other applications, such as remote medical consultations. Online access to and transfer of clinical information has also been slow to evolve, in part because of concerns about privacy and confidentiality. Chapter 7 reviews in greater detail the use of information technology to improve the quality of health care and some of the barriers to its more widespread adoption. Over the last decade, the primary impetus for change has been a desire to slow the rate of inflation of health care costs. During the coming decades, cost pressures will remain, but the health care system will also be shaped dramatically by broader forces transforming society in general, most notably the growth of the Internet and changing population needs for chronic care. Research documenting safety and quality concerns has been mounting for over a decade. Successful quality improvement initiatives are very slow to spread, and rarely adopted on a widespread basis. For these reasons, the committee believes that a more intense and far-reaching effort will be needed. Substantial improvement in quality over the coming decade can be achieved only by engaging the support of patients, clinicians, governing boards and managers of health care organizations, private and public purchasers, state and federal policy makers, regulators, researchers, and others. Change is needed at all levels, including the clinician and patient relationship; the structure, management, and operation of health care organizations; the purchasing and financing of health care; the regulatory and liability environment; and others. This report offers general principles, not a detailed blueprint, for the building of a new system. In part, the committee cannot foresee all the new organizations, forces, technologies, needs, and relationships that will develop even in the early years of the 21st century. More than that, however, the committee has come to Copyright National Academy of Sciences. In reshaping health care, local adaptation, innovation, and initiative will be essential ingredients for success. With these precepts in mind, the committee proposes the following agenda designed to bridge the quality gap: That all health care constituencies, including policymakers, purchasers, regulators, health professionals, health care trustees and management, and consumers, commit to a national statement of purpose for the heath care system as a whole and to a shared agenda of six aims for improvement that can raise the quality of care to unprecedented levels. The succeeding chapters of this report detail in turn the elements of this agenda. Specifically, the report: Sets performance expectations or aims for improvement for the 21stcentury health care system (Chapter 2). The current health care delivery system is not robust enough to apply medical knowledge and technology consistently in ways that are safe, effective, patient-centered, timely, efficient, and equitable. As we strive to close this gap, we must seek health care solutions that are patient-centered, that is, humane and respectful of the needs and preferences of individuals. And, most important, we must build a 21st century health care system that is more equitable and meets the needs of all Americans without regard to race, ethnicity, place of residence, or socioeconomic status, including the nearly 43 million people who currently lack health insurance (U. Studies of Educational Interventions and outcomes in Diabetic Adults: A MetaAnalysis Revisited. Intervention Strategies to Improve Adherence among Hypertensives: Review and Recommendations. Interpreting the Volume-Outcome Relationship in the Context of Health Care Quality. Collaborative Management to Achieve Treatment Guidelines: Impact on Depression in Primary Care. Compliance With National Asthma Management Guidelines and Specialty Care: A Health Maintenance Organization Experience. Requiring Physicians to Respond to Computerized Reminders Improves Their Compliance with Preventive Care Protocols. An Overview of Interventions to Improve Compliance with Appointment Keeping for Medical Services. Improvement of Hypertension Care by a Structured Treatment and Teaching Programme. The Determinants of Hypertension Awareness, Treatment, and Control in an Insured Population. The Effect of Patient-Carried Reminder Cards on the Performance of Health Maintenance Measures. Crossing the Quality Chasm: A New Health System for the 21st Century 2 Improving the 21st-Century Health Care System As discussed in Chapter 1, the American health care system is in need of major restructuring. As a statement of purpose for the health care system as a whole, the committee endorses and adopts the phrasing of the Advisory Commission on Consumer Protection and Quality in the Health Care Industry (1998). It is helpful to translate this general statement into a more specific agenda for improvement-a list of performance characteristics that, if addressed and improved, would lead to better achievement of that overarching purpose. Recommendation 2: All health care organizations, professional groups, and private and public purchasers should pursue six major aims; specifically, health care should be safe, effective, patient-centered, timely, efficient, and equitable. The committee believes substantial improvements in safety, effectiveness, patient-centeredness, timeliness, efficiency, and equity are achievable throughout the health care sector. This opportunity for improvement is not confined to any sector, form of payment, type of organization, or clinical discipline. Problems in health care quality affect all Americans today, and all can benefit from a rededication to improving quality, regardless of where they receive their care. The committee applauds the Administration and Congress for their current efforts to establish a mechanism for tracking the quality of care. Section 913(a)(2) of the act states: "Beginning in fiscal year 2003, the Secretary, acting through the Director, shall submit to Congress an annual report on national trends in the quality of health care provided to the American people. Fortunately, many of these encounters are effective and result in good outcomes, but such is not always the case. The following scenario, based on the composite experience of a number of patients, illustrates some of the serious problems facing patients and clinicians, problems that persist despite the widespread dedication of clinicians to providing high-quality care. Martinez, a divorced working mother in her early 50s with two children in junior high school, was new in town and had to choose an insurance plan. She had difficulty knowing which plan to select for her family, but she chose CityCare because its cost was comparable to that of other options, and it had pediatric as well as adult practices nearby. After receiving some recommendations from a neighbor and several coworkers, she called several of the offices to sign up. Although she knew nothing about the practice she finally found, she assumed it would be adequate. When she called for an appointment, she was told that the first available nonurgent appointment was in 2 months; she hoped she would not run out of her blood pressure medication in the interim. When she went for her first appointment, she was asked to complete a patient history form in the waiting room. She had difficulty remembering dates and significant past events and doses of her medications. Martinez called a site listed in her provider directory and was given an appointment for a mammogram in 6 weeks. Somehow, the films were never sent, and distracted by other concerns, she forgot to follow up. She hated even to think about having cancer in her body, especially because an older sister had died of the disease. For weeks she did not sleep, wondering what would happen to her children if she were debilitated or to her job if she had to have surgery and lengthy treatment. She was reluctant to call her mother, who was likely to imagine the worst, and did not know her new coworkers well enough to confide in them. It turned out that a possible abnormal finding had been circled the previous year, but neither she nor her primary care physician had ever been notified. Martinez had her appointment with the surgeon, and his office scheduled her for a biopsy. The biopsy showed that she had a fairly unusual form of cancer, and there was concern that it might have spread to her lymph nodes. She felt terrified, angry, sad, and helpless all at once, but needed to decide what kind of surgery to have. It was a difficult decision because only one small trial comparing lumpectomy and mastectomy for this type of breast cancer had been conducted. Martinez needed to have bone and abdominal scans to rule out metastases to her bones or liver. When she arrived at the hospital for surgery, however, some of this important laboratory information was missing. The staff called and hours later finally tracked down the results of her scans, but for a while it looked as though she would have to reschedule the surgery. This meant she had to see the surgeon, an oncologist, and a radiologist, as well as her primary care physician, to decide on the next steps. Wherever she went for care, the walls were drab, the chairs uncomfortable, and sometimes she would wait hours for a scheduled appointment. Martinez experienced many acts of consideration, empathy, and technical expertise for which she was grateful. Martinez, who had excellent health insurance and was seen by well-trained and capable clinicians, the system did not work and did not meet her needs. Neither she nor her previous primary care doctor had been notified of an abnormal finding on her earlier mammogram. Martinez was never confident that those directing her care had all the information about her previous care and its results. Prior to her surgery, critical laboratory information Copyright National Academy of Sciences. She was repeatedly required to tell her story, which became longer and more complex as time passed. It was not clear that her follow-up care consistently used the most up-to-date protocols. Yet treatments tried and proven futile in one admission would be recommended in the next as if they were fresh ideas. She had little assistance or information to help her understand the implications of choices about her surgery, radiation therapy, or chemotherapy. Although office and hospital staff focused on immediate medical problems, her discomfort, fear, and uncertainty were never addressed, and she was offered few resources to help her. Finally, her care was not efficient because much of its complexity and expense came from treating a tumor at a later stage than should have occurred. This is true even for patients with excellent insurance, in fine institutions, cared for by conscientious and well-trained clinicians. Common, too, is frequent inability of patients to make their needs understood, to be treated with respect and compassion, to learn what to expect about their health condition and treatment, and to have caregivers and institutions they can trust. These patients tell stories of fragmented care in which relevant information is lost, overlooked, or ignored; of wasted resources; of frustrated efforts to obtain timely access to services; and of lost opportunities. When clinicians and their families and those steeped in health management become patients, they, too, find that there appears to be no one who can make the systems function safely and effectively (Berwick, 1996, 1999; Khan, 2000; Singer, 2000). In this chapter, the committee puts forth six specific aims for improvement: health care should be safe, effective, patient-centered, timely, efficient, and equitable. These specific aims are intended to aid in achieving the overarching purpose stated in Recommendation 1 above. These aims are not new; they are familiar and have been valued, arguably for decades, among health care professionals, patients, policy makers, and communities. Yet American health care fails far too often with respect to these aims, despite its enormous cost and the dedication and good efforts of millions of American health care workers. After careful consideration, the committee has concluded that fundamental changes are necessary if our current health system is to achieve these aims. In its current forms, habits, and environment, American health care is incapable of providing the public with the quality health care it expects and deserves. The committee asserts, without reservation, that our health care can and should be far better than it is today, but it would be futile to seek that improvement by further burdening an overstressed health care workforce or by exhorting committed professionals to try harder. Instead, the improvements outlined here will require significant changes in the ways health care is organized, in the accessibility and usefulness of clinical evidence, in the environment of payment, and in other incentives that set the context for delivery of care. A redesigned care system can offer the health care workforce what it wants-a better opportunity to provide high-quality care. The ultimate test of the quality of a health care system is whether it helps the people it intends to help. This rather simple statement, as expanded upon in the following detailed discussion of the six aims for improvement set forth earlier, represents a major shift in thinking about the purpose of health care-a shift in attention from what is done to patients to what is accomplished for them. The committee believes the health care system should define safety, effectiveness, patient-centeredness, timeliness, efficiency, and equity using measures determined by the outcomes patients desire, although clinicians should not be asked to compromise their ethical values. Desirable personal health outcomes include improvement (and prevention of deterioration) of health status and health-related quality of life, and management of physical and psychological symptoms. The committee recognizes that the health of the public could be greatly improved by attention to and investment in a variety of areas, such as reducing violence and substance abuse and improving nutrition and transportation safety. This report, however, is focused specifically on the improvement of health care services to individuals. Safety Patients should not be harmed by the care that is intended to help them, nor should harm come to those who work in health care. The earlier report by this committee, To Err Is Human: Building a Safer Health System (Institute of Copyright National Academy of Sciences. In health care these errors include, for example, administering the wrong drug or dosage to a patient, diagnosing pneumonia when the patient has congestive heart failure, and failing to operate when the obvious (as opposed to ambiguous) signs of appendicitis are present. Processes also should not harm patients through inadvertent exposure to chemicals, foreign bodies, trauma, or infectious agents. The health care environment should be safe for all patients, in all of its processes, all the time. This standard of safety implies that organizations should not have different, lower standards of care on nights and weekends or during times of organizational change. To be safe, care must be seamless-supporting the ability of interdependent people and technologies to perform as a unified whole, especially at points of transition between and among caregivers, across sites of care, and through time. Specifically, in a safe system, information is not lost, inaccessible, or forgotten in transitions. Knowledge about patients-such as their allergies, their medications, their diagnostic and treatment plans, and their specific needs-is available, with appropriate assurances of confidentiality, to all who need to know it, regardless of where and when they become involved in the process of giving care. Ensuring patient safety also requires that patients be informed and participate as fully as they wish and are able. Patients and their families should not be excluded from learning about uncertainty, risks, and treatment choices.

