The 2^nd Annual Amyot Lecture - "Evidence-based Health Care: A Reality Check"

Presented by Dr. David Naylor, Dean, Faculty of Medicine, University of Toronto
Wednesday, October 18, 2000 at 6:30 p.m.
Theatre, Sir Frederick G. Banting Building
Tunney's Pasture, Ottawa

Summary

There is general support for the concept of evidence-based decision-making in medicine and health care. However, the current reality is that Canada's publicly-funded health systems are not yet organized in ways that enable research evidence to drive decisions, either at the clinical level or at the institutional level. Instead, historical happenstance, local circumstance, blissful or wilful ignorance, and simple belief systems continue to shape our health care.

This situation is particularly worrisome as consumer expectations rise, and as a new era of expensive 'post-genomic' medical technologies looms.

There are no 'magic bullets' to create a more sustainable and evidence-based system. However, some features of the current system undermine rational decision-making. A particular priority for future investment must be improvements in health information - including information content, info-structures and processes, and uptake of relevant information by the public.

Presentation in Full

1. Introduction
2. From Evidence-based Medicine to Evidence-based Health Care
3. Benchmarking the Performance of Our Health Care System
4. Evidence-based Health Care System in 2010: A Fantasy
5. A True Health System

Dr. C. David Naylor, Dean, Faculty of Medicine, University of Toronto

Dr. David Naylor is Dean of the Faculty of Medicine at the University of Toronto, with cross appointments in the graduate Faculty of Nursing and in several departments in the Faculty of Medicine. Dr. Naylor also serves as one of the inaugural members of the Governing Council of the Canadian Institutes of Health Research.

As founding CEO of the Institute for Clinical Evaluative Sciences (ICES) (1992-1998), Dr. Naylor is internationally recognized as a leader in the field of health services research. His personal research focusses on the effectiveness and efficiency of cardiovascular care. More generally, Dr Naylor's colleagues and collaborators at ICES and across Canada have worked to translate evidence from clinical and health services research into improvements in medical practice, health administration, and policy-making.

Dr. Naylor graduated from the University of Toronto with a Medical Degree and later earned a D.Phil. at Oxford as a Rhodes Scholar in social and administrative studies. He has co-authored over two hundred articles and editorials in medical journals, and co-authored several monographs. Dr. Naylor has received several national and international awards for research and served on the editorial boards of several medical journals. He was also special advisor to the Health Services Restructuring Commission in Ontario.

1. Introduction

Let me start by thanking David Dodge for that extraordinarily generous introduction and, of course, to thank the Deputy, Assistant Deputy and all of you for your hospitality. It is indeed a pleasure and privilege to be delivering the second annual Amyot lecture to an audience of Canadians dedicated to improving the health of our citizens.

In choosing a topic for tonight's discussion, I considered the prominence that evidence-based decision-making has achieved in the language of health care reform in this country. It seemed to be time to take stock, to consider in a sober fashion just how far we have or have not moved towards achieving the goal of an evidence-based health care system. Once I've moved through that area, in a somewhat optimistic vein, I'll share with you a fantasy about the form and function of an evidence-based health system as it might exist in Canada in the year 2010, a decade from now.

Given the intelligence, expertise and backgrounds of the individuals in this audience, I am going to move fairly quickly. In some cases, I'll leave you to read from the slides while I comment on them.

Let me begin with a definition, not of an evidence-based health care system, but of one of its key building blocks, an evidence-based approach to clinical practice within the system. Evidence-based clinical practice has being defined as shown on this slide. Note the key terms,an awareness of evidence and an awareness of the strength of the evidence. The definition, I confess, sounds sanctimonious, perhaps even judgmental...as if much of medicine in the past has been non-evidence-based. As one colleague of mine put it, in a somewhat heated exchange some years ago, "What do you think we've been doing for the past 100 years, Naylor?"

