UBC Faculty Research and Publications

Threats to quality : illusion or reality? : conference summary 13th Annual Health Policy Conference,… McGrail, Kimberlyn, 1966-; Cardiff, K. (Karen), 1953-; Helmer, Diane 2001

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


52383-McGrail_K_et_al_13thHPconf.pdf [ 844.1kB ]
JSON: 52383-1.0048416.json
JSON-LD: 52383-1.0048416-ld.json
RDF/XML (Pretty): 52383-1.0048416-rdf.xml
RDF/JSON: 52383-1.0048416-rdf.json
Turtle: 52383-1.0048416-turtle.txt
N-Triples: 52383-1.0048416-rdf-ntriples.txt
Original Record: 52383-1.0048416-source.json
Full Text

Full Text

   Threats to Quality:   Illusion or Reality?  Conference Summary 13th Annual Health Policy Conference November 10, 2000  Edited by: Kimberlyn M McGrail Karen Cardiff Diane Helmer   HPRU 01:15D December 2001    Health Policy Research Unit Research Reports           THE  UNIVERSITY  OF  BRITISH  COLUMBIA    The Centre for Health Services and Policy Research was established by the Board of Governors of the University of British Columbia in December 1990.  It was officially opened in July 1991.  The Centre’s primary objective is to co-ordinate, facilitate, and undertake multidisciplinary research in the areas of health policy, health services research, population health, and health human resources.  It brings together researchers in a variety of disciplines who are committed to a multidisciplinary approach to research, and to promoting wide dissemination and discussion of research results, in these areas.  The Centre aims to contribute to the improvement of population health by being responsive to the research needs of those responsible for health policy.  To this end, it provides a research resource for graduate students; develops and facilitates access to health and health care databases; sponsors seminars, workshops, conferences and policy consultations; and distributes Discussion Papers, Research Reports and publication reprints resulting from the research programs of Centre faculty.  The Centre’s Health Policy Research Unit Discussion Paper series provides a vehicle for the circulation of (pre-publication) work of Centre faculty, staff and associates.  It is intended to promote discussion and to elicit comments and suggestions that might be incorporated within revised versions of these papers. The analyses and interpretations, and any errors in the papers, are those of the listed authors.  The Centre does not review or edit the papers before they are released.  A complete list of available Health Policy Research Unit Discussion Papers and Reprints, along with an address to which requests for copies should be sent, appears at the back of each paper.   Conference Summary  November 10, 2000  Table of Contents   Table of Contents ....................................................................................................................... 1  Foreward....................................................................................................................................... 2   Reviewing The Emperor’s New Clothes: What Can Be Learned From The Erosion Of Trust And The New Policy Focus On Variations In Activity And Outcome In Medicine  Alan Maynard ......................................................................................................... 3  Waiting for Care: Effects of Wait Lists on Quality   Claudia Sanmartin.................. 12  Lessons from the Western Canada Waiting List Project  David Hadorn .................. 17  Facing the Big Downer at the Front Line: Quality of Care in the Primary Care Treatment of Depression  Elliot Goldner........................................................................... 21  Sharing Care for Depression in Primary Care  Nick Kates............................................ 27  Using Report Cards to Improve the Quality of Acute Myocardial Infarction Care  Jack Tu ........................................................................................................................................ 33  Assessment and Accountability for Surgical Indications and Outcomes  Charles Wright.......................................................................................................................................... 42  Closing Rapporteur  Rick Roger........................................................................................... 49   - 1 - Conference Summary  November 10, 2000 Foreword  The news media in Canada is constantly filled with articles alluding to the alleged fact that Canada’s health system is in an ever worsening ‘crisis’ – unable to provide consistent quality in the care delivered to meet accepted standards. Numerous threats are cited including long delays for surgery and hospitalization, perpetually overcrowded emergency rooms, overworked and increasingly scarce healthcare workforce, a lack of state-of-the-art health technologies, and uneven performance by health personnel.  The 13th annual policy conference organized by the Centre for Health Services and Policy Research aimed to explore tensions at play driving the increased focus on quality of care, how mechanisms to improve quality have been employed, and the evidence that these responses are really improving quality. The quality field is wide, so the conference concentrated on a few key areas – waiting lists and waiting times, primary care for the treatment of depression, and criteria that might help evaluate appropriateness of care. Complementing these themes is an overall assessment of the state of health care emphasizing the role of trust, and the need for greater patient involvement in the delivery of health services.  This volume includes a selection of papers delivered at that conference. We hope it will be of interest to anyone concerned about quality and about the future of health care in Canada.  Kimberlyn McGrail Karen Cardiff Diane Helmer  Vancouver, BC June 2002  - 2 - Conference Summary  November 10, 2000  Reviewing The Emperor’s New Clothes:  What Can Be Learned From The Erosion Of Trust And The New Policy Focus On Variations In Activity And Outcome In Medicine Alan Maynard   The series of papers in this volume are about threats to quality, which immediately prompts the question “What on earth is quality?”  Is it a process measure or an outcome measure?  The health care industry is remarkable; lots of money is spent on inputs, lots of things are done to people, but there are no measures of what is produced in terms of improvements in the health status of the population.  There are no standard, routinely used measures of health. The question is why we are so reluctant to use what is available and begin to understand our success.  The title of this paper, Reviewing the Emperor’s New Clothes, refers to what the evidence is beginning to show; we must be very uncertain and concerned about the success of the enterprise in which we are involved.  We have twenty years of evidence of major variations in medical practice, but what is now emerging is evidence of considerable failures in the processes and the fact that perhaps the Emperor, that is, the medical profession, is beginning to be seen for what it is.  We have had trust in it, but the clothes have fallen off and what we see is not a wholesome sight.  There is a lack of evidence that the profession has really gripped this issue as hard as we as consumers and taxpayers would have liked it to.  One theme of this paper is the importance of trust, and more specifically the demand for trust, the supply of trust, the agency relationship and the placebo effect.  This will point out deficiencies in the governance of medical practice.  Following this are questions about where we are now and where we should go from here.  Can we find the emperor some new clothes?  What can we do with the evidence?  And what can we do better about the regulation of the profession?  Trust is defined as “a firm belief in honesty, veracity, justice and strength; reliance on truth of statement without examination.”  That is what we do when we go to a doctor.  We are expecting a trust relationship, and it is an important part of the exchange between the patient and the doctor, as it is in all human transactions.  We expect doctors to be vigorous in their diagnosis and treatment.  Trust means that we do no have to regulate people, we do not have to have data systems, we do not have to monitor and evaluate everything that moves.  All of that type of apparatus is very expensive and amounts to regulation.  Trust is an alternative to regulation, and the questions we have to ask are whether it is an efficient alternative and where the correct balance between trust and regulation lies in a market where patients will have imperfect information.  Trust may be the principal determinant of transactions in most if not all markets.  We trust our motor mechanic; he sometimes lets us down.  We trust our doctor, we trust the guy who sells us real estate.  Trust is very much related to duty.  Most doctors believe in the Hippocratic oath, which says, “First, do the patient no harm.”  The very first thing we might wish to measure, then, is what evidence we have that doctors do not harm patients.  Do we have a reporting system for errors?  Can we avoid the simplistic persecution of people who unavoidable and humanly err so that we learn from the errors and induce change?  - 3 - Conference Summary  November 10, 2000  The political economist Adam Smith wrote a book called The Theory of Moral Sentiments, which contrary to the way it is sometimes portrayed, is not all about self-interest and greed.  Smith actually wrote: “Those general rules of conduct when they have been fixed in our mind by habitual reflection, are of great use in correcting the misrepresentations of self-love concerning what is fit and proper to be done in our particular situation…  The regard of those general rules of conduct, is what is called a sense of duty, a principle by which of greatest consequence in human life, and the only principle by which the bulk of mankind are capable of directing their actions.”  In other words, much of what we do is not based on greed and self-interest but on duty and trust.  Trust and duty are essential ingredients in the health care market, both in the public and private sector; patients expect and demand that these relationships are strong.  In a social exchange for this trust and duty, we give providers a great deal of independence (self-regulation) and high rewards (remuneration and prestige). In The Wealth of Nations Adam Smith wrote: “We trust our health to the physician; our fortune and sometimes our life to the lawyer and attorney.  Such confidence could not be safely reposed in people of a very mean or low condition.  Their reward must be such, therefore, as may give them rank in the society which so important a trust requires.”  Patients demand trust, but the marketplace may or may not honour and meet this consumer preference.  Some of the potential problems are the asymmetry of knowledge economists have talked about for many years.  Pioneering work by Bob Evans 25 years ago acknowledged that doctors know more, we hope, about diagnosis, about treatment options and about the prognosis.  That gives them a considerable amount of power, and implies that while the initial decision to consume health care belongs to the patient, once in the system, the people who are actually ‘demanding’ health care are the physicians.  They take control of demand.  This is not the simple, usual economic model of exchange.  The problem is that this gives providers an opportunity to breach trust and create demands for health care services that may be of little worth or benefit to the patient, but of considerable worth to the doctor.  They may decide they want a new Porsche, and so recommend treatments that are not necessary.  The scope for manipulation of demand by the profession is furthered by third-party payments.  Neither patients nor doctors pay directly for health care and thus neither has an immediate incentive to be efficient or to economise.  There is a relationship between trust and the placebo effect.  Placebo basically means ‘I will please’, and has been shown to have a considerable impact on people’s health.  The placebo effect is as high as 25-30% in some trials; that is, the act of a doctor providing a pill – in this case a pill of no consequence – actually seems to make patients better.  The effect is psychological, and produces some level of satisfaction in patients.  Markets had this problem of placebos, of quack remedies, in the late 19th and early 20th centuries.  At one point, expert opinion asserted that male ejaculation was exhausting; surgeons offered men vasectomies because they believed it would rejuvenate them.  Freud and W.B. Yeats actually had vasectomies based on this set of hypotheses, and the intervention may have produced a placebo effect.   - 4 - Conference Summary  November 10, 2000 This is from a 1988 editorial in the Lancet:   “Why is it deceitful to give a placebo if a large element of modern therapeutics is no better than a placebo? Is the gullibility of good hearted doctor preferable (and more ethical than) the skepticism of one whose prescription is pharmacologically inert, when the results are the same?”  As soon as we begin to go into the knowledge base, we begin to have some considerable problems.  Does the trust relationship allow doctors to use placebos?  What is their appropriate role, particularly when the placebos may have more effect, according to some trials, than pharmacological substances that are believed to be active?  Turning to the supply of trust, in recent decades it has become increasingly apparent that much is unproven.  This does not mean interventions do not work, but we simply do not know if they do.  We also know that a lot of what is proven to be effective is not delivered to patients who would benefit, especially the poor.  These things have been shown to occur even in supposedly ‘egalitarian’ health systems like in the UK and Canada.  For example, a recent study of the social class variation in the delivery of appropriate technologies in Ayr and Arran, an area of former coal mines and industrial wreckage, showed very clearly that interventions such as beta-blockers and statins were not being delivered to the poor after myocardial infarction, but were being delivered to the middle-class.  These are cheap proven technologies, and we know that the poor would benefit and live longer, but we are much better at providing them to the middle classes.  At the same time, we are relatively reluctant to actually put those sorts of measures into the system and use them thoroughly.  There is an awareness of variation in practice and there is awareness of variation in activity, partly due to public/private choices.  The UK experience of having doctors who work in large numbers on both sides of the public/private fence is of two-fold variations in the activity levels of doctors.  If we look at general surgery or urology or orthopedics, this variation exists even after adjustments for case mix and intensity. A cut in these variations or even a slight shift in the mean of the distribution could have a large impact on waiting list problems. But either physicians or managers are not using these results in the system. Instead, in 1998 our government decided to increase medical school intake by 20% on a base of 5,000, and in August 2000 announced a further increase of 1,000 training spots.   It is by no means clear that these extra doctors are needed, and the potential resource consequences downstream are quite considerable.  Routine measurement of variation in process and activity is not very thorough, and waiting lists may be a product of them.  The problem with the UK system is the enormous temptation for physicians to generate waiting lists because they, in turn, generate private practice income.  The failure to manage that, the failure of the Royal Colleges to actually govern these sorts of issues is noticeable; a conspiracy of silence surrounds the measurement of activity and outcomes.  In the mid-1970s the Rand Insurance experiment used a questionnaire, which is now called SF36 (Short Form 36).  It has 36 questions in it, is well validated, has been used in many trials, has been translated into about 20 languages and validated within different cultural situations, and yet we do not routinely use it in health care systems.  Its purpose is to measure changes over time in  - 5 - Conference Summary  November 10, 2000 the health or functioning of patients, measured in terms of simple questions about physical, social and psychological well being.  If health is the product of the health care process, why is it not measured?  We tend to look at mortality data, but that introduces the problem of adjusting mortality data for case mix severity.  We might also consider measuring functional status.  We have SF36 and the EuroQoL, or EQ5D, which some of my colleagues in York developed with people in northern Europe.  Again this is well validated in research, translated into numerous languages, and has been used on some large populations including some in primary care.  The use of these tools has been more readily adopted in the private sector than in the National Health Service in the UK.  The private sector has taken the SF36 and told the consultants (who, interestingly enough, are NHS consultants working in the private sector) that they must administer the questionnaire when the patient comes into the hospital and three months after discharge.  The private sector provides non-emergency elective procedures, including hip replacements, hernias, cataracts, and things like that.  They have not yet published anything, but they now have 50,000 measures three months out and are beginning to detect considerable variation in the success of doctors.  It seems that three months out, many people who have had their cataracts removed still have not gotten their vision back and are very depressed following the care they have received.  Such effects show up on the SF36.  Some orthopedic surgeons are less successful than others when physical functioning, social functioning and psychological functioning are measured three months after surgery.  One reason that the physician profession or Ministries of Health have not introduced something like the SF36 is the potential cost of administration.  Private insurance companies in Britain, however, have found that it is actually very cheap.  Doing the surveys at entrance to hospital and subsequently at 3 months costs about the same as a full blood test.  The question that deserves some exploration is the addition to knowledge and improvement in diagnosis you get from SF36 compared to a blood test.  Behind the SF36 issue is another problem that is emerging worldwide -- medical errors.  Interest in this topic arose partly from the Institute of Medicine report DH (Institute of Medicine, “To err is human” National Academy of Sciences, Washington DC, 1999), but there is lots of work in other countries as well.  We should accept that we all make mistakes, and we should accept that doctors and all other health professionals make mistakes and that sometimes patients will die as a result.  The question is when does the making of errors that we accept as part of life’s experience become negligence?  It may be that negligence is the failure to take proper care by responding to clusters of events.  But the numbers of events may be very small and thus quite difficult to pick up.  How can we develop systems to really begin to understand what is occurring?  The Shipman case in the UK is an interesting (if tragic) recent example.  This particular mass murderer was a General Practitioner who was convicted of administering overdoses of morphine to 15 people and killing them.  We suspect he actually killed 200, but part of his care was to advise the survivors to cremate the dead, which means there can be no autopsies.  This man was the most popular GP in his town, so by the usual satisfaction criteria he seemed to be providing good care!  The local purchaser was aware of his popularity and recognized that people were  - 6 - Conference Summary  November 10, 2000 queuing up to get onto his list.  The detection of his crimes was very difficult, because he killed these people slowly over a number of years.  If you go back and compare him to his peers, he shows up not in the extreme of the distribution, but in the middle.  Shipman has gone to prison but his discovery was not the result of the health service but because the local morticians became suspicious that a significant number of his patients were dying in his office.  Further examples include the Bristol crisis, where children with cardiac problems were treated with a very high-risk technique.  About 29 of these kids were killed by deficiencies in the surgical technique.  Other surgeons and locations used the same approach and also killed people, but they had lower volume, so it has been Bristol that has attracted all the attention.  This example was a classic example of a complete failure of the system.  An internal report was written, a report went to the Royal College of Surgeons and a report went to the Department of Health and nobody did anything about it.  Then we have a physician who formerly practiced in British Columbia and Ontario and was struck off there.  Neale left Canada and returned to the UK where he practiced in Northallerton in North Yorkshire.  There he continued to damage scores of women with deficient gynecological surgery.  We have another man called Ledwood, another gynecological surgeon, who did a lot of work in the private sector.  He used to call himself “the fastest cutter in the East”.  He may be right, but in the process he damaged scores of women very seriously.  