How Surgeons Make Decisions

Decisions for the care of our patients should be based on sound, reliable, and up-to-date evidence. This seems to be a truism for the members of the European Surgical Association (ESA), a group of Academic surgeons of whom I have the honor to have served as president.

In this year's presidential address, I will try to analyze how we as experienced academic surgeons make decisions about the diagnosis and treatment of an individual patient and how we could do better, at least how we could stimulate the surgical trainees to do better in the future.

According to the constitution of the ESA, a member has “...a reputation as a surgeon from contributions as a practitioner, author, teacher and/or original investigator....” We might even think about ourselves as opinion leaders in surgery.

Opinion Leaders

Australian surgeons, asked to identify opinion leaders in their field, defined them as clinicians who have an influence on the education and surgical practices of their colleagues.¹ The most respondents could only identify a few colleagues with such a national profile. Of the potential attributes to confer status as an opinion leader, expertise and teaching skills were rated as very important by 80% to 90% of the respondents. The opinion leaders were thought to have and could demonstrate up-to-date command of medical knowledge and could demonstrate a high level of clinical expertise, and they were familiar with scientific evidence, underpinning surgical practice, and enjoyed and were willing to share the knowledge. Opinion leaders were rated as very influential by significantly more surgeons than clinical audit and clinical practice guidelines. It may be disappointing for the members of the ESA that academic and professional contributions were not important attributes to the status of an opinion leader. The authors conclude that there is an important role for the opinion leader in promoting evidence-based surgery.¹

Authority and Surgery

Whereas many experienced surgeons, opinion leaders or not, would be happy to play a role in evidence-based surgery, others such as the surgical trainees will identify them just as authoritarian surgeons. And it is true that some degree of obedience, or compliance, is a requirement of training.

The authoritarian surgeon-leader can be especially recognized in academic departments of surgery in some European countries with a tradition of strong leadership and hierarchy. Those leaders have jurisdiction over house staff, house staff has authority over students, and all these people are seen as authorities by patients. Authority in solving clinical questions is acceptable as long as the authority's social power derives from expertise and knowledge and from information given in a convincing manner.² At this time, most surgeons realize, to paraphrase David Sackett,³ that the right answer to a clinical question cannot be found in the page of an authoritative textbook or at the feet of an authoritative attending surgeon. Thus, clinical controversies do not resolve to appeal to this authority (ie, the clinician with the most experience or the academic with the most prestigious research record). We should make decisions for the individual patient on a basis of objective, unbiased, and systemic analysis of scientific knowledge. We should practice evidence-based surgery or at least teach this to our students.⁴

Evidence-Based Surgery

Surgery lacks behind general medicine in the use of the principles of evidence-based medicine, but this shortcoming is recognized lately.^5,6 However, to practice evidence-based surgery, we need evidence.⁵ And although the amount of publications regarding clinical studies is overwhelming, most are observational studies, some with exciting data from the leading surgical centers of the world.^7,8 But the level of evidence of these studies is low when compared with randomized controlled trials (RCTs) and their subsequent systematic reviews, which represent the highest level of evidence (www.cebm.net/levels_of_evidence.asp).

RCTs are thought to be the best method to compare the effectiveness of different therapeutic interventions and thus more helpful to make decisions for the care of an individual patient.⁹ The number of RCTs in surgery is rather limited; and for that reason, the level of surgical research also has been criticized.¹⁰ Only 3.4% to 7% of the publications in the leading surgical journals are RCTs, depending on the definition of leading surgical journal and other factors.^10,11

Of these so-called surgical RCTs, 25% to 50% deal with the comparison of 2 or more interventions,^10,11 whereas the others report nonsurgical adjuvant treatment or other interventions in combination with surgical procedures. Layton Rikkers, Editor-in-Chief of Annals of Surgery and honorary member of ESA, was more optimistic when he concluded in a presidential address for the Society of Surgeons of the Alimentary Tract that the percentage of RCTs of all manuscripts in Annals of Surgery increased from 2% in 1973 to 12% in 2000.¹²

