It’s time to start development….. Since my talk at CSM (viewable here) I have been promoting the idea that eventually there should be a “repository” of DAGs for developers to post, and users to use with an overall spirit of collaboration to follow when it comes to developing, using, and disseminating graphical causal models as representations of clinical knowledge.

After all, DAGs as graphical causal models make assumptions about fundamental causal structures clear, and represent clinical knowledge developed iteratively through an empirical-rationale process and utilized iteratively in a rationale-empirical process of clinical reasoning for practice. So a mechanism for collaborating on development and dissemination makes sense.

DAGs make design, analysis and interpretation of trials less error prone (that is less biased); they provide a framework for systematic reviews and practice guidelines and they are very useful for teaching students about the reasoning they will do in practice (highlighting the knowledge they need to obtain to be clear thinkers when faced with clinical problems). While they are simply pictures, they are filled with meaning and their structure is built upon decades of research about logical inferences that can be drawn when the structure represents the truth of the causal relations depicted.

After months of investigating and considering the best approach, I have started a GitHub  ”organization” that is publicly available to keep a (hopefully) growing body of repositories with Physical Therapy related DAGs for physical therapy scholarship (broadly defined on the organization site).

For anyone interested it is located at:

I am currently working on developing tools to help equip people interested in contributing; who can then hopefully help me develop tools to equip other developers as well as potential users. If you are physical therapist or faculty member in a physical therapy (or related) department and you want to be involved - let me know:

If you know anyone that might be interested - please spread the word, share this post, share the link to the GitHub page, and/or share my email.