The problem is how to develop detection systems that will actually identify these providers.  Do we have any data from which we might be able to calculate error rates?  Data from Australia and from the US (Utah and Colorado and the Harvard group) suggests that about 5-10% of hospital admissions are associated with medical errors.  You may have a one in ten chance; it seems, of going in hospital and being affected by a medical error.  And both the Australian and the US data are saying there is about a 0.3% chance of admissions being associated with error that results in mortality.  The US and Australian data were collected slightly differently, and as soon as adjustments are made for that differences begin to emerge.  The Americans may be much more conservative in measurement of errors because of negligence claims and the legal system. The Australians used two reviewers; the Americans used one generalist.  All of this is the start of a literature, which says that the systems do have detectable problems.  But we are not currently detecting them in the UK or in Canada, and in Australia and the US they are being detected only in small area studies.  If we are concerned about quality, medical errors should come pretty high on the agenda of beginning to develop systems and beginning to think through what goes on and how errors can be reduced cost effectively.  The North Staffordshire case of a physician accused of poor care of children in relation to his research work is a good example of how difficult it is to determine whether the errors people make are defensible.  As the story started coming out, the parents had extreme reaction, as you might expect, the politicians were exposed to a considerable amount of media pressure, so they set up an inquiry.  The inquiry took evidence and condemned the physician.  The British Medical Journal wrote an editorial and condemned the physician.  The doctor who was condemned then went to the Medical Defense Union and the Medical Defense Union hired Ian Chalmers from the  - 7 - Conference Summary  November 10, 2000 Cochrane Collaboration.  I call Chalmers “Forensic Fred” because he goes through everything extremely carefully.  In September of 2000, there was an article in the British Medical Journal from Chalmers and his colleague about this inquiry, which demonstrated that due process was absent and the accusations were ill founded1.  The inquiry team was under intense pressure from the media and from the families and did not investigate the situation carefully enough.  They made accusations such as there was no peer review of the research protocol.  There was; it had gone to the Medical Research Council, had been reviewed by eight people, had been given an A+ award but not funded by the MRC; it had been funded regionally.  It was asserted in the inquiry that the parents did not give consent for their children to be randomized.  They went through all the records and found in the records that all the parents had actually accepted the randomization.  The BMJ (British Medical Journal) published an article showing there should be an inquiry into inquiries, and the editor was very sad that he had jumped the gun based on poor evidence.  The purpose of this anecdote is to illustrate that if we are going to address the issue of medical errors, we have to accept that they are extraordinarily difficult to measure and investigate.  It is easy to jump the gun and condemn people before having a full understanding of what is going on.  The concept of medical error has been documented at great length for decades, yet we have not done much with it.  Part of the difficulty is that even if you have evidence it can be tricky to translate it into practice.  Hospital-acquired infections are a good example.  The Institute of Medicine report and Australian data show that one of the major causes of medical errors and disaster is hospital-acquired infection.  Hospitals are dangerous places to be.  A simple, cost-effective technology would reduce infections -- washing hands.  In October 20002 the Lancet published a big study from Switzerland – a before-and-after design, but reasonably robust.  This study showed the positive impact of a sustained attempt to get doctors and nurses to wash their hands on general infection rates and methicillin-resistant staphylococcus aureus (MRSA).  The problem is how to get doctors and nurses to change their behavior.  As the Swiss emphasized, it is the doctors who are more reluctant to change than the nurses.  An American study last year ‘empowered’ patients by training the patients to say every time a nurse or doctor came near them, “Go and wash your hands!”  This approach appears to be successful but the outcome measures were not infection rates.  They instead used a much more crude outcome of a 30-40% increase in the use of soap (a finding replicated by John Radcliffe in Oxford).  This forces a return to the issue of how much trust should we have, and can we do things better?  The medical errors described in the literature are:  wrong drug, wrong dosage, wrong surgical procedure and infection control issues.  All of these are amenable to better measurement and to                                                  1 E. Hey and I Chalmers, Investigating allegations of research misconduct: the vital need for due process, British Medical Journal, 200, 321, 2000  2 Effectiveness of a hospital wide programme to improve compliance with hand hygiene, D. Pittet, S. Hugonnet, S. Harbarth, P. Mourouga, V. Sauvan, S. Touveneau, T.V. Pemeger and members of the Infection Control Programme, Lancet, 356,1307-12,2000  - 8 - Conference Summary  November 10, 2000 the detection of clusters.  But if clusters emerge, is there research evidence available about alternative ways of dealing with them?  We have both an absence of systematic appraisal and an avoidance of evidence-based incentive systems.  In the UK situation, purchasers have the potential power to contract and say, “as part of my buying surgery from you I want to see these characteristics for your infections control policies.  And if you don’t put them in I won’t pay you!”  How can we begin to encourage intelligent contracting between organizations, and how can we begin to change the incentives at the ordinary level at the ward?  We know about medical practice variations.  We know we have failed to measure outcomes.  And we have this new industry called medical errors.  We also have regulatory structures.  In the UK, the General Medical Council registers physicians and is currently putting in place re-validation procedures to occur every five years.  They are also working on developing portfolios for practitioners that would include activity rates in the public and private sectors and perhaps outcome measures in due course.  These sorts of mechanisms need to be looked at very carefully in the light of our concerns about quality.  We also need to look at educational bodies such as the Royal Colleges.  The Royal Colleges in Britain were initiated and started by people like Henry VIII.  The founding documents are written in consumer protection language -- language that attempts to protect people from quacks.  Why aren’t the Royal Colleges now offering us the degree of consumer protection that we should expect?  Why are there 17 Royal Colleges instead of one?  The Royal Colleges invited me to go and talk to them about how to reform the Royal Colleges.  One of the proposals I made was that there should be three Royal Colleges; one doers – surgeons; one thinkers – physicians; and one technos – radiologists and that sort, and that they should be involved in a much more systematic way with performance management.  They should use the activity and outcome data available, and be more involved in the question of quality.  This raises the question of who – what organization or entity – is in charge of quality.  One goal for our systems is to measure much better the medical errors that take place.  Another is to measure much more carefully the quality rather than just the process of what our systems do.  However, even to the extent that we actually measure process, we should be much more systematic in the management of that and pay attention to incentive structures.  We know there are enormous variations, but not always which rates are correct.  Being ‘low’ or ‘high’ is not enough information particularly as variation is normal to some extent in all human activity.  Many people will not report errors for fear that there is going to be blame.  We need to begin to create a culture where errors are reported because we think we (including the ‘perpetrator’) can learn something from them.  When a former chief executive of the hospital was a patient I went to see him and was there when a nurse came in to give him his medication.  He said, “this is the wrong medication!”  She said, “No it’s not, this is the correct medication.”  A search through records showed that it was the correct medication, but the previous day he had been given double the dose of this rather active ingredient.  I asked the organization whether the incident was logged.  It was certainly not an isolated incident.  What could be learned from this?   - 9 - Conference Summary  November 10, 2000 Economists will step back and say capital substitution may have a role when labour practice exhibits variations.  There was enormous variability on the production lines for cars, so humans were replaced with robots.  They paint them better, they screw the doors on, and they can be programmed.  There were variations between airline pilots and high rates of airline accidents in the 70s and 80s.  Pilots are now re-accredited every six months to ensure they are not alcoholics, they are not drug addicts and they can still fly the plane.  Computers were also put in cockpits so that when you fly to Britain or Canada you can be pretty sure that for most, if not all, of the journey, the computer is in charge.  We do not do anything similar with doctors.  We just assume that they are not addicted and that they can still practice.  In the US, in Singapore and in Europe there is now significant investment in robotics in medicine.  Within 10 years it will be possible to do hip replacements by robotics.  The computer will do it – scan you, make the incision, do the joinery and put you back together again.  Experiments show that the robots can perform these procedures with much less blood loss and a faster recovery time.  Given that we have failed to measure quality, to detect and remedy errors, we are going to have to invest in capital, in systems which control prescribing and reduce variability in surgery.  And it may well be that it will take capital substitution and changes in the technology which are, as yet, just hidden away in relatively esoteric university departments.  The other thing we are clearly going to have to invest in is measuring health status and linking such measures to activity and funding.  Such systems are essential to the management of waiting lists.  These are clearly the political must-do’s.  But it is really quite difficult because there is clearly interdependence among waiting lists, waiting times, emergency activity and outpatient service use.  As soon as you scratch the surface currently covered by waiting lists, you begin to see that this is just a thin veneer over a range of more strategic issues.  Meanwhile, the media focus is on waiting lists and this reduces your capacity to deal with the more fundamental problems.  If we could deal with our medical errors, we could perhaps deal with our waiting lists better.  And if we could reduce our activity variation, we could deal with our waiting lists better.  The constant focus on this waiting tip of the iceberg does not let us get down to the fundamentals of managing the health care system.  There are failures in the governance of medicine.  They have failed to measure, monitor and manage errors and failures.  It is everyone’s responsibility to try and work to do that better.  The remarkable thing is that despite all this discussion about errors, individual practitioners and the medical profession as a whole remain very much respected, so the trust relationship is still there.  In Europe, doctors are still more trusted than journalists and politicians – they are some of the most trusted people in European societies.  But self-governance appears to be ineffective and the risks of indolence in reform are clearly quite great.  Voltaire is said to have remarked:  “The role of the doctor is to amuse the patient while nature takes its course.”  Doctors are clearly important in the system.  Patients can be relaxed and assured if the emperor finds new clothes, but without change, the profession may lose its prestige and its high payment.  Patients will lose the placebo effect and will acquire considerable anxiety about the relative skills of different doctors.  If that is the desire of dilatory reformers, to undermine that trust relationship, we seem to be going very well in that direction.  A greater appreciation of the underlying problems is emerging, but I do not think we  - 10 - Conference Summary  November 10, 2000 should underestimate those problems because the political agenda is one thing, and improving health care and remedying deficiencies in quality is quite another.  - 11 - Conference Summary  November 10, 2000 Waiting for Care:  Effects of Wait Lists on Quality Claudia Sanmartin   Waiting lists and waiting times are hot button issues in health care.  But what is the relationship between waiting and quality of care?  In order to understand that relationship, we need to consider first how ‘waiting’ is being defined and interpreted, second the quality of the data and measures used to define waiting, and third whether there is any evidence of an association between waiting and quality.  This latter question can be further divided into whether waiting is a symptom, a problem, or both.  Most of the information presented here is derived from the work of the Western Canada Waiting List Project and recent analyses of the BC Ministry of Health waiting list data.  Newspapers across the country have been filled with reports about waiting lists and waiting times.  The headlines offer an excellent portrayal of how waiting lists and waiting times are presented to the public as a potential threat to access to and quality of health care.  For example, one headline in BC accuses the Ministry of Health of not producing a second waiting list report to hide an increase in waiting times.  A Montreal newspaper declares “Cancer Waiting Lists Underline Health Care Mess” and  “Province to Rush Cancer Patients Treatment in the US”.  Finally, an announcement by Minister Rock of increased funding was depicted as a means of addressing waiting list problems.  Health care organizations are also beginning to interpret and report on waiting lists.  The Canadian Medical Association, in conjunction with Angus Reid, has been conducting regular reports on access to health care.  They use public opinion polls to determine whether the public feels there has been deterioration in access to health care, with deterioration measured in part as waits in hospital emergency rooms, waiting times for surgeries, and waiting times for tests.  A 1998 report prepared for Health Canada assessed the waiting list situation across the country.3  Interviews with representatives from provincial Ministries of Health revealed that they did not feel waiting lists were their most urgent problem, but they did see waiting lists as a means to monitor the performance of the health care system.  The Canadian Institute for Health Information released its first report on health care in Canada in 2000.4  The report discussed the use of waiting times as a measure or indicator of the health care system.  Unfortunately, they were mostly not able to match this discussion with results because there is a lack of comprehensive and good quality data across the country that would enable such analysis.  Despite this, waiting lists and waiting times continue to be used as indicators and measures of performance within the system.  The media and system experts view waiting lists and waiting times as problems and specifically as threats to access to care that require action.  Governments across the country have started taking on this challenge.  For example, millions of dollars have been allocated to reduce waiting lists for cardiac care in Ontario.  In British Columbia, money has been allocated to reduce wait lists for a whole range of services.  In Quebec, budget debates have shown an opinion split                                                  3 McDonald P, Shortt S, SanMartin C, Barer M, Lewis S, Sheps S.  Waiting Lists and Waiting Times for Health Care in Canada: More Management!! More Money??  Ottawa (ON): Health Canada; 1998. 4 Canadian Institute for Health Information (CIHI). Health care in Canada : a first annual report. Ottawa (ON): Canadian Institute for Health Information (CIHI); c2000.  - 12 - Conference Summary  November 10, 2000 between using the surplus to help reduce the deficit and injecting it into the health care system to remedy problems that have evolved over the years.  It is clear that those who argue for the latter believe that without more money there will be further deterioration of the system and further increases in waiting lists.  Meanwhile, the federal government announced an infusion of money to take place over the next five years.  One television advertisement that quickly followed this announcement specifically said that this money was going to reduce waiting lists and waiting times.  While there may be a problem, it is difficult to quantify without knowing something about the quality of the measures used.  As mentioned previously, there is a lack of comprehensive data on waiting lists and waiting times across Canada.  Researchers and policy makers do not always know how the data are collected, how waiting times are calculated, or who, in fact, the waiting lists represent.  A few of the more common definitions include waiting times to see specialists or to receive diagnostic or surgical services.  But even these are not standardized or universally accepted.  Some argue that diagnostic waiting times should be measured from the date a general practitioner requests a test, while others promote measuring from when a specialist requests a test.  The lack of a universally accepted definition(s) of waiting times means that there is no guide in the collection of data and thus no comparability across jurisdictions.   It is important to consider clinical pathways and who exactly are counted in these data.  Patients can take various routes when seeking and undergoing treatment; they may experience various waiting times in the care process, so there is a need to ensure the data are collected consistently.  For example, some general practitioners request diagnostic tests in addition to referring the patient to specialists so that by the time the individual gets to the specialist, the results of the diagnostic tests are there and the decision to treat can be made.  Other general practitioners refer patients to a specialist, the specialist then refers the individual for a test, and the individual then needs to wait for that information before the decision to treat can be made.  It is important, therefore, to be clear about which waiting times we are interested in, how they are measured and what individuals are represented.  A recent study using the BC Surgical Wait List registry data examined when individuals were placed on the list during the course of care with specific attention to the consistency across hospitals and physicians.  Four common procedures were included: hip and knee replacements, coronary artery bypass grafts and cataract removal.  Physician payment data were used to create trajectories of care from the first contact with the specialist physician.  The majority of patients were placed on lists following their last consult, if they had more than one, or following their one and only consult.  A natural follow-up study would be to make these implicit (and reasonably consistent) decisions explicit and start to create standards about how the data are collected – in this case, when patients are placed on the waiting list.  An assessment of data quality needs to precede any assessment of the relationship between quality and waiting lists and waiting times.  Standard definitions and guidelines need to be used to determine when, and if, patients should be placed on waiting lists.  This first step in developing valid and reliable measures has been one focus area of the Western Canada Waiting List Project.  - 13 - Conference Summary  November 10, 2000 The second issue around quality is benchmarking.  It is very difficult (if not impossible) to determine whether waiting lists and waiting times represent threats to quality if it is not clear where the goalposts are.  How long does waiting for healthcare threaten the outcome?  There are very few universally accepted standards for maximum waiting times for selected procedures, though a myriad of approaches have been used to try and establish them.  Retrospective surveys of patients may ask “Was the length of time you waited acceptable?” or attempt to establish willingness to pay by asking  “Would you agree to having your taxes increased or pay out of pocket in order to have your waiting time decreased?”  Surveys asking physicians their opinions have been used by the Fraser Institute and the British Medical Association.  Clinical panels, such as the Cardiac Care Network in Ontario, convened to examine the clinical evidence, develop an urgency rating scale separating urgent and emergency cases and set waiting time benchmarks for each.  