DTIC ADA561535: Effects of Combat Deployment on Risky and Self-Destructive Behavior Among Active Duty Military Personnel

0.79 MB·English

by  Defense Technical Information Center

Effects of Combat Deployment on Risky
and Self-destructive Behavior Among
Active Duty Military Personnel

Cynthia J. Thomsen
Valerie A. Stander
Stephanie K. McWhorter
Mandy M. Rabenhorst
Joel S. Milner
Naval Health Research Center

Report No. 09-03

The views expressed in this article are those of the authors and do not nec-
essarily reflect the official policy or position of the Department of the
Navy, Department of Defense, nor the U. S. Government.
Approved for public release; distribution is unlimited

Naval Health Research Center
140 Sylvester Road
San Diego, California 92106
JournalofPsychiatricResearch45(2011)1321e1331
ContentslistsavailableatScienceDirect
Journal of Psychiatric Research
journal homepage: www.elsevier.com/locate/psychires
Effects of combat deployment on risky and self-destructive behavior among active
duty military personnel
Cynthia J. Thomsena,*, Valerie A. Standera, Stephanie K. McWhortera, Mandy M. Rabenhorstb,
Joel S. Milnerb
aBehavioralScienceandEpidemiology,NavalHealthResearchCenter,140SylvesterRoad,SanDiego,CA92106-3521,USA
bCenterfortheStudyofFamilyViolenceandSexualAssault,NorthernIllinoisUniversity,125PresidentsBlvd,DeKalb,IL60640,USA
a r t i c l e i n f o a b s t r a c t
Articlehistory: Althoughresearchhasdocumentednegativeeffectsofcombatdeploymentonmentalhealth,fewstudies
Received29September2010 have examined whether deployment increases risky or self-destructive behavior. The present study
Receivedinrevisedform addressed this issue. In addition, we examined whether deployment effects on risky behavior varied
25February2011
dependingon historyof pre-deployment risky behavior,and assessed whether psychiatric conditions
Accepted7April2011
mediatedeffectsofdeploymentonriskybehavior.Inananonymoussurvey,activedutymembersofthe
U.S.MarineCorpsandU.S.Navy(N¼2116)describedtheirdeploymentexperiencesandtheirpartici-
Keywords:
pation in risky recreational activities, unprotected sex, illegal drug use, self-injurious behavior, and
Combatdeployment
suicide attempts during three time frames (civilian, military pre-deployment, and military post-
Self-harm
Suicideattempts deployment). Respondents also reported whether they had problems with depression, anxiety, or
Illegaldrugs PTSDduringthesamethreetimeframes.Resultsrevealedthatriskybehaviorwasmuchmorecommon
Risk-taking in civilian than in military life, with personnel who had not deployed, compared to those who had
Unprotectedsex deployed,reportingmoreriskybehaviorandmorepsychiatricproblemsascivilians.Forthecurrenttime
Depression period,incontrast,personnelwhohaddeployed(versusneverdeployed)weresignificantlymorelikely
Anxiety toreportbothriskybehaviorandpsychiatricproblems.Importantly,deploymentwasassociatedwith
PTSD
increases in risky behavior only for personnel with a pre-deployment history of engaging in risky
behavior.Althoughpsychiatricconditionswereassociatedwithhigherlevelsofriskybehavior,psychi-
atricproblemsdidnotmediateassociationsbetweendeploymentandriskybehavior.Implicationsfor
understandingeffectsofcombatdeploymentonactivedutypersonnelanddirectionsforfutureresearch
arediscussed.
PublishedbyElsevierLtd.
1.Introduction aconstellationofbehaviorsthatareeitherovertlyself-destructive
orthatplacetheindividualatsubstantialriskofharm.
