Knees: Just the Facts, Doc

If asked, most orthopedic surgeons would say that while they are not scientists, they are grounded in—and have faith in—the sciences. But there is little that is scientific, says our knee expert, about how most knee specialists respond to patient queries.

 Dr. Donald Shelbourne, founder of the Shelbourne Knee Center in Indianapolis, has performed more than 6, 000 anterior cruciate ligament (ACL) reconstructions…and he has data on every single one of them. He doesn’t “think” he knows something…he can prove it. Dr. Shelbourne:

If a patient asks the average orthopedist, ‘What are your outcomes for this procedure?’ I can almost guarantee that the doctor will say, ‘pretty good.’ If the patients says, ‘Well, how many complications have you had?’ the surgeon will most likely respond, ‘Not many.’ As physicians, we should be able to be much more precise about our work.

Tinker Gray, Research Director at the Shelbourne Knee Center, has been along for the nearly 30-year ride of data collection and analysis. She states, “Dr. Shelbourne planned to collect data from the outset of his practice in 1982. Sports medicine was just starting out, and when I came here in 1984 we established a more systematic way of following up with patients, i.e., we graduated from using old computer cards. We started out asking rather basic questions and then as Dr. Shelbourne found that he needed additional information, we added more and more data points.”

Ahead of his time, all those years ago a younger Dr. Shelbourne had the radical idea that he wanted to know how good of a job he was doing. “I wanted to have a concrete sense of my effectiveness, where I was faltering and why, and what I needed to change. Even back then I could see that the utility of being able to definitely state, for example, ‘XYZ is the best treatment for this type of meniscus tear’ would be fantastic. That way, patients and colleagues would know that I am not giving my opinion, I am stating facts.”

Although Dr. Shelbourne’s work is scientific, it turns out that sometimes, it “just” involves a bit of horse sense. “Years ago I began thinking about the typical way that we approached rehabilitation after ACL surgery, namely, putting the patient on a restrictive program with crutches and braces. I did a compliance study because I suspected that our rehab directives were unrealistic. It turned out that the patients who did not listen to me fared better than those who did. The medical student collecting the data told me that almost all patients were not following my mandate that they sleep in the brace (although they were telling me that they were)."

"In the end, we found that those people who did what they felt comfortable doing—and followed their common sense—had a better medical outcome. When I presented the study there was a lot of hubbub, with people saying, ‘I don’t agree with you, ’ to which I replied, ‘This is not my opinion…these are the data speaking. These are the facts.’”

Tinker Gray adds, “At the time, it was accepted that 10% of patients who underwent ACL surgery had to have surgery later for range of motion problems. When the medical student was calling postop patients he also found that those who did weight bearing exercises earlier than Dr. Shelbourne had advised had better ROM. That gave some insight into that fact that ACL rehab wasn’t where it needed to be.”

Dr. Shelbourne is inflexible when it comes to creating flexible, healthy knees. He says, “Let’s take kneecap dislocations. Our results were not consistently good with these conditions, so we wanted to dig in and see why. I set the bar at what I thought was a normal goal—that both knees would be symmetric. Our treatment objective is to make the patient feel like they have two normal knees (which is what they want). So we found that it makes sense to start by asking, ‘What is different about the knee that feels comfortable to them, i.e., the healthy knee?’”

 “Our initial results were 75% to 80% (as far as making knee mirror knee), but the patient populations weren’t homogenous. People with dislocated kneecaps can have a congenital predisposition for the condition; there could also be a trauma dislocation on top of that. So in our database section on patellar realignment we have seven different types and can subclassify things. Since we have begun this process our success rate on patellar dislocations has gone up to 90%. The problem overall is that surgeons don’t have good success rates with operative treatment. Patients present seven different ways, but many surgeons only operate one way.”

From her view inside the workings of the Shelbourne data machine, Tinker Gray says, “I have seen over the years that surgeons tend to think, ‘Oh, well, XYZ patient hasn’t returned to see me so they must be doing fine.’ No, no, no…we have learned…you must find the patients and ask how they are doing.”