There have also been broader national policy approaches adopted in countries like the UK and Sweden.  The UK Patient Charter, for example, guaranteed patients access to in-patient services within 18 months.  In Sweden, a policy introduced in 1992 guaranteed access to 12 selected procedures within three months provided that patients met the clinical standard for those procedures.  Policies involving wait lists are constantly evolving, but the establishment of benchmarks to guarantee acceptable and maximum waiting time on a per procedure basis is essential.  Without them, there is no means of assessing the quality of care, or ensuring that it is not compromised.  Ideally, this type of exercise should be based on the evidential material from a variety of sources, including patient opinions, clinical expert panels and clinical evidence.  This is the second area that the Western Canada Waiting List Project is currently exploring.  There are at least two approaches that could be used to look at the relationship between waiting lists and waiting times as a symptom of other problems in the health care system.  Waiting lists and waiting times can be thought of as existing between the potential access to services and the realized access to (or utilization of) services.  The literature on access to health care tells us that there is a whole range of factors at the physician, hospital and patient levels that affects availability of services, including .the number, and distribution of hospitals and physicians, individual physicians’ access to operating room time, or the number of beds available for elective surgery.  All these factors can indirectly affect waiting lists and waiting times through the availability of personnel or services that are required to provide the service – they have an effect on potential access.  Other effects on waiting lists and waiting times come from explicit waiting list management policies adopted by either hospitals or physicians.  These include whether patients are placed on a waiting list, when they are placed on a list, how they are prioritized and how they are managed while they are on that list.  If we start to uncover which factors affect waiting lists and waiting times, other problems within the health care system may become more obvious.  The second approach is to look at the more direct (realized) effects of waiting lists and waiting times, either on patients as seen through their utilization of services, or through the health policy and planning route.  Numerous governments have adopted various policies to deal with waiting lists and waiting times; we need to think about how those policies affect the quality of care in other sectors within the health care system.  There is very little empirical evidence about which factors affect waiting lists and waiting times.  Despite this lack, it is known that there are clinical factors, such as whether a case is urgent or  - 14 - Conference Summary  November 10, 2000 emergent, that may have a significant effect on a patient’s wait time.  Non-clinical factors such as employment status or physician preference may also determine wait times.  Some studies have found hospital factors which change wait times such as whether the hospital has residents, the number of beds it has and its surgical rates.  Unfortunately, the studies conducted to date are based on small sample sizes and so are not generalizable.  A comprehensive model is needed, using appropriate modeling techniques, to provide the necessary evidence to begin to make policy and other changes that will improve the care process.  Analyses of the BC Waiting List data show that total waiting time is longer if a patient is placed on a waiting list at the beginning of the course of care rather than at the end of care.  This is as logic would dictate, but illustrates how management policies dealing with waiting lists can affect individual waiting times.  Physician factors that had a significant effect on waiting times were:  access to hospital services, the average surgical rate or the average number of cases per month and their own personal surgical wait list size.  Important hospital variables that tended to increase waiting times were the total number of operative cases, and the proportion of those cases being done on an urgent (versus emergent) basis.  Total waiting times also varied with time of year of placement on the list, with traditionally slow periods resulting in longer overall waits.  The processes that are involved in delivering care are responsible for some of the variation in waiting times.  A better understanding of the factors involved could provide more specific information about where the problems are within the health care system that result in longer waits and longer queues.  The direct effects of waiting lists and waiting times on quality also lack a great deal of empirical evidence.  There have been some UK surveys looking at patients’ experiences while awaiting orthopaedic care.  These patients report pain and restricted mobility as well as stress and anxiety because of the uncertainty surrounding the timing of their treatment.  Research on outcomes related to waiting times has concentrated in the area of cardiac care; little to no work has been done on non-urgent cases or procedures such as hip and knee replacements and cataracts.  It is not clear whether long waits affect the efficacy of the procedure or the intensity of care required.  The uncertainty around such questions needs to be addressed to determine whether (or when) waiting lists have a detrimental effect on the quality of care for patients.  There may also be negative effects on physicians, for example if physicians have to help patients cope with the stress and uncertainty of waiting.  But there is virtually no information about this either.  As new policies are implemented to try and deal with the problem of wait lists, an examination of the follow-on effects is essential.  Hospital departments and institutions are using waiting lists and waiting time measures to allocate resources, namely operating room resources, across specialties or within specialty groups.  This may have a positive effect for patients who are waiting on those particular lists, but it may have a negative effect on the access to care for patients not waiting for those procedures.  Furthermore, this approach shifts the debate towards making sure that the provision of services meets the guidelines for appropriate wait rather than trying to ensure that appropriate services for a population are being provided.  Government policies can have various effects on the quality of care.  Many countries, in addition to provinces in Canada, have adopted policies of targeted funding, where very specific procedures are given large sums of money to reduce waiting lists.  Once again, this might have a  - 15 - Conference Summary  November 10, 2000 positive effect for the people waiting for those specific procedures.  It might mean, however, less money for people waiting for other services.  It might also mean less money to invest in strategies such as information systems and management systems that might have a more long-term effect on dealing with the issues of waiting lists and waiting times.  There is very limited evidence to suggest that targeted funding policies are a positive solution to wait lists.  However, there is some evidence to suggest that the effect of targeted funding policies is very short lived.  Once the money has run out, waiting lists and waiting times start to increase again.  Some countries have adopted alternative policies.  Sweden has set a maximum waiting time of three months for patients who meet certain criteria including needing urgent or semi-urgent care.  The result was increased waits for people not covered by the guarantee.  Waiting times for cases where the patient was referred from general practitioners to specialists also increased.  The increase happened because patients were not referred to specialists until the specialist could, in fact, take them and meet the maximum weight time guarantee.  This is a clear example of some the potential unintended effects of setting policies; the policies will not exist in a vacuum, but instead will produce effects on other patterns of care.  There are clearly many groups and media sources that want us to believe that waiting lists and waiting times are a threat to access to and quality of health services in this country.  The lack of comprehensive and comparable data and benchmarks makes it difficult to know to what degree that potential threat is real.  There is very little evidence regarding the effects of waits, not only on patients, but also on their families and on health care providers.  Policies have and will continue to be introduced with or without appropriate information to back them up.  We should consider the development of this information a priority, but we should also be evaluating the effects of these policies on waiting lists and waiting times as well as their effects on other parts of the health care system.  - 16 - Conference Summary  November 10, 2000 Lessons from the Western Canada Waiting List Project David Hadorn   There has been a lot of work in Canada on the issue of waiting lists, including a large, recent report sponsored by Health Canada.5  The recommendations of that report led directly to the Western Canada Wait List project.  Another inspiration for the project was work done in New Zealand on wait lists specifically and prioritization more generally.  There has been long-standing agreement amongst the various players that something like the Western Canada Wait List project had to happen.  Partners in the project included medical associations, research centres, regional health authorities, ministries of health and the Canadian Medical Association; this type of partnership and cooperation is unprecedented.  My role, similar to my function in New Zealand, was to work with the doctors and get them to own the project.  The intent was not merely that they put up with it, but that they actually come to believe it was a positive move for them, their patients and for the health care system, and not some cynical ploy by government functionaries to further cut costs on the backs of working men and women.  It is no surprise that there is a lot of distrust between doctors and government.  In New Zealand, this was an important issue because there is a private sector there, and they use the criteria to decide who is going be treated in the public and private systems.  One of the first decisions to be made for the Western Canada Wait List project was how best to proceed to establish a 'nervous system' for the waiting list system.  The only existing knowledge was some very crude measures, like what the diagnoses were, where doctors were, how many patients were on a waiting list at a given time.  That does not provide much (if any) information about who is actually waiting for services, their level of pain, their level of disability, and the implications for their life expectancy if they are not treated.  What was missing was a way to develop a case-mix measure that would consider pain, disability and so on.  This would provide some texture, some clinical feel, and would also allow for comparison to other lists and over time.  The system used now in hospitals in New Zealand for prioritizing surgery cases is based on point counts, with points accumulated based on pain and disability.  But when we started the Western Canada Wait List project in 1998 it was not at all clear that doctors in Canada were going to buy this.  It turned out, however, that they did embrace the idea.  An example is criteria developed for hip and knee replacement.  These included basic information on what joint is involved and who the doctor is, as well as the various criteria through which patients accumulate points, including: pain on motion, such as walking or bending; pain at rest, while sitting, lying down or pain that prevents sleep; abnormal findings on physical exam; potential for progression of disease, and so on.  Doctors developed these criteria by reporting the facts they consider when deciding the urgency for a particular operation -- its relative priority.  There were six areas considered by the project: hip and knee replacement surgery, cataract removal, general surgery procedures, magnetic resonance imaging and children’s mental health.                                                  5 McDonald P, Shortt S, SanMartin C, Barer M, Lewis S, Sheps S.  Waiting Lists and Waiting Times for Health Care in Canada: More Management!! More Money??  Ottawa (ON): Health Canada; 1998.   - 17 - Conference Summary  November 10, 2000 In each area the criteria developed initially through the judgment analysis were used to follow hundreds of patients.  The clinical information from these patients was fed into a regression analysis that provided us with the relative points to be awarded for each of the criteria, scaled so that total possible value was 100.  The scale was then tested for face validity with doctors and did very well; in other words, the point total derived from the criteria matched well with the overall urgency the patient would have received if they had not been using this system.  When the project first started, the general surgery procedure considered was going to be inguinal hernia.  The developers wanted to consider a simple, common, banal procedure to see how that worked, but the doctors wanted to tackle something that was more of a challenge.  The ultimate desire is to create a system that can look at all operative procedures to prioritize both within and across surgical areas.  This is a limitation with the other interventions considered, although by and large the criteria for hip and knee replacement (for example) could probably be used more broadly for all orthopaedic procedures and there is some interest among orthopedic surgeons in testing that.  The general surgery criteria, in contrast, are so generic they could be used for anything.  The factors included are inputs into the so-called quality of adjusted life year concept. It could potentially be made even more generic.  For example, life expectancy implications of the condition without procedure and expected improvement in life expectancy could be applied to almost anything.  One of the original ideals for the Western Canada Wait List project was that we would try to look across procedures, to advance the methods for comparing hips with cataracts and so forth.  While the project has not been able to address this as much as desired, this might be a good starting point for future work.  The notion of developing point counts was adopted in part because it is one with which doctors are familiar.  The Apgar scale for newborns includes five criteria -- colour, respiratory effort, heart rate, muscle tone and reflex irritability -- each with score of 0 to 2.  You add these up to get a total score.  This is still the best indicator of neonatal stability.  The APACHE score is used in intensive care units; MPM mortality prediction models scores provide a number that reflects your chance of dying within a certain period of time; the Glasgow coma scale assesses reflexes, pupils and level of consciousness of people who have suffered head injury to predict subsequent outcome.  One basic function of these WCWL point count measures is to prioritize among patients on waiting lists.  That is the obvious and most immediate function; somebody with more urgency should ideally receive treatment before someone with a less urgent situation.  This is where we start touching on the notion of quality of care, because part of quality has to do with fairness.  It is only fair that people who are suffering the most, who have the most to gain, should have to wait less than those who are suffering to a more minor extent.  There is also an efficiency issue: we want to eliminate as much suffering and disability as we can for the dollars that we have to spend.  In New Zealand, the waitlist software developed takes into account how many patients have come in over the last six months, how many are currently on the waiting list, and the average emergency case load, all so that operating theatre allocations and so forth can be factored in.   - 18 - Conference Summary  November 10, 2000 Another function of the point count measures, one that comes more from a research and analysis perspective, is to serve as a case-mix measure for assessment and comparison.  Two intensive care units, two surgeons, or two hospitals that perform coronary artery bypass grafting will virtually always have raw outcome mortality rates that are different.  The claimed explanation for the higher rate will inevitably be that the patients were sicker, rather than that care was sub-par, as is sometimes suspected.  Case mix measures can help settle this question empirically.  A January 1997 British Medical Journal article reported on the work in New Zealand.6  A single nurse trained in a coronary care unit went around to all four centres in New Zealand that performed coronary artery bypass graft surgery and re-scored all the patients on the waiting list that had already been scored by the doctors.  The purpose was to create a standardized score following a set of rules for classifying various symptoms.  For example, when is angina severe angina?  Where there were disagreements, she went over them with the doctor on a case-by-case basis.  The result was a fairly normal bell-shaped curve for severity, which is, incidentally, what tends to result.  A separate analysis of each hospital showed there were different peaks for different hospitals; there were differences in the case-mix across location.  This also allowed a comparison of waiting times by severity across hospitals.  In some cases the differences were significant, which indicates inequity.  This signals a need to move resources from one place to the other.  Of course, the mention of a resource shift makes people start to get nervous and to reassess the information they provided.  This is why gaming always becomes an issue.  Once the rules and point system are understood, the expectation is that people are just going to write down what will make them look the best.  The threat is probably exaggerated, and there are ways of minimizing the potential impact.  When you have the information produced from the point system, you can actually then calculate how many operations need to be done in a given week to get the priority down to a particular level.  It is furthermore possible to estimate the costs associated with accomplishing that. Physicians in New Zealand were able to assess the existing patient load and estimate the cost of reducing the priority level to one they deemed reasonable.  Service for anyone below that level was considered unnecessary; they could purchase care in the private sector if they desired.  The cost estimate was provided to the politicians, and the politicians were astonished that the price tag was only about $5 million -- much less than the anticipated 10s of millions.  They judged this to be a small price to pay to keep the doctors happy.  So the process actually helped the doctors make their case for more money.  In fact, probably the number one reason for the doctors' participation in the Western Canada Wait List project was that they felt that this information would be more powerful at making a case for increased funding than anything they have had access to in the past.  Patients can be stratified by their point count total.  This in turn makes it possible to describe the pain and disability suffered by the average patient at an average point count, thus providing a point of understanding.  This approach was very effective in New Zealand.  A presentation to the Minister of Health impressed her, and made her believe cabinet would be convinced when she                                                  6 Hadorn DC, Holmes AC. The New Zealand priority criteria project. Part 2: Coronary artery bypass graft surgery. BMJ. 314(7074):135-8, 1997 Jan 11.  - 19 - Conference Summary  November 10, 2000 went to make a case for increased funding.  In fact, more money did follow in New Zealand.  Initially, $130 million was allocated for clearing waiting lists, followed by another $180 million.  This represents the end of Phase I for the Western Canada Wait List project.  There have been some lessons learned.  For example, doctors accepted a point count, a quantitative measure of severity, and thus a very powerful analytic tool.  Surgeons accepted generic criteria, which bode well for the ability to compare across and not just within surgical procedures.  Mental health professionals gave preliminary acceptance to the criteria for paediatric mental health.  These criteria will be subject to further testing.  Magnetic resonance imaging was not surprisingly the toughest one; it is a diagnostic procedure, so requires estimates of the chances it will find something clinically useful, as well as the likelihood that treatment will change because of that information.  This one perhaps has the highest likelihood of failure in its present form, and will be undergoing further testing.  It is now up to the regional health authorities, the doctors, and others to work together to actually implement the lessons.  The Western Canada Wait List project undertook more performance evaluation than was done in New Zealand.  Many more patients filled out the forms and had an overall urgency rating completed.  We did tests of inter-rater reliability, asking different doctors to look at and rate the same patients.  The results were a little uneven, but pretty good overall.  Cost of the procedures has not yet entered consideration in the priority framework, but at some point, in theory, cost-effectiveness ratios could be developed, though this is clearly not of much interest to the doctors.  