Abstract
It has been an honor to be a member of the Congenital Heart Surgeons’ Society, to serve on many of its committees, and especially to serve as its president. I must say that I never expected to be nominated as president, considering the prominent stature and reputation of my predecessors as well as that of my eminent successor, John Brown. However, I suppose my dedication to the society and its goals may have had something to do with my nomination. Nevertheless, when your colleagues bestow upon you such an honor, the appropriate action is to accept graciously, to try to serve with distinction, and to come up with a presidential address.
During my tenure, I have sought to keep its members engaged and to continue to elevate the stature of our society in the eyes of its own members as well as those of other health care professionals. For this presentation, I decided to begin with a little bit of my own history and a little about the history of this society. I will then touch upon the public’s growing demand for medical excellence and hopefully connect that with what we do as congenital heart surgeons and as a society of congenital heart surgeons.
Some Personal History
I became a member of the Congenital Heart Surgeons’ Society in 1995. At that time, I was the only congenital heart surgeon in Louisville and one of only two in the state of Kentucky. In addition to the prestige of being a member, the Congenital Heart Surgeons' Society (CHSS) offered me the opportunity to meet and befriend its other members, but most importantly to learn from them. If you are fortunate enough to have a mentor with whom you work regularly, your pathway to knowledge and skill is greatly facilitated. Having never had a readily available mentor myself, I have tried to learn from my CHSS colleagues as well as from a series of role models through my career.
My decision to become a surgeon was most influenced by my exposure to Dr Francis (Franny) Moore, when I was a medical student rotating on the surgical service at the Peter Bent Brigham Hospital in Boston. A true “surgeon-scientist,” Dr Moore exhibited a level of excellence that influenced many other medical students besides me to become surgeons. I returned to the South for most of my surgical training at Duke University Medical Center under Dr David Sabiston. During that time, I recognized the excitement of cardiac surgery and the special challenges of congenital heart surgery. I should also recognize Dr Newland Oldham who played a significant role in my exposure to congenital heart surgery while I was at Duke.
I truly value the time I spent as a resident at Duke, but more important than learning how to be a surgeon, it was there that I met my wife, Scott, and three of my four sons were born there. As most of you know, the demands of a congenital heart surgeon seriously challenge any balance between work and family. I owe most of what I have accomplished since that time to Scott’s love, tolerance, and understanding. I am proud to say that we have now been together for more than 36 years, a far greater tribute to her than to me. And also thanks to her, I can be proud of four grown sons, who learned to accept the limitations of their father’s availability, but who deserved much more than they received. I am thankful that despite the times I missed with them, we remain close with mutual love, respect, and support. And thanks to them, we are blessed with two daughters-in-law, a daughter-in-law in waiting, and two grandsons.
After completing training at Duke, I joined my close friend, Ranny Chitwood, at East Carolina University where we began a brand new cardiac surgery program. I happily accepted the role of chief of pediatric heart surgery, but soon realized that to be good in that role, I would need more training. Thanks to John Mayer, who I had met at a Society of Thoracic Surgeons (STS) meeting, I was able to land a fellowship at Boston Children’s Hospital in 1987. There I had the privilege of working with John as well as with Richard Jonas, and, of course, with Aldo Castaneda, who at that time was probably the most eminent pediatric heart surgeon in the world. All these men taught me much and served as valuable role models, but when I returned to East Carolina, I was by myself without a mentor to look over my shoulder or to provide ready advice.
When I moved to Louisville, Kentucky in 1989, I continued on my own as a congenital heart surgeon, pretty well trained, but learning a lot of things the hard way. As you know, the practice of pediatric and congenital heart surgery has evolved significantly in the past 25 years. I owe much of my ability to keep up with that progress to my membership in the Congenital Heart Surgeons’ Society, where we have been able to freely exchange the latest techniques and concepts.
Some CHSS History
The history of the CHSS goes back to 1972 when Mr Eoin Aberdeen, who at that time was the chief of cardiac surgery at the Children’s Hospital of Philadelphia, sought to bring together a small group of accomplished cardiac surgeons that had a special interest in pediatric heart surgery. His idea was to pattern this group after the European Cardiac Surgeon’s Club, initially limiting membership to 16. Actually the number of founding members was 13, 10 of whom attended the very first meeting on September 7, 1973, at the Sonesta Beach Hotel on Key Biscayne, Florida. You will recognize most of these men (Figure 1). Essentially all of them have played a significant role in the progress of pediatric and congenital heart surgery.
