If you use credit to acquire assets, real estate for example, you won't move the needle on your total net worth. So, the overindebted in the US are those who used credit for expenses or for failed investments. They are poorer than those with zero debt.
The only exception that I can see, and which the article does not mention, is the use of credit for acquiring untangible assets, such as education. The correction would be to measure earning potential per individual, instead of net-assets, but that is basically impossible at large scale. We are stuck using net assets as a proxy for earning potential. Sum of wealth is an acceptable measure of income inequality.