I'm not sure why the CDC instead of the FBI or something similar was used. Regardless, that's what they used.
Someone was not shot every single time a firearm was used in self-defense. Very often, people are held at gunpoint while 911 is called. In addition, when a home defense shotgun is loaded up with less-lethal munitions (rubber slugs, rock salt, birdshot, etc), the likelihood of a fatality is very low, but it still counts as use of a firearm in self-defense.
Shooting someone who is trying to grab a purse and run is generally not defensible, morally or legally. In the case of an armed mugger, a rape, or a home invasion, though, there's a very real chance that you'll be killed even if you cooperate, so using lethal force in self-defense is generally legally defensible under such circumstances and in my opinion is morally defensible as well.
That is the media's fault for distorting things so badly with sensationalist news coverage. Mass shootings are a tiny, tiny statistic that gets massive, massive attention. It's like airplane crash fatalities versus automotive fatalities. The media spends most of their time talking about the former, but the vast majority of travel fatalities are the latter. Likewise, mass shootings are statistically insignificant when compared to day-in-day-out gun crime connected with drugs, gangs, domestic violence, and armed robbery.
Slate (a liberal, anti-firearm publication) published a suofftopicry of the findings of the study: http://www.slate.com/articles/health_and_science/human_nature/2013/06/handguns_suicides_mass_shootings_deaths_and_self_defense_findings_from_a.html
Guns.com (whose editorial bias should be obvious) published a suofftopicry as well:
http://www.guns.com/2013/06/27/cdc-releases-study-on-gun-violence-with-shocking-results/
Read both, and where the two articles overlap, you'll get a good sense of what the study said. Or you can pay $39 and get the entire study to read yourself: http://www.nap.edu/openbook.php?record_id=18319&page=R1