{"id":85568,"date":"2019-06-08T08:00:50","date_gmt":"2019-06-08T12:00:50","guid":{"rendered":"https:\/\/valorguardians.com\/blog\/?p=85568"},"modified":"2019-06-09T10:03:18","modified_gmt":"2019-06-09T14:03:18","slug":"veterans-suicide-rates-and-the-demographics-nobody-discusses","status":"publish","type":"post","link":"https:\/\/www.azuse.cloud\/?p=85568","title":{"rendered":"Veterans, Suicide Rates, and the Demographics Nobody Discusses"},"content":{"rendered":"

Back in March, Dave Hardin wrote an article about veteran suicides and studies relating to the subject<\/em><\/a>.\u00a0 I previously\u00a0wrote a follow-up article here<\/a><\/em>, where I took exception to NPR’s article and provided data refuting their implied conclusions.<\/p>\n

But while I was doing that, something else occurred to me:\u00a0 the demographics of the US veteran and nonveteran populations are probably quite different.\u00a0 So I checked.<\/p>\n

Turns out they are.\u00a0 Because of World War II, Korea, and Vietnam, there are proportionally far more veterans over 55 than is the case in the general population.\u00a0 And a far smaller proportion of females are veterans of the US armed forces than the general population (which is somewhat over 52% female).<\/p>\n

It also occurred to me that different demographic subgroups might well have very different rates of suicide.\u00a0 Indeed they do; for starters, the suicide rate varies substantially with age and gender (they also vary by ethnicity).\u00a0 But if you can find the proper demographic data \u2013 e.g., a breakout by age and gender of both the veteran population and the US general population\u2019s suicide rates \u2013 you can account for those demographic differences and see what \u201cright should<\/em> look like\u201d with respect to veterans.<\/p>\n

Specifically, if you have that data you can determine if the observed US veteran suicide rate is higher than that of nonveterans after accounting for the demographic differences<\/span>.\u00a0 And you can also see how much higher (or lower) it is, if it\u2019s in fact different.\u00a0 All that requires is a straightforward but somewhat tedious weighted sum problem.<\/p>\n

You can also see if different veteran age\/gender subgroups are doing better or worse than their nonveteran contemporaries with respect to suicide.\u00a0 Knowing that is important, too \u2013 and could be even more valuable than knowing the overall \u201cscore\u201d when it comes to allocating VA resources.<\/p>\n

So I decided to see if I could find the data necessary to do exactly that.<\/p>\n

Turns out the data was readily available for the years 2005-2016.\u00a0 The VA published that data in 2018 here<\/em><\/a> \u2013 broken down by age group and gender (it was not, unfortunately, further broken down by ethnicity).\u00a0 The data below is taken from the study’s Appendix on the linked page, which is both downloadable and in Microsoft Excel format.<\/p>\n

I\u2019d hoped to do a follow-up article well before now.\u00a0 Unfortunately, events in the real world made that impossible.\u00a0 But better late than never, I guess.<\/p>\n

What follows is a bit long-ish.\u00a0 And yes, there’s some math involved.\u00a0 (smile)<\/p>\n

Veteran Demographics<\/u><\/p>\n

In a comment to DH\u2019s original article, I indicated that the US veteran population is overwhelmingly male.\u00a0 In that comment, I guessed 80+%.<\/p>\n

My guess was low.\u00a0 The actual figure in 2016 was a bit less than 91% male.\u00a0 And since males have a markedly higher suicide rate than do women (it varies by age group, but is typically at least 2.5x or more higher) , that demographic difference between veterans as a group and the US population as a whole has to be taken into account to do any kind of fair comparison.<\/p>\n

From the VA study linked above, here is the 2016 US veteran breakout by age and gender.\u00a0 I\u2019ve also calculated some percentages that will be useful in determining what \u201cright should<\/i> look like\u201d later.<\/p>\n

 <\/p>\n\n\n\n\n\n\n\n\n
US Military Veterans, 2016<\/strong><\/td>\n<\/tr>\n
Age<\/strong><\/td>\n18-34<\/strong><\/td>\n35-54<\/strong><\/td>\n55-74<\/strong><\/td>\n75+<\/strong><\/td>\nTotal<\/strong><\/td>\n<\/tr>\n
Male<\/td>\n1,609,000<\/td>\n4,228,000<\/td>\n8,161,000<\/td>\n4,356,000<\/td>\n18,355,000<\/td>\n<\/tr>\n
\u00a0\u00a0\u00a0\u00a0 Percentage<\/td>\n7.96%<\/td>\n20.93%<\/td>\n40.40%<\/td>\n21.56%<\/td>\n90.86%<\/td>\n<\/tr>\n
Female<\/td>\n377,000<\/td>\n753,000<\/td>\n573,000<\/td>\n143,000<\/td>\n1,846,000<\/td>\n<\/tr>\n
\u00a0\u00a0\u00a0\u00a0 Percentage<\/td>\n1.87%<\/td>\n3.73%<\/td>\n2.84%<\/td>\n0.71%<\/td>\n8.90%<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

