Which surgeons are the best

How surgeons can learn from the best

There is a lot of talk in medicine about quality. A surgeon from Zurich takes it seriously and suggests benchmarks in his area of ​​expertise. Does this herald the future of quality assurance?

The topic of quality is omnipresent in medicine. But: What distinguishes good medical quality? And what is the best way to measure it? This question has been driving surgeon Pierre-Alain Clavien from the University Hospital Zurich for years. His solution is called benchmarking. The term comes from the corporate world and describes the comparative analysis of products or processes using a fixed reference value (benchmark). Company A knows how it stands with its “goods” in comparison to the competition.

The director of the Clinic for Visceral and Transplant Surgery also wants to apply this principle to the quality of surgical treatment. This is a revolutionary idea, because up until now, benchmark studies in health care have primarily examined treatment processes. After intensive study of the subject, the surgery professor came to the conclusion that this could also make the patient outcome comparable.

Goal: fewer complications

The beginnings go back to the 1990s. At that time, the classic "open" gallbladder surgery was competing with the laparoscopic method (keyhole surgery). Clavien wondered which surgical procedure would be better for the patient. In this case, “better” meant fewer complications. Because the same goal is achieved with both treatment methods: the removal of the troublesome gallbladder.

"Around half of the patients with complications have multiple complications."
Pierre-Alain Clavien, visceral surgeon, University Hospital Zurich

So the first thing was to develop a method with which the complications caused by the operation can be recorded simply and objectively. Objectivity was particularly important because until then only mortality had often been recorded. If non-fatal complications were counted, they were mostly classified as severe or easy on a purely subjective basis. Clavien solved the problem by assessing the complications in terms of their need for therapy (see table).

Severity of surgical complications

Grade 1: Any deviation from the normal course that does not require intervention. Simple medication for nausea, pain and the like as well as physiotherapy are allowed.

Grade 2: Complication requires drug therapy such as antibiotics or blood transfusions.

Grade 3:
Complication requires surgical, endoscopic or radiological intervention - without (3a) or with general anesthesia (3b).

Grade 4:
Life-threatening complication requiring intensive care treatment. Dysfunction in one (4a) or in several organs (4b).

Grade 5:
Complication leads to death of the patient.

Source: Classification according to Clavien and Dindo

This approach has the advantage that you only need to look at the medical history: Did the patient receive antibiotics to treat pneumonia caused by surgery? Did he need a transfusion because of a bleeding? Or did he have to be transferred to the intensive care unit after the procedure? "You would have to be very active in order to keep this information hidden in the medical history," says Clavien.

The "Clavien-Dindo classification", which was tested for reproducibility and acceptance on a cohort of over 6000 surgical patients - the then assistant doctor Daniel Dindo played a key role in the work - was presented in a publication in 2004. Today it is considered the worldwide standard for recording surgical complications - not only in clinical studies, but also in everyday medical practice - and is the most frequently cited surgical work of all.

Despite this success, Clavien saw further room for improvement. In the next step, he and his team converted the classification into a point value: from 0 (no complications) to 100 (death of the patient). With the Comprehensive Complication Index (CCI) presented in 2013, which can be calculated using a computer program, all complications and their severity are systematically recorded in patients for the first time.

“This takes into account the fact that around half of the patients with complications have multiple complications,” explains Clavien. And his medical counterpart at the University Hospital Lausanne (CHUV), the visceral surgeon Nicolas Demartines, adds: "The CCI is also a brilliant idea for research." This is because differences in the complication rate can be revealed on the basis of a comparatively small number of patients.

The potential of measuring instruments such as the Clavien-Dindo classification and the CCI can be seen in the third stage of Clavien's mission for better surgical treatment quality: benchmarking. The discussion began with a simple question: What is the best possible result for a particular operation? To answer this question in their own specialist area, Clavien's group carried out a study with twelve leading hospitals for liver surgery worldwide to determine relevant benchmarks. One of the top centers was the University Hospital Zurich.