One interesting note is that this process was adapted from New Zealand's experience, where there is a public/private mix in financing health services.  There, they have identified both a 'financially sustainable threshold' and a 'clinically desirable threshold' and now have the capacity to see what the gap is between the two.  Perhaps because of the origins of some of these ideas, there has been some lessening of reluctance to consider that maybe Canada could learn from that.  Maybe there is some merit to the notion of setting a threshold, providing service to patients who are at the higher levels of suffering, and dropping off those near the bottom.  “These people probably do not need the service anyway and if they did why not let them purchase it privately?”, some people ask.  These kinds of discussions have not been encouraged during the WCWL project, but they have come up, though the above view is obviously a very simplistic way to look at the health system.  Existing research shows that establishing a private system alongside the public system in this way does not make any sense from either an efficiency or equity perspective.  One idea the doctors really resisted was making comparisons across treatments.  We collected a fair amount of generic health information in different procedures, such as the SF36, that could be used to make these comparisons.  Major literature reviews provide details on studies, measures used, and a framework for doing this sort of comparison.  But the notion of transferring dollars from one area to another -- among specialties -- always raises concerns.  The fear is that some types of interventions, for example hernia operations, will always suffer in comparison to others, like cataract surgery.  This may be true, but at some point sooner rather than later, this needs to be tackled.  Perhaps the doctors' confidence will be won with proof of the sophistication of the techniques currently available for analysis.  Hopefully, that day will come, and this sort of work will facilitate getting there.  - 20 - Conference Summary  November 10, 2000 Facing the Big Downer at the Front Line:  Quality of Care in the Primary Care Treatment of Depression Elliot Goldner  The quality debate is, essentially, about the very nature of quality—What does quality mean in the context of health care delivery?  How is quality best measured?  What interventions are most likely to improve quality?  This paper targets the subject of quality as it relates to the delivery of mental health services, and in particular, the delivery of services to individuals suffering with depression.  There is a caveat emptor to this paper.  The realm of mental health—psychiatrists, psychologists, etc.—is sometimes perceived as being slightly eccentric.  Consequently, the mental health sector is often regarded as being located outside the general medical system.  Mental health professionals are viewed that way, the system experiences it that way, and the sector operates that way.  So there may be some unique aspects of the treatment of depression and mental health problems that are not entirely applicable to the other health care delivery issues under examination in this volume.  Mental illness creates a profound burden of illness, causing suffering, disability, hospitalization and suicide worldwide.  It has been estimated that approximately three percent of the adult population in Canada is affected by mental illness causing serious disability and requiring significant treatment intervention.  The World Health Organization (WHO) has determined that mental illness is one of the largest contributors to disability worldwide.  When loss of life and disability are estimated using Disability Adjusted Life Years (DALYs), more than ten percent of the total burden of human disease is attributed to mental disorders.  Of the ten leading causes of disability worldwide, five are mental disorders.7   Not surprisingly, this WHO ranking identifies quadriplegia as the most severely disabling condition.  The second is dementia, followed by active psychosis, paraplegia, blindness. major depression, drug dependence, HIV infection, alcoholism and total deafness.  Five of these were either mental illness specifically or drug or alcohol dependence.  In many ways this is not surprising – disorders that can completely impair a person’s ability to think, to emote, to have normal kinds of mental experiences and functions are, clearly, very disabling and have a tremendous impact on the quality of people’s lives.  In addition to the WHO ranking system, a study by Ormel and colleagues examined disability across a variety of cultures and found that psychopathology was the factor that best predicted disability.8  A prospective study undertaken by Broadhead and colleagues, involving a sample size of 3000 individuals, found that depressed individuals were approximately five times more                                                  7 7 Murray DJL, Lopez AD (Eds.). Global Burden of Disease: A Comprehensive Assessment of Mortality and Disability from Diseases, Injuries and Risk Factors in 1990 and Projected to 2020. World Health Organization. Harvard University Press, 1996. 8 Ormel J, VonKorff M, Ustun TB, Pini S, Korten A, Oldehinkel T. Common mental disorders and disability across cultures. Results from the WHO Collaborative Study on Psychological Problems in General Health Care. JAMA. 272(22):1741-8, 1994 Dec 14.  - 21 - Conference Summary  November 10, 2000 likely to be disabled.9  These individuals experienced difficulty functioning and had significant periods of time away from work.  Epidemiological studies done in different jurisdictions across Canada show mood disorders and depression are very common.  In any one-year prevalence figures for these illnesses are in the range of four to eight percent of the population.10  Murray and Lopez11 predict future trends in the burden of illness created by mental disorders.  Their work suggests that the burden of illness caused by mental disorders will continue to increase due to changes in the age of the population and to social and economic factors.  It has been estimated that, within the next two decades, depressive illnesses will become the second leading cause of disease burden worldwide and the leading cause in economically developed countries such as Canada.  Depression is a significant public health problem that warrants our attention.  But what do we really mean when we talk about depression?  Is depression really a medical condition or is it more of a social condition?  In clinical practice and epidemiological studies, depression is usually divided into five forms.  One form is major depression, which is usually marked by a very depressed mood and inability to sleep, and results in significant disability.  Dysthymia refers to a less severe form of depression in terms of the severity of symptoms.  However, dysthymia is usually chronic in nature, sometimes lasting for many years.  Bipolar disorder, formerly known as manic-depressive illness, is characterized not only by episodes of depression but episodes of elevated mood or mania, sometimes with psychotic episodes.  This is a disorder in which genetic factors are quite significant.  Post-partum depression occurs in women following childbirth.  It can be very severe in nature, sometimes resulting in suicide.  Finally, bereavement is the form of depression that follows the death or loss of a loved one.  Bereavement is usually time limited with no sequelae.  There are some important problems with these categories of depression.  In particular, these are constructs that have been pulled, to a large extent, from specialty practice and may not reflect, in fact, the forms of depression that are seen in primary care.  Our knowledge of depression, what it really means, what causes it, is still very primitive.  Typically, depression is diagnosed using a list of symptoms.  If an individual presents with a number of these symptoms, a diagnosis of depression is assigned.  This is an inexact and simplistic way of making a diagnosis; nonetheless, it is commonly used in both the mental health sector and the primary care setting.                                                   9 Broadhead WE, Blazer DG, George LK, Tse CK. Depression, disability days, and days lost from work in a prospective epidemiologic survey. JAMA. 264(19):2524-8, 1990 Nov 21. 10 Murray DJL, Lopez AD (Eds.). Global Burden of Disease: A Comprehensive Assessment of Mortality and Disability from Diseases, Injuries and Risk Factors in 1990 and Projected to 2020. World Health Organization. Harvard University Press, 1996. 11 Murray DJL, Lopez AD (Eds.). Global Burden of Disease: A Comprehensive Assessment of Mortality and Disability from Diseases, Injuries and Risk Factors in 1990 and Projected to 2020. World Health Organization. Harvard University Press, 1996.  - 22 - Conference Summary  November 10, 2000 A potential problem with this approach is that it may lead to an over-medicalization of depression, and result in unnecessarily treating individuals with pharmaceuticals, such as Prozac (fluoxetine) or other antidepressants.  Conversely, it has been proposed that the threshold for treatment of depression is currently too high and consequently many people who would benefit from treatment are not receiving adequate care.  The debate is ongoing.  There is some evidence to support the concern that a number of individuals are not receiving adequate treatment for depression.  Studies have shown that approximately 60% of the cases that would be picked up through a standardized interview process are missed in the primary care setting, suggesting that a substantial number of cases may not be receiving appropriate care.  However, careful examination of these study results shows that family doctors do identify the more severe cases of depression; the sensitivity in the severe cases is very good.  It is the cases of minor depression, the ones that are at the lower level of the spectrum, that family doctors will typically not identify or recognize as depression.  Because the efficacy of treatment interventions for minor depression has not been demonstrated, it is not clear that improvements in case-finding for minor depression will necessarily improve patient outcomes.  It is argued that many current treatments for depression are either unproven or have been shown not to be efficacious.  The two interventions that are supported with research evidence are the use of anti-depressant medications and cognitive behaviour therapy.  Unfortunately, there are currently very few primary care practitioners who are knowledgeable about, and trained to provide, cognitive behaviour therapy.  More often than not, primary care practitioners treat depression with anti-depressant medications.  However, evidence shows that many patients, if they are given a diagnosis of depression, will refuse anti-depressant medications.  These patients usually request counseling or some other psychological treatment option; a recent study in the US showed that the majority of patients would prefer the counseling or psychological option.  Research evidence currently supports both options and shows that the efficacy of either intervention, the anti-depressants versus the psychological intervention, is equivalent.  If the efficacy is equivalent, the cost is relatively equivalent and the outcomes are relatively equivalent, individuals should, ideally, have a choice of treatment options.  However, the current primary care system in Canada is not adequately equipped to provide that choice.  In most primary care settings patients will likely be offered anti-depressant treatment, but in some cases no treatment will be offered.  Furthermore, the literature shows that when medication is prescribed it is sometimes not the appropriate medication.  For example, some family doctors will prescribe benzodiazapines or other medications that have been shown to have little or no efficacy.  This raises important questions about the factors that interfere with the delivery of appropriate care based on the best research evidence.  Access to reliable outcome information and effective dissemination strategies may form part of the explanation.  What happens when someone with depression enters the health care system in British Columbia?  What is the quality of care?  What interferes with the delivery of quality care, and what are some possible solutions?  Consider the case study of Mary, a 42 year old married female, who works in a bank and is experiencing very low mood.  The low mood has persisted for a number of months.  Mary feels very blue, very sad, she has had very significant sleep disturbance, she finds herself waking at 3:00 or 4:00 in the morning, feeling very upset, crying, not able to get back to  - 23 - Conference Summary  November 10, 2000 sleep.  She has very poor concentration, to the point where she has not been able to perform her duties at work and she has recently had suicidal thoughts.  When she visits her family doctor, there is a bit of a discussion about possible precipitants to her illness, but nothing clear emerges from the discussion.  She has been experiencing some stresses over the recent past.  For example, Mary is concerned about a sister, recently diagnosed with cancer, who lives far away.  Mary also reports a history of marital problems that recur periodically, and her teenage son is having some problems at school.  The practitioner is left with the task of identifying the potential for depression, making an accurate diagnosis and implementing a treatment plan.  Ideally, a primary care physician will begin the patient assessment by obtaining a complete history from the patient and, ideally, will be in a position to collect information from other informants12.  Although a full spectrum of laboratory investigations is not necessary, the primary care practitioner will need to rule out possible medical causes of the symptoms.  However, in the hands of some family physicians, there might be an over-exuberant investigation of other medical causes, that may be worsened by patients who present with numerous somatic complaints—such as gastro-intestinal disturbances, headaches, abdominal pain—complaints that are frequently non-specific.  In these cases some primary care physicians may not feel comfortable until all possible physical causes for the complaints are thoroughly investigated and ruled out.  The scenario described above raises some key questions related to the capacity to deliver quality mental health services in the primary care setting.  In particular, what sorts of activities might support both the process of diagnosing and treating mental health problems in the primary care setting?  Furthermore, what combination of policies is required to successfully establish, maintain and monitor these activities?  There is growing interest in the merit of collaborative treatment. It has been shown that collaborative treatment can assist the primary care physician in both diagnosing and treating individuals with depression.  In the ideal situation, there would be some shared care or shared evaluation.  This means that mental health specialists would be involved in the assessment or treatment plan of the patient.  This is particularly important when there are concerns about the potential for suicide.  Yet to date, provincial population based data show that in the Vancouver/Richmond health region, 69% of patients presenting with mental health concerns continue to be seen by their family doctor alone.  In other health regions across the province up to 90% of patients with mental health related problems are only seen by their family practitioner in the primary care setting.  Mental health teams exist in the province of BC, but due to resource constraints these teams primarily look after people with severe psychotic illnesses.  One of the major obstacles to collaborative treatment is that psychiatrists are, quite simply, often not readily available.  There is also some confusion about the role of psychiatrists.  Are psychiatrists meant to provide primary, long-term care to people with problems, such as depression, or are psychiatrists meant to serve as consultants who provide information on a one-off visit and provide support to family physicians?  The effective sharing of care, not only among psychiatrists but also with other mental health professionals and primary care practitioners, is                                                  12 It is ideal if information can be obtained from informants since people with depression may have limitations with their ability to describe their situation and symptoms.  - 24 - Conference Summary  November 10, 2000 currently not working well in most locations throughout Canada.  There is a wide interest in evaluating models of mental health service that could utilize resources more effectively by coordinating collaborative assessment and treatment services linked to the needs of the population.  It has been proposed that targeted education of physicians could improve recognition and/or management of depression in the primary care setting, resulting in better outcomes.  A comprehensive study, known as the Hampshire Depression Project, was undertaken in the UK to study this approach and its results were published in the Lancet (2000) by Thompson and colleagues.13  The study involved 59 family practices or general practices.  There was a mixture of solo and group practices in the study, with a total of 170 physicians participating.  The practices were randomized either to the intervention group or to the control group (29 practices in the intervention group and 30 in the control group).  The intervention group received a four-hour education session that focused on providing information on best practices in the recognition and treatment of depression.  The education session materials were based on clinical guidelines that were carefully reviewed for their efficacy and validity, etc.  The educators providing the four-hour session were available to answer follow-up questions for a period of nine months after the initial intervention.  The intervention group received a resource package that would help them with the management of treatment and continued to practice without the advantage of participating in the four-hour education session.  At a later date the control group was also offered the identical educational program.  The intervention group demonstrated 39% sensitivity in the recognition of depressive symptoms.  In other words, 39% of people who were depressed were identified as such by the practitioners.  Sensitivity in the control group was 36%.  The difference between the intervention and control group was not significant.  There was also no significant difference in specificity, or correctly identifying an individual who was not depressed.  Furthermore, there was no significant difference in outcome, with observations taking place at both six weeks and six months.  These results are very disappointing.  The intervention did not result in a substantial difference in the recognition of depression or treatment outcomes.  Yet in the UK, policy decisions had already been made to support wide dissemination of these sorts of educational programs and, in fact, it was predicted that the suicide rate in Britain would drop as a result.  Of course, people are very keen to find solutions and are very excited about any possible solutions for problems like this, but the Hampshire Depression Study provides an important lesson for us in terms of proceeding carefully and critically.  The researchers involved in this study critically examined the results and tried to understand and explain the lack of significant findings.  A number of explanations were offered including possible problems with the guidelines for recognizing and treating depression; problems with the actual delivery of information during the four-hour education session; and possibly, problems with the study design itself.  In comparison to this study, targeted education projects undertaken in the US have demonstrated very positive effects.                                                   13 Thompson C, Kinmonth AL, Stevens L, Peveler RC, Stevens A, Ostler KJ, Pickering RM, Baker NG, Henson A, Preece J, Cooper D, Campbell MJ. Effects of a clinical-practice guideline and practice-based education on detection and outcome of depression in primary care: Hampshire Depression Project randomized controlled trial. Lancet. 355(9199):185-91, 2000 Jan 15.  - 25 - Conference Summary  November 10, 2000 The Mental Health Evaluation and Community Consultation Unit (MHECCU) at UBC attempts to grapple with key questions related to quality of care in the delivery of mental health services.  The optimal approach to the recognition and treatment of depression in primary care settings remains an important question.  In recent decades, patients have assumed a more active role in the health care delivery system.  In the delivery of mental health services, self-care protocols are being developed which can be implemented by patients together with their family practitioner.  MHECCU is interested in evaluating these approaches and in examining the interaction between those diagnosed with depression and the primary care practitioners involved with their care.  Finally, since depression is common, disabling and, to some degree, under-recognized and under-treated, and since treatment occurs, for the most part, in primary care settings, there is a substantial responsibility for mental health specialists and primary care practitioners to work collaboratively to enhance the quality of care for patients with depression.  