Obviously, since almost 91% of veterans in\u00a0 2016 were male, almost 91% of the overall veteran suicide rate for that year will be due to male veterans.\u00a0 Since somewhat more than 55% of the US nonveteran population over the age of 18 is female \u2013 and since males have a much higher suicide rate than do females \u2013 we should expect the aggregate veteran suicide rate to be different from that for the overall and nonveteran US populations on that basis alone.<\/p>\n

Nonveteran Demographics<\/u><\/p>\n

Along with veteran suicide data, the same source also included suicide data for the US nonveteran population.\u00a0 Like the veteran data, this data was broken out by both gender and age groups. \u00a0They appear below.<\/p>\n

 <\/p>\n\n\n\n\n\n\n\n\n
US Nonveteran Suicide Demographics for 2016<\/strong><\/td>\n<\/tr>\n
Age<\/strong><\/td>\n18-34<\/strong><\/td>\n35-54<\/strong><\/td>\n55-74<\/strong><\/td>\n75+<\/strong><\/td>\nTotal<\/strong><\/td>\n<\/tr>\n
Males<\/strong><\/td>\n36,056,080<\/td>\n36,874,513<\/td>\n25,233,311<\/td>\n3,995,165<\/td>\n102,158,069<\/td>\n<\/tr>\n
\u00a0\u00a0\u00a0\u00a0 Number of Suicides<\/strong><\/td>\n8,853<\/td>\n9,965<\/td>\n7,028<\/td>\n2,021<\/td>\n27,862<\/td>\n<\/tr>\n
Female<\/strong><\/td>\n36,602,964<\/td>\n41,310,522<\/td>\n36,166,904<\/td>\n12,039,192<\/td>\n126,119,582<\/td>\n<\/tr>\n
\u00a0\u00a0\u00a0\u00a0 Number of Suicides<\/strong><\/td>\n2,251<\/td>\n3,854<\/td>\n2,875<\/td>\n506<\/td>\n9,486<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

 <\/p>\n

Looking at the two tables immediately tells us three things.\u00a0 First:\u00a0 being\u00a0 nearly 91% male, and given that males have a suicide rate far higher than females, veterans collectively would be expected<\/u> to have a much higher overall suicide rate than either the nonveteran population (which is 55.2+% female) or the US general population (which is 51.5+% female).\u00a0 Second:\u00a0 since age matters greatly vis-\u00e0-vis suicide rate, the fact that nearly 2\/3 of the US veteran population is over the age of 55 \u2013 with more than 1\/5 of the total being over 75 \u2013 will also have an impact.\u00a0 And, finally:\u00a0 since the age distribution of both male and female veterans vary greatly from that of the nonveteran population, direct comparisons of suicide rates by gender will also be rather pointless.<\/p>\n

These three factors also make clear one other fact:\u00a0 any simple comparison of veteran and non-veteran suicide rates is NOT going to be particularly meaningful; rather, such a comparison will instead be misleading.\u00a0 The demographics of the two groups are simply too different, age- and gender-wise.<\/p>\n

Estimating Reality<\/u><\/p>\n

Well, so how do we sort this out?\u00a0 We have suicide rates by age category and gender for both veterans and nonveterans \u00a0\u2013 and we can also figure out the percentage of each veteran age\/gender subgroup in either the total veteran population or of that gender of veterans.\u00a0 So a weighted sum of the nonveteran age\/gender subgroup suicide rates, using as weighting factors each age\/gender subgroup\u2019s total percentage in the veteran<\/em> population, will tell us what we should expect as an overall veteran suicide rate if it is the same as that of the nonveteran population<\/u>.\u00a0 That in turn will tell us if a problem actually exists or not \u2013 a higher than expected aggregate rate for veterans would indicate a \u201cproblem\u201d, while a lower than expected rate would indicate absence of an issue. (As noted above, due to demographic differences knowing this won’t really be of much utility, but hey – the media loves single number metrics.)\u00a0 The magnitude of the difference will also tell us how big the \u201cproblem\u201d is, if it exists.<\/p>\n

