How do you know that?

After four years of medical school, six years of orthopedic surgery residency, and a year of subspecialty training in shoulder surgery, I went out into practice in 2013 with a healthy apprehension that there was still so much more to know than what I could have reasonably learned over the course of 11 years. Even the most experienced surgeon only knows a certain subset of the sum of all that is known, and all that is known is only a certain subset of all that is knowable, and all that is knowable is only a certain subset of the knowledge that would be helpful in diagnosing and treating patients. That doesn’t, however, mean that we don’t know anything. We do try to offer patients an honest assessment of what we think they have from a diagnostic standpoint, as well as what we think they should do about it from a therapeutic standpoint. Inevitably, we will get asked by patients and colleagues, “how do you know that?” — instead of getting defensive and falling back on a “do you know who I am?” type of justification, I think it is much more beneficial to explain, as simply as possible, exactly how we arrived at the diagnosis or treatment we are proposing.

The funny thing about a diagnosis is that it’s not always a good thing to have. Earlier on in practice I felt like every patient needed a diagnosis, and I have realized that this is not exactly true. Normal shoulders sometimes hurt. I know it sounds heretical and antithetical to the miracle of modern medicine that treats everything and anything with a new pill or procedure. But it’s true. Not every patient that comes to see me in the office for shoulder pain needs a diagnosis. Once we have gone through our checklist of both common and esoteric potential diagnoses, we may find that a specific patient has either none of these diagnoses, or a combination of several diagnoses without a clear primary diagnosis. This is not a problem to be solved, but a reality to be accepted. It is for this reason that I sometimes need to tell patients the following before I send them for an MRI:

“Regardless of what the MRI shows it will be good news and bad news. It may show a rotator cuff tear that needs surgery, or it will show that you do not need surgery. If we see a tear and decide to go ahead with surgery, the good news is that we have a good sense of what the problem is and a reasonable solution, even if that solution is painful and will consume the next 6–12 months of your life. If we do not see a tear, we may not be left with a very clear diagnosis, but at least the good news is that you don’t need surgery.”

There is a cost that comes with clarity. There is a benefit that comes from not falling into a neat category because, if you are seeing a specialist, anything “bad” it would be discovered, even if the evidence for the diagnosis is subtle. In any case, there are only a few diagnoses in the shoulder world that need immediate or urgent attention, and as long as those are not present, there is no harm in waiting for the problem to declare itself more clearly.

Now that we are sufficiently wary of the two edged sword of having a diagnosis, we can enter into the abyss of choosing the proper treatment. We can look here at shoulder arthritis as a simple patient driven decision making paradigm, acute rotator cuff tears in young patients as a simple surgeon driven decision making paradigm, and chronic rotator cuff tears in older patients as a very complex scenario that requires a proper shared decision making process.

There is nothing simpler or easier than talking to someone about their bone on bone arthritis. Every joint in the body allows for motion by having a layer of cartilage that serves as the smooth bearing between the two bones that meet at that joint. In the case of the shoulder the top of the arm bone is a sphere that sits on a golf tee called the glenoid, which is part of the shoulder blade. When the cartilage coatings of both sides of the joint wear away, the exposed surfaces of the two bones end up grinding on each other, often resulting in severe pain and loss of function. If the patient is doing well enough and managing with the shoulder in its current state, I would say that no treatment is necessary, unless the patient becomes more symptomatic. If there is a constant, achy pain that affects their quality of life and their sleep at night, a treatment of inflammation in the form of NSAIDs or an image guided cortisone injection may be helpful. If this gives them lasting relief, then that can be the mainstay in their treatment plan. Finally, if the patient has symptomatic bone on bone arthritis, that has failed conservative management, we can then discuss the possibility of a joint replacement. When a patient is looking for me to validate their decision to accept or decline surgery I always tell them “it’s your shoulder … if you can live with it, I can live with it, and if you can’t live with it, nothing would make me happier than to replace your shoulder when you are ready to go ahead.” The patients are always taken aback when I tell them how much I love doing shoulder replacements, since I have already spent the previous months/years talking them out of surgery. It is important for them to understand that the reason to move forward with surgery is because they need it, not because I like to do it.

