I agree that these are separate disciplines, but I think that one reason we're not seeing anyone who does security usability work well right now is that that the best practitioners in each field tend to be silo'd by their specialization.

A single assessment is not necessarily a single metric, and an assessment comprised of audits in each domain seems like a good first step toward building understanding of a common goal and measuring progress toward it. Putting these audits together will hopefully start to expose not only the tradeoffs, but also the synergies at play in designing secure systems.

It seems to me that at the root of the security usability problem is a failure in collaboration between developers with solid software engineering practices and designers with solid UX design practices. Talent on both sides is in high demand, but there are few organizations that are able to get both working together effectively on these hard problems.