Unfortunately, the answer would be, "Lots of smart and helpful things, and lots of things that were not." Too often, pathophysiological theories, or inferences about effectiveness based on clinical anecdotage, led to practices that were enormously harmful to patients in years gone by.

Let me give you a few examples.

Autointoxication. A theory derived from a rather primitive understanding of bacteriology. It was championed by a charismatic English surgeon with the rather unlikely name of Arbuthnot Lane. Based on this theory, a young woman who complained too frequently of fatigue might undergo entire removal of her colon. If you read Canadian medical journals from the earlier part of the century, you will rapidly appreciate that this theory had scores of disciples across this nation who, with the best of intentions, did enormous harm.

Insulin coma for schizophrenia. Now, we're into the 1930s, the era when medicine is supposedly more scientific. Notice what killed off this remarkably dangerous practice, a randomized trial..evidence from a reasonably rigorous experimental design, applied to actual clinical practice. Well, you might say, "Surely, once we realized the value of randomized trials, we stop making mistakes like this." I wish it were so.

Antiarrhythmics - now, we're into the modern era. An utterly logical strategy... Irregularities in the heartbeat do lead to sudden cardiac death. Stop the electrical pattern associated with irregularities, and outcomes must surely improve. Well, they didn't. Fortunately, we now have sound randomized trials with hard outcomes to support the use of the current generation of antiarrhythmic drugs. But, how many thousand patients worldwide were killed by the old compounds?

This concept of evidence-based clinical practice is, of course, in part, a social marketing device...simply a way of calling attention to the importance of ensuring that our bedside decisions are actually derived from evidence that, in turn, is grounded in rigorous studies showing that a particular decision is likely to lead to improved outcomes. And, by that, I mean real outcomes, not some arcane laboratory measure --- outcomes that matter to real people, not cells or animals, and that make a difference in real time. We have come a long way in 20 years. I'm amazed at how much of modern medicine can now be traced to outcomes-based scientific evidence. But let us do the first in a series of reality checks.

We have not only failed to compile all the available evidence about clinical practice, we also face continued gaps in our knowledge, even when we do synthesize best evidence. Two examples will suffice: Effective Care in Pregnancy and Childbirth. A magisterial publication that spawned the now famous Cochrane collaboration, which seeks to synthesize all randomized trial evidence in all areas of clinical practice. This was a compendium of randomized trials --- covering every trial ever done that bears on the care of individuals during pregnancy and childbirth. But, as you can see, a substantial number of the interventions remained in limbo. We can't throw them out; and we can't embrace them.

The Canadian Task Force on the Periodic Health Examination was in some ways a catalyst for the modern evidence-based medicine movement. Starting in the late 1970s, it meticulously sifted and weighed all available evidence, not just clinical trials, before providing advice to primary care providers about what should or should not be part of the periodic health exam. Notice the verdict: Fully 76 of the maneuvers that were scrupulously examined were assessed as follows. Decision-making had to be guided by something other than scientific evidence.

And, so, for the present, one of the biggest limits to evidence-based clinical practice remains insufficient clinical evidence. This challenge underscores the enormous importance of ensuring that patient-centred research, particularly clinical trials, find a home in the new Canadian Institutes of Health Research.

There are, I believe, two other important caveats about evidence-based clinical decision-making. The first is that all forms of professional practice should, in the words of Edmund Pellegrino, remain the most humane of the sciences, and the most scientific of the humanities. We need clinical practice to be evidence-based. But it should be more. Last year in the inaugural Amyot lecture, Dr. McMurtry highlighted the importance of patient-centred medicine. An approach to care that is compassionate and respectful, one that elicits the patient's preferences, and is sensitive to his or her values and circumstances. Clinical care, after all, is ultimately a cross-cultural activity. The clinician's job is to reach across barriers to communication created by social class, educational attainments, different ethno-linguistic backgrounds and, perhaps, most fundamentally of all, the barriers between the world of those who seek care, and those who give it.