Doing this will also allow something far more useful: direct comparison of the suicide rates of veteran and non-veteran populations by age and gender groups.\u00a0 So if there\u2019s a problem with a particular veteran age\/gender subgroup, it will disclose that fact \u2013 and identify where the VA should be concentrating its resources devoted to suicide prevention.<\/p>\n

About \u201cAge Adjusted\u201d Rates<\/u><\/p>\n

In healthcare literature \u2013 and in the linked VA report concerning Veteran Suicide \u2013 one often sees reference to \u201cAge Adjusted\u201d rates.\u00a0 So let\u2019s discuss them briefly here.<\/p>\n

BLUF:\u00a0 they\u2019re totally irrelevant when trying to determine if Veterans, as a group and by age\/gender subgroup, are doing \u201cbetter\u201d or \u201cworse\u201d than their nonveteran counterparts.\u00a0 Yet you often see them calculated and used (and IMO, grossly abused).<\/p>\n

Short version:\u00a0 age-adjusted statistics are a method that allows observed data from a population with nonstandard demographics to be mapped onto a standard population distribution.\u00a0 That allows data from a group that may be demographically nonstandard to be used to draw conclusions and\/or make planning decisions for healthcare services to be applied to the general population.\u00a0 This link<\/em><\/a> provides a description of the process and its uses.<\/p>\n

Going the other way?\u00a0 Um, not so much.\u00a0 If you\u2019re providing services to a population for which you have known<\/span> demographics, those known demographics should be used as the basis for planning and allocation.\u00a0 Age adjusting those known demographics to some standard population distribution simply doesn\u2019t do much of value.\u00a0 You have the known demographics for the population in question, and can project those into the future using actuarial and other techniques; you can do that instead.\u00a0 No \u201cage adjustments\u201d are required.<\/p>\n

Therefore, using or even discussing \u201cAge Adjusted\u201d statistics in the context of veteran suicide \u2013 and especially using them while making decisions about how to allocate VA resources devoted to suicide prevention \u2013 is, to be charitable, IMO not<\/em> particularly useful.\u00a0 IMO the terms \u201cutterly inappropriate\u201d or \u201cludicrous\u201d would be more accurate.<\/p>\n

We know the actual demographic distribution of veterans \u2013 and we know it\u2019s heavily skewed towards both males (just under 91% as of 2016) and individuals 55 and older (nearly 2\/3 of all veterans as of 2016).\u00a0 This demographic difference is critical in deciding how to allocate the VA\u2019s resources effectively in the short term \u2013 because the VA serves the veteran population, not the nonveteran or general populations.\u00a0 And the US veteran population is also demographically very different from the distribution specified for an \u201cage normalized\u201d distribution.<\/p>\n

Using an age normalized distribution for veteran suicide rates to allocate VA healthcare resources related to suicide prevention would therefore be like using Gas Buddy\u2019s national average gas price to plan how much money you need for fuel to take a 1,200 mile road trip in California<\/i>.\u00a0 In both cases, you\u2019re in effect using a type of \u201cnormalized\u201d data to make plans relating to a situation that is known to be vastly different from \u201cnormal\u201d; the normalized data is thus irrelevant.\u00a0 In both cases, you should be using the known, relevant data instead.<\/p>\n

A better strategy would be to do the following:<\/p>\n

    \n
  1. separate out the US veteran and nonveteran populations;<\/li>\n
  2. partition these two populations by gender;<\/li>\n
  3. partition each gender into same-age subgroups;<\/li>\n
  4. calculate the suicide rates for each resulting gender\/age-group suicide rates for both veterans and non-veterans, as well as the gender and overall rates (which will be of marginal real utility, but the media seems to love them anyway); and<\/li>\n
  5. then<\/u> compare corresponding pairs of these individual age\/gender subgroup suicide rates.<\/li>\n<\/ol>\n

    Doing this will allow a direct \u201capples-to-apples\u201d comparison, by age group, of the veteran and non-veteran populations regarding suicide rates.\u00a0 That comparison will show which groups of veterans \u2013 if any \u2013 are doing \u201cworse\u201d than their civilian counterparts. \u00a0(The group doing \u201cworse\u201d with respect to suicide rate will be the one in each corresponding pair that has the higher suicide rate.)\u00a0\u00a0 The resulting comparisons will also indicate where the VA should concentrate its resources regarding suicide prevention.<\/p>\n

