Early Symptoms Of Arthritis

The Quality of Healthcare

This Knol describes a revolution in healthcare -- our new understanding that the quality of care is not what it should be, and our recent efforts to measure quality and improve it. We review the challenges of quality measurement (including the famous Donabedian triad of process, structure, and outcomes) and consider the various tools for quality improvement (including public reporting ["transparency"] and pay-for-performance). The Knol concludes with a discussion of "value" -- namely, quality divided by cost.

Introduction

We would all like to believe that the quality of our medical care is excellent - particularly when we're in an Emergency Room with crushing chest pain or an anesthesiologist is placing a mask over our face and asking us to count backward from 100. And, in fact, modern medicine saves lives and preserves function millions of times each year.

Yet, we now know that the picture is far from the idyllic view we once naively (and perhaps wishfully) held. Over the past two decades, scores of studies have demonstrated that the quality and safety of modern healthcare leave much to be desired. This Knol begins by considering the issues surrounding healthcare quality - how we think about it, measure, it, and ultimately improve it. It concludes with some observations regarding value, where quality and cost intersect. Although I will touch on medical errors and patient safety (a subset of the broader issue of quality), they are covered more fully in a dedicated Knol.

What is Healthcare Quality?

Quality of care has been defined by the Institute of Medicine (IOM) as "the degree to which health services for individuals and populations increase the likelihood of desired health outcomes and are consistent with current professional knowledge." In a major report published in 2001 ("Crossing the Quality Chasm" (1)), the IOM set forth six aims for a quality health care system (Table 1): patient safety (freedom from medical errors), patient-centeredness (a focus on the patient's needs and wishes), effectiveness (the degree to which care is consistent with the best available evidence), efficiency (the degree to which care does not squander scarce resources), timeliness (the degree to which care is delivered as promptly as is required for a given situation), and equity (the degree to which care is fairly delivered, with no disparities, for example, by race, socioeconomic class, or insurance status).

Table 1. The Institute of Medicine's Six Quality Aims
1. Patient Safety
2. Patient-Centeredness
3. Effectiveness
4. Efficiency
5. Timeliness
6. Equity

Even though the "six aims" make it clear that quality of care goes well beyond simply providing evidence-based care, evidence does provide the scaffolding on which most quality measurement and improvement activities are built. Figure 1 shows the yearly growth in the numbers of published randomized clinical trials (a study in which alternative strategies are pitted against each other, with patients randomly allocated to the treatment "arms" to ensure that the results are due to differences in the clinical strategy rather than differences in the patients themselves), the gold standard for evidence-based medicine. This type of research has helped define evidence-based practices in many areas of medicine, from preventive strategies for a healthy 52-year-old healthy woman to the treatment of the 30-year-old man with liver failure or the infant born 10 weeks prematurely.

Our Growing Understanding of Healthcare's Quality Problem

The first major chink in the healthcare quality armor came in the form of studies authored by Dartmouth's Jack Wennberg (photo at left) beginning in the 1970s and 80s, which showed remarkable variations in patterns of care that were neither supported by evidence nor justified by outcomes. For example, children in one Vermont town had rates of tonsillectomies five times higher than children in neighboring towns. New Haven residents were more than twice as likely to receive coronary artery bypass surgery and 50% more likely to undergo hysterectomy than similar patients living in Boston. This line of research, known as "small areas variations" (2), raised a fundamental question: how can doctors and hospitals be practicing high quality, evidence-based medicine yet have such stunningly different approaches to the same problem?

Wennberg's studies helped launch a field we now call "outcomes research." A generation ago, the correct approach to the woman with uterine fibroids, the child with a sore throat, or the man with prostate cancer was determined by a handful of prominent experts (a process now deridingly termed "eminence-based medicine, " to contrast it with today's favored "evidence-based medicine.") Outcomes researchers sought to determine, through rigorous studies, what truly were the best approaches, and to codify this information in the form of "practice guidelines." This was not an entirely new concept - we've long had vaccination guidelines that tell parents that their child should receive tetanus, haemophilus-B influenza, and polio vaccinations at 2, 4, and 12 months, or others that recommend that people to take malaria prophylaxis before traveling to sub-Saharan Africa. But never before had researchers approached all of medicine from this philosophical perch: seeking evidence to determine which tests and treatments work best. Another way to look at outcomes research was that it attempted to answer the question so provocatively raised by Wennberg's variations research: which rate is right (3)?

