This appears to be really shoddy research and once again illustrates why a macro economic approach to this problem simply doesn't work and why public have stopped listening.
The researchers are trying to use a system level risk returns balance and using very high level average metrics like risk of reduction in lifespan. How about the distribution around the mean and how about the suffering of people who undergo cancer treatments and survive? Having seen a close relative battle cancer and the family suffering, i wouldn't wish it on my worst enemy and their families.
The other piece these macro economic approaches miss is that while a system level cost isn't much, for most people their property, their job and their neighbourhood social networks ARE the major assets they have in life and in the case of nuclear accidents, they lose everything. This fundamentally Behavioral Economics view of consequences is what drives the recent popular opposition to nuclear power and unless nuclear proponents start analyzing nuclear power through a Behavioral Economics lens, they're likely going to be ignored by the population. Comparing risk of death against statistical correlations with air pollution isn't going to cut it.
On nuclear, we're barking up the wrong tree by pushing for economic analysis of old style nuclear. We need to develop and qualify designs that are passively safe with no risk of driving people away from their property and society before the industry will be taken seriously, otherwise the only interest will be in weapons research.