Founding Members of the Congenital Heart Surgeons' Society. Top row, left to right: Eoin Aberdeen, Phil Ashmore, Doug Behrendt, Aldo Casteneda; second row, George Daicoff, Anthony Dobell, Henry Edmunds, John Kirklin; third row, James Malm, Dwight McGoon, Robert Replogle; bottom row, Albert Starr, George Trusler.
The field of congenital heart surgery was very different then from today. Most patients coming to surgery were at least a year of age. That was before the ready availability of prostaglandin and two-dimensional echocardiography.
Although the first meeting of the CHSS was at a resort, the group decided to forego the pleasant surroundings and chose the expediency and easy accessibility of an airport hotel at Chicago O’Hare. Needless to say, the number of members of the CHSS has grown over the years.
Some of us understandably miss the inherent intimacy of a smaller group, where discussions of untoward outcomes and failures are more comfortably shared and the conduct of the meeting can be less formal. When I joined this society, there were between 50 and 60 members. When we chose to incorporate in 2003, the limit was set at 100. By 2010, that number was increased to 120. Later today, the membership will be asked to increase the limit to 150; a number that may never require another increase.
Is that too many? Although that is more than can sit around a table, the number is still small enough to permit honest and unencumbered discussions in venues such as this one today.
How about our impact relative to other heart surgery societies? Fortunately, most of us are also members of the STS and the American Association for Thoracic Surgery, and to their credit, their leadership, which typically includes some of us, holds high regard for those of us dedicated to pediatric and congenital heart surgery. However, the CHSS is unique in at least two ways: it is the only organization with its own research center and its sole purpose is to advance the craft and science of congenital heart surgery. As our own identifiable organization, we have more influence with health care partners than we do as a small fraction of these larger thoracic surgery organizations. Thus, I support the increase in our numbers, as long as we continue to require dedication to and distinction in our field to achieve membership.
Demanding Medical Excellence
I am sure that all of our members as well as other health care professionals have become aware of the increasing scrutiny that the field of medicine is experiencing. There was a time not so long ago when the average patient assumed that his doctor always made the right diagnosis and always provided the right treatment. That time is gone. And appropriately so!
The rapidly increasing cost of health care combined with the information age of the Internet has brought the question of health care quality to the forefront. Doctors are not infallible. The Institute of Medicine clearly pointed out that in its often cited report, “To Err is Human.” 1 And, recognizing that the medical community must start to focus on quality, that same committee published a follow-up report, “Crossing the Quality Chasm.” 2 Probably, the best work describing the issue of quality in health care, at least in the United States, was written by a journalist named Michael Millenson. Its title, “Demanding Medical Excellence,” 3 should be a principle for all medical societies. I like to think it is a primary principle of the Congenital Heart Surgeons’ Society.
At one time, the public thought that the doctor was a scientist whose knowledge and decisions were based on previously tested and scientifically proven treatments. When the public started to notice that doctors did not always agree, the reverence for the all-knowing physician started to wane. Why did 70% of children in one Vermont town require tonsillectomy, whereas only 20% required this operation in another Vermont town? 4
Decision Making
So how do doctors make their decisions?
We like to think they are scientifically based, but aren’t they mostly based on what we were taught and what we have subsequently experienced? This shortcoming in the science behind our decision making has placed an emphasis in recent years on what is referred to as Evidence-Based Medicine. The types of studies that back our decisions are ranked in terms of strength of evidence. The strongest evidence is based on multiple randomized prospective trials, and the weakest evidence is based on expert opinion. Sadly, most of what we do on a daily basis is not even based on expert opinion but is more typically derived from personal experience, anecdotes, instincts, and prior training.