Sizable benefits can be derived in the near future from automating certain types of data cholesterol score explained 20 mg lipitor visa, such as medication orders foods lower bad cholesterol fast discount lipitor 40 mg fast delivery. Efforts to automate clinical information date back several decades cholesterol lowering super foods buy generic lipitor 10 mg, but progress has been slow (Institute of Medicine definition of cholesterol test purchase lipitor 20mg on line, 1991) cholesterol levels and stress buy lipitor 40 mg low price, in part because of the barriers and risks involved cholesterol medication discount 5 mg lipitor with visa. An important constraint is that consumers and policy makers share concerns about the privacy and confidentiality of these data (Cain et al cholesterol test method buy 10 mg lipitor with mastercard. The United States also lacks national standards for the capture cholesterol levels 45 year old male buy lipitor 5mg on line, storage, communication, processing, and presentation of health information (Work Group on Computerization of Patient Records, 2000). The challenges of applying information technology to health care should not be underestimated. Health care is undoubtedly one of the most, if not the most, complex sector of the economy. Sizable capital investments and multiyear commitments to building systems will be required. Widespread adoption of many information technology applications will require behavioral adaptations on the part of large numbers of patients, clinicians, and organizations. Yet, the Internet is rapidly transforming many aspects of society, and many health-related processes stand to be reshaped as well. In the absence of a national commitment and financial support to build a national health information infrastructure, the committee believes that progress on quality improvement will be painfully slow. The automation of clinical, financial, and administrative information and the electronic sharing of such information among clinicians, patients, and appropriate others within a secure environment are critical if the 21st-century health care system envisioned by the committee is to be realized. Aligning Payment Policies with Quality Improvement Current payment methods do not adequately encourage or support the provision of quality health care. Although payment is not the only factor that influences provider and patient behavior, it is an important one. For example, fee-forservice payment methods for physicians and hospitals raise concerns about potential overuse of services-the provision of services that may not be necessary or may expose the patient to greater potential harm than benefit. On the other hand, capitation and per case payment methods for physicians and hospitals raise questions about potential underuse-the failure to provide services from which the patient would likely benefit. Indeed, no payment method perfectly aligns financial incentives with the goal of quality improvement for all health care decision makers, including clinicians, hospitals, and patients. This is one reason for the widespread interest in blended methods of payment designed to counter the disadvantages of one payment method with the advantages of another. Too little attention has been paid to the careful analysis and alignment of payment incentives with quality improvement. The current health care environment is replete with examples of payment policies that work against the efforts of clinicians, health care administrators, and others to improve quality. The following example, presented at an Institute of Medicine workshop on payment and quality held on April 24, 2000, 2 illustrates how payment policies can work against the efforts of clinicians, health care administrators, and others to improve quality: A physician group paid primarily on a fee-for-service basis instituted a new program to improve blood sugar control for diabetic patients. Specifically, pilot studies suggested that tighter diabetic management could decrease hemoglobin A1c levels by 2 percentage points for about 40 percent of all diabetic patients managed by the physician group. Data from two randomized controlled trials demonstrated that better sugar controls should translate into lower rates of retinopathy, nephropathy, peripheral neurological damage, and heart disease. First, expenses to conduct the project, including extra clinical time for tighter management, fell to the physician group. Second, over time, as diabetic complication rates fell, the project would reduce patient visits and, thus, revenues as well. But the savings from avoided complications would accrue to the insurer or a self-funded purchaser. The committee believes that all purchasers, both public and private, should carefully reexamine their payment policies. Recommendation 10: Private and public purchasers should examine their current payment methods to remove barriers that currently impede quality improvement, and to build in stronger incentives for quality enhancement. Payment methods should: Provide fair payment for good clinical management of the types of patients seen. The risk of random incidence of disease in the population should reside with a larger risk pool, whether that be large groups of providers, health plans, or insurance companies. Rewards should be located close to the level at which the reengineering and process redesign needed to improve quality are likely to take place. Substantial improvements in quality are most likely to be obtained when providers are highly motivated and rewarded for carefully designing and fine-tuning care processes to achieve increasingly higher levels of safety, effectiveness, patient-centeredness, timeliness, efficiency, and equity. Examples of possible means of achieving this end include blended methods of payment for providers, multiyear contracts, payment modifications to encourage use of electronic interaction among clinicians and between clinicians and patients, risk adjustment, bundled payments for priority conditions, and alternative approaches for addressing the capital investments needed to improve quality. Preparing the Workforce A major challenge in transitioning to the health care system of the 21st century envisioned by the committee is preparing the workforce to acquire new skills and adopt new ways of relating to patients and each other. At least three approaches can be taken to support the workforce in this transition. One is to redesign the way health professionals are trained to emphasize the aims for improvement set forth earlier, including teaching evidence-based practice and using multidisciplinary approaches. Second is to modify the ways in which health professionals are regulated to facilitate the needed changes in care delivery. Scope-of-practice acts and other workforce regulations need to allow for innovation in the use of all types of clinicians to meet patient needs in the most effective and efficient way possible. Third is to examine how the liability system can constructively support changes in care delivery while remaining part of an overall approach to accountability for health care professionals and organizations. The new rules set forth in this report will affect the role, self-image, and work of front-line doctors, nurses, and all other staff. The necessary environmental changes will require the interest and commitment of payers, health plans, government officials, and regulatory and accrediting bodies. The 21st-century health care system envisioned by the committee-providing care that is evidence-based, patientcentered, and systems-oriented-also implies new roles and responsibilities for patients and their families, who must become more aware, more participative, and more demanding in a care system that should be meeting their needs. And all involved must be united by the overarching purpose of reducing the burden of illness, injury, and disability in our nation. The committee envisions a system that uses the best knowledge, that is focused intensely on patients, and that works across health care providers and settings. Taking advantage of new information technologies will be an important catalyst to moving us beyond where we are today. The committee believes that achieving such a system is both possible and necessary. Crossing the Quality Chasm: A New Health System for the 21st Century 1 A New Health System for the 21st Century Fundamental changes are needed in the organization and delivery of health care in the United States. The experiences of patients, their families, and health care clinicians, as well as a large body of evidence on the quality of care, have convinced the Committee on the Quality of Health Care in America that the time for major change has come. In that year, three major reports detailing serious quality-of-care concerns were issued. The report describes the problem as follows: the burden of harm conveyed by the collective impact of all of our health care quality problems is staggering. It requires the urgent attention of all the stakeholders: the health care professions, health care policymakers, consumer advocates and purchasers of care. The challenge is to bring the full potential benefit of effective health care to all Americans while avoiding unneeded and harmful interventions and eliminating preventable complications of care. Meeting this challenge demands a readiness to think in radically new ways about how to 23 Copyright National Academy of Sciences. Our present efforts resemble a team of engineers trying to break the sound barrier by tinkering with a Model T Ford. That report calls for a national commitment to improve quality, concluding: "Exhaustive research documents the fact that today, in America, there is no guarantee that any individual will receive high-quality care for any particular health problem. The health care industry is plagued with overutilization of services, underutilization of services and errors in health care practice" (Advisory Commission on Consumer Protection and Quality in the Health Care Industry, 1998). The report on those results substantiates the serious and pervasive nature of quality-of-care problems. In the fall of 1998, the Committee on the Quality of Health Care in America established a Technical Advisory Panel on the State of Quality to review the most recent literature on quality. The detailed results of this review, now covering 8 years and more than 70 publications, are included in Appendix A. For example, problems with breast cancer care include underuse of mammography for early cancer detection, lack of adherence to standards for diagnosis (such as biopsies and pathology studies), inadequate patient counseling regarding treatment options, and underuse of radiation therapy and adjuvant chemotherapy following surgery. In its first report, To Err Is Human: Building a Safer Health System, this committee reviewed the literature on a specific type of quality problem-medical errors. We found about 30 publications published during the last 10 to 12 years Copyright National Academy of Sciences. These quality problems occur typically not because of a failure of goodwill, knowledge, effort, or resources devoted to health care, but because of fundamental shortcomings in the ways care is organized. Each of these factors plays a role, and each exacerbates the effects of the others. Our current methods of