The mortality numbers in this paper appear to be extrapolated from an observational study of 9,000 cases from 1988-1994. That paper distinguished between "privately insured" and "not privately insured". This paper distinguishes "adequately insured", "underinsured", and "uninsured", and applies the original paper's hazard numbers to everybody but the "adequately insured", with a coefficient modifier for the "underinsured" that they came up with themselves.