Put another way, patients are afflicted with diseases as biological entities, but experience illnesses in a social context. That is one reason, incidentally, why the social sciences and humanities belong so integrally in the Canadian Institutes of Health Research. But we also need, of course, medicine that is reality-based. In that respect, clinical practitioners must understand the limits of modern medicine, the role of complementary medicine and the limits on available resources.

Well, it sounds like a tall order, and it is. And, unfortunately, the order gets taller still, as we try to move from evidence-based clinical practice to the next level, an evidence-based health care system.

2. From Evidence-based Medicine to Evidence-based Health Care

Just as individual clinical decisions rest on a blend of evidence, values, and circumstances, so also are broader health care decisions conditioned by those same factors. Moreover, the health care system is, as everyone in this room knows only too well, a highly complex quasi-system. So it is that clinical decision-making, which may be modified by incomplete information, or lack of evidence in the first instance, is inevitably shaped by a witch's brew of systemic influences, operating at the institutional, regional, provincial and national levels. No one should be surprised, accordingly, that when we aggregate clinical decisions into profiles of regional or institutional service delivery, some striking and rather inexplicable variations emerge.

Let me give you some examples from work that my colleagues and I have published over the last five years or so, focusing on cardiovascular care in the first instance. New York State does about twice as much coronary bypass surgery per capita as Ontario, even though the incidence of coronary disease is similar in the two jurisdictions. This slide shows the rates of bypass surgery in Ontario and New York, broken down by the severity of the underlying blockages, and their impact on life expectancy, and by the age of the patients.

To your right, where surgery is expected to prolong life, we see rather disturbingly that Ontario has much lower rates of surgery on a per capita basis. But, of course, the really big differences are on your left. What we see here is a sub-group of patients where there is low life expectancy gain anticipated: this is quality-of-life-enhancing surgery. New York State does 17 times as many bypasses as Ontario on patients in this subgroup aged 75 and over. I am all for active seniors, but the logic for this particular pattern of practice challenges my imagination, especially when one realizes that older patients have a definite risk of suffering a stroke during these operations.

Ontario has the highest provincial rate of bypass surgery in Canada --- New York State, almost the lowest of the 50 states in America. It is clear that this comparison, if anything, underestimates the dramatic differences in practice patterns between the two countries. Data like these quickly illustrate why America spends so much more than Canada on health care.

Now, at this point, you may raise a concern. The previous slide did show that Canadians with severe blockages in their coronary arteries had reduced access to bypass surgery compared to Americans. Doesn't that translate into differences in outcomes? Well, we examined all senior citizens in Ontario and the United States who had recently been hospitalized with a heart attack. We measured how often they received expensive tertiary services, such as coronary angiograms, balloon angioplasties, and bypass procedures. As shown in the slide, American seniors were up to eight times more likely to get these procedures soon after a heart attack. And even after six months, Americans were still four times more likely to have undergone a major procedure, a huge difference in resource consumption. What was the impact on mortality, when all of these patients were followed for a year? Absolutely none.

Now, this is again, as I said, a skeptical and knowledgeable audience, so you'll ask me, "Well, what about quality of life?" Death isn't the only outcome that comes. If it were, we'd probably shut down a lot of the Canadian health care system tomorrow.

Several studies do show that American heart attack patients have better quality of life outcomes, but it's not that simple. The most extensive study undertaken was led by investigators from Duke University, and involved researchers across the U.S., along with many of us in Canada. It examined the array of quality of life and economic outcomes shown on this slide, and followed a large group of Canadians and Americans who had been part of a major worldwide study of heart attack drugs.

Americans predictably underwent many, many more expensive procedures than the Canadian trial participants. They were also far more likely to see specialists than Canadians, who relied instead on family physicians for ongoing care. We did find that Canadians were more likely to have chest pain symptoms. They had more limitations in their daily activities. They were even more likely to be somewhat depressed. Not good. But there was a catch. When asked to rate their overall health, Canadians responded virtually identically to Americans. And when we assessed work status, Canadians were just as likely as Americans to be back working full time.