organizing and delivering care are unable to meet the expectations of patients and their families because the science and technologies involved in health care-the knowledge, skills, care interventions, devices, and drugs-have advanced more rapidly than our ability to deliver them safely, effectively, and efficiently (The Robert Wood Johnson Foundation, 1996). For more than five decades, investments in biomedical research have increased steadily, resulting in an extraordinary expansion of medical knowledge and technology (Blumenthal, 1994). Between 1994 and 1999, the budget of the National Institutes of Health increased from $10. Spending on research and development in the medical device industry, most of which comes from private sources, totaled $8. As suggested earlier, quality problems do not generally stem from a lack of knowledge, training, or effort by health professionals. Today, no one clinician can retain all the information necessary for sound, evidence-based practice. No unaided human being can read, recall, and act effectively on the volume of clinically relevant scientific literature. Since the results of the first randomized controlled trial were published more than 50 years ago (Cochrane, 1972; Daniels and Hill, 1952), health care practitioners have been increasingly inundated with Copyright National Academy of Sciences. Over the last 30 years, the increase in such trials has been staggering-from just over 100 to nearly 10, 000 annually. The first 5 years of this 30year period accounts for only 1 percent of all the articles in the medical literature, while the last 5 years accounts for almost half (49 percent) (Chassin, 1998), and there is no indication that this rate is slowing. Studies on the effectiveness of medical practice have also become increasingly sophisticated, involving complex issues of patient selection and statistical procedures. As the knowledge base has expanded, so too has the number of drugs, medical devices, and other technological supports. For example, the average number of new drugs approved per year has doubled since the early 1980s, from 19 to 38 (The Henry J. The cost of pharmaceuticals is the most rapidly growing component of health care expenditures. As clinical science continues to advance, the challenge of managing the use of existing and new pharmaceuticals and health technologies will intensify. Without substantial changes in the ways health care is delivered, the problems resulting from the growing complexity of health care science and technologies are unlikely to abate; in fact, they will increase. For example, work being done in genomics offers significant promise for disease diagnosis and, eventually, treatment. Engineering advances in miniaturization will place diagnostic, monitoring, and treatment tools directly into the hands of patients as science improves and costs are reduced. And the application of epidemiological knowledge to large populations and databases will enable us to understand more and more about the dynamics of wellness and disease. Increase in Chronic Conditions One of the consequences of advances in medical science and technology is that people are now living longer. Although health care is by no means the only factor that affects morbidity and mortality, innovations in medical science and technology have contributed greatly to increases in life expectancy. The average American born today can expect to live more than 76 years (National Center for Health Statistics, 2000). Roughly 1 additional year has been added to life expectancy every 5 years since 1965. Because of changing mortality patterns, those age 65 and over constitute an increasingly large number and proportion of the U. Today, this age group accounts for approximately 1 in 8 persons, or 13 percent of the population (National Center for Health Statistics, 1999). In 2030, when the large baby boom cohort has entered old age, 1 in 5 persons (20 percent) is expected to be in this age group. These demographic changes have important implications for the organization of the health care delivery system, but we have yet to address them in any Copyright National Academy of Sciences. One consequence of the aging of the population is an increase in the incidence and prevalence of chronic conditions. Chronic conditions, defined as illnesses that last longer than 3 months and are not self-limiting, are now the leading cause of illness, disability, and death in this country, and affect almost half of the U. About 100 million Americans have one or more chronic conditions, and this number is estimated to grow to 134 million by 2020 (The Robert Wood Johnson Foundation, 1996). About 1 in 6 Americans is limited in daily activities in some way as a result of a chronic condition (The Robert Wood Johnson Foundation, 1996). Disabling chronic conditions affect all age groups; about two-thirds of those with such conditions are under age 65. The majority of health care resources are now devoted to the treatment of chronic disease. In 1990, the direct medical costs for persons with chronic conditions was $425 billion, nearly 70 percent of all personal health care expenditures (The Robert Wood Johnson Foundation, 1996). The indirect costs-lost productivity due to premature death or inability to work-added another $234 billion to this figure. Providing state-of-the-art health care to a population in which chronic conditions predominate is complicated by the fact that many of those afflicted have comorbid conditions. About 44 percent of those with a chronic illness have more than one such condition, and the likelihood of having two or more chronic conditions increases steadily with age. In 1987, annual medical costs per person were more than twice as high for those with one chronic condition ($1, 829) as compared with those with acute conditions only ($817) (The Robert Wood Johnson Foundation, 1996). Annual medical costs per person increase much more for those with more than one chronic condition ($4, 672). Unlike much acute episodic care, effective care of the chronically ill is a collaborative process, involving the definition of clinical problems in terms that both patients and providers understand; joint development of a care plan with goals, targets, and implementation strategies; provision of self-management training and support services; and active, sustained follow-up using visits, telephone calls, e-mail, and Web-based monitoring and decision support programs (Von Korff et al. Much of the care provided to the chronically ill is given by patients and their families. Activities performed range from the provision of basic support care to active monitoring and management. Although some degree of collaborative management is essential to achieve desired outcomes for many chronic conditions, patients vary a great deal in the amount of information they want to receive on their condition and their desire to participate in treatment decisions (Strull et al. Nonetheless, the collaboration involved in much of the care provided to the chronically ill adds another layer of complexity to the delivery of health care to this growing segment of the population. In a survey of physicians practicing in community settings, nearly 40 percent were in onephysician practices, and more than four of five practiced in settings with fewer than ten physicians (American Medical Association, 1998). Hospital consolidation is occurring in many markets; of the more than 5, 000 community hospitals, 3, 556 belong to some form of network or system (American Hospital Association, 2000). The formation of physician organizations is occurring much more slowly, however (Kohn, 2000). The prevailing model of health care delivery is complicated, comprising layers of processes and handoffs that patients and families find bewildering and clinicians view as wasteful. Patients in a 1996 Picker Survey reported that the health care system is a "nightmare to navigate"-that it feels less like a system than a confusing, expensive, unreliable, and often impersonal disarray (Picker Institute and American Hospital Association, 1996). Care delivery processes are often overly complex, requiring steps and handoffs that slow down the care process and decrease rather than improve safety. These processes waste resources; leave unaccountable gaps in coverage; result in the loss of information; and fail to build on the strengths of all health professionals involved to ensure that care is timely, safe, and appropriate. In a population increasingly afflicted by chronic conditions, the health care delivery system is poorly organized to provide care to those with such conditions. The literature is replete with evidence of the failure to provide care consistent with well-established guidelines for common chronic conditions such as hypertension (Stockwell et al. Successful chronic care programs tend to be ones that incorporate guidelines and protocols explicitly into practice. Such reorganization generally involves the delivery of care through a multidisciplinary team, the careful allocation of tasks among the team members, and the ongoing management of patient contact (appointments, follow-up) (Wagner et al. A review of 400 articles, randomized trials, and observational studies of self-management support interventions (Center for Advancement of Health, 1996), revealed substantial evidence that programs providing counseling, education, information feedback, and other supports to patients with common chronic conditions are associated with improved outcomes (Brown, 1990; DeBusk et al. Evidence suggests that there are numerous ways to enhance access to such knowledge and expertise, including education of patients and primary care providers (Inui et al. Patient registries have been used effectively in many settings to issue reminders for preventive care and necessary follow-up, and to provide feedback to the provider practice on patient compliance and service use (Glanz and Scholl, 1982; Johnston et al. Mechanisms for sharing clinical and other information among all members of the care team, ranging from patientcarried medical records (Dickey and Petitti, 1992; Turner et al. Thus the American health care system does not have well-organized programs to provide the full complement of services needed by people with such chronic conditions as heart disease, cancer, diabetes, and asthma. Nor do we have mechanisms to coordinate the full range of services needed by those with multiple serious illnesses. And our current health system has only a rudimentary ability to collect and share patient information. A growing body of evidence for some procedures and conditions suggests that higher volume is associated with better outcomes (Hewitt, 2000). But the results are consistent with the conclusion that the growing complexity of health care necessitates more sophisticated and carefully designed care processes. The application of engineering concepts to the design of care processes is a critical first step in improving patient safety. Yet few health care organizations have applied the lessons learned by other high-risk industries that have led to very low rates of injury. These lessons include organized approaches to collecting data on errors and analyzing their causes, minimizing reliance on human memory, and standardizing routine aspects of care processes (Chassin, 1998; Institute of Medicine, 2000).

