It would seem that the problem with evidence in physiotherapy is quite often that:
- a) there isn’t enough of it
- b) what there is is of bad quality
- c) you can’t really blind people to a lot of the treatments
- d) a number of the outcome measures are subjective as opposed to objective which gives an inherent risk of bias.
- e) the outcome is nearly always “we need to do more research”.
And these are only the problems that I can pull off the top of my head on a Saturday evening.
Traditionally we see clinical trials- randomised double blind trials as pretty much the gold standard of testing. It works well with pills, they are very easily blindable- the doctor has no idea if they are giving you the real thing or not, you have no idea if you have the real thing or not, and only the people designing the test really know what is going on. The outcome is very much objective- are your bloods better/worse, have your symptoms cleared up etc. etc. It is pretty difficult to affect your bloods through willpower alone, and so these are fairly decent tests.
The issue with physio is this:
I can’t, as a therapist- be blinded to whether I’m giving you exercises or not. I either am, or I am not. The same can be said for massage, for acupuncture, or taping or whatever. I know what my expertise is, and so I am cannot be blinded from that.
You, as a patient cannot really be blinded from it either. The closest you get to being blinded is if someone puts on tape in what they consider to be an ineffective pattern, or pokes you randomly, or gives you sets and reps at a level which is technically therapeutically unuseful.
A lot of the things we test for are fairly subjective- pain being the main one. We try to be objective- in terms of ranges of movement, muscle strength etc. but there are massive inter and intra-testing reliability issues. Even with the best will in the world, our accuracy is awful. Not only that, but think about muscle strength. There are huge amounts of things which might affect you- the patient, which might change the outcome of the test. You might be angry that your spouse annoyed you this morning, the drive might have riled you up last time and not this, your sleep patterns could be totally different, heck you might not have had your weetabix this morning. All of these mean that even if we’re trying to be objective, that might not work because you just happen to have had a good or bad day.
Pain is also pretty interesting, We know that if you’ve had a stressful day, or are generally under stress you are more likely to be more susceptible to pain. We know that if you have been sensitised, the pain score is more likely to be higher. We know that colour can affect pain scores, what kind of a day you have had- even down to how the therapist interacts with you- is going to have a difference about how you report pain.
So you can see, even if the evidence is there, it might not be totally reliable because you haven’t had a good day. This is neither a good thing or a bad thing in terms of being a patient, but it’s a complete nightmare if you are a data scientist trying to work out an objective measure for pain, given that different doses of something might provoke different pain responses from the same person dependent on the day or time.
Which brings me onto another thing about evidence in physiotherapy- or in general- a lot of the evidence base is based on specific populations- the big, well powered studies are often done in the military, others are done with sports people, some are done with much smaller cohorts. However, there are always restrictions to the population studied. The researcher wants to be as specific as they can and try to cut out confounding issues. In an MSK scenario, those who have had a stroke, anyone with type II diabetes, or another metabolic disease or some other mitigating factor be excluded from the study. In runners, someone who has a pre-exisiting sprained ankle might be excluded from a knee study etc. We cannot do this in clinic- the patient in front of us may have exisiting co-morbidities that show up on no research paper ever…. rarely do our patients actually look like the population on which studies are done.
So what to do? How do we know that it was the exercises that we gave you that made you better? How do we know it wasn’t the demeanour of the therapist that you saw? The reassurance that it was nothing serious? How can we be sure that your pain score that has gone down is actually because of what we have done, or is it because the payout from the car accident you were in has finally come through and now you are less stressed as a consequence?
I really don’t know the answers, but we do know that a combination of expertise, good communication, researched backed knowledge and good therapeutic alliance goes a long way.