I don't have time to do more detailed analysis but, from the article, "In 2052, ...Rose would be approaching her 60s." So lets assume that Rose would be exactly 60 in 2052. So she was born in 1992.
According to projections (http://www.census.gov/compendia/statab/2011/tables/11s0103.p...), a black woman who was 15 in 2007 would be expected to live to 63. So there is a significant risk that Rose would not live to the end of her payout period. But if the expected payout is >>18% of the present value of her payments (which it almost certainly is) then she got a bad deal.
Someone with time could use the annual mortality probabilities (http://www.ssa.gov/oact/STATS/table4c6.html) to calculate the expected payout. I'm bad at analysis so I'd probably use a Monte Carlo simulation.