An imperative exists to improve the recognition and management of depression in the primary care setting with the goal of reducing morbidity and mortality and improving the quality of life for those who suffer with it.   - 26 - Conference Summary  November 10, 2000 Sharing Care for Depression in Primary Care Nick Kates   The successful treatment of depression in primary care depends in part upon the relationships between mental health specialists and primary care practitioners and their interactions.  This paper will provide an overview of these, explore the shared care approach to managing mental health problems in the primary care setting, and describe lessons learned in Canada about the application of this approach.   Elliot Goldner's paper provided convincing evidence that depression is a significant public health concern and cannot be ignored. (see Goldner, this volume)  It also emphasized that the primary care sector carries the major responsibility for the assessment, identification and treatment of depression in Canada and highlighted the challenges involved with this.  Family physicians play a central role in the treatment of depression, and the patients with mental health problems they see are very similar, with respect to the nature and severity of their problems, to the patients seen by mental health specialists.14  Results from the Ontario Mental Health Survey Supplement1, however raise important questions about the adequacy of treatment for depression.  The Supplement reports that 72% of individuals with mental health problems are currently not receiving any mental health care over the course of a year.  The health sector must identify new ways of contacting, identifying and caring for these individuals.  The solution will not be found by simply increasing the supply of mental health specialists working in clinics or private practice settings.  If the health sector is going to be effective in treating depression and mental health problems, the primary care setting must receive serious and careful consideration.  It may be the ideal, and perhaps the only, location for identifying many patients with mental health problems and initiating treatment.  Caring for individuals with depression can be complicated by the existence, or development, of co-morbid conditions, and can create special challenges for the family physician.  First of all, depression increases an individual’s risk of developing significant physical illness.  Controlling for all other risk factors, for example, an individual with depression is three times more likely to experience a myocardial infarction.  Moreover, there is an increased prevalence of psychiatric disorders, often depression, with all chronic illness.  If not treated appropriately, the co-morbid conditions can deleteriously affect both patient outcomes and medical costs.  For example, depressed post-myocardial infarction patients have a three to four-fold risk of death over the six months following myocardial infarction15 when compared to post-myocardial infarction patients who are not depressed.  Secondly, depressed individuals are less likely to follow through with treatment interventions if these require the patient to attend a variety of different programs.  Several studies have shown that depression has a substantial impact on self-management of chronic illness.  Depressed post-                                                  14 Ontario Health Supplement.  Ottawa (ON): Health Canada; 1997. 15 Frasure-Smith et al JAMA; 1993: 270:1819.  - 27 - Conference Summary  November 10, 2000 myocardial infarction patients are more likely to drop out of exercise programs.16  Depressed smokers are 40% less likely to quit smoking over a nine-year period.17  Depressed coronary artery disease patients are less likely to adhere to low-dose aspirin therapy.18  Poor compliance complicates the process of care, interferes with the delivery of adequate treatment and may influence patient outcomes.  Depression is not only a serious public health issue; it also creates serious challenges for the clinician and the health system.  Despite the high prevalence of mental health problems in primary care, current relationships between family practitioners and psychiatrists do not always work well.  A series of focus groups conducted across Ontario confirmed recurrent problems identified in other studies.  These include poor communication between family medicine and psychiatry, difficulty in accessing timely psychiatric consultation, cumbersome and inefficient intake procedures, and a general lack of support and respect for family physicians.  In Ontario, as in most Canadian provinces, the division between family medicine and psychiatrists is mirrored in the provincial planning process.  Mental health and primary care reform are taking place at the same time, but in relative isolation from each other and with very little cross-consideration of planning documents.  Mental health documents invariably overlook the role of primary care, and primary care documents rarely refer to mental health care.  Changing this situation requires a shift in the relationship between the two specialties so that care can be shared.  The concept of shared care is built on the belief that delivery of mental health services is a partnership between mental health specialists and primary care practitioners, where responsibility for care is shared.  Each partner is responsible for delivering services according to their respective abilities and according to the needs of the case.  The smooth operation of shared care can only exist in an environment where there is mutual respect and trust between professionals from different disciplines and the roles are well defined, coordinated and complementary.  The caregivers must be sensitive and responsive to the local context.  Shared care operates on the principle of shared responsibility and underpins the goal of achieving continuity of care.  It also supports the patient’s enduring relationship with a family physician.  Shared care projects can range from attempts to make mental health intake criteria more user friendly for referring family physicians or organizing joint rounds to integrating mental health providers within primary care settings.  An example of the latter is the Hamilton HSO Mental Health and Nutrition Program, which has successfully integrated counselors, psychiatrists (and dietitians) into the offices of 90 family physicians across Hamilton, a city of 460,000 people in Southern Ontario.                                                   16 Blumenthal JA, Williams RS, Wallace AG, Williams RB, Needles TL. Physiological and psychological variables predict compliance to prescribed exercise therapy in patients recovering from myocardial infarction. Psychosomatic Medicine 1982;44(6):519-27. 17 Anda RF, Williamson DF, Escobedo LG, Mast EE. Giovino GA. Remington PL. Depression and the dynamics of smoking: a national perspective. JAMA 1990;264(12):1541-5. 18 Carney RM, Freedland KE, Eisen SA, Rich MW, Jaffe AS. Major depression and medication adherence in elderly patients with coronary artery disease. Health Psychology 1995;14(1):88-90.  - 28 - Conference Summary  November 10, 2000 As more shared care projects are established, specific questions arise as to whether these new approaches can overcome some of the difficulties in mental health service delivery outlined above.  First, can shared care increase the detection of individuals with mental health problems at an early stage?  Second, can it enhance linkages, making mental health services more accessible; furthermore, can it facilitate linkages between the mental health and primary care sectors at the time of discharge?  Third, can shared care increase treatment rates?  Fourth, does it improve both physician education and understanding of issues related to the detection and treatment of depression?  Moreover, how can patient education be improved, and will this result in patients participating more actively in their treatment?  Finally, are there adjunctive treatments that influence treatment rates?  Detection rates  Goldner noted that the detection rate for depression might be as low as 40-50%. (see Goldner, this volume)  There are many reasons why mental health problems are not detected, including that people do not raise depression when they see their family physician.  A study by Bland found that 50% of people who visited their family physician because they were depressed did not mention depression unless they were asked.7  The data for substance abuse (i.e. alcoholism) shows that 90% of people do not raise the subject with their family physician until asked.  There are many reasons why individuals do not initiate discussions of mental health problems in primary care.  These suggest some important, yet straightforward, interventions that may provide solutions.  One is the use of screening instruments in primary care.  The second is to simply ask the patient specific questions.  If the family physician asks a direct question, the chances are that discussion of related issues will occur.  Studies have shown that a single question  "Are you feeling blue or depressed?" will pick up 90% of people who might be identified with a more comprehensive depression screening instrument.  Individuals want to talk about their problems, but they are often waiting for the physician to raise the topic.  Providing information for patients is another idea.  A simple example might involve putting a sign in the waiting room that reads "Dr. Smith is interested in counseling.  If you have a problem, please feel free to raise it during a session."  There is also evidence that the more support available to a family physician, the more likely it is that the physician will raise a problem.  If you know what to do with a problem, you are more likely to try and detect it.  Focus group work with physicians has confirmed that if physicians know they have access to readily available support, they are more likely to be willing to address a potentially complex issue.  Primary care physicians routinely monitor patients at risk for serious medical problems, such as individuals with a history of suspicious pap smears or previous carcinomas.  It might also be possible for a primary care physician to identify people who might be at risk for depression, for example, people with a chronic medical illness; the isolated elderly or the recently bereaved.  Routine follow-up every six months with these individuals, either in person or by phone, might help to identify those who are developing signs of depression.                                                   19 Bland R., Newman S., Orn H.  Help-Seeking for Psychiatric Disorders.  Can J Psych 1997;42:935-942  - 29 - Conference Summary  November 10, 2000 The other important question about early identification is does it make a difference to patient outcomes?  Unfortunately, with the exception of schizophrenia, where early intervention has been demonstrated to make a difference in terms of both severity and frequency of episodes, there is currently very little evidence about this.  Data from the Hamilton Program20 suggest that a shared care approach to the delivery of mental health services can lead to the earlier detection of problems and increase access for groups who traditionally underutilize mental health services.  Impact on family physician behaviour  The evaluation of the Hamilton program has not fully examined whether the presence of counselors in the primary care setting has had a substantial impact on family physician behaviour, especially with respect to changing their practice patterns.  In a Cochrane review, however, Bower and Sibbald found little evidence of changes in physician behaviour when counselors worked in their offices21.  This raises the question of how best to influence the behaviour of individuals who see themselves as independent practitioners.  It also highlights the importance of the counselors being fully integrated into the primary care practice from the outset.  Improved linkages and accessibility  The Hamilton program has been successful in increasing accessibility to mental health care with a 10-fold increase in the number of individuals being referred for treatment compared with the period before the program was established.  This rate has remained constant over the six years the program has been in operation.  The Program has also led to a significant reduction in utilization of outpatient facilities with a non-significant reduction in the use of inpatient units.  One area where family physicians continue to identify problems, however, is in discharge planning by mental health services.  This deserves greater attention as we attempt to better link mental health services with the primary care setting.  One solution, for example might involve arranging a patient’s final visit with the mental health team at the primary care physician's office, so that everyone involved knows what the plan is.  Treatment rates  The Hamilton Program’s evaluation has identified many benefits from engaging counselors in primary care, though the precise mechanisms of action that leads to successful outcomes remain unclear.  This has, in turn raised further questions. Can we begin to identify some of the factors that lead to successful outcomes or increased referral rates?  Which mental health problems are best managed in primary care and which are best managed in traditional mental health services?                                                   20 The Hamilton HSO Mental Health & Nutrition Program was established in 1994 and involves 90 family physicians in 36 practices at 51 sites, serving 40% of the population of Hamilton. Counselors are permanently attached to each practice and visit each practice periodically (between one to four times each month, depending on the size of the practice). In 1999 there were 4,037 referrals to the counselors in the practice. 21 Bower P, Sibbald B. Systematic review of the effect of on-site mental health professionals on the clinical behaviour of general practitioners . BMJ 2000;320(7235):614-17.  - 30 - Conference Summary  November 10, 2000 Clearly, not all mental health care can be delivered in primary care.  We need to know which problems are best treated where and to make sure that people reach those services easily.  Treatment outcomes  A study by Wells22 examined a program that aimed to improve the treatment of depressed patients being cared for in the primary care setting.  The program involved 46 primary care clinics in six HMOs.  The quality improvement program included a commitment on the part of practices to participate in an education program for family physicians and other individuals working in primary care, using techniques designed to improve early identification and rapid referral to mental health clinics.  The results suggest that the program improved the quality of care, alleviated symptoms and decreased prescription use.  Another study in Puget Sound involved a combination of educational intervention for family physicians and Cognitive Behaviour Therapy delivered by psychologists.23  The results showed there were some changes in the use of mental health services but no change in the use of medical services.  Individuals with major depression in this study showed more improvement than individuals with minor depression.  The educational intervention produced no long-term changes among physicians; when family physicians were observed six months after the program ended, their skills had returned to baseline levels.  In contrast to these findings, the Hamilton Program has found significant improvements in the majority of people treated. Using the Centre for Epidemiological Studies - Depression (CESD) and the General Health Questionnaire (GHQ), 75% have shown significant changes in their scores, while over 80% of individuals who exceeded criteria for being a case at referral were no longer cases at discharge.  A study published in the Archives of Family Medicine24 showed that adjunctive treatment in primary care in the form of telephone contact had a positive effect on patient outcomes.  The telephone care involved ten six-minute phone calls over a four-month period from the practice nurse on the team for patients who had been started on anti-depressants by their family physician.  The focus was on maintaining compliance with the prescribed medication and increased social activation.  This was not a labour-intensive or time-consuming intervention, yet there was a 50% improvement in the group receiving the adjunctive treatment approach.                                                     22 Wells KB, Sherbourne C, Schoenbaum M, Duan N, Meredith L, Unutzer J et al. Impact of disseminating quality improvement programs for depression in managed primary care: a randomized controlled trial.  JAMA 2000;283(2):212-20. 23 Katon W, Robinson P, Von Korff M, Lin E, Bush T, Ludman E, Simon G, Walker E. A multifaceted intervention to improve treatment of depression in primary care. Archives of General Psychiatry. 53(10):924-32, 1996 Oct. 24 Hunkeler EM. Meresman JF. Hargreaves WA. Fireman B. Berman WH. Kirsch AJ et al.  Efficacy of nurse telehealth care and peer support in augmenting treatment of depression in primary care. Archives of Family Medicine 2000;9(8):700-8.  - 31 - Conference Summary  November 10, 2000  Physician education  Several different approaches have been used for physician education programs.  These include: half-day sessions in the family physician’s office; models of mentoring; training that involves a close link between the family physician and a psychiatrist who will provide advice when necessary; training programs targeting specific problems (these are often supported by the pharmaceutical industry); and the development of guidelines and telephone information lines.  In the Hamilton program, education is case-based.  Each case offers a teaching opportunity.  The education is brief, but focused.  It is important to arrange the education opportunity so that it suits the time constraints of the physician’s day.  Most importantly, the education takes place as part of a continuing relationship.  In summary, to strengthen links between mental health and primary care services it is not sufficient to just locate additional services in the primary care setting (e.g. counselors, education opportunities, guidelines, etc.).  The process needs to be active and interactive; the specialized staff with mental health service skills must be well integrated into the primary care setting and fit within the time constraints of the physician’s daily schedule.  There is little evidence to support the use of educational interventions and guidelines in the absence of an integrated approach that involves an ongoing relationship with mental health specialists.   - 32 - Conference Summary  November 10, 2000 Using Report Cards to Improve the Quality of Acute Myocardial Infarction Care Jack Tu   Acute myocardial infarction, or AMI, is the clinical term for a heart attack.  The purpose of this paper is to share some of our experiences with the development and production of AMI report cards for Ontario.  The first part will cover the history of health care report cards and some of the lessons we learned from the development of other health care reports cards around the world.  The second part will review some of the findings from the AMI component of a health care report card on cardiac care in the province of Ontario, the Institute for Clinical Evaluative Sciences cardiac atlas.  The final part will provide some of the results of an evaluation of the impact of this report card on treatment of AMI patients in Ontario.  Florence Nightingale became famous as a nurse because of her service during the Crimean War.  She introduced many changes to hospitals during that time, including changes in the configuration, location and organization of hospitals in England and elsewhere to reduce the number of deaths due to unsanitary conditions.  She became quite well known for her advocacy to improve the state of hospitals.  What Florence is less well known for is her role in advocating the development and public distribution of health care report cards.  Table 1 is from one of the world’s first health care report cards, developed by Florence and her collaborators in England in 1861.  This was a comparison of mortality rates among different types of hospitals in England.  The classification used was five major categories: hospitals in inner London, hospitals in other large towns, county and other important provincial hospitals, other hospitals, and naval and military hospitals.  There was a huge range in mortality rates, from 91% in hospitals in inner London to 16% in naval and military hospitals.  The release of this report attracted a considerable amount of attention but turned out to be not entirely accurate.  For example, this calculation was based on the number of deaths over all of 1861 divided by the number of patients in hospital on just one particular day in 1861.  This grossly overestimated the mortality, which in fact was closer to 10%.  She also did not make any adjustment for case mix differences across the hospitals.  Patients admitted to hospitals in inner London would likely be very different from patients admitted to naval and military hospitals, so this oversight could have a large effect on the comparison.  Though Florence and colleagues had good intentions, their methods were flawed and point to some lessons for current work on report cards.  First, we have to be very careful with our statistics and ensure that we present accurate, not misleading data.  Second, we need to make sure that the data presented are truly comparable -- e.g. are case-mix adjusted.   - 33 - Conference Summary  November 10, 2000 Table 1:  Hospital Report Card  “Notes on Hospitals”:  1861   Number of Special INMATES on  8 April, 1861 Number of DEATHS registered in 1861 Mortality Rate:  Percentage of Inmates In 106 Principal Hospitals of England 12709 7227 57 24 London Hospitals 4214 3828 91 12 Hospitals in Large Towns 1870 1555 83 25 County and Important Provincial Hospitals 2248 886 39 30 Other Hospitals 1136 457 40 13 Naval & Military Hospitals 3000 470 16  More recently, report cards comparing hospital-specific mortality rates have been developed and publicly released in a number of jurisdictions around the world.  This work began in 1983, when the California State Department of Health released a report on heart attacks, comparing the mortality rates among various hospitals treating AMI patients.  In 1994, Scotland released a report comparing mortality rates among various hospitals in that country.  In 1996, Pennsylvania got into the act releasing physician-specific as well as hospital-specific mortality rates.  Sweden has also been active in this area with a report card comparing outcomes among their hospitals.  Ontario was the first jurisdiction in Canada to release a similar report card, learning from and building on previous experiences.  The motivation for this work came from the Heart & Stroke Foundation of Ontario, which, in 1996, commissioned a study of regional variations in cardiovascular death rates in the province.  There was a two to three-fold variation across counties; Northern Ontario, for example, showed mortality rates from cardiovascular problems double that of downtown Toronto.  This report -- “The Hotspots Report” -- attracted quite a bit of attention, but the underlying reasons and the factors contributing to this phenomenon were not well understood.  The Heart & Stoke Foundation then approached our group at the Institute for Clinical Evaluative Sciences (ICES) to conduct a study of factors that could potentially explain and contribute to these variations.  ICES is a non-profit research corporation funded by a large block grant from the Ontario Ministry of Health.  Its mandate is to conduct applied research that contributes to improving the effectiveness, quality and efficiency of health care delivery in the province of Ontario.  The publication that came out of this work was entitled “Cardiovascular Health and Services in Ontario - an ICES Atlas”, more often referred to as “The ICES Cardiac Atlas”.  This report, released in February 1999, contains information on the burden of cardiovascular disease in Ontario, regional variations in cardiovascular mortality rates, the prevalence of cardiac risk factors in different parts of the province, and treatments and outcomes of AMI patients in the province.  It also includes some information on heart failure patients in the province as well as the waiting times and outcomes of patients undergoing various cardiac procedures.  Many of these results come from collaborations with the Cardiac Care Network of Ontario.  The information was picked up by the media as well as hospitals, physicians and other health care  - 34 - Conference Summary  November 10, 2000 providers in Ontario.  In 1996, cardiovascular disease was the leading cause of mortality in Ontario, claiming approximately 30% of all residents who died in that year.  This was followed closely by cancer, at 28%, then respiratory diseases and stroke.  Trends in adjusted cardiovascular death rates between the early 1980s and the mid-1990s suggest significant progress over the past 15 years in our ability to treat heart patients in the province, with approximately 50% reduction in the number of deaths secondary to cardiovascular problems. (Figure 1)  An initial impression of this trend might be that we are winning the war against cardiovascular disease and that it no longer needs much attention.  Unfortunately, if instead of rates you consider the absolute number of people projected to die from cardiovascular problems over the next 20 years a different picture emerges (Figure 2).  The difference, of course, is the increasing number of elderly residents in Ontario that will cause an estimated doubling of the absolute number of people dying from cardiovascular problems, even though the death rate on a per capita basis is declining.  The burden of disease implies there is still more work that needs to be done to identify new treatments and improve the quality of cardiac care in Ontario.  Figure 1:     - 35 - Conference Summary  November 10, 2000 Figure 2:    The motivation for this report card was to identify potential opportunities and to stimulate physicians in hospitals to improve the overall quality of cardiac care in the province.  There are several reasons why we felt this was necessary.  First, a number of studies have shown that the outcomes of heart attack patients in the real world are much worse than those of patients enrolled in large clinical trials.  Second, there is wide variation in the treatment of AMI patients in Ontario.  One might assume that with universal health care all residents of the province should have pretty equitable access to the latest cardiac procedures as well as the latest treatments for heart disease.  This is far from reality.  Finally, there has been a general failure in the impact of practice guidelines on the delivery of cardiac care in Ontario.  We need to look at other interventions such as reports cards as a potential alternative method for reducing some of these practice variations.  The Ontario Myocardial Infarction Database Project, funded by a grant from the Medical Research Council, created a large database that links together information from several of the large health care administrative databases in the province of Ontario.  We began by identifying acute myocardial infarction patients from hospital discharges using ICD9 code 410.  These patients were linked to the Ontario Drug Benefit database, which provides information on the utilization of various cardiac medications such as beta blockers or the statins, for residents over age 65.  We also linked this database to the physicians’ services database so we know whether  - 36 - Conference Summary  November 10, 2000 patients have gotten invasive cardiac procedures such as angiography, percutaneous transluminal coronary angioplasty and coronary artery bypass graft surgery.  Finally, we linked to the Ontario Registered Persons database, for information on the vital status of all Ontario residents so we can track in-hospital, 30 day and one-year mortality in our AMI patients.  The result is a powerful resource for studying the quality of AMI care in Ontario.  One of the more critical issues in developing a report card is the accuracy of the source data.  Are we sure the patients identified as having an AMI truly had an AMI?  Or have they been mis-coded and actually have something else, such as unstable angina?  To address this issue, we identified all patients who met the inclusion and exclusion criteria for our study and then sent all hospitals in the province the list.  This study covered the period April 1, 1994 to March 31, 1997.  The overall sample size was about 60,000 patients.  This gave the hospitals an opportunity to internally validate whether they agreed that these patients truly were AMI patients.  We received responses from approximately 70% of the hospitals in the province.  All but one reported that their accuracy rates for AMI coding were 94% or higher, and thus we felt fairly confident that we were really dealing with AMI patients.  Patients the hospitals did not validate were excluded from further analysis.  The other important issue for these types of analyses is case-mix adjustment.  We developed a logistic regression model to adjust for case-mix differences across the various hospitals in the province.  This model includes factors such as the age of a patient, which is a very important prognostic factor in AMI patients, the gender of the patient, and the presence or absence of a variety of co-morbid conditions, including shock, diabetes, congestive heart failure, cancer, etc.  These co-morbidities were identified using the secondary diagnosis fields in the hospital separations data.  Using this model and its regression coefficients, we then went on to estimate the prognostic importance for AMI of each of these risk factors.  We compared outcomes both at 30 days and one year after the initial infarct.  One of the important issues with case-mix adjustment models is how well they predict outcomes -- what proportion of variation can they explain?  Table 2 shows the area under the receiver operator characteristic curve, a statistical construct that measures how accurately a model predicts an outcome (in this case, mortality).  Overall, we found 0.78 for 30-day mortality and 0.79 for one-year mortality, pretty good numbers.  These models performed better than the Charlson Index, and performed well in other jurisdictions including California and Manitoba.  Figure 3 shows the inter-hospital variation in risk adjusted mortality rates following AMI in Ontario.  Each circle shows the results for one particular hospital with 30-day risk adjusted mortality on the left and one-year risk adjusted mortality on the right.  The average mortality rate was approximately 15% at 30 days and approximately 23% at one year with a range from 0% to 32%.  The identities of each of the hospitals were contained in the atlas, which led to some consternation among the hospitals at the higher end.       - 37 - Conference Summary  November 10, 2000 Table 2:  30-day mortality ROC area 1-year mortality  ROC area  N Ontario rule Charlson index Ontario rule Charlson index    Derivation AMI cohort      Ontario  52,616 0.78 0.74 0.79 0.77        Validation AMI cohorts      Manitoba  4836 0.77 0.72 0.78 0.75 California  112,234 0.77 0.71 0.78 0.74  * ROC denotes receiver operating characteristic curve     Figure 3:    There are a number of factors that account, at least in part, for the extent of variation across hospitals.  One appears to be the type of hospital to which a patient is admitted.  In general, we found that patients admitted to teaching hospitals or large hospitals had lower mortality rates compared to patients admitted to medium sized or smaller hospitals.  Another factor that appears to contribute to variations in outcome is the specialty of the attending physician.  Patients treated by cardiologists have lower mortality rates in general than those treated by internists, who in turn have lower mortality rates than patients treated by family doctors or other specialties.  We also  - 38 - Conference Summary  November 10, 2000 found that patients admitted in poorer parts of the province have higher mortality rates than those admitted in wealthier parts of the province.  In addition to looking at mortality rates among hospitals in the province, we have also looked at here is a striking amount of variation for all of these medications.  Beta blockers, shown since s with mortality, one of the factors that appears to explain some of the differences in utilization e have also looked at the utilization of invasive cardiac procedures among heart attack e conducted a survey of physicians in hospitals throughout the province to determine what the the use of various secondary preventive medications.  In these studies, we restricted our analyses to residents over the age of 65 because of the availability of full prescription drug use.  We unfortunately do not have information on the in-hospital use of medications such as thrombolytic medications and aspirin.  In the atlas study we focused on four medications, three of which have been shown in multiple randomized trials to improve survival after a heart attack.  These include beta blockers, angiotensin converting enzyme (ACE) inhibitors, and statins, which should be used following AMI, and calcium channel blockers, which should not.  Tthe 1970s to improve survival after a heart attack, had a median utilization rate of about 53%, with a range from 24% to 80%.  This amount of variation is unlikely to be explained by factors such as the prevalence of contraindications to these medications, but rather must be the reflection of other factors at these hospitals.  They provide an important opportunity for quality improvement initiatives, particularly for hospitals at the lower end of the spectrum.  Analyses of ACE inhibitors and statin medications showed similar patterns; there is a tremendous amount of variation and lack of standardization in the treatment of heart attack patients.  The good news is that these trends are improving over time.  Aof these therapies across hospitals is the type of hospital to which a patient is admitted.  In general, patients admitted to teaching hospitals and large hospitals are more likely to receive beta blockers and statins after their heart attack.  This may help explain the lower mortality at the teaching and large hospitals, and represents an opportunity for quality improvement for smaller hospitals.  Specialty of the attending physician also influences use of medication.  Patients treated by cardiologists or internists are more likely to receive these medications than are those treated by family doctors or physicians in other specialties.  This too represents an opportunity for quality improvement, not only among family doctors and other specialists, but also among the cardiologists and internists where the ideal rates could be even higher than they currently are.  Wpatients.  We have compared the one-year post MI utilization rates for cardiac catheterization, a diagnostic procedure, PTCA spell out or angioplasty, used to clean out the coronary arteries, and bypass surgery, used to bypass blocked coronary arteries.  Once again the message is there is a lot of variation in the likelihood of receiving these interventions and there is no consistency in the treatment of patients among the hospitals.  Perhaps the most important factor in these differences is the type of hospital to which a patient is admitted.  Patients admitted to hospitals with the capacity to do these procedures were far more likely to receive them than were patients admitted to hospitals without that capacity.  This source of inequity has been one of the more contentious issues that has come out of the atlas and has prompted attempts to reduce these inequities in the future.  W - 39 - Conference Summary  November 10, 2000 impact of all this work has been.  We were curious if it was just a one-week media splash or whether it actually led to sustained quality improvement initiatives.  In February 2000, we mailed a survey to all cardiac specialists and all hospitals in the province.  The survey asked for opinions of the cardiac atlas, and what changes, if any, they had made at their hospitals in response to the release of this information.  Approximately 120 complete responses have been sent in, out of about 400 mailed, a disappointing response rate.  There is also the potential for double response because the survey was sent to both the hospital and the cardiologists who work at those hospitals.  So it is difficult to calculate the true response rates to the survey.  Regardless, most of the respondents felt the information in the atlas was at least moderately he survey also indicated that 57% of the hospitals started one or more quality improvement he most significant limitation of the cardiac atlas reported through the survey was that in spite ne particularly interesting finding from the survey was the lack of impact the atlas appeared to he vast majority of physicians (70%) supported the public release of hospital-specific AMI verall, the survey suggests that the development and publication of an AMI report card in useful and quite a significant proportion thought it was very or extremely useful.  Particularly interesting is the rating of the utility of the different types of information.  The respondents felt that, overall, mortality information was the least useful, while information on various process of care measures, such as the utilization of beta blockers or statin medications, or the waiting times for cardiac procedures at the hospitals, was most useful.  This is an important lesson for future report cards; it appears there is something more 'actionable' on their part in the process measures compared to the outcome measures.  Tinitiatives in response to the atlas.  The most commonly reported initiatives were conducting a review of thrombolytic use and 'door to needle' times among heart attack patients, followed by reviewing the medical records of AMI patients, and introducing new critical care pathways or standing orders.  A number of hospitals with existing critical care pathways or standing orders reported revising them in response to the release of the atlas.  Tof our attempts to validate the hospital separations data there could be coding problems with them.  Another criticism was the timeliness of the data; the report came out in 1999 and focused on data from 1994 to 1996.  Third, respondents thought the risk-adjustment methods were inadequate despite our attempts to develop the best model we could, using administrative databases.  Ohave on patients.  When the atlas was released it received extensive media cover in Ontario on TV shows, in newspapers and on the radio.  But the overwhelming majority of physicians (81%) indicated that not a single patient had discussed this information with them in the year after its release.  Tmortality such as was featured in the ICES cardiac atlas.  There was, however, a significant minority who disagreed because they believed the data were potentially misleading or inaccurate, that the public would not understand the data, and that there was potential harm to a hospital’s reputation.  OOntario has had a beneficial effect, and certainly has led to a number of quality improvement  - 40 - Conference Summary  November 10, 2000 initiatives in hospitals and organizations throughout the province.  It also appears to be important to develop report cards that focus not only on outcomes but also on multiple aspects of care and, in particular, process of care measures.  Finally, future work should try to improve both the quality and amount of information that goes into these reports as well as the speed with which they are developed.  With the methodology in place, we have the capacity to produce subsequent atlases in a timelier manner.  We also have the ability to release the information over the Internet which may improve its accessibility.    - 41 - Conference Summary  November 10, 2000 Assessment and Accountability for Surgical Indications and Outcomes Charles Wright   Alan Maynard’s paper (see Maynard in this volume) is a perfect accompaniment to this, which will cover some of the basics on the issues of appropriateness and outcomes.  This is a quality issue, notwithstanding that quality of care can be defined in many ways.  The definition used here is “the degree to which health services for individuals and populations increase the likelihood of desired health outcomes and are consistent with current professional knowledge”.25  The ‘wishbone’ diagram shown in Figure 1 shows the dimensions of quality.  We have paid a fair amount of attention to safety, efficiency and competence.  Access issues are continuously on the front pages of newspapers.  We think that we deal with effectiveness, but if you actually examine the outcomes of most effectiveness studies, you find that they are surrogate or substituted outcomes.  Appropriateness has received very little attention, and acceptability is just beginning to emerge as a topic of interest.  Most hospitals in Canada are now doing some kind of a systematic patient satisfaction survey, for example, which gets to the issue of acceptability, but the relationship between patient satisfaction and actual health outcomes is debatable.    Appropriateness gets to the question about a particular patient’s condition justifying a specific intervention at a specific time.  This is perhaps a slightly different angle than that put forward by David Hadorn (see Hadorn in this volume).  Appropriateness comes up in many aspects of health care.  We are not always sure, for example, about the need for a particular hospital stay.  We know that patients can often be looked after better in different circumstances.  Questions about the 'right' length of stay, the use of diagnostic tests, and tremendous variations are being raised continuously; it is a constantly changing field.                                                  25 Journal of Quality Improvement, July 1993, quoting from:  Medicare:  A Strategy for Quality Volume II:  Sources and Methods, KN    - 42 - Conference Summary  November 10, 2000  When I started as a surgeon 30 years ago, every patient going near the operating room had a chest x-ray, electrocardiogram, SMA 16, hemoglobin and about 20 other blood counts.  Over time, much of this has been eliminated, but we have to wonder how many inappropriate diagnostic tests are still being done, not to mention the enormous amount of repetition that goes on because of our arcane communication systems.  Any intervention you look at -- including procedures and surgery -- shows huge variations not explicable by differences in medical need.  The Institute for Clinical Evaluative Sciences was, I believe, the first organization in Canada to publish a systematic atlas of health care services, and it created a lot of press about the variations it found (see Tu in this volume).  Other surveys and research in Canada and elsewhere have found the same thing, no matter what type of service is examined.  Results of this work have found their way into the Globe and Mail and even USA Today where a front-page headline was followed by a two-page story.  We have the same data in British Columbia but for whatever reason they have not been used the same way.  