Economist sees failure, I see success that should be duplicated. The most important clue is that the mortality trend rate is not flat, it's a scatter. The convergence rate of 50 deaths per 100,000 is probably the best that can be done, given current treatments, and everyone deserves that kind of care.
It may be argued that screening is primarily used by people who can already afford to see the doctor but is not available to people who can't. That would explain why the 10 year mortality is scatter shot at lower screening rate counties and converges at about 50 in high screening rate, apparently affluent, counties
Absent some breakthrough in treatment, it will be difficult to improve survival rates when five year survival rates have been headed north of 90% over the last 20 years. http://seer.cancer.gov/statfacts/html/breast.html
We can turn that successful trend around if we don't make medicine available to everyone. Where an economist sees "over diagnosis", people in cancer centers I've worked for see a disturbing rise in later stage diagnosis because people can't make it to the doctor. Further cut backs in care will make things worse. Getting care to people who are not getting it may take care of those counties with four to six times the death rate of the best care counties.