HighlevelsofcontinuingU.S.militarydeploymentsinsupport “Riskyorself-destructive”behaviorisabroadcategory.Atone
of war efforts in Iraq and Afghanistan have heightened concerns endofthespectrumareactivitiesthatcouldresultinharmtothe
abouteffectsofcombatonmilitarypersonnel.Researchhasdocu- selfbutarenotnecessarilyintendedtodoso,suchasriskyrecre-
mented increases in psychological problems among personnel ation (e.g., extreme sports), risky driving (e.g., driving without
returning from combat deployment. Although most studies have a seatbelt or afterdrinking), risky sexual activities (e.g., sex with
focused on post-traumatic stress disorder (PTSD), increases in strangers, unprotected sex), and substance use or abuse; at the
relatedproblemssuchasdepressionandsubstanceabusealsohave other end of the continuum are behaviors explicitly intended to
beendemonstrated(Boscarino,1995;Hogeetal.,2004).Compar- harm the self, including deliberate self-harm without suicidal
ativelylessattentionhasbeenpaidtopotentialeffectsofdeploy- intent(“self-harm”),suicideattempts,andsuicide.Althoughthese
ment on other psychological and behavioral health issues. The behaviors may be subsumed under the rubric of “self-injurious
present study investigated the effects of combat deployment on thoughts and behaviors” (Silverman et al., 2007), they do not
constituteahomogenousset;thebehaviorslikelydifferinseveral
respects, including their prevalence, motivational underpinnings,
* Correspondingauthor.Tel.:þ16195537174;fax:þ16195538459. thepotentialseverityofadverseconsequences,andthenatureand
E-mailaddress:[email protected](C.J.Thomsen). magnitudeofbarrierstoperformingthem.Nonetheless,different
0022-3956/$eseefrontmatterPublishedbyElsevierLtd.
doi:10.1016/j.jpsychires.2011.04.003
1322 C.J.Thomsenetal./JournalofPsychiatricResearch45(2011)1321e1331
typesofriskybehaviorstendtoco-occur.Forexample,individuals and self-harm (e.g., Whitlock and Knox, 2007), links with less
whoengageinself-harmaremorelikelythanthosewhodonotto overtlyself-destructivebehaviorssuchasriskyrecreation,unpro-
attemptsuicide(e.g.,MuehlenkampandGutierrez,2007),useillicit tected sex, and illegal drug use are less clear. Nonetheless, risky
drugs(e.g.,Matsumotoetal.,2005),andbeinvolvedinotherrisky behaviorsthatarenotintrinsicallyself-destructivesometimesmay
behaviors(e.g.,Laye-GindhuandSchonert-Reichl,2005).Similarly, be undertaken, either consciously or subconsciously, with self-
bothheavydrinkingandheavysmokingareassociatedwithrisky destructive intent. Consistent with this view, accidental deaths
sexualbehavior(e.g.,Leighetal.,1994)andriskydriving(e.g.,Fear and suicides have been found to have common risk factors (Stea
etal.,2008).Moreover,differenttypesofriskybehavioroftenshare et al., 2002; Thoresen and Mehlum, 2004). Previous findings of
commonpredictors.Forexample,mosttypesofriskybehaviorare elevatedratesofaccidentaldeathamongcombatveterans(Catlin
more common among young men than among women or older Boehmeretal., 2004;Kangetal., 2002;Knapiketal.,2009)may
men(Nell,2002;Zuckerman,2007). thus reflect some instances in which individuals purposefully
Thepresentstudyexaminedeffects ofcombatdeploymenton engaged in risky behaviors for self-destructive purposes. In the
fivespecifictypesofriskyorself-destructivebehavior:dangerous present study, we examined whether psychiatric conditions
recreational activities, unprotected sex with someone other than (depression, anxiety, and PTSD) were associated with risky
aregularpartner,illegaldruguse,self-harm,andsuicideattempts. behavior. In addition, we examined whether these conditions
Ourprimaryhypothesiswasthatratesofeachriskybehaviorwould mediatetheimpactofdeploymentonriskybehaviors.
