20 January 2011

Breaking down the Army's suicide data

Part I: What does the data say?

There's a lot of grist to Wednesday's news briefing with Gen. Chiarelli on the 2010 Army suicide statistics. As usual, everything I write is my own opinion, unvarnished.

Looking past the headlines telling you some suicides were up and some where down I want to point out an underemphasized point that Gen. Chiarelli made during his briefing:
So the numbers [...] have really only focused on this group, both the Army Reserve and the Army National Guard, to collect this data for about five years.

Here's my calculation of the suicide breakdown from the news briefing and what papers are reporting (I'd guess the numbers are correct +/- 2, but Gen. Stultz's numbers were "inaudible" in the transcript so I can't verify). These include both confirmed and suspected suicides. There were 343 overall Army surveillance suicides. Of those, 157 were active duty soldiers (17 of whom were US Army Reserve [USAR] or Army National Guard [ARNG]), 101 non-AD ARNG, 40 non-AD USAR.

First let's talk about the ARNG/USAR suicides. One important note that ties back to Gen. Chiarelli's quote above: there was no focus on data collection for the ARNG/USAR until 2005. Here's the data series for non-AD ARNG/USAR suicides going back to calendar year 2003: CY03: 16, CY04: 15, CY05: 29, CY06: 33, CY07: 61, CY08: 43, CY09: ~651, CY10: 141. This leaves us with nearly a ten-fold increase in non-AD suicides over the past seven years.

It's impossible to separate how much of the increase is merely an artifact of the increased surveillence,2 but some part of it is. Supporting this assertion is the data for active duty suicides which has trended upwards at a much steadier rate since 2004 (04 was slightly lower than 03). In any case, we cannot attribute all of this increase to the surveillence (indeed, nothing substantive has changed between 2009 and 2010 in this respect) so what can we figure out?

To begin with, there is a substantial delay between returning from a deployment and the manifestation of psychiatric distress (PTSD, MDD, others)3. It's important to keep this in mind when judguing the success (or lack thereof) in the Army's ongoing efforts to treat these symptoms. Even with the draw down in Iraq increasing dwell times and lowering the operations tempo we cannot expect to see immediate reductions in suicide or associated conditions.

Also salient to this issue is a point Gen. Stultz brought up, "Less than half of [the soldiers who commited suicide] had deployed. It’s just what Ray said; it’s something else going on in their lives." In fact, looking at past years this isn't very strange. Among active duty suicides there has been a relatively consistent distribution between suicides that occur during deployment, post deployment, and in those with no deployment history. This leads me to a slightly different conclusion than Gen. Stultz and many media outlets: Increased suicidality may be a consequence not of any aspect of non-deployed military life, but instead an association with the demographic groups the Army draws from. An insightful column touching on this topic was written by a Marine who served four tours in Iraq. I'd urge you to read the whole thing, but it is worth quoting him at some length here:
Prior to signing up, most of my friends asked themselves how they could pay for college growing up in the poorest class. What if you are not a great student or a superb athlete? You probably won't get that education through McDonald's and you definitely won't get it from the school or your minimum wages of your dual working parents. As we all know, it is almost impossible to get a job now without a good-looking diploma from a decently named school. And how do you get healthcare without a decent paying job?
[...]
More than three quarters of the men I served with didn't have any choices if they stuck around their hometowns. I am trying to make you remember what life was like at 17 or 18 and you didn't think there was a way out of the situation you were born into. Here is what I saw and was told: Some young men fled gang life in poor areas like Chicago, Redlands, Compton, San Bernardino, Watts or Portland (crimes they committed or crimes to be committed on them); some wanted US citizenship after having arrived from Latin America, Europe, or Africa; some fled religious and sexual persecution (yes there are gay marines and some are from Texas); some got off the isolated encampments known as Reservations; I had Marines escaping child abuse; some guys hated the farm life; and the mediocre athletes knew they didn't have the NFL talent now required to play at even the lowest junior university.
[...]
The Marines I know didn't have the luxury of thinking hard about other choices like Pat Tillman. In retrospect, most said they had no other choices.

We must acknowledge that some percentage of military members feel that they joined the military, at least in part, due to a lack of alternatives. More than merely belie the notion that our soldiers, sailors, airmen, and marines are all uncoerced volunteers this does a great deal to characterize the socioeconomic circumstances affecting large swaths of this country. Indeed, the USAR and ARNG advertising campaigns explicitly cite the money, training, and other benefits attendant to joining up.