Letting n cholesterol medication with alcohol cheap lipitor 10mg fast delivery, the chance that the biased random walk on Z cholesterol levels range normal discount lipitor 10mg, when starting from 1 cholesterol jones and his band purchase 10mg lipitor amex, ever visits 0 equals cholesterol lowering foods shopping list generic 40mg lipitor otc. Therefore cholesterol ranges nhs cheap lipitor 10mg, ~ P0 {at least h - 1 returns of (Xt) to 0} = h-1 cholesterol in foods list 10mg lipitor amex, and consequently cholesterol lowering diet plans free safe lipitor 5mg, ~ P0 {Xt - St h} h-1 cholesterol blood test definition buy 40mg lipitor mastercard. The lazy random walk on the n-dimensional hypercube has a cutoff at (1/2)n log n with a window of size n. We write W for the stationary distribution of (Wt), which is binomial with parameters n and 1/2. The study of (X t) can be reduced to the study of (Wt) because of the following identity: Proof of (18. We construct now a coupling (Wt, Zt) of the lazy Ehrenfest chain started from w with the lazy Ehrenfest chain started from z. Provided that the two chains have not yet collided, at each move, a fair coin is tossed to determine which of the two chains moves; the chosen chain makes a transition according to the matrix (2. Supposing that (Zt, Wt) = (zt, wt) and > t, Dt+1 - Dt = 1 with probability (1/2)(1 - zt /n) + (1/2)wt /n, -1 with probability (1/2)zt /n + (1/2)(1 - wt /n). By the Markov property and because 1{ > t} is a function of (Z t, W t), 1{ > t}Ez, w [Dt+1 - Dt Zt, Wt] = 1{ > t}Ez, w [Dt+1 - Dt Z t, W t] (18. Thus, until time, the process (Dt) can be coupled with a simple random walk (St) so that S0 = D0 and Dt St. For a sequence of irreducible aperiodic Markov chains with (n) (n) (n) (n) relaxation times and mixing times , if tmix /trel is bounded above, then there is no pre-cutoff. Recall that we write an bn to mean that there exist positive and finite constants c1 and c2, not depending on n, such that c1 an /bn c2 for all n. The question remains if there are conditions which ensure that the converse of Proposition 18. Below we give a variant of an example due to Igor Pak (personal communication) which shows the converse is not true in general. Let be a family of transition matrices with trel = o(tmix) and with a cutoff. One step of the chain can be generated by first tossing a coin with probability 1/Ln of heads; if heads, a sample from n is produced, and if tails, a transition from Pn is used. If is the first time that the coin lands heads, then has a geometric distribution with success probability 1/Ln. As can be directly verified, 1 is an (n) ~ (n) ~ ~ eigenvalue of Pn, and:= (1 - 1/Ln) is an eigenvalue of Pn for j > 1. The lazy random walk on the n-dimensional hypercube has a separation cutoff at n log n with a window of order n. Recall that refresh is the strong stationary time equal to the first time all the coordinates have been selected for updating. Since, when starting from 1, the state 0 is a halting state for refresh, it follows that s1 (t) = P1 {refresh > t}. Let and dn denote the mixing time and distance to stationarity, respectively, for the n-th chain in a sequence of Markov chains. Show that the sequence has a cutoff if and only if n tn mix lim dn (ctn) = mix Notes 1 0 if c < 1, if c > 0. David Aldous (2004) created the chain whose transition probabilities are shown in Figure 18. The shape of the graph of d(t) as a function of t is shown on the bottom of the figure. Since the stationary distribution grows geometrically from left-to-right, the chain mixes once it reaches near the right-most point. It takes about 15n steps for a particle started at the left-most endpoint to reach the fork. With probability about 3/4, it first reaches the right endpoint via the bottom path. Thus the total variation distance will drop by 3/4 at time [15 + (5/3)]n, and it will drop by the remaining 1/4 at around time (15 + 6)n. Thus, the ratio tmix /tmix (1 -) will stay bounded as n, but it does not tend to 1. Recently, Lubetzky and Sly (2008) have announced a proof of cutoff for random regular graphs: Theorem (Lubetzky and Sly (2008)). Then with high probability, the simple random walk on G exhibits cutoff d at time d-2 logd-1 n with a window of order log n. Ding, Lubetzky, and Peres (2008b) analyzed the cutoff phenomena for birthand-death chains. For any 0 < < 1 there exists an explicit 2 c > 0 such that every lazy irreducible birth-and-death chain (Xt) satisfies tmix - tmix (1 -) c trel tmix (1). Then it exhibits cutoff in total-variation distance if and (n) only if tmix (n) tends to infinity with n. Furthermore, the cutoff window size is at most the geometric mean between the mixing time and relaxation time. Earlier, Diaconis and Saloff-Coste (2006) obtained a similar result for separation cutoff. This process can be modeled as a random walk on the wreath product G = {0, 1}V Ч V, whose vertices are ordered pairs (f, v) with v V and f {0, 1}V. There is an edge between (f, v) and (h, w) in the graph G if v, w are adjacent or identical in G and f (u) = h(u) for u {v, w}. In the configuration function f, zeroes correspond to lamps that are off, and ones correspond to lamps that are on. Let denote the transition matrix for the lamplighter walk, and let P be the transition matrix of the lazy simple random walk on G. That is, at each time step, the current lamp is randomized, the lamplighter moves, and then the new lamp is also randomized. We have used the lazy walk on G as the basis for the construction to avoid periodicity problems later. We write for the stationary distribution of P, and for the stationary distribution of. Since the configuration of lamps on visited states is uniformly distributed, allowing the lamplighter to walk for the cover time of the underlying walk suffices to Figure 19. In this chapter we study several connections between the underlying chain G and the lamplighter chain G. We have by now defined several time parameters associated with a finite Markov chain. Some measure mixing directly; others, such as the cover time and the hitting time, attempt to measure the geometry of the chain. In the next section, we prove that the relaxation time trel of the lamplighter walk is comparable to the maximal hitting time thit of the underlying walk (Theorem 19. The proofs of these results use many of the techniques we have studied in previous chapters. Then there exist constants c1 < c2 such that for sufficiently large n, c1 thit (Gn) trel (G) c2 thit (Gn). For the upper bound, we use the coupling contraction method of Chen (1998), which we have already discussed (Theorem 13. The geometry of lamplighter graphs allows us to refine this coupling argument and restrict our attention to pairs of states such that the position of the lamplighter is the same in both states. Let (Yt) be the Markov chain on G with initial distribution, so that Yt has distribution for all t 0. We write Yt = (Ft, Xt), where Xt is the position of the walk at time t, and Ft is the configuration of lamps at time t. To conclude that 2t +o(1) trel (G) (2+o(1))thit, it suffices to show that 2 hit 1/2. Note that for lamp log 2 1 configurations f and g on G, the norm f - g 1 is equal to the number of bits in which f and g differ. Let M = max f, g, x (Note that M is a restricted version of a Lipschitz constant: the maximum is taken only over states with the same lamplighter position. Couple two lamplighter walks, one started at (f, x) and one at (g, x), by using the same lamplighter steps and updating the configurations so that they agree at each site visited by the lamplighter. Because is an eigenfunction for, 2thit M = sup 2 sup 2thit (f, x) - 2thit (g, x) f -g 1 f, g, x K 2thit [(f, g, x), (f, g, x)] f, g, x f, g, x (f, x) - (g, x) f - g 1 f - g 1 f -g 1 M sup E f - g 1. Let (Gn) be a sequence of graphs with Vn, and let tcov be the cover time for lazy simple random walk on Gn. There exist constants c1 and c2 such that for sufficiently large n, the mixing times of the lamplighter family (G) n satisfy c1 t(n) tmix (G) c2 t(n). For a reversible chain, the separation and total variation distances satisfy Ї s(2t) 1 - (1 - d(t))2. By reversibility, P t (z, y)/(y) = P t (y, z)/(z), whence P 2t (x, y) = (y) P t (x, z)P t (z, y) = (y) (z) z z P t (x, z)P t (y, z). Subtracting both sides