behigherfollowingcombatdeploymentthanbeforedeployment. In addition to predicting that combat deployment would
Ofthefivetypesofriskyorself-destructivebehaviorconsideredin increase rates of risky behavior in general, we expected deploy-
thepresentstudy,onlysubstanceuseandsuicide-relatedthoughts menteffectstodifferdependingontheindividual’spriorhistoryof
and behaviors have received significant research attention as risky behavior. Specifically, effects of deployment in increasing
possible consequences of military deployment. Combat deploy- riskybehaviorwereexpectedtobemostpronouncedamongthose
ment has been associated with increased likelihood of drug and whohadengagedinthebehaviorpreviously.Priorengagementin
alcoholproblems(Browneetal.,2008;Calhounetal.,2008;Hoge risky behavior may be a marker for psychological difficulties or
et al., 2004; Koenen et al., 2003). However, empirical findings vulnerabilities that place the individual at risk of developing
regardingtheassociationbetweendeploymentandsuicide-related adversereactionstodeployment.Ataminimum,priorengagement
ideationsandbehaviorshavebeenweakandinconsistent(Centers inabehaviorindicatesthatitispartoftheindividual’sbehavioral
forDiseaseControl,1987;Hall,1996;Hansen-Schwartzetal.,2002; repertoire.Iftheindividualhasengagedinabehaviorinthepast,
Kaplanetal.,2009;O’TooleandCantor,1995;Thoresenetal.,2003; perhapsasameansofcopingwithdistress(e.g.,Klonsky,2007),he
Wojciketal.,2009;Wongetal.,2001). orshemayturntothesamestrategyagaintocopewiththestresses
Althoughweareawareofnopreviousresearchexaminingthe associated with deployment or with readjustment following
impact of deployment on risky recreational activity, unprotected deployment. We could locate no prior research investigating
sex, or self-harm, combat deployment has been linked with whetherdeploymenteffectsaremoderatedbypriorbehaviors.
increasesinriskydrivingbehaviors(e.g.,speeding,failingtowear Inthepresentstudy,wefirstusedwithin-subjectsanalysesto
one’sseatbelt;Fearetal.,2008).Inaddition,inasampleofArmy examinechangesinriskybehaviorfromciviliantomilitarylife(for
soldiers returning from Iraq, specific combat experiences were bothcombatdeployedandnon-deployedpersonnel)andfrompre-
related to scores on a scale assessing general preferences for deployment to post-deployment life (for combat deployed
engaginginriskyactivities(e.g.,seekingthe“thrillofdanger,”liking personnelonly).Next,wedirectlycomparedratesofriskybehavior
to drive fast; Killgore et al., 2008). However, because risk-taking amongactivedutypersonnelwhohadpreviouslycombatdeployed
preferences prior todeployment werenotassessed in this study, andthosewhohadnot.Inaddition,weexaminedwhethereffects
itisnotpossibletodeterminewhethercombatexposurechanged of combat deployment on risky behavior differed depending on
riskpreferences,orwhetherthosewhopreferredriskysituations whethertheindividualhadengagedinthatbehaviorinthepast.
soughtoutcombatexposure;indeed,someevidencesuggeststhat Wepredictedthateffectsofdeploymentonagivenriskybehavior
militarypersonnelwithhigherrisk-takingproclivitiesmaybemore would be stronger among individuals who had a history of
likelytodeployandtovolunteerforhazardousduties(Belletal., engaging in that type of behavior. Each analysis was conducted
2010;Bricknelletal.,1999;Jobeetal.,1983;Zuckerman,2007). across five different types of risky behavior (risky recreational
There are several mechanisms by which combat deployment activities,unprotectedsex,illegaldruguse,self-harm,andsuicide
might increase risky behavior. Joiner’s (2005) interpersonal- attempts), allowing for a determination of whether combat
psychological theory of suicidal behavior posits that repeated deployment had similar or distinct effects on different types of
exposuretopainorfear-inducingsituations(e.g.,throughcombat behavior. A final setof analyses examined patterns of psychiatric
exposure)resultsinhabituation,i.e.,greatertoleranceforpainand problems (depression, anxiety, and PTSD) before and after
lower levels of fear. In turn, this habituation increases the indi- deploymentandamongdeployersandnon-deployers,andtested
vidual’s capability to enact lethal self-injury. Using similar logic, whether deployment-related increases in psychiatric problems
Killgore et al. (2008) suggested that habituation to danger as might account for any effects of deployment on risky or self-
a result of long-term combat exposure may increase risk-taking destructivebehaviors.
