Ruling Out vs Ruling In
📝 Weekly paper summary
Functional Movement Screen Task Scores and Joint Range-of-motion: A Construct Validity Study (Hincapié et al., 2021)
Category
Construct validity study
Context
During movement assessments, researchers and practitioners are often interested in understanding the underlying nature of one's pain, injury, or limited performance. For example, many practitioners and organizations use the Functional Movement Screen (FMS) as a tool to identify aberrant movement patterns that justify future interventions. A central assumption of the FMS is that failure to complete an FMS task is indicative of a joint "mobility" or "stability" limitation. However, there is little data regarding the validity of the FMS tasks, or any other movement assessment battery, to measure joint range of motion. Although previous data suggest that limitations in ankle and hip range of motion (ROM) capacity may increase spine flexion or knee valgus during a squat, it's incorrect to automatically assume the converse (i.e., that squatting with excessive spine flexion or knee valgus means you have ROM limitations; this doesn't logically follow from the previous literature) or contrapositive (i.e., that squatting without excessive spine flexion or knee valgus indicates no ROM deficits; this does logically follow from the previous literature, but re-testing findings in different ways is always a good idea). Therefore, the purpose of this study was to assess the construct validity of the FMS.
Correctness
Again, this is one of our lab's studies, so I'll try to limit my bias as best as possible :)
Although we used raters with extensive training (physiotherapists, athletic therapists, etc.), the practitioners made all ROM measurements with a goniometer. We then averaged the results from three ROM measurements for each side. No matter how precise we were or how good our training was, there will be more noise in the measurements relative to a motion capture system. However, given that we had a sample size of 101 varsity student-athletes, I believe we alleviated this signal-to-noise ratio concern sufficiently.
Another limitation of this data is that, for some FMS tasks, the scores were very imbalanced. For example, consider that in the squat task, only one individual received a score of "0" (felt pain when performing the overhead squat). In contrast, 57 people scored a "1" (unable to perform the pattern as described), 38 people scored a "2" (completed the pattern with minor compensations), and five people scored a "3" (completed the pattern perfectly). In the rotary task, zero people scored a "0" or a "1". There is no hard and fast rule describing how unbalanced is too unbalanced, but this is a statistical limitation worth highlighting nevertheless. However, I can confirm that we got an identical answer when I ran the stats for this study with rank-ordered statistics (where this becomes less of an issue). Therefore, I think our findings are as robust as possible for this dataset. It would be nice to have a more well-rounded sample in the future, but you take what you can get when sampling from athlete populations :).
A final limitation to highlight is that we did not do any limb-specific analyses but instead took the lower average unilateral ROM measurement as the individuals' ROM capacity. It was necessary to do this given that the number of statistical tests would have doubled otherwise (and we would have surely had unknown false positives to deal with). Therefore, although the relationships between movement and ROM measures weren't as direct as possible, our testing was more conservative, which I believe is a positive. For example, if your left ankle has sufficient dorsiflexion, but your right doesn't, it's more conservative to say you have an ankle ROM limitation than to say you don't.
Contributions
- FMS task scores can "rule out" ROM limitations. In other words, if you score a "3", you certainly have a higher ROM capacity, but if you score lower, it's unclear whether you do or do not have a ROM limitation. Therefore, it's unfair to assume the converse statement highlighted above (that just because you move poorly means that you have ROM limitations). However, the contrapositive appears to be true (that squatting without excessive spine flexion or knee valgus indicates no ROM limitations).
- Just because someone cannot perform a task without excessive spine flexion or knee valgus does not mean they have a ROM limitation! This logic is a common mistake that (especially young) practitioners make when assessing movement patterns; they see someone flexing their spine during a squat and automatically assume that a hip and ankle ROM intervention is required. However, this data suggests that the practitioner should instead conduct more specific ROM tests following the movement task to "rule out" ROM limitations and identify the actual cause of the movement impairment.
- It's probably a good time to distinguish the differences between screening and assessment. Screening is used to identify the possible presence of a problem (i.e., excessive spine flexion when squatting; a yes-or-no outcome). In contrast, assessments are processes to understand why someone may be moving the way they do (i.e., what is causing this person to squat the way they do). Therefore, we can use movement tasks to screen for aberrant movement behaviours. However, understanding why those behaviours emerge to address them during training or rehabilitation requires much more nuance.
🧠 Fun fact of the week
Greenland sharks can live up to 400 years, and females are not ready to reproduce until they are around 156 years old! Sharks were already super cool, but learning more and more about aquatic wildlife, in general, makes me realize how little we (especially myself) know about what lies in the oceans (or how old the life there might be). Super cool!
🎙 Podcast recommendation
Pretty great conversation so far! I'm about halfway through and enjoy Tim's perspective and approach to answering Lex's questions.
🗣 Quote of the week
"Weakness is strong. I must be stronger"
- Jocko Willink