of the inequality from 1 and maximizing over x and y yields (19. For the lamplighter chain G on a finite graph G with vertex set V having V = n, the separation distance s (t) satisfies s (t) Pw {cov > t} (19. Since Pw {Xt = cov t} and are both probability distributions on V, we have Pw {Xt = wt cov t} (wt). Since the only way to go from all lamps off to all lamps on is to visit every vertex, we have Pw 2-n Pw {Xt = wt cov t} t ((0, w), (1, wt)) = (1, wt) 2-n (wt) Pw . Consider an irreducible finite Markov chain on state space with transition matrix P, and let cov be its cover time. If some states are missed in the first interval, then the probability that all are covered by the end of the second interval is at least 1/2, by the definition of tm. Note: throughout the proof, asterisks indicate parameters for the lamplighter chain. We will run the lamplighter chain long enough that, with high probability, every lamp has been visited and enough additional steps have been taken to randomize the position of the lamplighter. Define the probability distribution µun on G by s n (n) µun = P(0, v) {(Fun, Xun) cov = s}. For s un, conditional on (n) cov = s and Xs = x, the distribution of Fun equals n, the distribution of Xun is P un -s (x, ), and Fun and Xun are independent. Thus, µun = s xV (n) (n) P(0, v) {(Fun, Xun) cov = s, Xs = x}P(0, v) {Xs = x cov = s} (n) [n Ч P un -s (x, )]P(0, v) {Xs = x cov = s}. When Gn is the complete graph on n vertices, with self-loops, then the chain we study on G is a random walk on the hypercube- n although not quite the standard one, since two bits can change in a single step. The maximal hitting a o time is n and the expected cover time is an example of the coupon collector problem. Hence the relaxation time and the mixing time for G are (n) and (n log n), n respectively, just as for the standard walk on the hypercube. For the one-dimensional case, we note that HЁggstrЁm and a o Jonasson (1997) examined lamplighter walks on cycles. Here both the maximal hitting time and the expected cover time of the base graph are (n2)-see Section 2. Hence the lamplighter chain on the cycle has both its relaxation time and its mixing time of order (n2). For higher-dimensional tori, we have proved enough about hitting and cover times to see that the relaxation time and the mixing time grow at different rates in every dimension d 2. For the random walk (Xt) on (Z2) in which the lamplighter n performs simple random walk with holding probability 1/2 on Z2, there exist conn stants c2 and C2 such that the relaxation time satisfies c2 n2 log n trel ((Z2)) C2 n2 log n. These follow immediately from combining the bounds on the hitting time and the cover time for tori from Proposition 10. Notes the results of this chapter are primarily taken from Peres and Revelle (2004), which derives sharper versions of the bounds we discuss, especially in the case of the two-dimensional torus, and also considers the time required for convergence in the uniform metric. Random walks on (infinite) lamplighter groups were analyzed by Ka imanovich and Vershik (1983). They compute the spectra for the complete graph and the cycle, and use representations of wreath products to give more general results. Then for any vertices v, w and any lamp configuration f we have (n) t ((0, w), (f, v)) Pw {cov < (1 +)t(n) }2-n min P tcov (u, v), cov uVn (n) (n) (19. Definitions We now construct, given a transition matrix P, a process (Xt)t[0, ) which we call the continuous-time chain with transition matrix P. Continuoustime chains are often natural models in applications, since they do not require transitions to occur at regularly specified intervals. That is, each Ti takes values in [0, ) and has distribution function P = 1 - e-t 0 if t 0, if t < 0. Let (k) be a Markov chain with transition matrix P, independent of the rank=0 k dom variables (Tk). For an m Ч m matrix M, define the m Ч m matrix eM:= representation, Ht = et(P -I). Let P be an irreducible transition matrix, and let Ht be the corresponding heat kernel. Note that the above theorem does not require that P is aperiodic, unlike Theorem 4. Let P be an irreducible transition matrix, not necessarily ape~ riodic or reversible. Let P = (1/2)(I + P) be the lazy version of P, and let Ht be the heat kernel associated to P. Let Y be a binomial(4m, 1) random variable, and let = m be 2 a Poisson variable with mean m. P{ + m = 2m + j} 2m Here we write am bm to mean that the ratio am /bm tends to 1 as m, uniformly for all j such that j A m. Thus for large m we have j A m [P{Y = 2m + j} - P{ + m = 2m + j}] j A m P{Y = 2m + j}. First we show that shortly after the original chain is close to equilibrium, so is the continuous-time chain. Each of these steps is a step of P with probability 1/2 and a delay with probability 1/2; thinning a Poisson process of rate 1 this way yields a Poisson process of rate 1/2. We then claim that the continuous-time chain is close to equilibrium shortly after time k/2. P{ + m = k}P k, P{Y = k}P k, k0 P 4m = where is Poisson(m) and Y is binomial(4m, 1). Let P be a reversible and irreducible transition matrix with spectral gap = 1 - 2. Since u(0) = f 2, we conclude that 2 Ht f 2 2 2 2 = -2u(t), whence u (t) = u(t) f 2 -2t. The reader should check that (x)Ht (x, y) = (y)Ht (y, x), and so Ht fx (y) = Ht fy (x). Therefore, by Cauchy-Schwarz, Ht fx (y) - 1 = Ht/2 fx (z) - 1 Ht/2 fy (z) - 1 (z) 2 z Ht/2 fx - 1 the above with (20. The following gives good upper and lower bounds on tmix for this product chain. For a reversible transition matrix P with spectral gap, let n 1 ~ ~ P(n):= n i=1 Pi, where Pi is the transition matrix on n defined by ~ Pi (x, y) = P (x(i), y (i))1{x(j) =y(j), j=i}. The family of Markov chains with transition matrices P(n) has a cutoff at 1 2 n log n. To obtain a good upper bound on d(t) for product chains, we need to use a distance which is better suited for product distribution than is the total variation distance. For two distributions µ and on, define the Hellinger affinity as I(µ, ):= x (x)µ(x). If µ, then we can define g(x):= µ(x) 1{(x)>0}, and we also have the (x) identity dH (µ, ) = g - 1 2. The Hellinger affinity satisfies n n i=1 µ(i) and:= I(µ, ) = i=1 I(µ(i), (i)), and therefore n d2 (µ, ) H d2 (µ(i), (i)). The total variation distance and Hellinger distance satisfy If µ, then where g(x) = µ(x) (x) 1{µ(x)>0}. Let P be an irreducible reversible transition matrix, and let Ht be the heat kernel of the associated continuous-time Markov chain. Therefore, for x, y, n Px {X t = y} = Ht (x, y) = e(t/n)(Pi -I) (x, y) = i=1 n i=1 n i=1 i=1 (i) Ht/n (x(i), ), (i) ~ n Px {Xt/n = y (i) }. We outline below an alternative proof that Nt has a Poisson distribution with mean t; fill in the details. The chance of at least one transition in each subinterval is and the chance of more than one transition is O((t/)2). The number of transitions recorded in subintervals are independent of one another. Therefore, as 0, the total number of arrivals tends to a Poisson distribution with parameter t. Show that if µ = I(µ, ) = n i=1 n µi and = n i=1 i, then I(µi, i), i=1 n and therefore d2 (µ, ) H d2 (µi, i). Mixing of product chains is studied in detail in Barrera, Lachaud, and Ycart (2006). The Hellinger distance was used by Kakutani (1948) to characterize when two product measures on an infinite product space are singular. A classical example is the simple random walk on Zd, which we have already met in the case d = 1 in Section 2. This walk moves on Zd by choosing uniformly at random among her 2d nearest neighbors. There is a striking dependence on the dimension d: when d 3, the walk may wander off "to infinity", never returning to its starting place, while this is impossible in dimensions d 2. We still think of P as a matrix, except now it has countably many rows and columns. The matrix arithmetic in the finite case extends to the countable case without any problem, as do the concepts of irreducibility and aperiodicity. The joint distribution of the infinite sequence (Xt) is still specified by P along with a starting distribution µ on. Let Ak be the event that the walk started from zero reaches absolute value 2k before it returns to zero.

Discount lipitor 5 mg overnight delivery. HEALTH TIPS IN URDU/ HOW TO REDUCE CHOLESTEROL LEVELS QUICKLY BY HOME REMEDY.