An example of a rotator cuff tear for which I would recommend surgery is a 55 year old manual laborer that routinely lifts 50 pounds above her head, injured herself at work, and her previously perfect asymptomatic shoulder has now sustained a full thickness rotator cuff tear that affects her daily activities and wakes her up at night in severe pain. In this case the likelihood of returning to heavy overhead lifting without surgery is unlikely — not impossible, but not likely. Furthermore, the likelihood of having a good result with surgery is highest if it happens within the first six weeks after injury. Therefore, since we are unlikely to get the desired result without surgery, and there is a time limitation for getting an optimal result with surgery, this becomes one of the few situations in which I would recommend surgery to someone the first time I meet them. Even though surgery comes with risks, and there is no guarantee that the patient will be able to go back to her physically demanding job, there is still enough evidence to suggest that this is the best choice for most patients in this situation. It is just as simple and straightforward as the arthritis example above, but instead of being very simple in a way that allows the patients to choose any path at any time based on their symptoms, it is simple in that there is one clear treatment that makes sense for most people.

If my entire practice was primary joint replacement and acute rotator cuff tears, my life would be very easy. However, in my practice I have the privilege of helping patients with very difficult problems make the best possible decision with incomplete information. One area of significant controversy that does not lend itself to simple recommendations is the world of chronic rotator cuff tears. These are tears of the rotator cuff that we see on MRI, but we do not have a clear timeline on what specific event caused the tear, such as we do with the acute rotator cuff tears that are related to a specific traumatic event that we presume to be the cause of their tear. Just so we do not get into the weeds trying to make the sometimes somewhat unclear demarcation between an acute and chronic tears, let us say we are dealing with a 65 year old gentleman with many years of shoulder pain, no specific injury, who has failed conservative management and comes in to review his MRI, which shows a full thickness retracted tear and some early muscle atrophy, which is an indication that this tear is not new. Since the patient has already done 12 weeks of physical therapy with faithful adherence to his home exercises without any improvement, it is now reasonable to say that the patient should be offered further treatment, since continued therapy is unlikely to help. The options are to repair the rotator cuff, or to do a reverse shoulder replacement. Both of them are good treatments, but each one comes with significantly divergent positives and negatives. For example, a rotator cuff repair is intended to restore the patient’s normal anatomy, while a reverse replacement permanently converts the ball to a socket and the socket to a ball. A rotator cuff repair requires six weeks of immobilization in a sling followed by six more weeks of activity restriction, while a reverse shoulder replacement allows the patient to wean out of the sling and return to light activities as tolerated. A rotator cuff repair has a significant risk of retear and failure, while a reverse shoulder replacement should be a fairly predictable and reproducible operation. Surgeons should at least consider both of these treatment options, even if they do not necessarily offer both options every time. The appropriateness of one treatment over the other must be considered by the surgeon, and this perspective should be conveyed clearly to the patient as a critical part of their decision making process. There are times where I think the patient would almost certainly have a dismal failure with a rotator cuff repair, and I will tell them that it is not a good option for them. There are other times where I think the patient will be unhappy with the limitations of a reverse shoulder replacement, and I will tell them that even if the rotator cuff repair may be less predictable, it is worth doing since that is their best chance of getting the function that they want to have in their shoulder. Finally, there are times where I truly think both options are reasonable to consider, and in these cases we spend quite a bit of time discussing the expectations in each case. I try to frame it for patients by describing the two surgeries like they are two movies that I have seen before many times, and I am simply trying to help them pick which movie to watch.

As time passes, we will have more information on which to formulate both more clear diagnoses, as well as safer and more successful treatment options. Perhaps there will also be treatment algorithms that use artificial intelligence to give a greater degree of certainty on the correctness of diagnoses and a greater degree of clarity on the likelihood of success of a certain treatment. That being said, we will still be left with a certain degree of murkiness that comes with complex medical decision making, and it will always be the clinicians job to help the patient navigate these waters. My only hope here is that I have framed for you how truly difficult that is even when the clinician is well trained and well intentioned, so that you are not disappointed with answer if if you ask “how do you know that?”