Now, one interpretation is that we are simply more stoic than our flamboyant friends to the south. And perhaps that's part of the answer. Then, again, as H.L. Mencken once said, "For every complex problem, there is a simple solution and it is almost always dead wrong."

So, let me offer another interpretation that is more complex and, I suspect, more accurate. It goes like this. Modern medicine is full of halfway technologies, defined as shown on this slide. Much of what we have to do helps only some patients some of the time. And, because we are often uncertain about which patients will actually be helped, modern medicine has become a great game of probabilities and uncertainties.

It turns out, for example, that even when bypass surgery is done on patients who are at medium risk of shortened life expectancy, the vast majority of those undergoing surgery would still be alive ten years later if they had simply been given drugs. Another way to think of this phenomenon is in terms of the number of patients who must be treated with a particular technology, to prevent a bad outcome for one patient. Literally scores of asymptomatic patients with moderately high blood pressure must be treated to prevent one adverse cardiovascular event. Or, more dramatically, 70 previously well middle-aged men with high cholesterol, would together take well over 200,000 pills over the course of several years to prevent one fatal, or non-fatal, heart attack. Modern medicine, you can see, even when evidence-based, is still a very young and rather primitive science.

The consequence of marshalling halfway technologies, in a climate of uncertainty within highly variable local health systems, is a great deal of variation in practice patterns and resource consumption. The other consequence is that when these technologies are used for ever-wider clinical indications - as is true when contrasting American clinical decision-making with Canada's more conservative practice patterns - we rapidly hit a point where health care has diminishing marginal impacts on health status.

We may start with a near linear relationship between service investments and improvements in population health status. But as service intensity rises, more and more resources are poured into achieving smaller and smaller gains. Our great challenge, from the health policy standpoint, is deciding where along the flattening part of the curve, we are going to stop expending public resources on health services per se. At what point would we be better off making investments in public infrastructure or in post-secondary education, or in social policy initiatives to improve the life chances of our children?

Well, let's retreat from those tougher questions to some simpler issues.

I've highlighted for you some of the profound transnational variations in how health care systems function. It would be nice to believe that inside our universal Canadian health care systems, such variations are not apparent. Unfortunately, we all know that's not the case. Since 1994, when we published the first in an ongoing series of atlases of health and health services, the group at the Institute for Clinical Evaluative Sciences has repeatedly demonstrated dramatic variations in processes and outcomes of care by institution and by region. Many other research groups nationally have demonstrated the same phenomena.

If you fall ill, your chances of receiving a particular diagnostic or therapeutic procedure, or having a good outcome from a particular illness or procedure, are still influenced to a disturbing degree by where you live or where you happen to be treated. It is sometimes assumed that when variations are found, more care means bad care. At local hospitals where physicians are providing residents of county A with three or four times as much of a particular service as residents of county B, then it is sometimes surmised that the hospitals in county A are wasting resources --- or that the physicians involved are ripping off our publicly funded health system. Not so. Here is an example.

Hip and knee replacement surgery has a remarkable impact on quality of life for patients with arthritis. When we examined service profiles in the early 1990s, we found that southwestern Ontario had much higher rates of hip and knee surgery than in the metro Toronto area. And all of these differences are dramatically statistically significant.

This was unsettling as the relative differences in surgery rates had been more or less constant for a decade. We therefore reviewed a random sample of patient records from hospitals in the two regions. We hypothesized that patients in the southwest might have milder symptoms, or that more of them would be receiving surgery for weak or even inappropriate indications. The records were reviewed by experts who were deliberately blinded to the patient's hospital or the region of residence. As it turned out, ages and symptom status were identical in the two regions. Furthermore, when explicit criteria were used to assess the appropriateness of surgery in each region, there was absolutely no difference.