In this case cholesterol test dubai generic 40mg lipitor free shipping, one approach is to start a random walk at v cholesterol test in bangalore purchase lipitor 20 mg fast delivery, allow the walk to explore the graph for some time cholesterol in chicken eggs 5mg lipitor mastercard, and observe whether the node w is ever encountered 2.8 cholesterol ratio good lipitor 5 mg fast delivery. If the walk visits node w cholesterol definition in hindi discount 40mg lipitor fast delivery, then clearly v and w must belong to the same connected component of the network cholesterol levels canada purchase 10mg lipitor with amex. On the other hand cholesterol chart uk generic lipitor 10 mg without a prescription, if node w has not been visited by the walk by time t cholesterol test water purchase lipitor 20mg otc, it is possible that w is not accessible from v-but perhaps the walker was simply unlucky. In particular, when w is connected to v, we desire an estimate of the expected time until the walk visits w starting at v. Given a Markov chain with state space, it is natural to define the hitting time A of a subset A to be the first time one of the nodes in A is visited by the chain. To connect our discussion of hitting times for random walks on networks to our leitmotif of mixing times, we mention now the problem of estimating the mixing time for two "glued" tori, the graph considered in Example 7. Let V1 be the collection of nodes in the right-hand torus, and let v be the node connecting the two tori. In view of this, it is not surprising that estimates for Ex (v) are useful for bounding tmix for this chain. For an irreducible Markov chain with state space, transition matrix P, and stationary distribution, the quantity Ea (x)(x) x does not depend on a. Observe that if x = a, Ea (x X1 = y)P (a, y) = (1 + hx (y)) P (a, y) = (P hx)(a) + 1, y so that If x = a, then + Ea (a) = y (P hx)(a) = hx (a) - 1. Hence we can define the target time of an irreducible chain by t:= Ea (x)(x) = E x (the last version is a slight abuse of our notation for hitting times). Since t does not depend on the state a, it is true that t = (x)(y)Ex (y) = E. For an irreducible Markov chain with state space and stationary distribution, thit 2 max E (w). Note that for a transitive chain, for any b, t = E = Hence we have Corollary 10. When the underlying chain is not transitive, it is possible for thit to be much larger than t. Consider the example of simple random walk on a complete graph on n vertices with a leaf attached to one vertex (see Figure 10. Let v be the leaf and let w be the neighbor of the leaf; call the other vertices ordinary. The first return time to v satisfies both + Ev v = Ev w + Ew v = 1 + Ew v (since the walk must take its first step to w) and + Ev v = Ea (x)(x) = x x, y (y)Ey (x)(x) = E (b). Clearly Ev v = 0, while Ev w = 1 and Ev u = 1 + Ew u, where u is any ordinary vertex. By conditioning on the first step of the walk and exploiting symmetry, we have 1 Ew u = 1 + (Ev u + (n - 2)Ex u) n 1 = 1 + (1 + Ew u + (n - 2)Ex u) n and 1 (Ew u + (n - 3)Ex u). Solving yields Ex u = 1 + n2 - n + 2 n2 - n + 4 = O(n) and Ex u = = O(n) n n (we only care about the first equation right now). Commute Time the commute time between nodes a and b in a network is the time to move from a to b and then back to a: a, b = min{t b: Xt = a}, (10. The commute time is of intrinsic interest and can be computed or estimated using resistance (the commute time identity, Proposition 10. In graphs for which Ea (b) = Eb (a), the expected hitting time is half the commute time, so estimates for the commute time yield estimates for hitting times. The following lemma will be used in the proof of the commute time identity: Lemma 10. Recall that R(a b) is the effective resistance between the vertices a and b in a network. Note that Ea (b) and Eb (a) can be very different for general Markov chains and even for reversible chains (see Exercise 10. A network G, {c(e)} is transitive if for any pair of vertices x, y V there exists a permutation x, y: V V with x, y (x) = y and c(x, y (u), x, y (v)) = c(u, v) for all u, v V. The reader should check that a random walk on a transitive network is a transitive Markov chain. The resistance R satisfies a triangle inequality: If a, b, c are vertices, then R(a c) R(a b) + R(b c). For a random walk on a transitive connected graph G, for any vertices a, b V, Ea (b) = Eb (a). For any three states a, b, c of a reversible Markov chain, Ea (b) + Eb (c) + Ec (a) = Ea (c) + Ec (b) + Eb (a). A natural approach to proving this is to assume that reversing a sequence started from a and having bca = n yields a sequence started from a having cba = n. For example, the sequence acabca has bca = 5, yet the reversed sequence acbaca has cba = 3. It turns out that it is much easier to start at stationarity, since it allows us to use reversibility easily. Recall that we use E to denote the expectation operator for the chain started with initial distribution. We say that if and only if is a subsequence of, that is, there exist indices 132 10. A binary tree after identifying all vertices at the same distance from the root 1 i1 < < im n with (k) = (ik) for all 1 k m. We let n denote the number of vertices and note that the number of edges equals n - 1. Identify all vertices at level j for j = 1 to k to obtain the graph shown in Figure 10. Using the network reduction rules, this is equivalent to a segment of length k, with conductance between j - 1 and j equal to 2j for 1 j k. Thus the effective resistance from the root to the set of leaves B equals k R(a B) = j=1 2-j = 1 - 2-k 1. Let x and y be two points at distance k 1 in the torus Zd, and let y be the time of the first visit to y. There exist constants 0 < cd n Cd < such that c2 n log(k) Ex (y) C2 n log(k + 1) 2 cd nd Ex (y) Cd nd 2 uniformly in k if d 3, if d = 2. There is a constant c1 so that for j c1 k, the ~ ~ box j is a cutset separating x from y. If the points x and y are the diagonally opposite corners of a square, the upper bound in (10. Now consider the case where x and y are in the corners of a non-square rectangle, as in Figure 10. Suppose that x = (a, b) and y = (c, d), and assume without loss of generality that a c, b d, (c - a) (d - b). The line with slope -1 through x and the line with slope 1 through y meet at the point z= By Proposition 9. When (c - a) and (d - b) have opposite parities, let x be a lattice point at unit distance from x and closer to y. Applying the triangle inequality again shows that R(x y) R(x x) + R(x y) 1 + 4 log(k + 1) 6 log(k + 1). Bounding Mixing Times via Hitting Times the goal of this section is to prove the following: Theorem 10. Consider a finite reversible chain with transition matrix P and stationary distribution on. Part (i) says that the total variation distance to stationarity starting from x, for reversible chains, can be made small just by making the return time to x close to its stationary probability. However, a similar inequality for the Cesaro mixing time t mix (defined in Section 6. Let P be the transition matrix for a finite reversible chain on state space with stationary distribution. Then K is reversible with stationary measure K given by K (x) = (x)/2 for x and K (mxy) = (x)P (x, y). Consider a finite irreducible aperiodic chain with transition matrix P with stationary distribution on. Thus, the constant sequence with every element equal to (x) is the convolution of the sequence with the sequence {uk + (x)}, so its generating function k=0 k=0 m = (x)(1 - s)-1 equals the product of the generating function F m=0 (x)s with the generating function [um + (x)]sm = U (s) + (x) m=0 sm = U (S) + m=0 (x). Differentiating the last equation at s = 1, we obtain that 0 = F (1)(x) - U (1), and this is equivalent to (10. Mixing for the Walk on Two Glued Graphs For a graph G = (V, E) and a vertex v V, define the set W = {(v, i): v V, i {1, 2}}, (10. Let H be the graph with vertex set W and edge set {{(v, i), (w, j)}: {v, w} E, i = j}. Then there is a coupling of two H random walks on H which has a coupling time couple satisfying u, vH G G H max Eu, v couple max E couple + max Ex v. If Xt = v, then toss a coin, and define Xt+1 = 0 (Xt+1, 1) if heads, 0 (Xt+1, 2) if tails. Until time G 0 couple:= min{t 0: Xt = Yt0 }, 0 define (Xt)tcouple and (Yt)tcouple by lifting the walks (Xt) and (Yt0) to H via G G the procedure described above. G G Until time G (v, 1):= inf{t couple: Xt = (v, 1)}, H couple (Yt) to (Xt) by setting Yt = ((Xt), 2). The expected difference couple - couple is bounded by maxxG Ex (v), whence for u, v H, H G Eu, v (couple) E(u), (v) (couple) + max Ex (v). Consider the problem of waiting for the sequence T T T to appear in a sequence of fair coin tosses. These waiting times are hitting times for a Markov chain: let Xt be the triplet consisting of the outcomes of tosses (t, t + 1, t + 2). Let G be a connected graph on at least 3 vertices in which the vertex v has only one neighbor, namely w. Consider simple random walk on the binary tree of depth k with n = 2k+1 - 1 vertices (first defined in Section 5. Write hm (k) for the expected hitting time from v to 0 for simple (that is, not lazy) random walk on the hypercube. Deduce that both mink>0 hm (k) and maxk>0 hm (k) are asymptotic to 2m as m tends to infinity. Show that Ea (bca) = [R(a b) + R(b c) + R(c a)] ce, eE where bca is the first time that the sequence (b, c, a) appears as a subsequence of (X1, X2. Show that for a random walk (Xt) on a network, for every three vertices a, x, z, Px {z < a } = R(a x) - R(x z) + R(a z). In the latter case the path from x to a to z involves an extra commute from z to a beyond time z. Suppose that is a sequence with generating function k A(s):= k=0 ak s and is a sequence with generating function B(s):= k l k=0 bk s. Let be the sequence defined as ck:= j=0 aj bk-j, called the convolution of and . Prove that the inequality tmix (1/4) 8 max E x + 1 x holds without assuming the chain is lazy (cf. Show that for simple random walk (not necessarily lazy) on Zn, with n odd, tmix = O(n2). That is, at each time t 1, the particle walks one step clockwise with probability p (1/4, 1/2), stays put with probability 1/2, and walks one step counterclockwise with probability 1/2 - p. Show that tmix (1/4) is bounded above and below by constant multiples of n2, but t (1/4) is bounded above and below by constant multiples of n. Suppose that is a strong stationary time for simple random walk started at the vertex v on the graph G. Show that starting from any vertex x in H, the random time v + is a strong stationary time for H (where is applied to the walk after it hits v). It is also instructive to give a general direct argument controlling mixing time in the graph H described in Exercise 10. For t > 2khmax + tmixG we have in H that Indeed for all x in H, we have Px {v > 2hmax } < 1/2 and iterating, Px {v > 2khmax} < (1/2)k. The mean commute identity appears in Chandra, Raghavan, Ruzzo, Smolensky, and Tiwari (1996/97). The cover time cov of (Xt) is the first time at which all the states have been visited. More formally, cov is the minimal value such that, for every state y, there exists t cov with Xt = y. We also define a deterministic version of the cover time by taking the expected value from the worst-case initial state: tcov = max Ex cov. It can be large enough for relatively small chains to arouse mathematical curiosity. For instance, we might view the progress of a web crawler as a random walk on the graph of World Wide Web pages: at each step, the crawler chooses a linked page at random and goes there. The time taken to scan the entire collection of pages is the cover time of the underlying graph. Lovґsz (1993) gives an elegant computation of the expected cover time tcov of simple random walk on the n-cycle. The walk on the remainders has covered all n states exactly when the walk on Z has first visited n distinct states. Let cn be the expected value of the time when a simple random walk on Z has first visited n distinct states, and consider a walk which has just reached its (n - 1)-st new state. The visited states form a subsegment of the number line and the walk must be at one end of that segment. The Matthews Method It is not surprising that there is an