The contribution of the members of the ESA to the same journal is relatively large. From 1996 to 2004, 17 of 90 (19%) papers published in the yearly issues that are in part dedicated to the meetings of the ESA were RCTs. This hopefully shows the scientific potential of the Association and the attention that the members give to clinical trials. Almost half of the RCTs from the ESA came from The Netherlands, and this might reflect the important place that was given to evidence-based surgery in this country in recent years. It may also be a reflection of and the interest of Dutch surgeons to challenge novel therapies invented or introduced by much more innovative other surgeons.¹³

Whereas surgical scientific activity is frequently measured using the contribution of RCTs to surgical journals, this is only one part of the story since many surgical RCTs are published in high-profile medical journals. In the last 10 years, 226 of 828 RCTs in the New England Journal of Medicine were identified as “surgical,” although only a minority of these studies compare surgical interventions.

Complexity of Surgical RCTs

Much can be said about the complexity to perform RCTs in surgery and how much more difficult it is to compare 2 surgical interventions than 2 drugs.¹⁴ No more than half of the therapeutic questions in the gastrointestinal surgical literature could be answered by a controlled trial if the clinical settings were ideal.¹⁵ There are 3 important factors that especially hamper the setup of a surgical RCT. First, there is often a strong preference of the patient for one arm of the study (eg, when minimal invasive surgery is compared with open surgery or when surgery is postponed by neoadjuvant nonsurgical treatment). Second, many conditions are too infrequent to allow a single-center study, and many centers have to participate to have sufficient patient accrual. These multicenter RCTs carry the risk of large variation in surgical quality in the different participating centers, often depending on the volume of interventions per surgeon or hospital.¹⁶ Third, the surgeon can have a strong opinion in favor of one of the alternatives. This is very often the reason for surgeons and surgical groups not to participate. What is experimental in one place or country can be standard therapy in the other. This is the case, for instance, when extended lymph node dissection added to resection of gastrointestinal cancer is studied. Transthoracic esophagectomy and lymph node dissection was compared with transhiatal esophagectomy for adenocarcinoma of the distal esophagus in a RCT in 2 centers in one country,¹⁷ whereas the extended resection was standard therapy for some years in another unit¹⁸ and participation to this RCT by the latter unit was thought to be unethical.

Evidence-Based Surgery in Clinical Practice

There is only limited scientific evidence on the majority of clinical problems, and the surgical community is not always aware of the evidence that is present. A study among French gastrointestinal surgeons showed that less than half of their current practice was against existing evidence regardless whether these were academic surgeons or surgeons that practiced in other settings.¹⁹ Also, Australian colorectal surgeons did not recognize in all guidelines the level of evidence. But guidelines based on level I and II studies were much better recognized and accepted.²⁰

Evidence-based practice appeals probably much more to the young surgeon, especially when trained in EBM as is usual in the Department of Surgery in the Amsterdam Medical Center. Surgeons can use their authority to ask a trainee to solve a clinical problem according the rules of EBM or EBS as they learned in these courses.⁶

I have confronted a surgical trainee, with a track record in surgical science, with the ongoing discussion whether extended lymph node dissection should be added to gastrointestinal cancer resection. Although this discussion is somewhat settled in our unit and in The Netherlands with regards to gastric cancer^21–23 (not indicated) and distal esophageal cancer¹⁷ (most likely indicated in Barrett's cancer), there is much debate whether extended lymph node dissection is of benefit in pancreatic cancer. She (the surgical trainee) performed analysis according to the rules of EBM (www.cebm.utoronto.ca/practice/).

Practicing Evidence Medicine

The clinical problem was phrased in a way that could direct the search to relevant and precise answers to the question whether for a 61-year-old Dutch man with a pancreatic head mass without evidence or local ingrowth and distant metastases that undergo pancreaticoduodenectomy, extended lymph node should be added to the operation.