Once these evidence-based practices were determined for hundreds of symptoms and diseases, it wasn't much of a leap to begin assessing how often doctors get it right. The answers were sobering. A seminal 2003 study in the New England Journal of Medicine found that medical practice comported with evidence-based guidelines 54% of the time (4). That's right - nearly half the time, what doctors did for their patients was inconsistent with the best published evidence! Undoubtedly, some of these whiffs represented situations in which patients didn't quite fit the mold; perhaps in others, doctors were trying out new ideas. But more often, it appears that physicians and healthcare organizations fail to provide the "correct" care, perhaps because they do not know what it is, or because there are obstacles to delivering it reliably.

Whatever the explanation, we now understand that, despite yearly expenditures in the U.S. of more than $2 trillion, patients in our healthcare system (and this isn't unique to America - similar findings have come from studies in other countries) have a chance of receiving evidence-based care that is only slightly better than a flip of a coin. And this is not a simple case of failing to cross the "t's" - adherence to evidence-based practice generally correlates with ultimate clinical outcomes, such as mortality rates (5).

How Can Healthcare Quality be Measured?

Avedis Donabedian (1919-2000, photo at right) was a pioneering University of Michigan health services researcher who thought deeply about how to measure the quality of care. His conceptual framework, now known as "Donabedian's Triad, " has become the underpinning of most quality measurement activities. The Triad divides quality measures into structure (how is care organized), process (what was done), and outcomes (what happened to the patient) (6).

Each element of the Triad has important advantages and disadvantages when used to measure quality (Table 2) (7). In recent years, most widely used healthcare quality measures have been process measures, as clinical research has established the link between these processes and improved outcomes. For example, once strong evidence connected the survival benefit of giving heart attack patients aspirin when they arrived at the Emergency Room, measuring the rate of aspirin administration in this situation became a (reasonable) process measure of quality.

Adapted from Shojania, 2001, with permission.

However, in areas in which processes are less relevant and the science of case-mix adjustment (collecting and statistically adjusting for a series of variables, such as patient age or other illnesses, to understand the patient's burden of baseline illness) is suitably advanced, outcome measurement is often used. One area in medicine where this is the case is cardiac bypass surgery, where we have a pretty good handle on how to adjust for how sick the patient is prior to the surgery, which permits legitimate comparisons of patient outcomes after surgery (8).

In still other areas, in which the processes are quite complex and case-mix adjustment is not sufficiently advanced, structural measures are used as proxies for quality. Examples here include whether a hospital employs specialized "intensivists" to staff critical care units, whether it has a specialized service for stroke patients, and whether it uses computerized physician order entry (CPOE) rather than handwritten medication orders.

A helpful way to think about the Donabedian Triad is to apply it to another high stakes decision that many of us make under considerable uncertainty: having a child apply to college. Assume that you are trying to decide whether to encourage your child to go to a private university or a public, state-run institution. In the United States, the average yearly cost differs by about $25, 000, so the decision to attend a private college will cost about $100, 000 more over four years. Let's say that you're not Bill Gates: coming up with this extra cash represents a true hardship. But you want the best for your child, so you resolve to make the decision based on "value": namely, is the extra $100, 000 for the private school worth it? What factors would you consider in making this decision?

First, you might look at structure: how things are organized at the different schools. Is there a strong history department? A foreign language requirement? A jobs placement unit? Good Internet connectivity in the dorms? And what is the student-to-teacher ratio in the classes? As you consider these kinds of things, you realize that they tell part of the story, but there's much more that you'd like to know.

So you consider processes: what is actually done at the school. Do the students take a humanities course? Do they join clubs? Fraternities? In thinking about both structures and processes, you recognize that these are both merely proxies for what you really care about: what actually happens to the kids after they graduate (outcomes). So you might seek evidence that links them together. Are there any data, for example, that a foreign language requirement is associated with later propensity to travel? Or that students who join fraternities have higher incomes a decade after graduation? Probably not (and you might favor these things for their inherent value; after all, one of the key outcomes you care about is that your child has a positive experience during college), but these are the kinds of process-outcome or structure-outcome links you'd try to find.