essentially monotone relationship between hitting times and cover times: the longer it takes to travel between states, the longer it should take to visit all of them. A walk covering all the states can visit them in many different orders, and this indeterminacy can be exploited. Randomizing the order in which we check whether states have been visited (which, following Aldous and Fill (1999), we will call the Matthews method-see Matthews (1988a) for the original version) allows us to prove both upper and lower bounds. Choose an arbitrary initial state x and let Sn be a uniform random permutation, chosen independently of the chain. For any two distinct states r, s, define the event A2 (r, s) = {(1) = r and (2) = L2 = s}. Conveniently, A2 = r=s A2 (r, s) Just as conveniently, Ac is the event that (2) is visited before (1). Thus Ex (T2 - T1) = Px (A2)Ex (T2 - T1 A2) + Px (Ac)Ex (T2 - T1 Ac) 2 2 1 thit. Thus Px (Ak) = 1/k and Ex (Tk - Tk-1) = Px (Ak)Ex (Tk - Tk-1 Ak) + Px (Ac)Ex (Tk - Tk-1 Ac) k k 1 thit. The proof above strongly parallels the standard argument for the coupon collecting problem, which we discussed in Section 2. For random walk on a complete graph with self-loops, the cover time coincides with the time to "collect" all coupons. Define cov to be the first time such that every state of A has been visited by the chain. When the elements of A are far away from each other, in the sense that the hitting time between any two of them is large, the time to visit just the elements of A can give a good lower bound on the overall cover time. Fix an initial state x A and let be a uniform random permutation of the elements of A, chosen independently of the chain trajectory. For example, when the underlying graph of (Yt) contains a leaf, tA = 1 for any set A containing both the leaf and its (unique) neighbor. Consider simple random walk on the rooted binary tree with depth k and n = 2k+1 -1 vertices, which we first discussed in Section 5. The maximal hitting time will be realized by pairs of leaves a, b whose most recent common ancestor is the root (see Exercise 10. For such a pair, the hitting time will, by symmetry, be the same as the commute time between the root and one of the leaves. Fix a level h in the tree, and let A be a set of 2h leaves chosen so that each vertex at level h has a unique descendant in A. Notice that the larger h is, the more vertices there are in A-and the closer together those vertices can be. If their closest common ancestor is at level h < h, then the hitting time from one to the other is the same as the commute time from their common ancestor to one of them, say a. We discuss the case of dimension at least 3 first, since the details are a bit simpler. Then tcov tA min 1 + 1 1 + + 2 A - 1 cd dnd log n(1 + o(1)), which is only a constant factor away from our upper bound. In this case the Matthews upper bound gives E(cov) 2C2 n2 (log n)2 (1 + o(1)), (11. To get a good lower bound, we must choose a set A which is as large as possible, but for which the minimum distance between points is also large. Assume 2 for simplicity that n is a perfect square, and let A be the set of all points in Zn both of whose coordinates are multiples of n. Here we examine the time required for all 2k patterns of length k to have appeared. In order to apply the Matthews method, we first give a simple universal bound on the expected hitting time of any pattern. Consider the Markov chain whose state space is the collection = {0, 1}k of binary k-tuples and whose transitions are as follows: at each step, delete the leftmost bit and append on the right a new fair random bit independent of all earlier bits. We can also view this chain as sliding a window of width k from left to right along a stream of independent fair bits. Black squares show the states unvisited by a single trajectory of simple random walk on a 75 Ч 75 torus. In the coin tossing picture, it is natural to consider the waiting time wx for a pattern x {0, 1}k, which is defined to be the number of steps required for x to appear using all "new" bits-that is, without any overlap with the initial state. Now fix a pattern x of length k + 1 and let x- be the pattern consisting of the first k bits of x. The additional time required is certainly bounded by the time required to construct x from entirely new bits.

Identification A technique used in ceremonial speaking to enhance feelings of familiarity and closeness cholesterol diet foods to avoid buy lipitor 40 mg free shipping. Keynote Address the keynote address represents the keynote of a larger idea taking place at a conference or exposition usually organized around a central theme cholesterol test that measures particle size cheap 5 mg lipitor free shipping. Magnification A technique in ceremonial speaking used to give benefit to the audience cholesterol food chart pdf buy discount lipitor 5mg, amplify emotion cholesterol i shrimp purchase lipitor 20mg with amex, and exceed audience expectations cholesterol lowering dog foods purchase lipitor 20 mg line. Presentational Speech Also called a speech to present an award cholesterol & your eyes buy lipitor 5 mg line, the presentational speech serves to highlight the merits of the award recipient and to point out the purpose and significance of the award being given cholesterol test understanding buy cheap lipitor 40mg online. Roast A roast is a variation of the toast in which the speaker pays tribute to a person by poking fun at her or him in a friendly way cholesterol test vap cheap 10mg lipitor overnight delivery. Special Occasion Speech A special occasion speech includes one of several kinds that celebrate an occasion. More specifically, it might introduce a speaker, entertain an audience, or inspire people. Speech of Introduction A speech of introduction is a brief presentation used to introduce the main speaker of an event and to inspire the audience to listen to that speaker. An experiment in collective wisdom from hundreds of undergraduate commencement speeches. Then you ask the question: "Is there ever a non-busy time for assembling a group together for a presentation? The combined expertise of several individuals is becoming increasingly necessary in many vocational (related to a specific occupation) and avocational (outside a specific occupation) presentations. Identify the differences between a small group, a team, and a speaking group Evaluate your individual presentation skills Describe the four coordination elements of group presentations List the four common types of group presentations Apply chapter concepts for coordinating group communication Discuss techniques for coordinating a group assignment Plan speech organization for the intended audience Practice effective group delivery chapter objectives chapter outline Introduction Communicating about group interaction o Interaction Roles o Decision Making o Conflict Resolution Preparing all Parts of the Assignment o Type of Group Presentations o Establishing Clear Objectives o Logistics for Group Members o Agreed Outcomes & Debriefing Organizing for Your Audience o Content o Structure o Packaging o Human Element Delivering Your Presentation as One Conclusion Review Questions and Activities Glossary References 2. Individual commitment to a group effort - that is what makes a team work, a company work, a society work, a civilization work. Also, the government, private, and public sectors have many committees that participate in briefings, conference presentations, and other formal presentations. It is common for group presentations to be requested, created, and delivered to bring together the expertise of several people in one presentation. Thus, the task of deciding the most valuable information for audience members has become a coordination task involving several individuals. All group members are responsible for coordinating things such as themes, strong support/evidence, and different personalities and approaches in a specified time period. Coordination is defined in the dictionary as harmonious combination or interaction, as of functions or parts. This chapter focuses on how the group, the speech assignment, the audience, and the presentation design play a role in the harmonious combination of planning, organization, and delivery for group presentations. In group presentations, there is often a tendency to put the focus on "presentation. Successful group work begins with something more than simply exchanging contact information. Thus, all your individual actions and words, including silence or no response, communicate something to others. This is why group members are disappointed when other members do not attend group meetings. Although "group" and "team" are often used interchangeably, the process of interaction between the two is different. Beebe & Mottet (2010) suggest that we think of groups and teams as existing on a continuum. On one end, a small group consists of three to fifteen people who share a common purpose, feel a sense of belonging to the group, and exert influence on each other (Beebe & Masterson, 2009). On the other end, a team is a coordinated group of people organized to work together to achieve a specific, common goal (Beebe & Masterson, 2009). Many-perhaps even most-vocational and avocational group members and size are determined by those who requested the group presentation. Whereas, vocational and avocational teams are guided by defined responsibilities for team members. For example, a public relations campaign team typically includes an account executive, research director, creative director, media planner and copywriter/copy editor. You may be most familiar with casual groups and social groups such as your fraternity or sorority or even your neighborhood. However, there are many types of groups formed everyday including committees, educational groups, problem-solving groups, task forces, work groups, and even virtual groups. In presentational speaking it is important to view the group as a speaking group, which is a collection of three or more speakers who come together to accomplish message content goals. In short, all groups require individuals to build harmony and rapport with one another but successful speaking groups are known more for their message continuity between speakers not the harmony between group members. Whether a negative or positive mental image, the image may be accurate of the past, but may have nothing to do with the current assignment. This icebreaking conversation can play a powerful role in your group, establishing a communication plan for cohesiveness, or the tendency for a group to stick together and remain unified in the pursuit of its instrumental objectives (Carron, Brawley, & Widmeyer, 1998) and minimizing social loafing, the decreased effort of each individual member as the number of a group increases (Tubbs, 1995, p. The conversation also will aid your group in a discussion concerning what communication vehicles and content will have priority for this speaking group. You will find, if you are honest, that 90% of the trouble is traceable to loafing. Benne and Sheats (1948) proposed a classification of roles in three broad categories: (1) task roles, (2) groupbuilding and maintenance roles, and (3) individual roles. Communication related to preparation include such things as guidelines for electronic information retrieval, sharing research information and visual aid content, and the scheduling of milestone appointments such as draft due dates and rehearsal 18-2 Group coordination is key in building message continuity. At its most basic level, group coordination focuses on group communication, "the process of creating meanings in the minds of others" (Tubbs, 1995, p. Such coordination requires establishing shared meanings about interaction roles, the decision-making process, and conflict resolution. In short, the purpose of group coordination is to assist you in establishing a communication plan. For many people, the mental image that forms when they hear they have been assigned to a group features some of their worst experiences or a quick Chapter 18 Group Presentations Task roles emphasizing organization focus on script development-cohesive language, transitions, and consistent graphics. Therefore, the group