The researchers chose living liver donation as the first operation. This is because patients who give away part of their liver are usually young and healthy. Safety is therefore the top priority in this intervention. For their work published in 2016, the scientists analyzed the treatment results of more than 5000 patients - three and six months after the procedure. In addition to the Clavien-Dindo classification and the CCI, they also considered various definitions of liver failure.

Constantly monitor quality

The scientists defined 75 percent of the median value from the twelve top clinics as benchmarks (for each treatment result). For example, the benchmark for the general complication rate was 31 percent. This means that as long as no more than 31 percent of the patients in a clinic suffer any complications from this procedure, the clinic works well. For serious incidents, the researchers determined a benchmark of 9 percent, for the CCI a benchmark of 33.

"The 75 percent hurdle for the benchmark is chosen arbitrarily," says Clavien. The idea behind it is to define a benchmark that is geared towards the best, but does not only accept the “top of the top”. The doctor emphasizes that Roger Federers are not needed in surgery. If you achieve a result like the best 75 percent of the clinics, you are doing very well. On the other hand, anyone who misses the benchmark as a clinic, department or individual surgeon has to rethink their work and look for opportunities for improvement. This is what benchmarking is all about.

Romand Demartines is already implementing this idea of ​​continuous monitoring of treatment results in his clinic. "Every week, we as a team discuss all patients and look at their complications," he says. This process can be further systematized with benchmarks and the treatment results can be compared between the clinics.

“You don't need Roger Federers in surgery. If you achieve a result like the best 75 percent of the clinics, then you are doing very well. "
Pierre-Alain Clavien, Visceral Surgeon, University Hospital Zurich

Like Clavien, Demartines is convinced that benchmarks are essential for an objective discussion about medical quality. That would finally stop comparing apples with oranges, he says. The doctor recalls the case of a Swiss university hospital that performed only a few liver transplants but had a zero mortality rate. What some saw as proof that a small number of cases does not speak against high quality, for others it was a sign that the clinic only treated patients with a low surgical risk.

Benchmarking could defuse this misunderstanding. Because since March benchmarks have also been available for liver transplants - as well as for other surgical interventions. They have been identified for precisely defined patients with a low surgical risk. So you could look at the results for this group of patients at Clinic X. At the same time, you would then also know how many patients were treated with an increased risk; with them "worse" results are to be accepted. How bad these can be could be evaluated with further benchmark studies, in which only high-risk patients would be included.

Provide the necessary pressure

The benchmark concept is particularly impressive for large and difficult interventions, says Clavien. This is because the clinics could learn from the best worldwide or nationally (depending on whether the benchmarks are defined internationally or nationally) instead of, as is usually the case today, based on the national or regional average. In addition, patient-relevant outcome data are more meaningful than the quality indicators most commonly used today, such as the number of cases and mortality rates (see graphic).

Where the risk of dying after an operation is lowest

Postoperative mortality in Europe compared to the UK

For Clavien it is therefore clear that benchmarks should be used for the political control of the treatment offer. That is much fairer than the current system, which only requires a minimal number of interventions. According to Demartines, both case numbers and benchmarks should be used as criteria. Because there is no question that a certain amount of practice is required for a good quality of treatment. This not only affects the surgeon, but the whole team, which has to be well trained in complicated operations.

Clavien is aware that his ideas do not trigger storms of enthusiasm in everyone. He says: "There are different perspectives in health care - we represent those of the patient." In Demartine's experience, the large hospitals tend to be in favor of benchmarks, while the small ones tend to be against. The doctor says that the latter are often afraid of losing their offer. But that doesn't have to be: "If the small clinics do a good job with simple cases and pass on the difficult cases, then everything is okay."

