Beginning with the beginning:
Last year, more active-duty soldiers committed suicide than died in battle. This fact has been reported so often that it has almost lost its jolting force. Almost.: This paragraph characterizes completely the reporting of said data. At no point is it systematically laid out. Nowhere are absolute numbers given. Time periods shift at random and the actual data that is reported seems to be randomly chosen. Anyone who thinks this is a sufficient amount of disclosure of new data from which they are going to draw conclusions is a fool, a narcisssist, or both. This incomplete and inadequate reporting is hugely problematic throughout the essay.
Worse, according to data not reported on until now , the military evidently responded to stress that afflicts soldiers in Iraq and Afghanistan primarily by drugging soldiers on the front lines. Data that I have obtained directly from Tricare Management Activity, the division of the Department of Defense that manages health care services for the military, shows that there has been a giant, 682 percent increase in the number of psychoactive drugs — antipsychotics, sedatives, stimulants and mood stabilizers — prescribed to our troops between 2005 and 2011. That’s right. A nearly 700 percent increase — despite a steady reduction in combat troop levels since 2008. 
: I have no idea how data which shows a massive increase in psychoactive medication prescriptions can be turned into data saying that those prescriptions were given to troops who were "on the front lines." I'm not saying that there are no troops who receive psychoactive medication "on the front lines," but the construction of this entire sentence gives the misleading impression that the entire increase in psychoactive prescriptions was accounted for by troops "on the front lines" with absolutely no attendant evidence.
While I think many people may find it intuitive that troops "on the front lines" not be taking psychoactive medication, a bit more thought will raise some interesting questions regarding that intuition. Should all troops with a diagnosis befitting treatment with psychoactive medications be removed from "the front lines"? Would we be better off with those same troops, with those diagnoses, not receiving medication? How would the answers to these two questions affect the military's ability to meet the objectives set forth by our civilian leadership?
: In this passage Friedman somehow, and without any subsequent explanation, asserts that a drop in combat troop levels ought to have lead to a decrease in psychoactive prescriptions, rather than the "giant" increase that we have seen. I'm serious, that's what he said, go read the passage again.
The first error here is that Friedman is using different temporal spans for the period during which psychoactive prescriptions grew, and the period during which combat troop levels have been decreasing. This is why his blithe reporting of this new data is such a problem. Setting aside the preposterous assumption that falling combat troop levels should lead to a decrease in prescriptions (more on this in a moment); how much of the rise in prescriptions came between 2005 and 2008 when combat troop levels were not being steadily reduced? Without knowing this it's impossible to assess whether his last sentence is wrong about the data, or simply wrongheaded by its assumptions. (While we're at it I would not characterize troop levels at any point during the Iraq or Afghanistan wars as "steadily" changing, in either direction).
Why should falling combat troop levels lead to a stable, or falling number of prescriptions for psychoactive medications? It may seem intuitive that fewer people experiencing combat equals fewer people to be diagnosed with psychiatric illness, but that's unfortunately not how it works. In the case of traumatic brain injuries and other physical injuries that may (or may not) have attendant psychiatric manifestations it may be true (though it's arguable). In the case of psychiatric illnesses such as depression, anxiety, and post-traumatic stress disorder the association is not as clear cut. Moreover, many of these conditions take years to manifest themselves, other times it may take years for sufferers to seek help for a psychiatric condition. This time delay is a well known and documented aspect of PTSD, particularly among people involved in violent conflicts. Friedman's blasé glossing of this fact is inexplicable to me.
The much shorter version of the preceding three paragraphs: It does not at all logically follow, as Friedman says, that falling combat troop levels should not lead to an increase in psychoactive prescriptions. (Moreover what constitutes "combat troops" in irregular wars?)