    We can even go a step further.\u00a0 Given known (as of 2016) veteran demographics, we can calculate what the male and female veteran suicide rates would have been if each veteran age\/gender subgroup had exactly the same suicide rate as the corresponding nonveteran group<\/u>.\u00a0 Comparing that calculated figure with the observed rates for male and female veterans then indicate if male and females veterans, as a group, are doing \u201cbetter\u201d or \u201cworse\u201d with respect to suicide than their nonveteran contemporaries.\u00a0 (Hey, the media loves that kind of stuff \u2013 even if it\u2019s of little real utility.\u00a0 It\u2019s easy to do, and I\u2019ll have the data.\u00a0 So why not?)<\/p>\n

    Below, I\u2019ve done exactly that.<\/p>\n

    Data, Calculations, and Rates<\/u><\/p>\n

    From the Appendix at the VA study page linked above, here is the pertinent data for US veterans and nonveterans, by gender and age, for 2016.\u00a0 I\u2019ve also calculated some percentages that will be used as weighting factors for later calculations.<\/p>\n\n\n\n\n\n\n\n\n\n
    \u00a0Male US Veterans, 2016<\/strong><\/td>\n<\/tr>\n
    Age Range<\/strong><\/td>\n# Suicides<\/strong><\/td>\nPopulation<\/strong><\/td>\nSuicides\/ 100k<\/strong><\/td>\n% Male Veterans<\/strong><\/td>\n% All Veterans<\/strong><\/td>\n<\/tr>\n
    18-34<\/td>\n840<\/td>\n1,609,000<\/td>\n52.2<\/td>\n8.766%<\/td>\n7.965%<\/td>\n<\/tr>\n
    35-54<\/td>\n1,527<\/td>\n4,228,000<\/td>\n36.1<\/td>\n23.035%<\/td>\n20.930%<\/td>\n<\/tr>\n
    55-74<\/td>\n2,180<\/td>\n8,161,000<\/td>\n26.7<\/td>\n44.462%<\/td>\n40.399%<\/td>\n<\/tr>\n
    75+<\/td>\n1,270<\/td>\n4,356,000<\/td>\n29.2<\/td>\n23.732%<\/td>\n21.563%<\/td>\n<\/tr>\n
    Total<\/td>\n5,822<\/td>\n18,355,000<\/td>\n31.7<\/td>\n100.000%<\/td>\n90.862%<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

     <\/p>\n\n\n\n\n\n\n\n\n\n
    \u00a0Male US Nonveterans, 2016<\/strong><\/td>\n<\/tr>\n
    Age Range<\/strong><\/td>\n# Suicides<\/strong><\/td>\nPopulation<\/strong><\/td>\nSuicides\/ 100k<\/strong><\/td>\n% Male Nonveterans<\/strong><\/td>\n% All Nonveterans<\/strong><\/td>\n<\/tr>\n
    18-34<\/td>\n8,853<\/td>\n36,056,080<\/td>\n24.6<\/td>\n35.294%<\/td>\n15.795%<\/td>\n<\/tr>\n
    35-54<\/td>\n9,965<\/td>\n36,874,513<\/td>\n27.0<\/td>\n36.096%<\/td>\n16.153%<\/td>\n<\/tr>\n
    55-74<\/td>\n7,028<\/td>\n25,233,311<\/td>\n27.9<\/td>\n24.700%<\/td>\n11.054%<\/td>\n<\/tr>\n
    75+<\/td>\n2,021<\/td>\n3,995,165<\/td>\n50.6<\/td>\n3.911%<\/td>\n1.750%<\/td>\n<\/tr>\n
    Total<\/td>\n27,862<\/td>\n102,158,069<\/td>\n27.3<\/td>\n100.000%<\/td>\n44.752%<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

     <\/p>\n\n\n\n\n\n\n\n\n\n
    \u00a0Female US Veterans, 2016<\/strong><\/td>\n<\/tr>\n
    Age Range<\/strong><\/td>\n# Suicides<\/strong><\/td>\nPopulation<\/strong><\/td>\nSuicides\/ 100k<\/strong><\/td>\n% Female Veterans<\/strong><\/td>\n% All Veterans<\/strong><\/td>\n<\/tr>\n
    18-34<\/td>\n53<\/td>\n377,000<\/td>\n14.1<\/td>\n20.423%<\/td>\n1.866%<\/td>\n<\/tr>\n
    35-54<\/td>\n121<\/td>\n753,000<\/td>\n16.1<\/td>\n40.791%<\/td>\n3.728%<\/td>\n<\/tr>\n
    55-74<\/td>\n83<\/td>\n573,000<\/td>\n14.5<\/td>\n31.040%<\/td>\n2.836%<\/td>\n<\/tr>\n
    75+<\/td>\n4<\/td>\n143,000<\/td>\n2.8<\/td>\n7.746%<\/td>\n0.708%<\/td>\n<\/tr>\n
    Total<\/td>\n257<\/td>\n1,846,000<\/td>\n13.9<\/td>\n100.00%<\/td>\n9.138%<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