Finally, you turn to outcomes themselves: what becomes of kids after graduation? Here, you might consider the kinds of graduate schools that seniors were accepted to. Or the net worth of graduates at age 40. As you consider these sorts of measures, you realize that - even if you believed that these were important outcomes (should we measure income or happiness at age 40?) - you'd need to ask an important question: are the kids who went to the private vs. public colleges really comparable; in intelligence, drive, aspirations, family income, and family connections? In other words, before making too much of outcome differences among colleges, you'd somehow need to understand these baseline demographic differences (which might explain the ultimate outcomes like post-graduation income or other measures of success) and adjust for them in your analysis before concluding that it was the type of college that made the difference. (Interestingly, studies that have done just that have generally shown no advantage for private or more prestigious colleges.)

So, let's return to healthcare quality. Measures of structure (does the hospital have a dedicated stroke unit or a computerized pharmacy?) are great if robust research links these things to important outcomes. Ditto processes, such as did patients receive evidence-based treatment for a heart attack or pneumonia?

But what we really care about is the ultimate outcome. Did the patient live or die? Could the patient walk and read three months after his or her stroke, or was the patient back to work or the golf course a month after a hip replacement? However, to make sense of these outcomes (and use the information to decide whether to go to Hospital A or B, or see Doctor X or Y), you'd need to know a lot about how sick the patients were before the procedures, and adjust for any important baseline differences.

Presently, most well known and publicly available healthcare quality measures are process and structural measures, largely because the science of case-mix adjustment is very young and because collecting outcomes is very laborious (particularly when the outcomes are ones that may not occur for months or years after a clinical encounter, such as 1-year survival rate after surgery). This situation is beginning to change, in part because electronic health records will markedly facilitate the collection of outcome data, and because there is a substantial research effort going into improving case-mix adjustment. The end result is likely to be more and better publicly available outcome data. In a few years, patients may be able to answer questions like these: I'm trying to decide which doctor to see for my rheumatoid arthritis. For patients like me, which doctor gives his patients the best chance of being fully functional and pain free after 2 years of treatment? Or, I was just diagnosed with stomach cancer. Looking at similar patients who are my age and who also have diabetes, where do I go to maximize my chances of being alive in 5 years?

Levers for Quality Improvement

For physicians, policymakers, administrators, and patients, the evidence of major quality deficits in healthcare has led to a recognition of a number of problems that often stand in the way of high quality care. These problems include: the lack of information regarding provider or institutional performance, the absence of incentives for quality improvement, the difficulty for practicing physicians to stay abreast of modern evidence-based medicine, and the absence of system support (such as information technology) for quality. Each will need to be addressed if we are to markedly improve the situation.

The first step in quality improvement begins with quality measurement, which is why an understanding of the Structure-Process-Outcome triad is so important. Efforts to measure quality have a long history in medicine, and, as you might suspect, they have not always been welcomed. In 1911, Ernest Codman (photo below left), a surgeon at Massachusetts General Hospital, promoted his idea to measure the outcomes of every patient after surgery. For this heresy, he was booted off the MGH medical staff. He later started his own hospital - which he dubbed the "End Results Hospital" - and helped found the American College of Surgeons and a hospital accreditation organization that today is The Joint Commission (9).

Despite the interest of a few visionaries, though, change has been slow. As recently as a decade ago, there were only a handful of quality measures supported by strong evidence, such as whether patients with a heart attack received aspirin (which help thin the blood) or beta-blockers (medicines that slow the heart down, decreasing its workload) in the Emergency Department. In recent years, literally scores of such measures have been developed, and subsequently promulgated by a variety of organizations, including payers (such as the Centers for Medicare & Medicaid Services), accreditors (such as the Joint Commission), and medical societies. These measures have identified many opportunities for improvement for individual physicians, practices, and hospitals.

A problem has emerged with this burgeoning field of outcomes research: no individual physician can possibly remain abreast of all the evidence-based advances in his or her field. Practice guidelines, such as those for the care of a patient with a stroke or with severe back pain, aim to synthesize evidence-based best practices into a set of summary recommendations (10). Although concerns about "cookbook medicine" linger (probably more among doctors than patients, who likely find the notion of systematizing high quality care reassuring), there is increasing consensus that best practices should be "hard wired" if possible, and that the most consistent way to do this is through the use of guidelines, checklists, and other features of "highly reliable organizations."

But the guideline business is a tricky one. With so much new information entering the healthcare field every year, guidelines rapidly become outdated (11). Moreover, most guidelines are written with a patient with a single disease ("guideline for the treatment of the patient with high cholesterol" or "for the patient with a heart attack") in mind. Yet many patients present with multiple, and sometimes overlapping, illnesses (12). In these situations, the appropriate treatment for one disease may be quite harmful for another.