The reason I bring this up is that there are many papers examining the relationship of suicide (and suicidality) to socioeconomic circumstance. Data from the Department of Defense's Suicide Event Report found that the demographic traits most associated with suicide were (in suicides per 100k individuals):
High school diploma/GED or less (29.1), AA degree or tech cert (25.1), HS grad (16.4), 4 year degree (9.8)
Being divorced (27.6), Married (15.9), Never Married (15.2)
Rank E1-E4 (20.1), Rank E5-E9 (14.8), Officer (10.2)
Age < 25 (20.1), 40+ (15.7), 26-29 (13.7), 30-39 (12.1) White (17.4), Black (11.9)
Regular (16.9), N. Guard (14.4)4
Putting all that together it becomes evident that the same risk factors one finds in society at large are mirrored in the increased suicide rates of ARNG/USAR members. It may be that when Gen. Carpenter said, "We are really [...] the canary in the mine shaft, especially for us in the National Guard because we recruit in local communities and we are a reflection of those communities" he was understating the case. This may seem like an easy conclusion to draw (and one I wasted a lot of words on), but it has been completely overlooked in nearly every discussion of these numbers I have read beyond the Gen. Carpenter quote above.

Leaving the discussion of ARNG/USAR soldiers without deployment who have committed suicide, I want to move on to those who had deployed. In a brief post yesterday I asserted, "Combat causes PTSD," as it pertained to my point in that post it was true enough. As usual it's actually much more complicated, but I think it is useful to look seperately at those who commit suicide (or suffer PTSD, MDD, etc) after/during a deployment ex their never-deployed peers. This is a rudimentary way of examining whether there is something peculiar to the experience of deploying to war, and its attendant stresses, that changes occurrence rates, or if Army suicides merely occur at a similar rate to the general population because that's where they draw their members from.

Using Gen. Stultz's figure5 of appoximately half ARNG/USAR suicides being among those who had deployed we get approximately 70 suicide victims among Gen. Chiarelli's 156 to 181,000 ARNG/USAR soldiers mobilized each year. Choosing the average (168.5k) leaves us with a suicide rate of 41 per 100,000 people. Doing a similar calculation for the active duty Army (and using Gen Chiarelli's estimation of 2/3 of soldiers who commit suicide being deployed or having a history thereof) gives us 105 suicides out of 569,000 for a rate of 18 per 100,000 people. Clearly something is going on here.

Unfortunately, I don't know what it is. I'm just a blogger though, so I can offer some conjecture and point to some questions that need to be asked in figuring out why their rate is nearly twice that of their regular Army peers. Demographically nearly 13% more regular Army soldiers are married than either their ARNG or USAR peers. I can't find data that shows the number who are divorced, but there is a wealth of data showing that young single males are the most at risk group for suicide. Considering the regular Army's marriage "advantage" over ARNG/USAR is most pronounced among the enlisted ranks it stands to reason this has some effect. In a similar vein, unemployment has been associated with increased suicide rates, a stressor that afflicts ARNG/USAR soldiers but not regular Army soldiers; Gen. Carpenter brought this up int he context of the 81st Brigade, "[W]hen they came back, they had an unemployment rate of around 33 percent." There may yet be other demographic trends that could yield more insight.

Clearly there's no single explanation for the sky high suicide rate among ARNG/USAR soldiers who have deployed, it does seem reasonable that the various risk factors exist among them in a synergistic fashion though. They may not have regular contact with peers who share similar experiences, they may feel alienated as they reintegrate, the acceptability of help for psychological trauma preached in the regular Army is largely absent from public discourse, and help may not be as readily available to them as it is to those in the regular Army. We need to begin asking quesitons about ARNG/USAR soldiers' access to health care, support groups, and so on.

I realize this post may be unsatisfying as it raises more quesitons than it answers. However, I think it's very important to have a clear understanding of what this report (and past ones) do and do not say in advancing treatment of suicidal soldiers.

Check back in a few days for part II, wherein I will talk about some of the Army's recent inititives to not only help soldiers in acute mental distress, but avoid putting them there in the first place. In that post I will also talk about the suicide rate's plateau in 2010 (describing a -3% change as a decrease is overly generous), its history, and what we might expect in the future.

1 For the life of me I can't find Army data for CY09, but everyone is reporting that CY10's 141 was nearly double, so I'll ball-park it at 65.

2 It would be incredibly useful to compare ARNG/USAR suicide data to that of the general population (especially among those who had never deployed). Unfortunately it can't be done because of the very long lag time in population wide surveillance data. The CDC's most recent suicide data is from 2007.

3 Throughout this post I'm going discuss suicide specifically but also use indicators such as suicidality and psychiatric conditions (e.g. PTSD and MDD) when talking about the mental health of troops. As I have mentioned in other posts suicide itself is very rare (in the general population there are approximately 19 per 100,000 people annually) and by its vary nature makes retrospective causal assessment very difficult. There are very strong associations between the metrics I will use, but if you think I've erred please let me know in the comments or email me.

4 I suspect that this data is different because of its age (2008) as well as it being DoD wide.

5 These figures are estimates, and pretty crude ones at that.

No comments:

Post a Comment