"The allocation of services is not decided on the basis of the quality of treatment, but rather on the basis of the number of cases and the structural quality in the clinics."
Dieter Hahnloser, visceral surgeon at the Lausanne University Hospital

What upset Demartines, on the other hand, is the attitude taken by individual surgeons that everything used to be operated on and it always went well. "These people don't look at their treatment results properly." This is also due to the ego of some surgeons who do not accept restrictions, says Demartines. Another "brake" for a real quality discussion is the fact that many medical professionals still live in the "comfort zone". "It is far too seldom asked: What are we not doing well, what can we improve?"

The results of selected benchmarks could provide the necessary pressure here, Clavien is convinced. Nevertheless, he advocates careful and sensible handling of the “hot” numbers - also with regard to their publication. "Such information does not belong on Facebook," says the doctor. At the same time, however, it must be ensured that a hospital that does not achieve the benchmark goals improves or stops its services.

Bad data quality

That is still a long way off. However, according to Clavien, such a benchmark system could be introduced relatively quickly in Switzerland. Because the data on complications required for this are already collected routinely today. The Working Group for Quality Assurance in Surgery (AQC) plays an important role in this. The organization, founded in 1995, collects patient data on a voluntary basis and compiles statistics on hospital admissions and surgical interventions. According to the company, this should enable comparisons among the participating doctors and hospitals - with the aim of increasing safety and transparency in surgery through improved quality.

What sounds good, however, is still far from a lived reality. According to Luzi Rageth, head of the AQC office, the data quality of the AQC database is currently insufficient to make reliable statements about case-related, postoperative complications in Switzerland. The main problem lies in communication within the clinic about complications and communication after leaving the hospital. In addition, some hospitals lack honesty. Another problem inherent in the system is that the “black sheep” in the industry do not participate in the AQC, explains the HSG economist.

In this regard, it should look better with highly specialized visceral surgery, which includes complicated interventions on the liver, esophagus, pancreas and colon. In this area, the clinics are obliged to provide outcome data. In addition, the database is audited, which means that the completeness and correctness of the data are randomly checked. Despite these measures, the quality of the transmitted data is still too poor for a serious hospital comparison, explains visceral surgeon Dieter Hahnloser from the CHUV on request.

In addition to technical problems, the doctor sees another reason that some hospitals do not take the issue seriously enough. This also has to do with the fact that the allocation of services in highly specialized visceral surgery by the Conference of Health Directors (GDK) was only provisional in some cases. The decision was not made on the basis of the quality of treatment, but primarily on the basis of the number of cases and the quality of the structure in the clinics, says Hahnloser. That shows where the priorities are currently.

"It is far too seldom asked: What are we not doing well, what can we improve?"
Nicolas Demartines, Visceral Surgeon University Hospital Lausanne

Hahnloser hopes that the GDK will definitely make its allocation this year. After that, it is important to move forward with quality promotion including benchmarks. This is also the right path for Stefan Breitenstein, chief physician for visceral surgery at the Winterthur Cantonal Hospital. As head of the quality management department at the Swiss Society for Surgery, he is convinced that benchmarking will be used in all areas that go beyond basic surgical care. In order to guarantee good data quality, the data delivery in these areas must be declared mandatory and the data must be checked by an independent party.

There is another motivation for Clavien to quickly introduce the benchmark system into surgery. Studies show that the complication rate correlates better with treatment costs than, for example, the patient's surgical risk. This has to do with the fact that the treatment of serious complications quickly becomes very expensive, explains the doctor. Increasing surgical quality is therefore the best way to save a lot of money in the health system.

That should actually be of interest to politicians. Hahnloser waves it away. Health politicians in Switzerland - due to the nature of the system - looked primarily for their canton, he says. Clavien also relies less on politics than on the patient in promoting the quality of surgical treatment. Because they are likely to want to know more and more often where the quality of treatment is good and where one should rather not go. Associated with this is the insight that you have to travel a bit for a major elective intervention, says Clavien.

Follow the science department of the NZZ on Twitter.