     <\/p>\n\n\n\n\n\n\n\n\n\n
    \u00a0Female US Nonveterans, 2016<\/strong><\/td>\n<\/tr>\n
    Age Range<\/strong><\/td>\n# Suicides<\/strong><\/td>\nPopulation<\/strong><\/td>\nSuicides\/ 100k<\/strong><\/td>\n% Female Nonveterans<\/strong><\/td>\n% All Nonveterans<\/strong><\/td>\n<\/tr>\n
    18-34<\/td>\n2,251<\/td>\n36,602,964<\/td>\n6.1<\/td>\n29.022%<\/td>\n16.034%<\/td>\n<\/tr>\n
    35-54<\/td>\n3,854<\/td>\n41,310,522<\/td>\n9.3<\/td>\n32.755%<\/td>\n18.097%<\/td>\n<\/tr>\n
    55-74<\/td>\n2,875<\/td>\n36,166,904<\/td>\n7.9<\/td>\n28.677%<\/td>\n15.843%<\/td>\n<\/tr>\n
    75+<\/td>\n506<\/td>\n12,039,192<\/td>\n4.2<\/td>\n9.546%<\/td>\n5.274%<\/td>\n<\/tr>\n
    Total<\/td>\n9,486<\/td>\n126,119,582<\/td>\n7.5<\/td>\n100.000%<\/td>\n55.248%<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

     <\/p>\n

    Examining the above tables, we can draw the following conclusions:<\/p>\n

      \n
    1. Male veterans <55 years old have substantially(35-54) to markedly (18-34) higher suicide rates than their nonveteran contemporaries.<\/li>\n
    2. Male veterans >55 years old have slightly (55-74) to markedly (75+) lower suicide rates than their nonveteran contemporaries.<\/li>\n
    3. For male veterans, the suicide rate peaks in the youngest age group, with a sharp drop in the next age group and a generally declining trend afterwards (there is a slight uptick for the 75+ age group, but the overall trendline appears to be declining).<\/li>\n
    4. The opposite pattern is seen in male nonveterans, with the youngest age group having the lowest suicide rate and the rate rising with each succeeding age group, ending with a sharp increase in suicide rate for those nonveterans 75 and older.<\/li>\n
    5. Female veterans under the age of 75 have markedly higher suicide rates than the corresponding nonveteran group.<\/li>\n
    6. Females 75 and older have a markedly lower suicide rate than the corresponding nonveteran group.<\/li>\n<\/ol>\n

      Due to gross demographic differences between US veteran and nonveteran populations, it is meaningless to compare directly the overall male and female suicide rates for veterans and nonveterans.\u00a0 However, it is possible to determine whether male veterans and female veterans collectively have a higher or lower suicide rate than they would if their suicide rates by age group were the same as their nonveteran contemporaries.\u00a0 Doing so is nothing but a simple weighted-sum problem, where it is assumed that veterans have the same suicide rates as their nonveteran contemporaries but veteran demographics are used to provide the weighting factors.\u00a0 Those calculations are presented next.<\/p>\n

      What \u201cRight Should<\/i> Look Like\u201d<\/u><\/p>\n

      In the previous section, we have both veteran and nonveteran suicide rates by age group and gender.\u00a0 We also have the percentages that each veteran age\/gender group constitutes for both their gender and of veterans overall.\u00a0 That means we have enough information to calculate what \u201cright should<\/em> look like\u201d \u2013 e.g., we can calculate what the male, female, and overall veteran suicide rates would be if veterans had the same suicide rate as their nonveteran contemporaries<\/u>.<\/p>\n

      It\u2019s a simple if perhaps tedious weighted-sum problem.\u00a0 The individual veteran subgroup percentage is used as a weighting factor; the general population suicide rate for that age\/gender subgroup is the multiplied by that weighting factor.\u00a0 Summing these produces a weighted sum that tells us what \u201cright should<\/em> look like\u201d for male and female veterans – that it, what the suicide rate for male and female veterans would be if the two populations had exactly the same suicide rates as did nonveterans.<\/p>\n

