Accuracy in Physiotherapy Diagnosis

[Laura Thornton]

Pretty great article once again collaborating physical therapist and orthopedic surgeon practice. Eric suggested to review this article in BJSM by our friend Mark Jones et al. and share it for everyone on the discussion board.

Does Physiotherapy diagnosis of shoulder pathology compare to arthroscopic findings?

Mary Magarey, Mark Jones, Chad Cook, Michael Hayes
BJSM 2015; 0; 1-7

This study analyzed diagnostic accuracy of the physiotherapist musculoskeletal examination by comparing diagnosis with arthroscopic diagnostic examination for common shoulder pathologies.

First of its kind to look at a comprehensive shoulder examination with interview and physical tests/measures instead of one or cluster of physical tests.

The authors found that the physiotherapists had consistently higher specificity and lower sensitivity in diagnosing specific shoulder pathologies. They were also better at identifying the structure, but not the pathology of the structure at fault. There was marginal, non-significant influences of physiotherapy diagnosis on the actual results of the arthroscopy.

The authors make a strong point about how the PT assessment is not only based on pathological tissues but in focusing on how and why the shoulder and surrounding region is impaired in movement and function. Therefore we have an inherent error in the ability to identify the specific pathology within the specific structure at fault. We can sacrifice this ability to see the exact pathology to focus on the impairments and participation restrictions that guide our treatment plan.

What does everyone think about the applicability of this study?

[Laura Thornton]

Attached the article:

[ABengtsson]

It’s a really interesting article and I like the very objective reflection on our limitations, as well as strengths when it comes to evaluation.
The main thing I take away from this is, similar to what we talked about during our shoulder weekend (and in general), is that we really need a good cluster/combination of tests to determine how to treat a pt.
I think it would be interesting to know what Mark Jones (or Chad Cook) would have to say about how and if these results significantly change the way they evaluate and treat patients.
One of the hardest evals for me is still the shoulder, especially when it comes to figuring out structure at fault/interpreting clusters, even more so when it’s a very irritable pt and the majority of tests are (+).

Thanks for posting this!

[omikutin]

Thanks for sharing! It’s interesting that we have a lower sensitivity. I sure hope we are able to rule out any serious underlying pathology, VBI, systemic, UMN, LMN. I believe ruling out the red flags are most important. Like you said as a PT our goal is to look at movement and function and correct the impairments. I still get stuck trying to figure out how do I treat a patient with low back pain if I don’t know the specific pathology? Specifically patients with low back pain, I have found it helpful putting them in classifications per Dellito and treating them accordingly. Alex- I also find the shoulder tricky. I try to rule out first anything systemic and then from the subjective I see which test clusters I shoulder use (impingement? instability?etc.) It’s a lot to think about.

[Nick Law]

Halley,

Thanks so much for sharing and posting. I must admit that this article was a little hard for to with the large proportion of statistics and how the measured agreement/accuracy. However, after spending a little bit of time with it I think I have a decent understanding of the results.

A few thoughts:

– Oksana – > the patients included in the study were those who had undergone standardized routine examination by a physician including the use of other imaging modalities; only patients deemed appropriate for arthroscopic investigation were included. That is, the ruling out of red flags had already occurred by the time the patient made it to the PT/arthroscopists, so that this study really can’t be used to help determine if we are/aren’t able to identify more serious red flags via our comprehensive clinical examination, though you are right to indicate that this certainly is important. I would like to think that if studied we would have better accuracy at determining mechanical vs. non mechanical shoulder pain.

– It certainly makes sense to me that we are better able to identify the structure at fault compared to the specific pathology of that structure. I think that this has larger clinical relevance as well. That is, my treatment approach differs more if my working hypothesis is glenohumeral hypermobility due to a passive constraint lesion than some form subacromial disorder in the absence of hypermobility, than it would if my hypothesis was supraspinatus tendinopathy compared to subacromial impingement.

– I don’t suppose myself to be a better therapist than any of the clinicians participating in this study; however I would be curious to know what elements were included in the standardized history and examination, and the diagnostic reasoning form that was used. Perhaps I missed it, however I did not see any reference to an appendix that might include such information. I would be interested to know how their examination and reasoning process compares to the one we utilize.

– “The finding of structural pathology at arthrosocopy does not necessarily mean that the pathology is the source of the participants presenting symptoms.” I think that is a highly relevant statement, one that must be considered when interpreting the results of the study. How would the results have differed if the MD diagnosis was made on the basis of both history and exam + arthroscopy results vs. arthroscopy alone?

