Is there a graph of the decline in overdoses with time?

Your explanation suggests an exponential decay (ignoring aggravating conditions, like seasonal temperature, violence, ...)