The Ministry of Health has produced a publicly available database called PURRFECT - Population Utilization and Referral Rates for Easy Comparative Tables.  This database provides information on everything that is done to everybody in a hospital everywhere in BC.  In 1988-1999, rates of total hip replacements by local health area show that Campbell River and Prince George have very high rates (Figure 2).  The latter is particularly interesting because it is one of the hot button areas in the province.  Overall, there is more than a four-fold variation across the local health areas of BC after adjustment for age and sex, when services are attributed to the patient's area of residence rather than location of hospital.  This last adjustment is crucial; without it, for example, the citizens of Vancouver would have the highest rate of neurosurgery in the world because that is where neurosurgery services for the entire province are provided.    - 43 - Conference Summary  November 10, 2000  Consideration of these rates immediately raises the question of what constitutes necessary medical care.  Surgeons must constantly keep in mind the balance between the potential benefits and the harm that is certain to result from surgery.  When the decision is to proceed, it is assumed that the benefits are going to accrue and that the harm will be both temporary and not severe, but this is not always the case.  Two to three percent of people having coronary artery bypass surgery die on the operating table.  One percent of patients develop a pulmonary embolism after major surgery.  But the benefits are huge for certain groups of patients and the question is:  how do we better isolate the groups of patients more likely to benefit and less likely to be harmed?  Following the consideration of necessity is the issue of costs.  The classic example is elective orthopedic surgery for hips and knees, which has little if anything to do with longevity or survival or mortality; but everything to do with quality of life.  These are wonderful operations, but at what point in the spectrum should a patient, particularly in a publicly-funded system, get a total hip replacement?  You wake up one morning with a twinge of pain in the hip; this is clearly not the right time to operate.  Twenty years later you have severe continuous pain, you are completely disabled, you can not climb stairs and you can no longer take care of basic personal necessities like bathing and shopping.  This is clearly too late. The point in between when operations tend to take place depends on the resources dedicated to orthopedic surgery; there is no question about that.  In other words, judgment is required to determine at what point a major intervention at major cost with major risk becomes justified.  This takes us back to the issue of appropriateness.  Another very common surgical intervention is cataract removal.  The variations for this intervention are not quite as large as with other procedures, but some areas still show twice the frequency as others.  It remains an important procedure to consider because it is so common.  Similar to hip and knee surgery, there is a great deal of judgment involved; there is no point at which a switch is flipped and a patient with a cataract needs an operation.  With interventions like this there is a range rather than a specific point of appropriateness.  Operations on the spinal canal are another example, and these show six-fold variation across British Columbia.  This is very difficult to explain on the basis of medical requirement.  To validate these data, we might want to consider a procedure for which no judgment is needed.  Examples might be an obstructed bowel with a cancer or cancer of the breast.  These procedures exhibit what could be called 'the colour of life variation', a range from one to one and a half times the rate from the lowest to highest areas.  It is unreasonable to expect perfect equity.  A fascinating feature of effectiveness is the huge emphasis on surrogate outcomes.  Examples include the number of cases, diagnoses, stages of disease, imaging studies that show decreases in tumour size, laboratory results that show a favourable change in whatever is measured, infection rates, re-operation rates, and patient's nutritional status.  There are literally hundreds of surrogate outcomes, and interest in each will vary by specialty.  For example, orthopedic surgeons are interested in range of motion in joints, joint pain and so on.  These are certainly important in providing a quality process of care.  But does any of that ultimately matter?  The only results of  - 44 - Conference Summary  November 10, 2000 an intervention that ultimately matter to a patient are health related quality of life and longevity (or the converse, mortality).  We should be measuring these outcomes.  The measurement of health related quality of life is a relatively young discipline.  There are certainly ifs, buts and maybes, and all kinds of questions about the validity of measures currently used, but as Alan Maynard points out (see Maynard, this volume), there are now some well-validated standardized instruments.  Why are we not using them?  More recently, they have been used in population health and with specific groups, but in the evaluation of interventions we are way behind.  Longevity and mortality have always been measured, in fact we have paid tremendous attention to these outcomes.  But we need to find better outcome measures because of the variations we find, and the tools to do so are available.  This is gaining in importance because we are currently doing all kinds of interventions that were not even dreamt of 30 years ago, and many of these are exclusively related to quality of life rather than longevity.  Systematic measurement of quality of life outcomes has the potential to help clinicians make better decisions; in other words, they would be able to feed back the results of prior patients into the decision-making process at the front end.  There is no question that we have consumers that are going to be much more active and payers, whether it is a public system or private system, are rightly asking questions like “What are we getting for our investment in health services?”  That is a very difficult question to answer, except with numbers that show interventions and outcomes.  “Trust me, I’m a surgeon” only goes so far these days.  Don’t tell me, show me.  Measure these things and then we will be in a better position to make a case for the allocation of resources.  These are going to be survival strategies in the future because of the emergence of the notion of accountability.  The Regional Evaluation of Surgical Indications and Outcomes project was funded by the Health Transition Fund operated by Health Canada.  It was a large project and operated under the aegis of the Vancouver/Richmond Health Board.  Its goal was to determine the feasibility of measuring the appropriateness and outcomes of surgical procedures on a regional basis and then of using these measures to influence clinical, administrative and policy decision-making.  The first objective was to get information about appropriateness using some kind of tool, recognizing that the 'perfect' tool does not exist.  We chose to study six high-volume elective surgical procedures since this kind of work does not lend itself to emergency situations.  Second, we needed to measure the outcomes using standard instruments.  Ultimately, of course, we wanted to disseminate the results to clinicians and decision makers and help them use them.  There are all kinds of off-the-shelf guidelines for assessing the appropriateness of indications for procedures.  These tend not to be very useful unless they are operationalized and built into the system in such a way that they have to be used.  That introduces a policy decision that we have not yet made.  We chose the InterQual software package, which is the biggest and most widely used.  They produce standard packages for all types of things such as the appropriateness of imaging studies, and the appropriateness of primary care management.  We used the Indications for Surgery and Procedures package.  These packages are clinically rich; the guidelines are  - 45 - Conference Summary  November 10, 2000 designed by specialty surgeons for specialty surgeons.  They are supported with clinical references, but many aspects of the guidelines are controversial and debatable.  Nevertheless, 700 procedures are covered in the series with very elaborate algorithms.  And, of course, they are used widely in the United States by HMOs to either grant or deny care.  These are only guidelines; they are not commandments, and the way they are used is important.  Clinicians will make exceptions to an algorithm or a guideline if, in his or her judgment, it is necessary to do so.  The whole technique is based on indications for surgery, clinical findings, symptoms, imaging, and a trial of clinical (rather than surgical) management.  This last is often where the case fails, or does not match the criteria; there has not been an adequate trial of alternative therapy.  The outcome is binary -- it is yes or no, appropriate or not appropriate, according to the guidelines.  If the nurse reviews the clinical information available and it matches the criteria, the patient is booked for surgery.  If not, the case is referred to a physician.  The physician then performs his or her own review, and if in his judgment it meets the criteria, the procedure is booked.  It is often perceived that the application of these guidelines denies people necessary care, but data from InterQual suggest that only 10% to 15% of cases do not match criteria – and that is before physician review.  The ultimate result in applying these guidelines is a relatively small number of cases that “do not match the criteria” and, therefore, the appropriateness has to be questioned.  Not surprisingly, different groups of surgeons have different rates, but the tool does permit you to investigate why that might occur.  A lot of the procedures we do require physician judgment -- they are subjective: arthroplasties, lens replacement, prostate surgery, hysterectomies, tonsillectomy, on and on and on.  These are the procedures chosen for this study. In each case we applied the InterQual ISP guidelines preoperatively, and the quality of life questionnaires (the SF-36 and a variety of disease-specific instruments) both before and after operation.  The disease-specific instrument is important because the SF36 is a generic questionnaire; the former is likely to be much more interesting to surgeons.  There are many well-known and well-validated scales available for measuring disease-specific issues.  We encountered quite a few difficulties along the way.  It was difficult to generate enthusiasm for the study - everybody is busy, there are a lot of pressures, morale is low, and the last thing anybody wants to see right now is another form.  It was difficult to get at the necessary information in part because we did not have a mandate to do this -- participation was entirely voluntary.  We were dealing with five hospitals in the region with five different systems.  We did end up with superb cooperation from the hospitals that had very little interest in cooperation historically.  The indication guidelines are controversial; ten surgeons in a room will never agree completely on any particular detailed course of therapy.  There was a definite perceived threat to surgeons’ autonomy, a perception of second-guessing the decision-making process in the surgical office.  We had to deal with selection bias because we did not manage to achieve 100% participations for the selected procedures.  The process is resource intensive; this was an expensive study.  Patient response to the mailed survey was about 50 - 60%, probably as good as could be expected.  And there were language problems, although we did have facilities to deal with Punjabi and Mandarin because of the population in the region.  - 46 - Conference Summary  November 10, 2000  In total, 6,500 patients were enrolled in the study, across all five major hospitals and all five procedures.  The availability of the clinical indication form -- the voluntary contribution from the surgeons indicating whether the patient matched the criteria -- varied from a low of about 42% in the case of hip replacement to over 90% for spinal disc surgery.  The latter is undoubtedly because there is a special interdisciplinary clinic through which all patients are treated.  Slightly more than half (56%) of all pre-operative surveys were returned.  The post-operative surveys are still coming in.  A chart review confirmed that this type of study would not be possible from the hospital chart alone.  There is insufficient information in the chart to be able to match any set of criteria for performing a procedure; there was some variance by hospital but less than 50% of charts overall contained the information required.  In other words, policy decisions would be required to operationalize the process.  This was a feasibility study, so it looked in detail at what does and does not work, what would have to happen if this were worth doing on an ongoing bases, the problems encountered and so on.  Only half of patients in some specialties had the information provided, so what about the other half?  There probably was a selection bias.  Accuracy was also a question.  There was one particular group for which it was clear that the clinical indication forms had been duplicated -- the same completed form was sent for all patients.  Problems like this make evaluation difficult, but do speak directly to feasibility.  It is possible to argue that there are many potential explanations for variation, that quality indicators do not need to include health-related quality of life outcomes and therefore there is no need to operationalize this kind of evaluation.  Contrary to this, we believe there is an issue here, that wide variation in rates is a serious concern.  Patient-reported outcomes are ultimately the only outcome that matters.  Data about meeting indications for surgery and outcomes of surgery should be used as indicators of performance, of quality.  And it is not terribly difficult to figure out the implications for questions of access and waiting lists as discussed in other papers in this volume (see Maynard, Hadorn and Sanmartin, all in this volume).  What should we do if we are not able to collect outcomes data on patients?  If it is not systematized, and if physicians do not wish to participate, what should the chairman of a hospital board say if a prominent citizen or anyone else for that matter, comes to him and says “Could you tell me what your outcomes are for these 5,000 cataracts you’re doing every year?”  And he says or she says, “Oh, I’m sorry, I don’t have the slightest idea; nobody does, you’ll have to ask the 36 ophthalmologists who are doing the procedure.”  Is that an acceptable response?  Whose responsibility is this?  It is absolutely, certainly the surgeons’ and they have taken that responsibility for years.  But it has not been systematized, but rather has been taken on individually.  The result is that it is subject to all the varieties in surgeons’ approaches to interventions and intervention decisions.  So it must be the organization’s responsibility as well.  This brings us back to policy decisions.  Getting physicians to buy in to any changes is a major issue that is going to require a lot of attention.  At this point it is not obvious how it will work. The resource implications suggest that implementation on some sort of structured sampling basis  - 47 - Conference Summary  November 10, 2000 would make the most sense.  For example, it might be possible to put everyone on notice that a program evaluation is coming; it might not be this year, but it might be next.  The difficulties of our study suggest that an evaluation should not be voluntary.  Woody Allen once said  “More than at any other time in history, mankind faces a crossroads.  One path leads to despair and utter hopelessness; the other to total extinction.  Let us hope that we have the wisdom to choose correctly.”  We must have the wisdom to choose correctly or the system is going to be in very serious trouble.  We have to take this issue of accountability for quality, appropriateness and outcomes seriously.   - 48 - Conference Summary  November 10, 2000 Closing Rapporteur Rick Roger  A primary underlying purpose of the collection of papers presented at this conference was to identify real quality issues audible through the rhetoric and noise surrounding the Canadian health care system.  The October 19, 2000 headline from the National Post provides a good illustration of the need for considered and objective appraisal against a background of embellished discordance.  The banner headline leaps from the page, “Half of Canadian Radiology Units Obsolete”.  Following in the article itself are the now almost obligatory references to Greece and Turkey, revealing we are told, how Canada has fallen behind other developed nations in our high-ticket diagnostic machine capacity.  The crudest possible ratio is deployed, machines per population, a step down from tests per thousand, a step sideways from accessibility measures for various clinical conditions, and a few steps behind what should be measured in terms of clinical outcome and the need for care.  The headline directs attention to a study that has been misused innumerable times, yet like the proverbial bunny it keeps on going.  Do we need more and better research with better communication of findings concerning real quality issues?  The affirmative answer is more than obvious, but to drive home the media agenda reality, let us return to our headline story where a sidebar talks about the provinces hitting what is labeled as “the Medicare Wall”.  Saskatchewan, where 40% of the provincial non-debt service budget goes to health care, while the provincial road system and other services are starved for funds, is selected as the Post’s case in point.  Unrecorded in any of this coverage is the notable consideration that the people the health care system serves are aging with more care and expenditure requirements while the groups served by the public educational system, as just one example, are starting to shrink as the echo boom peaks.  There is expert, albeit under-reported, evidence concluding that overall government expenditures in total will not increase in excess of population growth or the rate of economic growth (for example, Denton and Spencer, Population Aging and its Economic consequence, McMaster University, 2000).  A careful look at demographic trends suggests that we could indeed reach 45% or more of provincial public expenditures going to health care, but that this might be affordable if we concentrate on immigration and other measures to bolster the work force in younger age groups.  A full range of pertinent demographic information is not offered in mainstream media commentary, nor is our attention directed to the public versus private sector expenditure dynamic.  A dollar spent in either sector has a similar “opportunity cost,” but what will certainly change if the public insurance scheme is abruptly dismantled is the source, by income quintile, of the sustaining funds.  The juxtaposed National Post articles, one identifying system short-comings, crudely measured, without reference to achievement, the other speculating that the provinces cannot afford even the existing system, provide a front-page display of opinion dressed as fact.  We have a market based reform and privatization agenda clothed as objective reporting.  Deliberate deception or simply an incomplete rendering of available information?  Only the Post editors know for sure.    - 49 - Conference Summary  November 10, 2000 The system is indeed broken, or at least strained, as it struggles with an international nursing shortage, but it is not at the point of crisis that some would have us believe.  Strong and influential forces are at play and it is not all that difficult to discern the biased messaging involved.  The call goes out that the system is not sustainable, that the public wants choice, wishing to treat health care more like a commodity to be bought and sold without such annoying obstacles as the Canada Health Act.  There are 10 million baby boomers working their way through the aging process, or so we are told, and the only logical response is to introduce a “market model,” a widely used and remarkably simple code phrase for 2-tier medicine.  But does the public buy this?  Consider the last federal election where the poster boy for market driven medicine and significant (if not wholesale) privatization at the beginning of the campaign was denying, by voting day, that such thoughts had ever crossed his mind.  Disappointing, no doubt, for those attempting to set the agenda in advance.  There is no need to dwell at this point on why these issues are raised in the National Post and elsewhere.  Bob Evans summarized the reasons for us very well in 1997 in two articles about why particular interests coalesce when it comes time to debase the public health system.26  And just today Alan Maynard noted that public policy development is the synthesis of evidence and values (see Maynard, this volume).  In Canada, we need to understand that the bulk of the population values the single payer, publicly financed medical care system, implicitly recognizing universal access as a dimension of quality, despite the rhetoric in the media.  Our agenda as managers in the system should be clear enough:  we should challenge the noise on the line, deal with the rhetoric, and concentrate on quality as an instrument of non-market based reform.  