       <\/p>\n\n\n\n\n\n\n\n\n\n\n
      Expected Male and Female Veteran Suicide Rates, Corrected for Known Veteran Demographics<\/strong><\/td>\n<\/tr>\n
      Male Veterans<\/strong><\/td>\nFemale Veterans<\/strong><\/td>\n<\/tr>\n
      Age Group<\/strong><\/td>\n% Male Veterans<\/strong><\/td>\nNonveteran Suicide Rate<\/strong><\/td>\nExpected Subgroup Contribution<\/strong><\/td>\nAge Group<\/strong><\/td>\n% Female Veterans<\/strong><\/td>\nNonveteran Suicide Rate<\/strong><\/td>\nExpected Subgroup Contribution<\/strong><\/td>\n<\/tr>\n
      18-34<\/td>\n8.766%<\/td>\n24.6<\/td>\n2.16<\/td>\n18-34<\/td>\n20.423%<\/td>\n6.1<\/td>\n1.25<\/td>\n<\/tr>\n
      35-54<\/td>\n23.035%<\/td>\n27.0<\/td>\n6.22<\/td>\n35-54<\/td>\n40.791%<\/td>\n9.3<\/td>\n3.79<\/td>\n<\/tr>\n
      55-74<\/td>\n44.462%<\/td>\n27.9<\/td>\n12.41<\/td>\n55-74<\/td>\n31.040%<\/td>\n7.9<\/td>\n2.45<\/td>\n<\/tr>\n
      75+<\/td>\n23.732%<\/td>\n50.6<\/td>\n12.01<\/td>\n75+<\/td>\n7.746%<\/td>\n4.2<\/td>\n0.33<\/td>\n<\/tr>\n
      Expected Male Veteran Suicide Rate:<\/strong><\/td>\n32.79<\/td>\nExpected Female Veteran Suicide Rate:<\/strong><\/td>\n7.82<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

       <\/p>\n

      The observed suicide rates for male veterans and female veterans are 31.7 and 13.9, respectively.\u00a0 So that means that male vets have zero problem, while female vets have a big issue \u2013 right?<\/p>\n

      Yes . . . and no.\u00a0 And it points out just how misleading reducing the problem to a single number really can be.<\/p>\n

      As seen in the previous section above, male veterans 18-34 have a huge issue regarding veteran suicide \u2013 that group\u2019s suicide rate is literally more than twice that of their nonveteran contemporaries.\u00a0 But the fact that veterans 55 and over have a slightly to markedly lower<\/em> suicide rate than corresponding nonveterans – coupled with the fact that these older veterans constitute over 2\/3 of male veterans – completely obscures that issue. (It also obscures the fact that male veterans aged 35-54 have a substantially higher suicide rate than their nonveteran counterparts.)<\/p>\n

      The female veteran overall suicide rate number, on the other hand, appears indicative of a real problem.\u00a0 Only female veterans 75 or older have a lower suicide rate than their nonveteran contemporaries; female veterans younger than 75 have a markedly higher suicide rate than their nonveteran contemporaries.\u00a0 Since female veterans 75 and older make up less than 8% of the female veteran population, in this case their lower rate doesn\u2019t hide the larger problem.<\/p>\n

      For completeness, here\u2019s the overall veteran calculation.\u00a0 The methodology is the same; the weighting factors are % of veteran population vice % of veteran gender group.<\/p>\n\n\n\n\n\n\n\n\n\n\n
      Expected Overall Veteran Suicide Rate, Corrected for Known Veteran Demographics<\/strong><\/td>\n<\/tr>\n
      Male Veterans<\/strong><\/td>\nFemale Veterans<\/strong><\/td>\n<\/tr>\n
      Age Group<\/strong><\/td>\n% of Veterans<\/strong><\/td>\nNonveteran Suicide Rate<\/strong><\/td>\nExpected Subgroup Contribution<\/strong><\/td>\nAge Group<\/strong><\/td>\n% of Veterans<\/strong><\/td>\nNonveteran Suicide Rate<\/strong><\/td>\nExpected Subgroup Contribution<\/strong><\/td>\n<\/tr>\n
      18-34<\/td>\n7.965%<\/td>\n24.6<\/td>\n1.96<\/td>\n18-34<\/td>\n1.866%<\/td>\n6.1<\/td>\n0.11<\/td>\n<\/tr>\n
      35-54<\/td>\n20.930%<\/td>\n27.0<\/td>\n5.65<\/td>\n35-54<\/td>\n3.728%<\/td>\n9.3<\/td>\n0.35<\/td>\n<\/tr>\n
      55-74<\/td>\n40.399%<\/td>\n27.9<\/td>\n11.27<\/td>\n55-74<\/td>\n2.836%<\/td>\n7.9<\/td>\n0.22<\/td>\n<\/tr>\n
      75+<\/td>\n21.563%<\/td>\n50.6<\/td>\n10.91<\/td>\n75+<\/td>\n0.708%<\/td>\n4.2<\/td>\n0.03<\/td>\n<\/tr>\n
      Expected Overall Veteran Suicide Rate:<\/strong><\/td>\n30.51<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