And how do they manage this gargantuan process? Tinker Gray: “Each year during the anniversary month of someone’s surgery a computer program sends a letter and survey to the patient via email. If there is no response, then we send out a paper survey. All of the surveys are validated statistically, and measure pain, stability, everyday functioning, and activity level in terms of sports or work activities. The surveys are very specific, i.e., there is one for older patients with osteoarthritis, another for patients who are heavily involved in sports, etc. People write a lot of comments on our surveys and Dr. Shelbourne reads every one of them. This has been an important lesson: the doctor has to be involved. One might be tempted to hire an outside company or just say, ‘My research department is going to handle that, ’ but taking one of these routes means that your understanding of the results will not be as clear.”

Gray, the one who dots the ‘I’s and crosses the ‘T’s when it comes to the database, has advice for those surgeons and/or practices embarking on substantial data collection. “Think it through from the beginning as far as the type of factors you want to evaluate. For example, for ACLs there are different types of meniscal tears and varying treatment approaches. It is best to collect this specific information at the time of surgery and record it in a manner that allows you to sort it properly.”

And while you may think that you have a simple system, and can wait awhile before undertaking the data analysis, that could lead to a hornet’s nest of issues down the line. Gray states,

Once you start to collect data don’t wait a year or two to analyze it. That way you can best detect any problems with coding, sorting, etc., and can rectify them. Remember that as time goes on you will be getting more curious and experienced, and will be asking more questions—meaning that specificity in coding is also critical.

"For example, if there is a lateral meniscus tear you should indicate whether the tear was in the middle third, anterior part, etc. You would also want to code for the way the tear is different from others, as well as the different ways of repairing it. Not going to this level of detail means that you can’t answer the question, ‘Is ABC type of tear more amenable to repair than XYZ tear and how exactly it should be treated?’”

Among other things, Dr. Shelbourne sees 200 ACL tears and 20 patellar dislocations per year. Such volume naturally creates the opportunity for detailed knowledge. “The average orthopedic surgeon sees 20 ACL tears a year and one or two patellar dislocations. That is just not enough to recognize patterns. But I am surprised to be so alone in the woods."

"The fact that very few people are collecting this data is rather surprising. It’s a bit like playing basketball and taking shots, but not looking to see if they go in. If a 15-year-old kid asks me, ‘What will my knee be like in 10 years?’ I can’t envision telling him, ‘I don’t know.’”

But most surgeons are not able to provide such information…it just doesn’t exist, says Dr. Shelbourne. “A couple of years ago we got a call from a well known knee surgeon who said, ‘I am supposed to give a talk on failure after ACL surgery, but I can’t find any data. Can you help me out?’”

“Unfortunately, once most orthopedic surgeons walk out of the OR they don’t have much interest in postop…they are moving on to the next case. If the patient is truly motivated, however, then he or she is going to want some answers. If the surgeon says, ‘Go home and elevate your leg 8 inches’ there are patients who will say, ‘Why not 12 inches?’ If you say, ‘Do your rehab exercises two or three times a day’ some patients will ask, ‘Why not four?’ We surgeons are the ones in charge of this whole process and yet so often we can’t give specific answers. I actually feel sorry for patients. Orthopedists used to be good at treating musculoskeletal problems—now most of us are only good at operating on musculoskeletal problems.”

And the barriers to changing this situation?

Think about it, ” says Dr. Shelbourne, ‘any funding out there is coming from companies to sell you things to use in the OR. They are not exactly motivated to further nonoperative or postoperative treatment.

"There is more likelihood that laboratory or animal research will be funded because that is usually a relatively quick one or two year study. To follow patients long term, however, is a substantial commitment of resources…not to mention the issues with obtaining IRB approval for even simple follow-up of patients.”

But Dr. Donald Shelbourne and Tinker Gray get the job done…and provide data and guidance to anyone seeking to provide the ultimate in data-based knee care for patients.