The PICO structure was used (Table 1): Searching for the best evidence in the literature revealed 3 publications about 2 relevant RCTs^25–27 (Table 2). Two RCTs were recently published as abstract.^28,29 The assessment of the published studies is depicted in Tables 3 and 4.

TABLE 1. The Clinical Scenario in the PICO Structure

TABLE 2. Process of Literature Search

TABLE 3. Worksheet According to the Principle of EBM for Pedrazolli et al²⁶: Results From RCT 1

TABLE 4. Worksheet According to the Principle of EBM for Yeo et al²⁵ (Shorter Follow-up in Same RCT): Results From RCT 2

Postoperative mortality and morbidity and survival after 1 and 3 years were used as outcome parameters. According to the practice of evidence-based medicine, event rates in the control arm (CER, no lymph node dissection) in the experimental arm (EER, lymph node dissection), and the relative risk increase (RRI) and absolute risk increase (ARI) and numbers needed to harm or treat (NNH or NNT) were given.

After critical appraisal of the 2 level I studies, the evidence could be applied to our patient, and the conclusion was that extended lymphadenectomy could not provide surgical benefit in this patient but lead to longer operative time and increased chance of delayed postoperative gastric emptying²⁵ or decreased quality of life.^29,30 The 2 RCTs that have appeared as abstract led to a similar conclusion that pancreaticoduodenectomy with extended node clearance²⁹ or para-aortic node dissection²⁸ did not provide any survival benefit in pancreatic cancer.

Our surgical trainee warns us that the reported trend to a survival benefit in a subgroup in one study was methodologically unjustified²⁷ and the number of extra lymph nodes retrieved with extended lymph node dissection was moderate (13.3 versus 19.8).²⁷ This finding casts some doubt on the completeness of lymph node dissection and on the definition of the operative technique.³¹ This would obviously be a justified criticism from Japanese surgeons who are more experienced in lymph node dissection added to gastrointestinal cancer surgery. Therefore, the preliminary negative results of the new Japanese study²⁸ are even more convincing, although the definitive report should be awaited.

For the time being, we conclude that, as in gastric cancer,³² lymph node dissection in pancreatic cancer is most likely more important for staging than for therapy.

Future

The systematic approach used for the above-mentioned patient with a pancreatic tumor demonstrates that a complex clinical problem can be solved when the principles of evidence-based surgery are followed. Experienced and authoritarian surgeons and young surgeons or surgical trainees can work very well together in this process, especially after being trained in the technique of critical appraisal in well-structured courses as are offered. In the evidence-based surgical practice, the surgeon will step down from the pedestal and discuss clinical questions with surgical trainees on an equal level. This will positively affect quality of surgical training and patient care. This effect seems optimal as long as surgical practice is not tyrannized by evidence-based guidelines and as long as there is a place for experienced-based and innovative ideas. Therefore, authorities in the field of surgery are challenged to promote evidence-based-surgery, as is currently the practice of the Royal College of Surgeons of England in programs for young surgeons.³³

In an atmosphere of evidence-based practice, the paucity of evidence at the highest level will oblige us to adhere to lower evidence-based guidelines³⁴ or, more importantly, will stimulate us to set up randomized controlled trials as a basis for new guidelines. The perceived increase of RCTs within this association shows that we are on the right track. Although much has been said about the gloomy future of surgery and the decreasing interest in surgery as a profession partly because of the excessive working hours,³⁵ the shift from authority-based to evidence-based surgical practice will compensate for many negative effects and makes surgery attractive for young surgeons of both sexes. Also, these future surgeons will be better trained and interested in surgical research. When surgical research has been mocked as a comic opera, at least this comic opera has a happy ending.

ACKNOWLEDGMENTS

The author thanks Drs. Tjarda van Heek and Liesbeth Jansen for turning their systematic approach to a clinical problem and all other surgeons in training at the AMC for challenging the author's authority.