– The final paragraph in the discussion prior to the limitations was most helpful to me. Two things in particular. First, in our examining and treating patients we must evaluate treat the entire patient and not just the pathoanatomics. Second, although identification of specific tissue pathology may help with safety and prognostic value in certain cases, our treatment is based heavily upon impairments identified contributing to functional limitations vs. tissue specific pathology.

[Laura Thornton]

I was the same way with the statistics…I looked over this one for awhile.

Alex and Oksana I think you guys make good points about when a patient is highly irritable and a lot of the tests are positive, it’s hard to differentiate the structure at fault. The authors talk about one of the limitations of this study is that a lot of these patients probably are in the highly irritable side of the spectrum if they are appropriate for arthroscopic assessment and intervention so how clouded are our tests at this stage?

That’s a great example with the shoulder and treatment of 1. passive restraint vs. subacrominal disorder than 2. impingement vs. supraspinatus tendinopathy. It’s hard to tell how much our treatment approach would change for #2 if we could make a specific diagnosis like that.

I think with a lot of our PT interventions, we are affecting more structures around the shoulder joint than an arthroscopic intervention so the importance of specific pathology in specific structure might not be so crucial for our intervention. For example, with postural strengthening exercises to improve shoulder mechanics during elevation with subacrominal disorder, we are changing not only the muscle coordination/control and strength of contraction of the target muscles, but training the positioning of the shoulder and taking stress of painful subacromial structures (might not know the exact one). With an arthroscopic intervention, they are targeting one or two very specific pathologies of the shoulder and not much else.

[sewhitta]

The authors do a nice job of highlighting the fact that there is evidence to show pathology can be present without pain. I feel this is the discrepancy in the results. I say the reason the PT didn’t identify as many RC tears as the arthroscope did is because they weren’t symptomatic and there is a risk their operation will be on a structure that’s not even causing them pain. I would argue that knowing a pathology that is present will bias your treatment approach. We see this in the clinic all the time. For example, you may have seen a patient for an initial evaluation and the script from the MD reads “MRI reveals L5-S1 disc herniation, please eval and treat with extension exercises and core stabilization.” This doesn’t mean it’s symptomatic. Now I’m going into my eval with disc herniation in my mind, my judgement may be clouded and there’s a chance I could overlook other possibilities. O’Sullivan states in his 2005 paper on classification of CLBP, “even when a specific pathoanatomical diagnosis can
be made, there is still a need to classify the disorder based on the mechanism(s) that drive the pain disorder to ensure appropriate management.” As Nick stated, it’s nice to know precisely what structures are at fault and the pathology present to make predictions on healing and prognosis and to educate our patient’s, but our goal is to improve functional movement and our approach is based on symptom irritability. This is more challenging when a patient has been informed of abnormal imaging findings. Patient’s have an incredible desire to know exactly what’s wrong with them. When a patient feels pain, they need to know why. They don’t like to hear their therapist run through a list of possibilities and seem uncertain. They need to know what’s going on and what can be done about it. When their doctor can look at an image and say with certainty “this is what you have and I can fix it with surgery”, that’s very concrete and easy to conceptualize. However, this is the problem with healthcare. I feel patient’s would have much better outcomes with therapy if physicians stuck to just performing physical examinations because, as research shows, there is much better agreement between clinicians. I nice follow up study to this one would be to follow these patient’s after their operations and throughout therapy and recovery and perform another physical examination and history and compare the findings.

[Nick Law]

Halley – I completely agree with the notion of affecting multiple structures, not just the structure at fault. There could be one or multiple subacromial structures serving as the pain generating, and yet in my PT intervention I am focusing on a whole host of factors and tissues to improve mechanics and function.

I would want to add that, although at one level identifying and treating a highly specific tissue (supraspinatus tendonitis vs. subacromial bursitis) may not make our treatments different in LARGE proportion, it may make a small difference, and those small differences may in some ways be significant. That is, although our accuracy may not be all too good, I don’t think we should completely abandon the attempt itself to try and be as specific as possible, recognizing full well our limitations. My guess is that expert clinicians have learned to better (though not perfectly) detect subtle differences in pattern recognition and because of this they are slightly more pointed and focused in their treatment.

Sean – I completely resonate with your example. I simply think we need to be so confident and convincing in our explanation of the impairments we find, how such impairments are contributing to the patients pain and specific functional limitations, and how we can adequately address them that the patient has trust and confidence in us even if we haven’t given them a specific pathology. Easier said than done for me.

[Author]

Posts