The problematic side of prediction - the PT workforce

I recently attended a workshop that discussed estimation of workforce for physical therapy. It made me think about the problematic aspects of prediction. You should not be surprised to learn that I think of this as a causal model allowing predictive inferences. The model is based on currently understood cause - effect relationships; but also based on projected cause - effect relationships. The extent to which our predictions are based on currently understood vs. projected cause effect relationships will certainly vary. As anyone that is making such predictions knows, there are lots of assumptions either way.

My favorite story regarding prediction is the Great Horse Manure Crisis of 1894 - discussed effectively and thoughtfully at this blog (here).  It is clear that in this crisis, innovation and creativity curtailed all predictive concerns - though created some new ones.

On the one end of the spectrum - only based on current cause - effect relationships - the assumption is that new causal forces will not emerge. The other end of the spectrum - adding causal - effect relationships - the assumption is that the model has effectively predicted the causal - effect relationships that will emerge. Both are assumptions and both clearly influence the accuracy of predictions from the predictive model.

There is a requirement for DPT programs - existing and new - to justify the number of students (and existence for new programs) based on predicted work force needs. I do not have a problem with this generally. It is important to keep this in mind. However, such balancing typically occurs in a market economy through market forces (Adam Smith’s “invisible hand” - which really just refers to cause-effect relationships we cannot see). For example - if the workforce starts to require less physical therapists there is a response, fewer students seek the expensive and time consuming education towards becoming a physical therapist. Programs will shrink out of necessity, some may close, some may not start. I do realize there can be delays in such a response, but overshoots in the number of therapists but such perturbations are not necessarily bad.

I have put together a few directed graphs (not necessarily DAGs as some of them have cycles that I did not want to unfold due to time and space constraints).


In all models the availability of a job is based on supply and demand. The real questions come from what are the causes of supply and demand. The supply side is easier - supply is subtracted from with attrition (any reason to leave the supply) and added to with graduates (new professionals entering the supply). Demand is influenced by insurance, health care, public health and demographics - as rather large conceptual categories.

dagitty-model (1)

In this model I have added a cycle that could be unfolded if we wanted a DAG but for space reasons I have stuck with a directed graph. The cycle here is simply that the availability of a job influences students enrolling which then influences graduates and rights supply. There is likely a time delay so it is not a perfect negative feedback control system but it does prevent supply from growing too out of hand.

dagitty-model (2)

Here I have added University Programs - they certainly influence Graduates. More programs would mean more graduates assuming students willing to be enrolled does not limit graduates. However, University Programs should also be having an influence on Demographics, Public Health, Health Care and Insurance. The scholarly productivity of a University program should influence these variables, in some ways reducing demand, in other ways increasing demand through the identification of new approaches, new markets, innovative practice opportunities.

Of course - a sufficient supply of physical therapists is really quite critical to several features of demand. In other words, without supply then demands eventually dissipate. dagitty-model (3)

They find other solutions to the problems that are not being met due to a small supply.

Supply is needed. So if we use the availability of jobs to determine our workforce needs, and we intentionally reduce our supply to attempt to be a “visible” hand that manipulates the market to avoid too much supply, are we going to thwart the ability to implement innovative and creative solutions to the problems of public health, health care and insurance? Are we going to shrink demand through a shrinking supply? Has this happened already in the profession?

To be innovative, creative and to truly “Transform society through movement to improve the human experience”,  - I suggest we need an over supply. We need more physical therapists than we currently have and that we can currently predict we need because we cannot know for sure the extent to which a lack of supply diminishes innovation, creativity and the solving of problems.

So, when predicting workforce needs, never underestimate the value of a robust supply to the generation of ideas.

Leave a Comment