On this basis, we suggested that it was time for an increased investment in joint replacement surgery for the greater Toronto area. And subsequent studies have confirmed that some areas of Toronto were, and probably still are, seriously under-served with this quality-of-life-enhancing technology.

None of this is rocket science. And if we don't do it, the Canadian health care system will continue to be buffeted by provider rhetoric and media sensationalization. In that latter regard, many of you will recall that 10 years ago, Canada's newspapers repeatedly ran front-page headlines about patients dying waiting for heart surgery. The American business media, the American Medical Association and then-President George Bush, all got into the game, warning Americans about the deadly evils of socialized medicine in Canada.

At the time, a group of researchers, clinicians, administrators and dedicated public servants developed the Cardiac Care Network of Ontario, with the world's first evidence-based queuing guidelines, and a comprehensive registry that tracked every patient waiting for open-heart surgery. Variations on that system have been adopted in several provinces, and in a large number of countries worldwide. As we argued then, rationing by queue is inherently more equitable than any system rationing by income or insurance, such as exists currently in the United States.

But, the goal of rationalizing waiting lists is not just to ensure that the sickest patients, or those at highest risk, come to the head of the line. A waiting list for any health services is analogous to a price in the marketplace. It's a signal of unmet demand, which may of may not correspond to actual needs. It's also a terrific opportunity to engage providers and the public in a dialogue about expectations and responsibilities. I'm delighted that Health Canada has funded the Western Canada Waiting List Project, but we can and should move ahead with many more such efforts to measure, manage, and demythologize queues in our health care system.

3. Benchmarking the Performance of Our Health Care System

The Minister has been a wonderful champion for this shift in thinking. The government's investment in the Canadian Health Services Research Foundation and in the Canadian Institute of Health Research, along with its support for improvements in health info-structure, are all accelerating the transformation of the culture of our health care system. I'm particularly pleased that we are moving toward an era in which the provinces and federal government may finally co-operate in benchmarking the performance of our system.

Let me suggest in this latter regard, how important it is that we consider and apply a wide range of different measures and put the reports on them into language that our citizens and taxpayers can understand.

Surveys, for example, can be surprisingly powerful tools. Here are some results from a recent five-nation survey of randomly chosen households by a Harvard team. Respondents were asked to estimate the duration of their last visit to a primary care physician. Americans clearly led the way with long visits. But when the researchers asked the citizens if they were getting enough time with their doctor, Americans were again leading the way! They were significantly more unhappy, as you'll see here.

A reasonable conclusion --- one that those here probably appreciate all too well --- is that managing public expectations may be as important as managing resources in our health care system. From the same survey come these data on access to care. Intriguingly, these results generated very little publicity when they were released in late 1998, but I believe they are extremely striking. Canadians were sending a clear signal that they were not happy with their ability to access outpatient specialist care across this country.

This is soft evidence, I know. But I wonder how these results would look if we were able to compare Newfoundland to Saskatchewan and British Columbia, or to do comparisons within provinces. And how might these data map onto information on physician supply by province, or by region?

While on the subject of health care report cards, let me emphasize that no sector should be immune from scrutiny. We all know that the last decade has seen dramatic restructuring of the health care system, with a shift in resources, as shown hereaway from hospitals, and towards community-based care and prescription drug expenditures. I am not an expert on home care, but the last time I looked, the rate of growth in home care spending had outstripped all other major areas. About half of that spending is for long-term community care, yet the indications for, and availability of, such services continue to vary dramatically from one province to the next.

The other half of the spending is on so-called post-acute care, following a hospitalization. When Peter Coyte and his team examined post-acute home care in Ontario, the results were sobering. The hospitalizations here are for cataract surgery, a variety of respiratory diseases, and for musculo-skeletal procedures, most notably, hip and knee replacements. As you can see on the slide, the percentage of patients who received home care services after discharge from hospital varied even more wildly than some of the discretionary surgical services that have frustrated planners and managers, and policy makers for decades. If we are to continue moving from a hospital-based sickness care system, to a more integrated health system - and I believe we must do so - the culture of assessment and accountability, and the general requirement for evidence-based decision-making, must be applied urgently to the community and home care sector.

