Knowledge based practice: Functional movement systems

Ramez Antoun recently connected the KBP blog posts on differential diagnosis to the Functional Movement Systems, Selective Functional Movement Assessment (SFMA) break outs. This peaked my interest and reminded me that the 2010 book by Gray Cook et al was one that I had wanted to read - it was time. I do recall Ramez and his classmate Connor Ryan explaining the concepts to me while they were students - it actually led to some discussions about heart rate variability (HRV) and stress as the book introduces people to HRV that might otherwise have never considered it a biomarker, and my dissertation was on job stress and HRV indices. I have been generally aware of the FMS approach and, of course, the FMS screen itself. The system generally resonates with me given my background including a solid course on qualitative movement analysis, a highly integrated neurological PT course sequence with a full year of PNF in all of it’s glory, and a doctoral dissertation on HRV and stress that included lots of readings in systems theory generally, and complexity theory in particular. As a cardiopulmonary PT, it is common to incorporate PNF and functional approaches in the examination, evaluation and interventions of patients that are quite ill, and several years ago we added an entire section to the cardiopulmonary lab based on what are considered neuromuscular techniques for breathing (by a section I mean all students get a full 1/3 of a semester on these methods).

At this point I am simply 3 chapters into the book but I must say I am very impressed with the system Gray Cook and his colleagues have put forward. It is a spectacular example of the reasoning the profession needs. It resonates quite well on several levels. It is an excellent example of the balance that KBP is attempting to promote between rationalism and empiricism with a critical realist epistemology. At it’s core it is a system derived logically from first principle a priori’s regarding movement. These principles are a massive culmination of knowledge - integrated from several fields of study that are synthesized into a causal structural system which can then put forth hypotheses for testing, but more importantly principles to develop abductive and deductive inferential processes (algorithms) for practical use. These can be modified should there be new knowledge based on further empirical evidence that inductively leads to modifications of the underlying causal structure.

The system also highlights the multi-level (hierarchical) complexity of the science and knowledge that physical therapy practice is based. I have commented on this before when describing the limitations of randomized controlled trials. FMS (as a total system that includes the two tools FMS and SFMA) articulates an excellent system that gives a structure to much of the complexity and variabilty that therapists face in clinical practice. This is the sort of variability and complexity that is highlighted in my 2005 letter to the editor of the PT journal about using complex systems to help clinical research become more clinically useful (here). FMS (again, the full system, not just the screening tool) is a step in the right direction for coming up with not only knowledge based practice approaches to exam, evaluation and treatment, but approaches to research study design that are isolated for the sake of making unbiased inductive inferences, but connect to a larger, well described though not fully understood, causal structure. With the full causal structure described we can discuss the “known knowns” the “unknown knowns” the “known unknowns” and perhaps most importantly, the “unknown unknowns” associated with the hierarchy of causes of movement that can lead to dysfunction. Of these, the “unknown knowns” usually raises an eye brow or two - an “unknown known” is knowledge (a conclusion) that can be deductively determined from a valid deductive form and premises that we know (or accept) as true that entails the conclusion (thus a sound conclusion); but that we had not considered prior to reasoning (rationalizing) about the causal structure.

Ramez and I are currently starting to collaborate on translating the knowledge of FMS into a system of DAGs - directed acyclic graphs as representations of graphical causal models, as a representation of the underlying knowledge of the FMS system that Cook et al have articulated so well.

Based on FMS - the big picture, highest level - more coarse grain DAG might look something like this be:

dagitty-model-2With System Traits and System States being high level abstractions (emergent properties)  that cause movement patterns (observable actions, effects).

Realizing movement patterns are then causes of system traits and states we must unfold the loop over time – for an explanation see this post – but here is what it would look like with the FMS big picture would be:

FMS-BigPicture-Core-IteratedWith each iteration happening in the future - thus not depicted with a cycle as traditionally depicted with a feedback look, but depicted across time as it happens in reality.

Special thanks to Ramez Antoun for also reading through this post to check for factual or reasoning errors regarding my understanding of the functional movement systems approach!

Leave a Comment