propensities among veterans. In support of these formulations,
recentevidenceindicatesthatcombatexposureisassociatedwith 2.Methods
self-reported habituation to pain and decreases in fear (Brenner
et al., 2008; Bryan et al., 2010). A related possibility is that 2.1. Participants
returning combat veterans may engage in risky behavior in an
attempt to create an “adrenaline rush” similar tothose that they The Naval Health Research Center (NHRC) Combat Stress and
experiencedincombat(Vaughan,2006). Substance Use survey was conducted between August 2006 and
Increases in risky behaviors following deployment also may August 2007. Participants were active duty military personnel
resultfromdeployment-relatedincreasesinpsychiatricsymptoms serving at U.S. Marine Corps (USMC) installations in Southern
such as depression and PTSD (e.g., Hoge et al., 2004). Although California and Arizona within three major commands: Marine
negative emotional states have been linked with both suicidality CorpsAirStationYumaArizona(primarily3rdMarineAirWing);
C.J.Thomsenetal./JournalofPsychiatricResearch45(2011)1321e1331 1323
1st Marine Logistics Group, Camp Pendleton, California; and 1st assessed bya single item; the other types of behavior wereeach
MarineDivision,CampPendletonand29Palms,California.These assessedbytwoitems:oneaskedaboutengaginginthebehavior
commandsrepresentair,support,andinfantryunits.Atotalof2612 whiledrinking,andtheotheraskedaboutengaginginthebehavior
personnel attended survey recruiting sessions. Of these, 2539 whilenotdrinking(e.g.,“attemptedsuicideafterdrinkingalcohol”;
participants completed some part of the survey, yielding a 97% “attemptedsuicidewhenyouwerenotdrinkingalcohol”).Ineach
participation rate. The present sample included only participants case, respondents reported whether they had engaged in the
whoprovideddataabouttheirdeploymentsandwhocompletedat behavior(a)asacivilian,(b)whileinthemilitarybeforetheirfirst
least7ofthe8itemsassessingriskybehavior(N¼2116;response combatdeployment,and(c)afterreturningfromtheirfirstcombat
rate¼81%).Mostparticipants(93%)wereintheUSMC;theother deployment. Finally, for each of the same three time frames,
7%wereNavy(USN)personnelembeddedwithinUSMCunits.As respondents were asked whether they had ever “been told by
expected,themajorityofparticipantswereyoung(M¼24.09yrs, adoctorormentalhealthcounselor”thattheyhad(1)depression,
SD ¼ 5.30), male (92%), and junior-ranking enlisted personnel (2)anxiety,or(3)PTSD.
(E1(cid:2)E5¼85%);only4%werecommissionedofficers.Forty-seven
percent reported at least some college or other post-secondary 2.3. Analyticstrategy
technical training. Approximately equal numbers were single
(47%)andmarriedorcohabiting(46%),withtheremaining7%being Priortotestingourhypothesesregardingtheeffectsofdeploy-
divorced, separated, or widowed. Compared to members of the mentonriskybehavior,weprovidedescriptiveinformationabout
USMCasawhole(MarineCorpsCommunityServices,2007),USMC bothcombatdeploymentsandratesofeachtypeofriskybehavior
participantsinthepresentsampleweresignificantlyyounger(76% withinoursample.Thisincludesanexaminationofdemographic
sample vs. 67% USMC aged 25 or younger), lower in rank (44% factorsassociatedwithbothdeploymentandriskybehavior,aswell
samplevs.39%USMCrankedlowerthanE4),lesslikelytobeoffi- asassociationsbetweendifferenttypesofriskybehavior.Wenext
cers(4%sample vs.11%USMC),andmorelikelytobefemale(8% investigatedchangesinoverallratesofriskybehaviorfromcivilian
samplevs.6%USMC),ps<0.01.USMCpersonnelinthesampledid tomilitary/pre-deploymentlife,separatelyfordeployersandnon-
notdifferfromthegeneralUSMCintermsofmaritalstatus(44%of deployers. The hypothesis that risky behavior would increase
sampleand45%ofUSMCmarried). following deployment was examined in two ways. First, we
examinedchangesinriskybehaviorfrompre-deploymenttopost-
2.2. Surveyprocedures deploymentamongpersonnelwhohadcombatdeployed.Second,
we compared the risky behavior of deployers and non-deployers
This research was conducted in compliance with all federal duringdifferenttimeperiods (asa civilian, whilein themilitary/
regulationsregardingtheprotectionofhumansubjectsinresearch, pre-deployment, and currently). For deployers, current behavior
and was approved by an Institutional Review Board at NHRC was defined as post-deployment behavior; for non-deployed
(protocol #NRHC 2006.0008). Commanders referred personnel personnel, current behavior was defined as pre-deployment
fromparticipatingunitstomeetwithciviliansurveyadministrators behavior(astherewasnodeploymentperiod).Inthissetofanal-
at prearranged times and locations, based on their availability yses,deploymenteffectsonriskybehaviorareindicatedbydiffer-
withinthetrainingschedule.Groupsrangedinsizefrom28to408. ences between deployers and non-deployers in current behavior
Atthebeginningofeachsession,thesurveywasintroducedasan (thatwerenotapparentpriortodeployment).Afterexaminingthe
anonymous questionnaire concerning the relationship between overallimpactofdeploymentonriskybehavior,additionalanalyses
combat stress and substance use. Informed consent information wereconductedtodeterminewhethertheimpactofdeployment
was read aloud. Consent information included the following: on each specific type of risky behavior was moderated by prior
participationwas voluntary, with no repercussion for nonpartici- engagementinthattypeofriskybehavior.Afinalsetofanalyses
pation;participantscouldleaveblankany questionstheydidnot examined whether deployment effects on risky behavior were
want to answer, and could quit at any time; surveys were mediatedbymentalhealthissues.