Recently, a group of pediatric cardiologists at Boston Children’s Hospital looked at their decision making over a period of one week, grading the strength of evidence behind each decision. 5 In that time period, ten cardiologists made a total of 1,188 decisions. Almost 80% of those decisions could not be related to any published data. Less than 3% were based on a study specific to the question at hand. Although this study comes from a pediatric cardiology practice, it is representative of doctors in general, including congenital heart surgeons.
So why are so few of our decisions based on scientifically derived studies?
Medicine is complicated!!! The human body is complex. Performing studies is expensive. In the area of congenital heart disease, many abnormalities are rare, and even within a single diagnosis, there is often a spectrum of deformity. Thus, designing statistically valid clinical experiments comparing one treatment to another is challenging at best, and impossible in most cases. That is not to mention the ethical issues and limited financial support for studies of uncommon medical conditions.
Monitoring Outcomes
So, if the public is demanding medical excellence, and we want to provide medical excellence, how can we determine whether we are doing the right thing and doing the right thing right?
If we cannot set up the experiments we would like, we must pay attention to what we are already doing. We all like to think that what we provide is of the best quality. On a day-to-day basis, we think we do a pretty good job. But if we don’t regularly sit down and review our results and compare them to some ideal of quality, we don’t really know.
The whole idea of measuring and evaluating outcomes goes back at least to 1913, when Ernest Codman, a surgeon at the Massachusetts General Hospital, introduced the “End Result Idea.” 6 His idea of tracking surgical successes and failures makes sense to us now, but it was not embraced by his colleagues; in fact, it was disparaged and essentially ignored. Quality of medical care was simply assumed by most practitioners until 1966 when a physician named Avedis Donabedian began to study health care quality and outcomes. 7 He was the one who recommended breaking health care delivery into three elements: structure, process, and outcome.
The importance of monitoring quality became painfully evident to the adult cardiac surgery community in 1987 with the publication of Medicare mortality rates for coronary bypass surgery. 8 The publication of these outcomes proved embarrassing to many centers and served as the impetus to the development of the STS Adult Cardiac Surgery Database.
At that time in pediatric heart surgery, we were more focused on achieving survival in our more complicated patients and failed to recognize the importance of continuously monitoring our overall results. In the early 1990s, however, monitoring quality in pediatric cardiac surgery came to our attention. Formal inquiries into two separate pediatric heart surgery programs, one in the United Kingdom and one in Canada, revealed unexpected high levels of mortality. 9,10 Several similar but less publicized instances of excessive pediatric heart surgery mortality were also occurring in the United States.
Marc de Leval, at the 1993 meeting of the American Association for Thoracic Surgery, courageously reported on his own experience with a cluster of surgical failures. 11 He admitted in this published report that had “a mechanism of continuous monitoring…been in place…the unfavorable trends…would have been detected earlier” and lives would have been saved. To his credit, he decided on his own to retrain in the performance of this operation.
Collecting pediatric heart surgery data had actually begun in the mid-1980s. We can be proud that the CHSS was among the first to collect outcome information from multiple institutions. 12
The importance of continuous quality assessment that became evident in the 1990s led to the development of two major congenital databases: the European Association for Cardio-Thoracic Surgery (EACTS) and STS databases, which have grown substantially since their inception. These databases, which are in essence, registries, allow us to determine what real outcomes are and not just those outcomes that warrant publication in a scientific journal. These databases allow each center to compare its results with those of other centers.
We have to thank several members of our two societies for the substantial progress made in the development of these databases. Dr Gus Mavroudis and Dr Jeff Jacobs deserve much of the credit for the STS Congenital Database 13 and Dr Bohdan Maruszewski for the EACTS Congenital Database. 14 It is fitting to note that these two databases were not developed completely independently, but with a significant amount of collaboration and cross fertilization across the Atlantic. It is especially gratifying that the number of patients and procedures in these databases has reached the point to allow clinical research, providing descriptive analyses of real results, not just exceptional ones.
If you are a participant in the STS Congenital Database, you are periodically provided with a report of your outcomes comparing them to the outcomes of the other participants (Figure 2). In a separate letter, you are given a code to tell which one you are. If you are one of the low mortality centers you are proud and relieved, but if you are one of the higher mortality centers, you should take note. Using this database as a self-assessment tool is a primary reason for its development. But, I am struck by another feature of this example: the degree of variation in mortality between centers.