Guidelines should be contrasted with clinical pathways, which try to enumerate a series of steps, usually time-based (on day one, do the following; on day two) (13). That makes pathways potentially more useful for medical processes that have predictable trajectories, or for the care of patients after procedures, such as the post-operative care of patients after colon cancer or cataract surgery.

The Changing Policy Environment for Quality

Although one could argue that professionalism and medical ethics should generate more than enough incentive to provide high quality care, the recognition that such care often depends on a system organized to translate research into practice means that it will take significant investments (i.e., in physician education, hiring case managers or clinical pharmacists, building information systems, and developing guidelines) to deliver the right care every time. The traditional payment system, which compensates physicians and hospitals equally whether quality is terrific or terrible, provides no incentive to make the requisite investments. This is changing rapidly.

Recent years have seen a variety of policy initiatives to catalyze quality improvement. Most involve a series of steps: defining reasonable quality measures (evidence-based measures, capturing appropriate structure, processes, or outcomes), measuring the performance of providers or systems, comparing these results to local or national norms or expectations, and using these results to promote improvement. This final imperative creates the greatest degree of uncertainty and experimentation in the policy environment. Although these activities are going on in healthcare systems around the world, my comments regarding specific interventions will focus on those in the United States.

Although one might hope that simply giving individual providers feedback about their quality of care will generate significant improvement, decades of experience have demonstrated that this strategy leads to only modest change. Increasingly, a more aggressive strategy of disseminating quality results to key stakeholders ("transparency") is gaining favor (14). In some cases, transparency is achieved through public reporting, with the hope that providers will find public display of their quality gaps to be sufficiently concerning that they will be motivated to improve (perhaps by seeking out lessons from the better performers, whose results are also publicly displayed). The most visible incarnation of this strategy is Medicare's "Hospital Compare" Web site (www.hospitalcompare.hhs.gov), where every hospital's performance on approximately 20 quality measures is publicly reported (Table 3). One of the biggest surprises of the still-young quality revolution is that public reporting is generating significant improvements, despite the fact that relatively few patients are presently using the data to guide their decisions (i.e., to go to one hospital or one doctor over another) (15). Professional pride appears to be a very powerful motivation.

Table 3. Examples of Publicly Reported Quality Measures

Acute myocardial infarction measures
Aspirin at arrival
Aspirin at discharge
ACE inhibitor or ARB for LVS dysfunction
Beta-blocker at arrival
Beta-blocker at discharge
Thrombolytic agent received within 30 min of hospital arrival
Percutaneous coronary intervention (PCI) received within 120 min of hospital arrival
Smoking cessation advice/counseling

Heart failure measures
Evaluation of LVS function
ACE inhibitor or ARB for LVS dysfunction
Discharge instructions
Smoking cessation advice/counseling

Pneumonia measures
Oxygenation assessment
Initial antibiotic timing
Pneumococcal vaccination
Influenza vaccination
Blood culture performed in the ED prior to initial antibiotic received in hospital
Appropriate initial antibiotic selection
Smoking cessation advice/counseling

Surgical care improvement/surgical infection prevention measures
Prophylactic antibiotic received within 1 h of surgical incision
Prophylactic antibiotics discontinued within 24 h after surgery end time

A newer strategy is to tie payments for service to quality performance ("Pay for Performance, " or "P4P") (16). There are dozens of P4P experiments going on throughout the United States. While P4P appears to result in modest additional improvement over that generated by simple transparency (17), it also raises a host of concerns and challenges. For example, substantial proportions of quality data, particularly that drawn from billing records, are inaccurate or misleading (this would be a concern for public reporting as well; it's just that the stakes are higher with P4P). Then there is the question of whether P4P payments should go to the best performer or the greatest improver (most P4P programs have decided that both need to be compensated in the name of fairness). There are concerns about whether reimbursement cuts to poor performers, which are often the source of P4P bonuses for high performers, will unfairly disadvantage rural hospitals or those that care for indigent populations. Finally, as the financial stakes grow, the potential for "gaming" (focusing on checking a box indicating that you performed smoking counseling, for example, without really doing so) grows, as does the possibility that P4P will create undue focus on measured practices, accompanied by relative inattention to other important process that are not being compensated ("playing for the test") (18).

