Assessing Shoulder Pain Part 1: Special tests are nothing special

Part 1: Special Tests shouldn’t be so special. In fact most of them suck. 

Special tests for the shoulder are contentious—rightly so.

I’ll contend here that some special tests have value, but most are just plain old rubbish.

But before we get into the thick of it, let’s start with the premise of a special test: A special test adjusts the probability that a pathological condition is either present or absent. Differently said, it helps you to rule-in and/or rule-out a specific pathology.

Are we always sure what a test is doing though? For instance, do you know if Empty Can is ruling sub-acromial impingement in or out? or maybe this test is actually for rotator cuff pathology (and would that be for complete tears, partial tears or for tendinopathies?), or is it giving information on labral integrity? What about biceps pathologies? Do you know?

Turns out, that the empty can test has predictive power for all of these conditions. Problem is, the predictive power is in every case weak. All the investigations into this test come out with sad little likelihood ratios that give us next to no idea what to do with a positive or negative result. Add to this the conflicting findings between studies that look at the empty can test and the issue gets cloudier still. Click here, here, here, here, here, here, here, here, here, here, here, here, here, here, and here to look at some of the studies that have investigated the psychometric properties of Empty Can. Take a preemptive aspirin. 

As a new grad, I did empty can and all the other tests based on what I learned in school. I didn’t have a clue what Sensitivity, Specificity, or Likelihood Ratios meant. When I sat there, like an idiot, not knowing what to do with the results of the 20 shoulder tests I just went through, I would scratch my head, give up, and say something impressive sounding like, “you have a supraspinatus tendinopathy, with some concomitant subacromial bursitis.” It was the ‘concomitant’ that made them think I new what I was talking about.

What a load of crap.

I didn’t start to learn what the psychometric properties were for special tests for a year or so after starting practice. It didn’t cross my mind that this was something to learn. When I found out the likelihood ratios that came out of a positive Hawkins Kennedy, Neers, Painful Arc, Empty Can, Horizontal Adduction, etc I was disgusted. None of them were strong enough to allow me to arrive at a diagnosis with any degree of confidence. Resisted ER was slightly better for SAIS but still fairly impotent with a +LR of 4.39 (representing about a 15% increase in the chance that the person has impingement).

after ruling out red flags, I don’t care what the precise pain generating structure is.

As part of this inquiry, I learned about test item clusters and the good work done by a fellow named Park, and more good work by another fellow named Michener. Park and colleagues found that when Painful Arc, Infraspinatus test, AND Hawkins-Kennedy were all positive then you can finally get a satisfying LR of 10.56 (or +45% chance) that there’s some sort of impingement going on. Even when you get the strong likelihood ratio pointing at impingement syndrome, what does that mean clinically? Or in other words, “How does this information guide treatment?”

My argument, is that it doesn’t. At least, it doesn’t any more than a comment in the history along the lines of, “it hurts my shoulder when I do overhead activities.”

The reason I say this, is because, after ruling out red flags, I don’t care what the precise pain generating structure is. I’ll use other strategies to guide treatment and prognosticate. But that’s for next weeks post. For now, I have to hedge a wee little bit. So we’re clear…. I’m not advocating that we throw out the physical exam. I’m also not saying we should get rid of all the special tests.

Some special tests are well powered to rule in/out complete RC tears (bear hug, hornblowers, IR lag sign, lift off test); GH instability (load and shift, apprehension-relocation); and labral tears (Biceps Load, Crank, O’Brien’s). I wouldn’t call the information from these tests all that precious though; it is, at best, a small part of what guides treatment and prognosis.

With that said, the physical exam does have a ton of utility and usefulness. Take a look at the list! The physical exam lets you:

  1. Screen the region (CSpine, TSpine, ribs, the rest of the shoulder girdle).
  2. Get some baseline measures.
  3. Identify impairments
  4. Identify asteriks signs
  5. Identify symptom modifiers
  6. Impress upon the patient your attention, caring, thoroughness, and expertise. (Contrast this to a brief visit with their GP who may not have had time to even expose the area)

So let’s not throw the pearls out with the shells. Physical Exams are brilliant. I just want to urge you to investigate shoulder special tests further and decide for yourself how useful they are.

Fun fact teaser for next week: 1) Frost and colleagues found that MRI findings of pathological supraspinatous tendons are both very common and related to age. Interestingly, they found that pathological supraspinatous tendons are NOT related to impingement and/or clinical status... Yet another blow to the biomedical model. 

Next week , part 2 will investigate an alternative strategy to guide treatment. We’ll talk about impairments, asteriks signs, and symptom modification. We’ll also examine prognosticating on shoulder pain patients. It will be a satisfying dessert following this special test roast.