completely anonymous and would be processed by civilian
researchers, not military personnel; and individual data would 3. Results
never be presented. To better ensure that participation was
voluntary,officersandenlistedpersonnelparticipatedinseparate 3.1. Combatdeployment
locations.Toincreasethesenseofanonymity,participants sealed
theirown completed surveys in stamped envelopes addressed to More thanhalf(58%;n¼ 1225)ofrespondentsin thepresent
Northern Illinois University (NIU); envelopes weremailed imme- samplehadbeencombatdeployed.Mostdeployers(57%)hadone
diatelyfollowingthesession.Personnelwhochosenottopartici- deployment, 29% had two, and 14% had three or more. Among
patewereaskedtositquietlyandthensealtheirblanksurveysin combatdeployers,97%haddeployedinsupportofOperationIraqi
theirmailingenvelopesasthoughtheyhadparticipated. Freedom or Operation Enduring Freedom (91% OIF,19% OEF,13%
Thesurveyassesseddemographiccharacteristics(sex,age,rank, otheroperations[e.g.,DesertShield/Storm]).Themodaldeployment
education level, marital status, number of children) and combat lengthwas7months,with35%ofdeploymentsbeingshorterand
deployment history (e.g., dates, locations, duties). A combat 19%beinglonger(M¼6.86;Mdn¼7;range:1e19months).Three-
deploymentwasdefinedbythereceiptofimminentdangerpayor quartersofdeployershadreturnedfromdeploymentwithinthepast
combat zone tax exclusion benefits; non-combat deployments year(45%inthepast6months).Men(58%)andwomen(53%)were
werenotconsidered.Eightitemsassessedriskybehavior.Respon- equallylikelytohavedeployed,c2(1,N¼2066)¼2.09,n.s.However,
dents were asked to indicate whether they had engaged in five personnel who were married or cohabiting (69%) or divorced,
different types of behavior: risky recreational activities (“done separated, or widowed (80%) were more likely to have deployed
thingsforfunthatweresodangerousyouwerelikelytobeinjured than those who were single (45%), c2 (2, N ¼1921) ¼ 130.38,
orkilled”),unprotectedsexwithsomeoneotherthanone’sregular p < 0.001. USN personnel (70%) were more likely to have been
partner, illegal drug use, self-harm (“hurt yourself physically to deployed than USMC personnel (57%), c2 (1, N ¼2058) ¼ 9.93,
calmdownorfeelbetterwhenyouweretense,anxious,orupset”), p<0.01.Inaddition,deployers,comparedtonon-deployers,were
and attempted suicide. Self-harm and illegal drug use were each significantly older (Ms ¼ 25.67 vs. 21.90 years), higher in rank
1324 C.J.Thomsenetal./JournalofPsychiatricResearch45(2011)1321e1331
Table1 constitute two underlying clusters, one representing deliberate
Associations(crudeoddsratios)betweenlifetimeengagementinhigh-riskbehav- self-harm and one representing behaviors that may increase the
iorsanddemographiccharacteristics.
likelihoodofselfharm,butarenotnecessarilyintendedtodoso.