The system we have now will be held together only by considered, patient and consumer focused reform, not through market based restructuring masquerading as the only remedy for deficient care or service.  Health services research, or in today’s context, research efforts directed at the concepts of quality and population health, are symbiotic not antithetical, and it is time to move beyond the notions we may have once had of conceptual dissonance, or so the conference organizers tell me!  Research efforts need to be integrated.  Boundaries defining the careers of academics and researchers are not that useful at the management level, let alone at the care delivery coalface.  As demonstrated by many of the papers in this volume, there is an important connection between public policy and the research agenda.  There are further examples of mixed and missed messages worthy of mention in the context of quality improvement.  Consider, for example, the frequent argument that Canadians are seeking out-of-country care in significant and ever growing numbers.  The latest report on this from the Canadian Institute of Health Information states that the amount of money Canadians spend on health care in the US is one-tenth of one percent of the total expenditure on health, an amount hardly even measurable.  Yet the “we already have a second tier regularly seeking and receiving treatment elsewhere” contention continues unabated in the general media.  Can this be the mark of the true Canadian, so under-valuing our own achievements with an assumption that we would all go elsewhere, just given a chance?                                                   26 See: Journal of Health Politics, Policy and Law, 22(2) and Journal of Public Health Medicine, 19(1).  - 50 - Conference Summary  November 10, 2000 We should think also of under-reported news on important but perhaps admittedly mundane issues such as policy development.  The Ministry of Health in BC has launched laudable attempts at improved goal-based management reporting.  Very good work on balanced scorecards started in 1996 or 1997.  Why does the product of this effort never see the light of day?  Is it a threat to some group or another, or do we simply and consistently bypass the “good enough” in our quest for reporting system perfection, spinning our wheels without organizational traction control?  We could do so much more to get the real issues on the table if we were not afraid of rolling out some of these data sets; examples of this are shown throughout the papers in this volume.  BC has an integrated data repository that could serve as a model for Canada.  It is quite possible to determine, for example, how many people are discharged from hospital who do not get the appropriate follow-up drugs (see, for example, Tu in this volume), yet we do not flag this, nor do we report on how trends vary from region to region, from population to population.  There are dozens and dozens of repository-based indicators available to drive published reports, but we cannot seem to manage the potential information system through to production.  We also have, in this province, a robust array of “small area variation” statistics, but once again we do not do much of consequence with this data stream.  All of this with and by a Ministry of Health committed to quality as a strategic priority; it is truly sobering to contemplate so much unfulfilled potential for evidence-based policy development.  Maynard presents a compelling analysis of trust in economic terms (see Maynard, this volume).  He identifies the supply of and demand for trust in health care, and stresses the need for professionalism and honesty as an alternative or augment to regulation.  Maynard emphasizes the relationship between trust and duty, and the special status accorded those who deliver care for a self-regulated platform.  On the demand side for trust, we were provided with interesting commentary on the placebo effect and what happens when a person or group thinks something is working when it is not.  There was an excellent piece in the October 2000 Atlantic Monthly (Fisher, Better Living Through the Placebo Effect) about the persisting belief in the curative power of vitamin C despite compelling evidence in study after study indicating no benefit, except, importantly, to those convinced it will work.  Beliefs can be very difficult to change, but erroneous thinking is, evidently, not without utility.  One half decade ago, Greg Stoddart and John Lavis published a monograph directed at the typology of costs that should be considered in a health care system planning and resource allocation.27  An interesting read which should be obligatory for all CEOs or CFOs with a pretense of managerial moxie.  The first order of business in any system redesign effort should be to shed potentially harmful services.  Arguably this might include mammography screening under forty years of age, or prostate specific antigen (PSA) screening.  The evidence on PSA put together by the BC Office of Health Technology Assessment shows that this screening tool is not effective in improving health outcomes.  But do listen to this: in the last short while, the membership of the Health Association of British Columbia voted 354 to 6 in favour of a resolution to introduce PSA testing as a benefit for males under the age of fifty-five.  Even to an audience of relatively informed governors we have not communicated the simple knowledge of                                                  27 Can we have too much Health Care, CHEPA, 1994  - 51 - Conference Summary  November 10, 2000 what obviously works and what obviously does not, or worse yet, of what might indeed cause harm in the absence of effective guidelines for their use.  Up next for consideration in the Stoddart and Lavis context are those services that do nothing, or ‘the flat of the curve medicine’, so to speak.  Estimates of the proportion of services represented here go as high as 40%.  It is contended by many that inappropriate treatments and tests are starting to reduce, perhaps we are doing better, so name your number, and it could be as low as only 10 or 20 %.  Nonetheless, it is clear that there are many interventions in the system that could be restricted with the proceeds re-invested for better impact elsewhere.  It is also clear that we are never going to get on with that re-investment agenda as long as we are afraid to challenge some of the underlying issues and power dynamics.  Finally, the armchair economist in me wants to point out that a third group of concern are services costing more than the benefits yielded.  But even before getting to that last group, as Uwe Reinhardt has pointed out28 dealing with the first two represents a matter of rationalizing -- not rationing – the health care system.  Two of the papers presented deal with depression and mental health services more generally (see Goldner and Kates, this volume).  It is satisfying to witness research attention beyond surgical procedures and the other quantifiable services and interventions that appear most often in the literature.  Mental health is a fuzzier area; it is harder to measure both inputs and outputs, and there are important connections between primary and secondary care, with consequential requirements for the system to actually function as a system for effective results.  Chronic illness management is a key matter for research-led understanding, observing the parallel between mental health goals and treatments experienced over individual lifetimes, and the disease staging objectives for diabetes, asthma, and congestive heart failure to name just a few conditions of contemporary importance.  In the Capital Region in BC, my administrative home base, the population aged 65 - 74 is going to be 50% larger 15 years from now than it is now.  That population is now aged 50 - 60, the age when many chronic diseases start to manifest.  We need to discern how to manage chronic diseases and how to integrate primary and secondary care in much more effective ways.  This is another example of a non-market based reform imperative.  Jack Tu provides evidence that we are not doing everything we could to track proper use of health services (see Tu, this volume).  While it is difficult to talk about things that we are doing that we should not, it is equally difficult to engage in discussion about things we are not doing and should.  The provision of appropriate medication following heart attacks is a good example.  While managers in the system do understand the importance of research as an input to quality as we attempt to advance improvement efforts, we are often puzzled and even anxious as we try to maneuver the system forward.  An example might help illustrate the context in which quality is addressed.  In my home region, media reports last year featured an 85-year old woman who fell out of bed with an unfortunate complication that led to premature death.  An extremely regrettable situation and an opportunity for remedy-focused learning, but part of the learning must involve consideration of the treatment path, “root cause” analysis, so to speak.  The                                                  28 See: The Importance of Quality in the Debate on National Health Policy, in Health Care Quality Management for the 21st Century. James B Couch, editor, American College of Physician Executives, 1991.  - 52 - Conference Summary  November 10, 2000 complications involved occurred following an angioplasty.  If we are to avoid future situations of this nature, we must consider that angioplasties on 85 year olds have not always been done.  Between 1994 and 1999 the number of angioplasties we did for people over 80 living on Southern Vancouver Island went up three times; the number of angiograms went up three times; the number of peripheral angiographies went up three times.  The province-wide trend line was much the same.  Talking with people who work in the neuroscience area provides anecdotal evidence showing that in many cases there were unanticipated health consequences to these interventions – strokes being the most frequently referenced example.  These are potentially damaging interventions, we know now that the age of potential benefit is increasing but we are not looking at trends with sufficient rigor and reporting on the prevalence of damage.  Perhaps more encouraging news comes from Dr. Wright’s paper (see Wright, this volume).  It is no surprise that there has been some resistance in the medical community; there is inevitable resistance from any of us as clinicians, managers, or even researchers, to being examined in particular ways.  But the project assessing appropriateness of surgery has been done in a fashion that is going to yield some very worthwhile results.  News of progress in understanding the issues around waiting lists and beginning to address those is also encouraging (see Sanmartin and Hadorn, this volume).  The Western Canada Wait List project is covering interventions amounting to about 10-15% of the cost of inpatient surgical activity, so making progress here is by no means a trivial undertaking.  Also, these are high variation services (in small area terms), so making modest changes could have significant positive impact on re-organizing service delivery.  The fact that clinicians and patients are prepared to talk about point scales is also a positive sign (see Hadorn, this volume).  A follow-on to this one that will be very difficult to implement will be denying service to people below a certain point count.  We have not addressed this very well in Canada, hiding behind the lack of definition in the Canada Health Act.  Finally, on the subject of waiting list management, it is impossible to overstate the importance of appropriate and sustainable balance between various service streams, for example hysterectomies and orthopedics.  There is no hospital or regional administrator in the country who has not witnessed one specialty group grinding against another, struggling to best respond to the public interest is an area of great consternation.  Indeed, consumers should be more involved than they have been.  Alan Maynard’s example of empowering people to demand that practitioners wash their hands before approaching is a great illustration of this (see Maynard, this volume).  Consumer involvement could be expanded in many directions, teaching consumers to question why they are not on drug X, why they are on drug Y, or why their family doctor is not asking about depression, and so on.  As system managers, we really do need to make better use of the knowledge and experience base of all care providers.  We may have found better ways to listen to doctors, but it is just as important to realize we need to reach out to paraprofessional groups and the nursing community who know very well what works and what does not in the situations that matter most.  For example, if we want to understand what yields benefit in the home support continuum, we should ask the nurses and support workers involved.  If we want to know what is working in a rehab program, we should ask the therapists and we should ask the patients.  It sounds trite, but it doesn’t happen all that often.  To our certain detriment, we are not doing enough to encourage  - 53 - Conference Summary  November 10, 2000 this sort of input either as managers or as researchers; it should be incorporated when thinking about where to go from here.  We do have a broken system right now, to that extent the National Post has a point; we do have significant problems and as system managers, we read about the all the time.  Critical appraisal cannot be escaped and consider again the context.  According to Karl Weicke29, the first step to take if you seek to collapse an organization is to thrust people into unfamiliar roles.  Consider that there were 1,100 new nursing jobs last year at Vancouver Hospital, the province’s jewel in the crown teaching hospital.  These were not all new nurses, but they were in new and unfamiliar roles.  The second step is to leave key roles unfilled – which also happened at Vancouver Hospital and elsewhere, including my own region.  The third step is to make some tasks ambiguous.  That fits for thousands of health care sector jobs.  The fourth step is to discredit whatever might have worked before; forget the past and dishonor it if you can.  Finally, create a context where a small mistake can have monstrous consequences.  This seems to describe the current health care system pretty well.  In the absence of the development and implementation of the thinking and the tools that we have heard about today, the public health system as we know it in this country will make one more federal election, but not two.  And many of us in this room will need to ask ourselves if we wish to be remembered as the generation that let Medicare go, not because we did not have the tools to make it better, to fix it, but because we were afraid to apply these tools in the public interest for the good of those we serve.  Our challenge is clear, we have the tools, we need to put what we know to work.                                                  29  Making Sense of the Organization, Blackwell Publishers, 2001  - 54 - Conference Summary  November 10, 2000 Biographies:  Elliot Goldner is the head of the Mental Health Evaluation and Community Consultation Unit (Mheccu) at UBC.  He is a psychiatrist and full-time faculty member in the Department of Psy-chiatry at UBC where he also heads the Division of Community Mental Health.  He is also faculty with the Centre for Health Services and Policy Research.  Dr. Goldner developed Canada's first province-wide program for people with anorexia nervosa and other severe eating disorders.  In 1995 he was invited by the Saskatchewan Ministry of Health to assist in their development of a provincial framework of services to people with eating disorders.  He has published articles on the implementation and teaching of evidence-based methods in psychiatry and mental health services.  Research activities include development of measures to assess treatment outcomes, evaluation of mental health information systems, and assessment of patterns of mental health care to people with severe and persistent mental illness.  David Hadorn obtained his MD from the University of California in 1976 and practiced in hospital emergency departments for 12 years.  In 1990 he completed a two-year fellowship in health services research at RAND.  Since then he has worked as a policy analyst and researcher for government health agencies in the U.S., Canada, and New Zealand.  He was Research Director, Western Canada Waiting List Project, University of Alberta, Edmonton, a federally funded project developing clinical criteria for use in managing waiting lists.  David Hadorn has submitted his PhD thesis in the philosophy of science, which concerns the nature of evidence, natural kinds, and randomness in health outcomes research.  Nick Kates is a psychiatrist and a Professor with the Department of Psychiatry & Behavioural Neurosciences at McMaster University, with a cross appointment in the Department of Family Medicine.  For the last six years he has been Director of the Hamilton-Wentworth Regional Psychiatry Program, and Director of the Hamilton-Wentworth HSO Mental Health Program.  For 10 years, he was Director of a Community Mental Health Clinic in Hamilton and spent five years as Director of the McMaster Psychiatry Residency Training Program.  He was a member of the Graham Committee that developed the Mental Health Plan for Ontario in 1988, of a provincial committee developing mental health legislation and of a provincial Committee developing a plan for primary care reform in Ontario in 1997.  Alan Maynard is Director of the York Health Policy Group in the Department of Health Studies at the University of York.  He is also Chair of the York NHS Trust, which provides care to a local population of about 350,000.  He was Founding Director of the York Graduate Health Economics Program (1977-1983) and Founding Director of the Centre for Health Economics at York (1983-1995).  He is a member of the Department of Health's HTA Commissioning panel and Founding editor of the journal Health Economics.  He is a member of the Board of Governors of the Royal College of Nursing Institute.  He has written extensively in professional journals and for the media and has worked as a consultant for the World Bank, the World Health Organization and the UK's Department for International Development.  Rick Roger is CEO of the Vancouver Island Health Authority in BC.  Prior to joining VIHA, he was the CEO of the Vancouver/Richmond Health Board, and prior to that he served as Vice-President, Corporate Services & Chief Financial Officer, for the Vancouver/Richmond  - 55 - Conference Summary  November 10, 2000 Health Board.  Rick Roger is a former Associate Deputy Minister of the Saskatchewan Department of Health and he also served a term as Director of the Policy Research and Management Services branch of Saskatchewan Health.  Rick Roger has participated in a number of short-term assignments for the World Health Organization reviewing hospital operations and national health programs.  He has worked at the provincial and national levels in the development of funding, reporting and risk management systems.  Mr. Roger received a Master’s in Health Administration from the University of Alberta in 1980 and completed an administrative residency with Alberta Mental Health Services as part of this degree.  Claudia Sanmartin completed a PhD in health services research in the Department of Health Care and Epidemiology at the University of British Columbia.  Her dissertation focused on waiting lists in British Columbia and the factors that affect the variation in waiting times and waiting list size.  Dr. Sanmartin worked with the Western Canada Waiting List Project, and is currently with Statistics Canada.  She obtained an MSc from the Department of Health Administration at the University of Toronto.  Jack Tu is a Senior Scientist at the Institute for Clinical Evaluative Sciences.  In addition to his position at ICES, he is also: Associate Professor with the Department of Medicine at the University of Toronto and Staff Physician in the Division of General Internal Medicine at Sunnybrook & Women's College Health Sciences Centre.  He also holds cross appointments as a Scientist at the Clinical Epidemiology Unit at Sunnybrook and with the Department of Public Health Sciences and Health Administration at the University of Toronto.  Dr. Tu earned a MD from the University of Western Ontario, a Master of Science degree in Clinical Epidemiology from the University of Toronto and a PhD in Health Policy from Harvard University.  Charles Wright is a Clinical Professor with the Department of Health Care & Epidemiology at UBC and is the Director of the Centre for Clinical Research & Evaluation at Vancouver Hospital and Health Sciences Centre.  His career has been largely in clinical and academic surgery as professor of surgery and Chair of the Department of Surgery, University of Saskatchewan and University Hospital, Saskatoon.  Dr. Wright was Chair of the Medical Care Insurance Commission of Saskatchewan in the 1980s, and from 1989 to 1996 was Vice-President, Medical Affairs, at Vancouver Hospital & Health Sciences Centre.  In September 1998, he was appointed Scientific Officer of the newly formed Canadian Health Services Research Foundation and he is also a medical surveyor with the Canadian Council on Health Services Accreditation.  He received his medical degrees from the University of Glasgow and holds a MSc from McGill University.   - 56 - 


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items