       <\/p>\n

      The VA study appendix data gives the overall veteran suicide rate as 30.1.\u00a0 The overall number, adjusted to take into account veteran demographics, would lead one to believe that veterans are doing better than the general population.\u00a0 That\u2019s true for some age\/gender groups, and false for others. Comparison here also fails to give any insight on reality, once again showing that attempting to reduce a complex situation to a meaningful single-number metric is quite difficult.\u00a0 But the media loves to use such simple \u2013 and often meaningless \u2013 single-number metrics.<\/p>\n

      Bottom Line<\/u>.<\/p>\n

      The issue of veterans and suicide is complex.\u00a0 No single number I\u2019ve found portrays it adequately; that\u2019s true for veterans collectively and for veterans partitioned by gender.<\/p>\n

      For males, the situation is both good and bad.\u00a0 Male veterans as a group actually have a lower suicide rate than male nonveterans; this is due to the grossly age-skewed nature of the US male veteran population vice the male US nonveteran population.\u00a0 But the situation is somewhat \u201cbipolar\u201d, if you will; older male veterans (55 and up) are doing slightly to markedly better than nonveterans with respect to suicide rate, while males veterans less than age 55 have a higher suicide rate than nonveteran contemporaries \u2013 and the 18-34 age group appears to be badly hurting on that score.<\/p>\n

      For female veterans, the situation is worse.\u00a0 Female veterans as a group have a substantially higher suicide rate than their nonveteran contemporaries.\u00a0 That\u2019s because female veterans have a substantially higher suicide rate than their nonveteran contemporaries at all ages below 75.\u00a0 Only female veterans 75 or older have a lower suicide rate than the corresponding group of nonveterans – and they comprise only about 8% of female veterans, so that doesn’t mask the overall problem.<\/p>\n

      In any case, if there\u2019s a significant problem or \u201ccrisis\u201d regarding veteran suicide the media appears to have it completely wrong.\u00a0 (No surprise there.)\u00a0 Any such \u201ccrisis\u201d is not<\/span> among elderly veterans.\u00a0 Rather, the serious problems appear to be among young male veterans and female veterans under the age of 75. These two groups together make up about 16.5% of the veteran population, and each has a suicide rate markedly higher than their non-veteran counterparts.\u00a0 Male veterans age 55 and up \u2013 a group that makes up roughly 62% of all veterans, and which the media seems to think are in \u201ccrisis\u201d \u2013 are actually doing better<\/i> with respect to suicide rate than their nonveteran contemporaries (in the case of veterans 75 and up, far better). The demographic differences between the veteran and nonveteran population obscures these facts unless one takes a more detailed look at the data than a cursory overall rate comparison.<\/p>\n

      This is not an attempt at minimizing or ignoring the issue of veteran suicides.\u00a0 But trying to reduce the issue of veteran suicide to a single number that can explain what’s going on is both misleading and foolish.\u00a0 It\u2019s misleading, because veterans demographics guarantees the overall veteran suicide rate will be higher than that of the nonveteran population \u2013 though for male veterans, it\u2019s actually lower than it should be if veterans mirrored the general population with respect to suicide rates.\u00a0 And it\u2019s foolish, because reality here is more complex than a single number can portray.\u00a0 Whether the statement that \u201cveterans commit suicide more frequently than non-veterans\u201d is true or false depends on precisely which subgroups of veterans you\u2019re talking about.<\/p>\n

      Why?<\/u><\/p>\n

      So, why don\u2019t we ever seem to see veteran demographics \u2013 and their expected impact due to the fact that veterans as a group are 90+% male and generally are older than nonveterans \u2013 discussed in detail in published articles concerning veteran suicide?\u00a0 I have my own theory.\u00a0 It can be summed up in three words: \u201cFollow the money.\u201d<\/p>\n

