With most energy generation schemes, including nuclear, many costs and risks are externalized to a third party -- namely the taxpayer.

I'm not familiar enough with the methodology to dispute it's finding regarding deaths, but I'd argue that evaluating risks solely on the number of attributable deaths is not a full accounting of risk.

What is the economic and social impact of the 20 mile exclusion zone around Fukushima? You're looking at a handful of deaths, but many lives have been affected nonetheless. If my rooftop solar panel kills my family somehow, that is a personal tragedy, but the scope in terms of people affected is nothing like a rare, but catastrophic nuclear failure.