Representative graphical depiction of discharge mortality from centers submitting data to the STS Congenital Heart Surgery Database 2003-2006. Dotted line = overall value for the fifty-eight centers. CI = confidence interval. (With permission from Jacobs et al. 23 )
Variation
Figure 2 also demonstrates the range in overall mortality at 58 different centers. A range of 1% to 7.5% is disturbingly broad, but that comparison may not be fair as it lacks adjustment for case mix. The range and degree of variation in outcome, however, can be appropriately displayed when looking at high-risk STAT (STS/EACTS Mortality Categories) 25 category 5 operations (Figure 3) and the Norwood operation (Figure 4). Variation in outcome does not only exist between centers but may also exist within a single center. Note in Figure 5 taken from a recent CHSS study that institution 18 was among the best with the Norwood operation and interrupted aortic arch repair but among the worst for operations for pulmonary atresia with intact ventricular septum. 16
Variation in mortality for STAT Category 5 Operations at centers submitting data to the STS Congenital Heart Surgery Database 2008-2011. See text for derivation of term STAT. The horizontal dashed line depicts aggregate STS mortality prior to discharge for all STAT Category 5 operations performed in 2008-2011: 17.0%. Dots represent the number of cases and mortality prior to discharge for individual STS Congenital Heart Surgery Database Participants (Centers). The bars depict 95% confidence intervals for each for individual STS Congenital Heart Surgery Database Participants (Centers). (With permission from Jacobs et al. 24 )
Variation in mortality for the Norwood operation at centers submitting data to the STS Congenital Heart Surgery Database 2005-2009. Mortality data for the Norwood operation is displayed as a funnel plot. The horizontal dashed line depicts aggregate STS mortality prior to discharge for the Norwood operation. The dashed lines depicting exact 95% binomial prediction limits were overlaid to make a funnel plot. Squares represent the number of cases and mortality prior to discharge for individual STS Congenital Heart Surgery Database Participants (Centers). (Derived with permission from Jacobs et al. 15 , pg 2188.)
Variation in mortality within single institutions. Six institutions submitted sufficient data to compare outcomes within themselves for four surgically managed diagnoses. Each institution is noted along the horizontal axis and its rank regarding mortality for the four diagnoses is represented along the vertical axis. The lower the rank, the more superior the performance. Note the significant difference in rank for institution 18 for the Norwood and the PAIVS groups. (PAIVS, pulmonary atresia with intact ventricular septum; TGA, transposition of the great arteries; IAA, interrupted aortic arch.) (With permission from Karamlou et al. 16 )
Why do these variations in outcome occur? Is it because we vary in our practices?
It turns out that there is a lot of variation in how we practice. For example, Burstein and colleagues published a article that looked at models of care delivery for patients with congenital heart disease at 94 centers in the United States and found significant variation. 17 Primary responsibility for postoperative care varied significantly among these centers with cardiothoracic surgeons assuming full responsibility in some centers, critical care physicians in some, cardiologists only in a few, and a mixture of specialists in others. Another article looking at preoperative and intraoperative practices for newborns with hypoplastic left heart syndrome also showed significant variation between centers. 18 Figure 6 depicts how varied the choice of initial procedure was among surgical sites, with some dedicated to the classical Norwood, some to the right ventricle to pulmonary artery conduit, some to the Hybrid and others with a mixture. And, of course, if practices vary between centers, it is not surprising that hospital costs have also been shown to vary significantly (Figure 7). 19
Center variation in choice of initial surgical palliation for hypoplastic left heart syndrome. Each center is noted along the horizontal axis. RV-PA, right ventricle to pulmonary artery. (With permission from Brown et al. 18 )
Center variation in unadjusted total hospital costs for patients undergoing four different congenital heart repairs. Each box represents one center with the middle line representing the median cost and the box ends the interquartile range. ASD, atrial septal defect; VSD, ventricular septal defect; Art, arterial. (With permission from Pasquali et al. 19 )
So is this like that situation in Vermont, with the tonsils?
I think it is especially if we fail to recognize that there is a lot of variation in what we do and what we get. Recognizing that there is significant variation in our field is a crucial step to improving its quality and providing excellence to our patients.