      Small differences don\u2019t usually get the American public\u2019s attention; they also don\u2019t justify increased budgets for studies or win Pulitzer prizes.\u00a0 But dramatic differences can get attention \u2013 and can be used as justification to garner more funding for research, to help \u201cfix the problem\u201d, or gain fame for journalists (and thus higher future incomes).\u00a0 That\u2019s true whether those \u201cdramatic differences\u201d are real or not.\u00a0 Now tell me:\u00a0 would journalists, bureaucrats, and researchers ever present dramatic cases based on misleading metrics that might not reflect the full story when they\u2019re asking for money or presenting a news story?\u00a0 (smile)<\/p>\n

      Further, doing something along those lines is virtually a \u201cno-lose\u201d scenario.\u00a0 Those doing so know full well that only rarely will someone double-check their math, methodology, or underlying assumptions.\u00a0 That’s true even if it’s possible to find the information necessary to double check; and finding the info often isn’t easy and sometimes isn\u2019t possible.\u00a0 So even when done intentionally instead of through ignorance, using misleading numerical data and to support a questionable or false conclusion (AKA \u201ccooking the books\u201d) is often a low-risk proposition.<\/p>\n

      The media seems to have focused on the \u201celderly veteran\u201d angle.\u00a0 Here, they\u2019re using a misleading number and making a comparison that is simply not valid because of major demographic differences between the populations.\u00a0 Why?\u00a0 My guess is that it\u2019s because it\u2019s sensational. (\u201cJust look at how we\u2019ve failed<\/em> those poor, helpless elderly vets.\u00a0 We need more money to fix the problem!\u201d)\u00a0 Maybe they simply didn’t know any better. But do you really expect them to admit they were \u201cout to lunch\u201d yet again regarding this latest bogus reporting on veterans and veterans’ issues any time soon?<\/p>\n

      In reality, a close examination of the data shows that it’s young male veterans<\/i> – not older ones – who are hurting the worst with respect to suicide; female veterans younger than 75 are a close second.\u00a0 For each of those groups, the suicide rate either approaches doubling or more than doubles that of the corersponding group in the nonveteran population.\u00a0 But the best I can tell, neither of those facts are much discussed in the media.<\/p>\n

      Why?\u00a0 Maybe it doesn\u2019t \u201csupport the agenda\u201d.\u00a0 Maybe it\u2019s because females and younger male veterans each comprise less than 10% of all veterans.\u00a0 Maybe the media hasn\u2019t noticed it because they don\u2019t \u201cdo math\u201d.\u00a0 Maybe there’s some other reason.\u00a0 Dunno.\u00a0 But best I can tell, the numbers say that each is a real issue.<\/p>\n

      Whether one or more political agendas are also at work, well, decide that for yourself.\u00a0 I have my own opinion here, but I prefer to deal in hard fact vice speculation.<\/p>\n

       <\/p>\n

      Author’s Note<\/span>:\u00a0 edited after publication to (1) correct an unintentional error in wording, (2) change one section title to better reflect the section’s subject, and (3) replace all occurrences of the word “peers” with “contemporaries”.<\/em><\/p>\n","protected":false},"excerpt":{"rendered":"

      Back in March, Dave Hardin wrote an article about veteran suicides and studies relating to the … Veterans, Suicide Rates, and the Demographics Nobody Discusses<\/span>Read more<\/a><\/p>\n","protected":false},"author":623,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[220,213,118],"tags":[],"class_list":["post-85568","post","type-post","status-publish","format-standard","hentry","category-the-floggings-will-continue-until-morale-improves","category-your-tax-dollars-at-work","category-veterans-issues"],"_links":{"self":[{"href":"https:\/\/www.azuse.cloud\/index.php?rest_route=\/wp\/v2\/posts\/85568","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.azuse.cloud\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.azuse.cloud\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.azuse.cloud\/index.php?rest_route=\/wp\/v2\/users\/623"}],"replies":[{"embeddable":true,"href":"https:\/\/www.azuse.cloud\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=85568"}],"version-history":[{"count":1,"href":"https:\/\/www.azuse.cloud\/index.php?rest_route=\/wp\/v2\/posts\/85568\/revisions"}],"predecessor-version":[{"id":87831,"href":"https:\/\/www.azuse.cloud\/index.php?rest_route=\/wp\/v2\/posts\/85568\/revisions\/87831"}],"wp:attachment":[{"href":"https:\/\/www.azuse.cloud\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=85568"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.azuse.cloud\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=85568"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.azuse.cloud\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=85568"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}