must plan on how they will identify and close gaps in content and support material. Finally, task roles at the level of delivery necessitates that the group communicate about assumptions, such as every individual is familiar with presentation software like PowerPoint or every individual is a regular user of the videosharing website YouTube. Other logistical challenges associated with delivery include planning the introduction of the group, where to stand, and equipment set up. Relationships within a group must be built and maintained simply because they are composed of individuals with different personalities, work styles, expertise, and availability. Your job as a group is to determine the best communication strategies for this speaking group. The strategies should support and enhance learning about and working with the differences. Although time restraints may limit the sophistication and quantity of your strategies, a communication plan for interaction roles should not be skipped. The best place to start is by selecting a group leader with the most appropriate leadership style to help the group maintain credibility within the group, among the audience, in the assignment and its assessment, and during the delivery. Thus, all group members should be aware of three small-group leadership styles - highly directive, participatory, and negligent (Brilhart, Galanes & Adams, 2001). A highly directive leadership style is where a leader uses an authoritarian method of dealing with group members. The participatory leadership style centers around a designated leader who offers guidance, suggestions, listening, and concern for members while also showing concern for completing the task. A negligent (or laissez-faire) leadership style is characterized by a leader who offers little guidance or direction. The group leader may guide the communication planning by first initiating a conversation about what communication media are accessible to group members. Some group members may not have access to a smartphone, text capability or all social networking sites such as Twitter, LinkedIn, and Facebook; and may not have consistent access to email or the Internet. For example, it is not uncommon for a student in a class to have Internet access only during open lab or library hours. Finally, keep in mind that some individual schedules or userstyles do not allow them to check email at the same daily frequency or dictate the same response style. All members should be careful not to criticize, judge or insult nonusers, limited users, and even overusers of technology. The gathering of contact information may be accomplished within the context of this conversation. The group leader can facilitate communication about member experience in the areas of presentation planning, organization, and delivery (see Table 18. Keep the focus on yourself by asking: What is my knowledge related to the specific assignment? Clearly think about: What degree of confidence do I need to develop about my own abilities? When you think about group coordination, decision-making is primarily about setting protocols-mutually agreed upon ways of interacting. As a group be very clear about how you will procedurally make decisions within this speaking group; and how the group will make decisions that require assimilating large amounts of information, exploring different ideas, Delivery techniques to be appropriate for the different stages of group presentations. Further, you will need to maintain ethical relationship boundaries with group members as appropriate to your interaction roles. Kelley (1992) suggests individuals be "skilled followers" who engage in two critical activities: (1) they are independent and critical thinkers, and (2) they actively engage in the work, rather than waiting to be told what to do. You can contribute best by being aware of and monitoring your strengths and weakness and the effect they have on group members. You will always have to apply and modify your individual knowledge, skills, and Yes/No & Either/Or: Most Common Types of Decisions Focus on whether a group should do something or not: Should we have handouts? Deciding between options: Should we use this inductive argument or that deductive argument? Should we use an operational definition or a logical definition to define this concept? Decisions put on hold until after certain decisions are met: Should we wait to determine visual aids until after we decide on how much technical language we use? Should we wait to determine the binding for the written document until after we know how many people will attend? The group may have to make decisions about the flow of information among members, proposed solutions, the quality of work, or even interpersonal relations among members. Finally, each group member should remain flexible and learn how to accept newness, incompleteness, and how not to blame others. Second, the amount of knowledge, understanding, and quality underlying a decision varies. Successful conflict resolution also involves developing a sound negotiating strategy, which involves the overall approach you take when you exchange proposals and counterproposals with another person when discussing a settlement to a conflict (Beebe & Mottet, 2010, p. By articulating a specific plan that addresses both conflict categories appropriately for this speaking group, group members gain a feel for what it will mean to balance between actively listening, doing his/her fair share, and soliciting comments throughout the process. The communication plan also may help your group reach consensus rather than engage in groupthink, which refers to a faulty sense of agreement that occurs when group members seemingly agree but they primarily want to avoid conflict (Beebe & Mottet, 2010, p. Overall group coordination will play a role in helping you reflect on group dynamics, plan for communication during group work, reinforce relationships, and establish a unified commitment and collaborative climate. Too many problem-solving sessions become battlegrounds where decisions are made based on power rather than intelligence. Wheatley conflict resolution Perhaps the greatest interpersonal skill needed is the ability to work compatibly with others, regardless of whether or not you like them personally (Lahiff & Penrose 1997). Just because you have worked in groups before does not guarantee you have experienced all types of conflict. The conflict of ideas and conflict of feeling (personality conflict) are most common among members. They include incompatible personalities or value systems; competition for limited resources Now it is time to think about the what of your presentation-the expected content. Many speaking groups are derived from an invitation to speak, and inherent in the invitation many times is a prescribed speaking 18-5 preparing all parts of the assignment Chapter 18 Group Presentations Thus, it is important to be clear whether the group is being assessed on product(s) or outcome(s) only or will the processes within the group-such as equity of contribution, individual interaction with group members, and meeting deadlines-also be assessed. Groups without a pre-assigned assessment rubric may use the three dimensions to effectively create a group evaluation instrument. The group should determine if the product includes both a written document and oral presentation. The written document and oral presentation format may have been pre-assigned with an expectation behind the requested informative and/or assignment-or topic. In group presentations, you are working to coordinate one or two outcomes- outcomes related to the content (product outcomes) and/or outcomes related to the group skills and participation (process outcomes). Therefore, it is important to carefully review and outline the prescribed assignment of the group before you get large quantities of data, spreadsheets, interview notes and other research materials. Finally, the symposium is a series of short speeches, usually informative, on various aspects of the same general topic. These four types of presentations, along with the traditional group presentation in front an audience or onthe-job speaking, typically have preassigned parameters. Therefore, it is important that all group members are clear about the assignment request. Not all group presentations require a format of standing in front of an audience and presenting. A structured argument in which participants speak for or against a preannounced proposition is called a debate. The proposition is worded so that one side has the burden of proof, and that same side has the benefit of speaking first and last. A panel consists of a group of experts publicly discussing a topic among themselves. Individually prepared speeches, if any, are limited to Failure comes only when we forget our ideals and objectives and principles. Sample Product Assessment Guide: Approach: Clarity: Development: Style: 18-6 Chapter 18 Group Presentations Although the two should complement each other, the audience, message, and format for each should be clearly outlined. Additionally, each group member should uniformly write down the purpose of the assignment. Yet the goal is for each member to consistently have the same outcome in front of them. This will bring your research, writing and thinking back to focus after engaging in a variety of resources or conversations. Once the assignment has been coordinated in terms of the product and process objectives, type of presentation, and logistics, it is important for the group to clearly write down the agreed outcomes. Agreed outcomes about the product include a purpose statement that reflects an agreement with the prescribed assignment. It also includes the key message or thesis to be developed through a presentation outline, a full-sentence outline of virtually everything the speaker intends to say. The outline allows the speakers to test the structure, the logic, and persuasive appeals in the speech (DiSanza & Legge, 2012, p. As a group decide the best way to leave enough time at the end to put all the pieces together and make sure everything is complete. If there is a written document, it should be completed prior to the oral presentation rather than at the same time. Next, the group can strategically add meeting dates, times, and venues to the action timetable. A meeting is a structured conversation among a small group of people who gather to accomplish a specific task (Beebe & Mottet, 2010, p. The more systematic a group is in these two areas, the more unified the process and the product. The system begins with each group member writing down the message, specific purpose, and central ideas for the group presentation. If these are still to be determined, then have each group member identify the areas of background information needed and basic information gathering. Next, simply create a general format for note-taking- whether typed or handwritten and what types of details should be included especially sources. Also with the increasing use of electronic databases be very clear on when related articles should be forwarded to group members. True genius resides in the capacity for evaluation of uncertain, hazardous, and conflicting information. All of this leads to the foundation of clearly defining the responsibilities of each group member. The group should be clear on what are individual, joint (involving more than one group member), and entire group tasks. The length of the presentation refers to your time limit, and whether there is a question and answer period involved. If there is a deadline, then the group begins by creating a schedule from the final deadline. As a group, create an action timetable explicitly listing all processes and outputs, as well as communication update points. So a schedule of who meets with whom and when is useful for planning work and agendas. For example, informational meetings may be called simply to update all group members; solicitation meetings are called to solicit opinions or request guidance from group members; group-building meetings are designed to promote unity and cohesiveness among group members; and problem-solving meetings result in making decisions or recommendations by the time the meeting convenes.