My guess is that P4P will remain an important part of efforts to improve quality, but that the dominant theme for the next several years will be that of increasing transparency. Particularly as patients (or their proxy decision-makers, such as insurers) become more comfortable with interpreting quality data and the data becomes more accurate and meaningful, public reporting is likely to lead to significant shifts in market share or public perception. When that happens, the additional bang for the buck from a formal P4P program run by the payer may not be worth its additional political challenges and hassles. Of course, a healthcare organization determined to improve its quality performance because of public reporting may well shift dollars around to reward such performance (i.e., giving the doctors or nurses bonuses when certain goals are achieved) - in essence creating internal P4P motivated by external transparency.

Quality Improvement Strategies

Although there are as many quality improvement strategies as there are quality problems, a few themes dominate the field. For practices that require predictable repetition, efforts to "hard wire" the practice or employ individuals who focus on the activity are often beneficial. For example, the best strategy to increase the rate of influenza vaccination of hospitalized patients is likely to be embedding the order in a standard order set, either paper-based or computerized. Having a nurse remove a patient's shoes before the doctor enters the room can increase rates of diabetic foot exams in an outpatient practice. Giving respiratory therapists the authority to adjust the settings on a mechanical breathing machine according to evidence-based algorithms helps patients get off the ventilator much more quickly than waiting for doctors to do it themselves (19).

In some areas, quality improvement involves much more complex and interdependent activities. In these circumstances, bringing teams together to examine their practices and participate in a shared quality improvement process is the most likely path to success. For example, a group of cardiac surgeons in the Northeastern U.S. developed such a program. First, they agreed on a set of best practices and then measured the degree to which their surgical teams adhered to these practices. They also looked at outcomes, including mortality rates, surgical infections, and re-hospitalizations. The surgeons even visited each other's operating rooms to observe both the technical aspects of the surgeries and the manner in which each surgical team did its work. The result was a stunning 24% reduction in the mortality rate of patients undergoing heart surgery (20).

Whether one is trying to improve quality by using a checklist, installing a new computer system, or training the doctors and nurses to communicate more effectively, we know that most initiatives do not work precisely the way they were planned. Some efforts will fail, others will have unexpected consequences - and shockingly few will follow the path that was outlined on a set of PowerPoint slides in a hushed conference room.

With this in mind, healthcare has adapted a number of techniques, many drawn from other industries or initially promoted by quality gurus like C. Edward Deming, to help guide quality improvement activities and assess their impact. Many institutions and individuals use a version of a "Plan, Do, Study, Act" (PDSA) cycle (Figure 2 ) (21), recognizing that QI activities must be carefully planned and implemented, that their impact needs to be measured, and that the results of these activities will often be imperfect and require re-tooling.

Figure 2. The PDSA (Plan-Do-Study-Act) Cycle

The "study" portion of the PDSA cycle can be as simple as collecting data on how many women received mammograms at the appropriate time to much more sophisticated statistical analyses. Whatever the measurement strategy, it is important to set goals and benchmarks. For example, what should the rate of hospital-acquired infections be? Or, what percentage of diabetics should have measures of long-term glucose control (HgbA1c, or glycosylated hemoglobin) of less than 7 mg/dL? Although the answers would seem to be zero and 100%, respectively, this may be unrealistic in the context of day-to-day medical care. So what are appropriate goals? Some organizations use benchmark - data from other comparable organizations, or perhaps, known top performers - to measure themselves against. Others examine their own performance over time, often using sophisticated statistical charts ("run charts") to tell whether worsening performance (or an improvement, for that matter) can be explained by statistical fluke or is significant enough that it implies that something has gone significantly wrong (or, right!).

Value: Connecting Quality to Cost

Outside of healthcare, we base most of our purchasing decisions on our perception of value: quality divided by cost. Who among us is rich enough to always buy the finest thing, or cheap enough to always buy the least expensive one? Most of us spend our purchasing lives trying to determine the value of the things we're considering buying - whether a car, a house, a caf latte, or a college education - by weighing the measured (or perceived) quality against the cost. As we do this, we ask a simple question: is the item or service worth the price?
Healthcare decisions have traditionally not been made this way, partly because of the limited ability of patients (and their doctors, for that matter) to make rational judgments about the quality of a given physician, surgeon, or hospital, and partly because one of the functions of healthcare insurance (usually) is to protect us from the full cost of our purchase. In other words, when you're buying your coffee or your car, you're well aware of the price and are in a perfect position to judge whether the quality is worth it. But in healthcare, the service is often "covered" or you're responsible only for a portion of the cost, in the form of a "co-pay."