Before I leave the topic of report cards, one cautionary comment. In a rush to simplify the indicators for general consumption, we should be very careful about what information gets lost or distorted. In this regard, Maclean's magazine has given Canadians a lot of first-class information about our health care system, but their composite ranking of regions is junk science. I am sorry to say that the World Health Organization's ranking of health systems is almost as bad, and is seen as something of a marketing ploy by many of us in the health services research field.

That said, no report cards, whether badly or brilliantly conceived, will change a health care system that is structured so as to be impervious to evidence. Responsiveness entails changes in behaviour at the micro- and meso-level. In turn, changes in behavior are most likely to occur when there is alignment of incentives. That much we have all learned from our colleagues who are skilled in what Thomas Carlisle once called the "dismal science of economics."

Unfortunately, the Canadian health care system continues to be structured in a fashion that ignores a body of evidence from both economics and organizational theory. Incentives are not aligned because the system fundamentally remains poorly integrated. Allan Rock put it well when he said, "I think it is questionable that we have even had a real system.

We've had more of a series of separate service delivery outlets." I accept that the emergence of health regions in nine provinces has been a very positive step towards better integration of services and potential alignment of incentives. However, payments for both physicians and prescription drugs have remained outside the budgetary envelope in those nine provinces. As well, two successive governments in Ontario have rejected regionalization and we have yet to develop an integrating strategy.

4. Evidence-based Health Care System in 2010: A Fantasy

Well, all this is too gloomy. So on a more upbeat and forward-looking note, I would ask that you join me in a fantasy about an evidence-based health system for 2010. Information and info-structure are absolutely essential. We must synthesize what we know and make it widely available, in useful decision support tools. And those tools should also be available to our patients, and to taxpayers. This, I think, is eminently achievable.

Integrated delivery networks, with aligned incentives, publishing balanced score cards on a range of clinical and efficiency indicators. Again, eminently achievable.

No system that fails to invest in research and development can ever be successful or sustainable. Fortunately, we have seen tremendous support for applied research at the federal level and in some provinces. But the next step is to ensure that that R & D function is built into the system on all levels.

More on info-structure... The electronic patient record is the Holy Grail of modern health care. We don't need smart cards or central computers. We don't even need a standardized format. There are many fine data-integrating engines out there on the market already. Service organizations could accordingly exchange patient level information on a strictly controlled and encrypted system that would actually enhance, not threaten the confidentiality of medical records. For example, the system could track all access by providers and report it back regularly to patients so they would know who had seen their medical records --- a major improvement on the current situation.

Integrated information systems and integrated health systems: a perfectly logical partnership. It would be so much easier to deliver care effectively and efficiently in this type of environment. It would reduce a lot of redundancy in testing, and enhance evidence-based decision-making.

George Hagel once wrote, "The only lesson of history is that governments and communities have never learned anything from history." He was unduly pessimistic, I believe. At least in the Canadian health care system, we are getting better at learning from each other's successes and failures. But we can do much more. Clearinghouses for best practices, with on-line availability of data for benchmarking could transform the evidentiary basis for what we do.

Provider payment --- always a thorny area. This comes down to incentives at the micro level. It is clearly time to implement blended compensation systems for physicians and other fee-for-service providers. Currently, with pure fee-for-service, the payment system reinforces the doctor's position as a pushcart vendor in an era of health care supermarkets.

This is self-explanatory. We cannot begin to maintain a high-quality and evidence-based system unless our providers can surf successfully on the rapidly-moving waves of modern medical science.

