The correlation between firearm deaths and the number of firearms owned by civilians is extraordinary. Also, in retrospect, the U.S. Federal Assault Weapons Ban of 1994-2004 not only prevented firearm sales escalation, but it also prevented firearm deaths, especially suicides, in the U.S.. This effect is evident in the entire population and in the older adolescent and young adult (AYA) population.
The firearm death rate, whether inflicted by others or self (suicide), is highly correlated with firearm prevalence among countries with available prevalence data. The firearm fatal homicide and suicide rates have less than a 1 in 1,000 chance of being due to chancer or to something other than firearm prevalence.
Whether analyzed by country, U.S. state, or calendar year, the firearm suicide rate is directly proportional to firearm prevalence. The firearm suicide rates have much less than a 1 in 10,000 chance of being due to something other than firearm prevalence.
For all ages, the evidence that the Ban reduced the firearm suicide rate while it was in effect and was followed by an increase after it expired is even more obvious.
That other methods of suicide were not related to the Ban is consistent with a causal association between the Ban and firearm suicides.
The Ban was also associated with a reduction in suicide deaths among adolescent and young adult (AYA) Americans and a subsequent increase after it expired. Since handguns and rifles are likely to be used for suicide more than assault weapons, the correlation may not seem plausible. There is evidence, however that handgun and rifle prevalence also declined during the Ban and abundant evidence that they increased after the Ban.
States that passed Red Flag Laws have prevented the escalation of firearm deaths that have occurred in the rest of the U.S. Connecticut was the first to do so, in 1999, followed by Indiana in 2014 and California in 2014. Together, the three have have had a statistically-significant decrease in their firearm rate; those without such laws have had a 21% increase (p << 0.05).