Much of the decade-long revolution in quality measurement and reporting can be seen as an effort to provide patients (or other interested parties, such as, in the United States, employers or insurers; in countries with single-payer systems, the government) with the information they need to make rational decisions about healthcare value. The hope is that consumer-driven demand for high value healthcare will be a uniquely powerful catalyst for improvements in both quality and efficiency (22). The accelerating pace of the quality movement provides grounds to believe that this strategy will bear fruit.

Conclusion

You'll be pleased to know that the recognition of these sobering safety and quality problems has led to remarkable changes in the thinking and practice of doctors, nurses, healthcare administrators, government officials, payers, and others, and the rapid adoption of new technologies, regulations, training models, incentive systems, and more. Even with the breathtaking changes in clinical science - with the advent of transplants of nearly every important organ, remarkable devices like implantable defibrillators, the mapping of the human genome - one is hard pressed to identify another area of medicine that has been so utterly transformed in the past decade. Although a large quality chasm remains, it appears to be narrowing in response to these efforts.

References

1. Committee on Quality of Health Care in America, IOM. Crossing the Quality Chasm: A New Health System for the 21st Century. Washington D.C.: National Academy Press, 2001.
2. Wennberg J, Gittelsohn A. Small area variations in health care delivery. Science 1973; 182:1102-8.
3. Wennberg J. Which rate is right? N Engl J Med 1986; 314:310-1.
4. McGlynn EA, Asch SM, Adams J, et al. The quality of health care delivered to adults in the United States. N Engl J Med 2003; 348:2635-45.
5. Higashi T, Shekelle PG, Adams JL, et al. Quality of care is associated with survival in vulnerable older patients. Ann Intern Med 2005; 143:274-81.
6. Donabedian A. The quality of care. How can it be assessed? JAMA 1988; 270:1743-8.
7. Shojania KG, Showstack J, Wachter RM. Assessing hospital quality: a review for clinicians. Eff Clin Pract 2001; 4:82-90.
8. Peterson ED, Coombs LP, DeLong ER, Haan CK, Ferguson TB. Procedural volume as a marker of quality for CABG surgery. JAMA 2004; 291:195-201.
9. Sharpe VA, Faden AI. Medical harm: historical, conceptual, and ethical dimensions of iatrogenic illness. New York, NY: Cambridge University Press, 1998.
10. Weingarten S. Translating practice guidelines into patient care. Guidelines at the bedside. Chest 2000; 118:4S-7S.
11. Shekelle PG, Ortiz E, Rhodes S, et al. Validity of the Agency for Healthcare Research and Quality clinical practice guidelines: how quickly do guidelines become outdated? JAMA 2001; 286:1461-7.
12. Boyd CM, Darer J, Boult C, Fried LP, Boult L, Wu AW. Clinical practice guidelines and quality of care for older patients with multiple comorbid diseases: implications for pay for performance. JAMA 2005; 294:716-24.
13. Every NR, Hochman J, Becker R, et al. Critical pathways. A review (AHA Scientific Statement). Circulation 2000; 101:461.
14. Berwick DM. Public performance reports and the will for change. JAMA 2002; 288:1523-4.
15. Williams SC, Schmaltz SP, Morton DJ, Koss RG, Loeb JM. Quality of care in U.S. hospitals as reflected by standardized measures, 2002-2004. N Engl J Med 2005; 353: 255-64.
16. Millenson ML. Pay for performance: the best worst choice. Qual Saf Health Care 2004; 13:323-4.
17. Lindenauer PK, Remus D, Roman S, et al. Public reporting and pay for performance in hospital quality improvement. N Engl J Med 2007; 356:486-96.
18. Wachter RM. Expected and unanticipated consequences of the quality and information technology revolutions. JAMA 2006; 295:2780-3.
19. Kollef MH, Shapiro SD, Silver P, et al. A randomized, controlled trial of protocol-directed versus physician-directed weaning from mechanical ventilation. Crit Care Med 1997; 25:567-74.
20. O'Connor GT, Plume SK, Olmstead EM, et al A regional intervention to improve the hospital mortality associated with coronary artery bypass graft surgery. JAMA 1996; 275: 841-6.
21. Plsek PE. Quality improvement methods in clinical medicine. Pediatrics 1999; 103: 203S-14S.
22. Porter ME, Teisberg EO. How physicians can change the future of healthcare. JAMA 2007; 297:1103-11.