Tele-health. I don't know about others in the audience, but I am amazed at the potential and quality of some of our current tele-health technology. Couple this technology with robotics for remote monitoring and even remote surgery, and we have entered a world in which Medicare, more than ever, can become a unifying force in our national life.

5. A True Health System

Thus far, of course, you will have noticed a glaring oversight. I've been talking about health services and sickness care and yet, as Bob McMurtry so brilliantly elucidated in last year's inaugural Amyot lecture, there is no reason why our health care system cannot move beyond the repair shop model to become a true health system.

Consider some of the lines of evidence here. We recently categorized urban dwelling Ontarians into five equal groups, based on the average household income of the neighborhood in which they lived. Then, we examined how many in each group were hospitalized with heart attacks. Comparing the lowest to the highest income group, the hospitalization rates were almost three times higher for those Ontarians who were poorer. More generally, we know that Canadians with lower incomes and less education live much shorter, unhealthier lives. They're more likely to smoke, to be sedentary, follow an unhealthy diet, and be overweight. A reasonable reply to that line of evidence might run as follows: "Yes, the poor are more likely to get sick and, therefore, they indeed have shorter and less healthy lives. But at least once the poor fall ill in our society, the Canadian Medicare system ensures that they receive similar services and achieve similar health outcomes to those who are comparatively wealthier or better educated." You could say that...and you would be wrong.

Several studies now show that, all things being equal, wealthier Canadians continue to be more likely to access specialist services than those who are economically disadvantaged. As to outcomes, this figure from a recent New England Journal of Medicine article, is sobering. We took 50,000 Canadians with heart attacks and followed them for a year. We again are using the method of dividing this population into five income groups. Each line on the slide represents an income quintile with thousands of patients. The highest median household income was about $29,000 --- the lowest, $16,000. Not a huge difference. But the difference in death rate shown here is indeed huge. There is no drug available today that comes anywhere close to conferring this degree of survival advantage on patients with coronary disease. In fact, a comprehensive statistical model suggested that for each increase in neighborhood income of $10,000, the relative risk of death for these patients fell by 10%.

The last line of evidence, it may be germane, is shown here. These are data from the National Population Health Survey. They summarize the actions that Canadians aged 15 and over reported that they were taking to improve their own health.

Let's start at the bottom and move up. These are all in thousands. There were a couple hundred thousand vitamin gobblers. We also have about 600,000 Canadians who wanted to quit or reduce smoking --- clearly, a very small fraction of those in the population who would have been current smokers. About 2.5 million Canadians, taking two groups together, wanted to either lose weight or eat better. Again, if you consider national dietary norms, this could apply to much of the population. Almost seven million Canadians intended to exercise more; whether they were or were not doing it is beyond me. But, then, of course, the crowning number - 12.2 million Canadians who proudly reported that they were doing nothing to improve their own health.

Let me suggest then that our fantastical evidence-based health care system must build the capacity to carry out groundbreaking studies on the broader determinants of population heath across the life course. As such, it would be superbly suited to partner with the Population and Public Health Institute in our new national funding agency for health research. More than that, in its decision-making, it could be a health improving system, based on a population wide perspective in which the full range of determinants of health could be considered and addressed.

Ladies and gentlemen, we've covered a lot of ground and it's time for me to close. I believe that we are emerging from the period in which Canadians' faith in our Medicare system has been badly shaken. Public confidence seems to be still lower than it should be. Fears about the system's non-sustainability are being fanned by those who have their own reasons for wanting to dismantle Medicare.

Some wise and generous investments in health care have been made both by this government and various provincial administrations. But I do not believe that those investments in themselves will take us where we need to go. What is required now is a pervasive shift in vision and culture. We must move rapidly towards a more evidence-based sickness care system, one that can go from evidence to action, from measurement to management, and from assessment to accountability.

But then, we must move resolutely forward again to a better integrated health improvement system that will enhance the well-being and productivity of all Canadians. I can imagine no better challenge or finer legacy for those who are assembled here tonight. Thank you for your attention.