When Jeff Jacobs presented at the 2011 STS meeting the variation in outcomes for benchmark operations in congenital heart surgery, I was struck with the importance of dealing with variation in our field. 15 It was that recognition that inspired me to suggest that the CHSS form a new committee to focus specifically on quality improvement and outcomes. Now that we know there is a lot of variation in our field, how do we deal with it?
First, variation in practices is not all bad. Without variation there would be little innovation or impetus for scientific investigation. In some cases, variation in practice is dictated by the resources we have to work with. But when there is significant variation in outcomes, it makes sense to look back at the variation in practices to determine which practices correlate with the best outcomes. Practices that do not correlate with the best outcomes create unnecessary variation and should be eliminated.
Raising All Boats
Eliminating poor performing programs results in only a modest improvement in overall quality (Figure 8). A more desirable approach is eliminating unnecessary variation: improving the outcomes at all centers while minimizing the variation in outcome between centers (Figure 9). Put differently, the goal is to raise all boats!!
Removing poor performing providers has minimal effect on improving overall quality. (With permission from Miles 26 .)
Applying a quality improvement initiative to all providers reduces variation among providers and improves overall quality. (With permission from Miles 26 .)
How to do that, however, is not readily obvious. Delivering care to our patients is complicated and multifactorial. Dissecting care into its many parts with statistical analysis is one approach, but that is slow and tedious, and may miss important factors that don’t reach statistical significance. In my opinion, we will improve outcomes faster if our programs spend more time sharing than competing. Let’s learn from the sites with the best outcomes. John Kirklin had hoped that would happen when he founded the CHSS Data Center in 1985. I am not sure that that ever happened. But it can happen now.
Because surgeons are the cornerstones of all congenital heart programs, we have the clout to make it happen. I would like to finish by telling you about an exciting although challenging initiative that our society is pursuing in collaboration with the STS. Over the past year, members of our Committee on Quality Improvement and Outcomes have been working with the STS and several pediatric cardiologists to develop a quality improvement collaborative for pediatric cardiac surgery.
Jeff Jacobs, on our behalf, and John Mayer, on behalf of the STS, have been the prime leaders in this effort. The idea is to create a learning collaborative based on the approach pioneered in adult cardiac surgery by the Northern New England Cardiovascular Disease Study Group. 20 That study group was created in response to the Medicare report of coronary bypass surgery in 1987. These hospitals recognized a striking variation in mortality and variation in care processes between centers.
Because they felt that Medicare reimbursement for heart surgery would be in jeopardy if they did not improve their results, they took the brave step of sharing data and comparing processes. They introduced the idea of having teams of key caregivers from each center visit the other centers looking for ways to decrease variation in processes with the hope of improving outcomes.
As a result of that approach, a dramatic reduction in coronary bypass mortality was experienced at all of the participating centers. Similar adult cardiac surgery initiatives have been applied in Virginia and Michigan with similar improvements in variation and outcome. 21,22 It is hoped that these principles can also be applied to congenital heart surgery.
The key elements proposed for our initiative include participation of multidisciplinary teams from different programs, the conduct of round robin site visits, the sharing of outcome data, and training in continuous quality improvement techniques. This is a complex and ambitious endeavor, but we have been able to enlist some well-qualified help, which includes surgeons, cardiologists, and an anesthesiologist. If we are able to secure funding for this program, I hope that you will be willing to include your institution in this collaborative, especially if your center is a high performer.
So, to summarize, congenital heart surgery has come a long way since that first meeting of the Congenital Heart Surgeons’ Society in 1973. We should be proud of the progress that has occurred in our field over the past 20 to 30 years. We should also be aware that the quality of our work must not be assumed but must be regularly assessed. With that assessment, we must expect that there will be variation in outcomes. How we deal with that variation will determine how much we improve the quality of our work and how we are perceived by the patients and families that we serve. Remember that they are demanding medical excellence. May our society and our sister society across the Atlantic be recognized for delivering medical excellence.
Footnotes
Presented at 3rd CHSS-ECHSA Meeting, September 14, 2012, Chicago, IL, USA.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.