Predictor Typeofhigh-riskbehavior Considering the demographic predictors of lifetime risky
Risky Unprotected Illegal Self-harm Suicide behavior,menweremorelikelythanwomentoengageinrisky,but
recreation sex druguse attempt notexplicitlyself-destructive,behaviors;womenweresignificantly
Unprotectedsex 5.42*** morelikelytoreportattemptingsuicide(seeTable1).Relativeto
Illegaldruguse 3.45*** 3.54*** USMC personnel, USN personnel were significantly less likely to
Self-harm 6.16*** 2.39*** 2.08*** engage in all of the behaviors except for suicide attempts. Age,
Suicideattempt 4.43*** 3.42*** 2.43*** 11.84***
education, and military rank were unrelated to the likelihood of
Male 2.91*** 1.42* 1.70** 1.02 0.59*
Olderage 0.93*** 0.99 0.94*** 0.91*** 0.95* engaging in unprotected sex, but were negatively related to the
Higherrank 0.81*** 0.99 0.78*** 0.77*** 0.75*** othertypesofbehavior.Withrespecttomaritalstatus,respondents
Highereducation 0.82*** 1.07 0.82** 0.75*** 0.82 whoweresingleorformerlymarried(i.e.,divorced,separated,or
USN(vs.USMC) 0.40*** 0.59** 0.50*** 0.56* 0.71 widowed)weregenerallymorelikelytoengageinriskybehaviors
Singlea 1.57*** 1.26* 1.15 1.42** 1.19
Div/sep/widoweda 1.13 1.91** 1.09 1.02 2.35** thanweremarriedorcohabitingrespondents(seeTable1).
USN¼Navy;USMC¼MarineCorps.Nsvaryfrom1901to2097.ORsgreaterthan1.0
indicateriskfactorsandORslessthan1.0indicateprotectivefactorsforhigh-risk 3.3. Timingofriskybehavior
behavior.
*p<0.05.**p<0.01.***p<0.001. Fig. 1 depicts the mean number of different risky behaviors
a Thereferencegroupwasmarried/cohabiting.
reportedbydeployedandnon-deployedpersonnelforeachofthe
threetimeframes(i.e.,civilian,pre-deployment,andcurrent).Itis
(Ms¼4.62[E4-E5]vs.3.24[E3]),andhigherineducation(Ms¼2.57 important to note that for non-deployed personnel, estimates of
vs. 2.44, where 2 ¼ high school diploma and 3 ¼ some college), ts pre-deploymentandcurrentratesofriskybehaviorarebasedonthe
(1974.8e2031.1)(cid:3)4.3,ps<0.001. samedata,becausebothtimeframesareequivalentforthisgroup.In
thesampleasawhole,thelikelihoodofeachtypeofriskybehavior
decreasedsignificantlyfromthecivilianperiodtothemilitary/pre-
3.2. Riskybehavior deployment period, McNemar c2 (1, Ns ¼ 2054e2116) (cid:3) 17.78,
ps<0.001.Althoughthedeclinewasmostpronouncedforillegal
Nearlythree-quarters(73%)ofparticipantsreportedengagingin druguse(from39%to5%),itwassubstantialfortheotherbehaviors
one or more types of risky behavior in their lifetimes (i.e., as as well: self-harm (13% to 6%), suicide attempts (5% to 2%), risky
acivilianorwhileinthemilitary).Half(52%)reportedengagingin recreation(38%to22%),andunprotectedsex(37%to29%).Table2
unprotected sex with someone other than a regular partner, and provides the percentage of respondents reporting each specific
half(49%)hadengagedinrecreationalactivitiesdangerousenough riskybehaviorduringeachtimeframe,separatelyfordeployersand
tocauseseriousinjuryordeath.Twoinfive(41%)hadusedillegal non-deployers.Declinesinriskybehaviorfromciviliantomilitary/
drugs, and one in five (19%) reported engaging in self-harm. pre-deploymentlifeweregenerallygreaterfornon-deployedthan
Attempted suicide was reported by 7% of respondents. Compari- fordeployedpersonnel.Fornon-deployedpersonnel,allformsof
sonsof lifetime ratesof risky behavior while drinkingversusnot riskybehaviordecreasedsignificantlyfromciviliantomilitary/pre-
drinkingalcoholwereconductedforthethreetypesofbehaviorfor deployment periods, McNemar c2 (1, Ns ¼ 874e891) (cid:3) 15.56,
whichthiswasassessed.Respondentsweremorelikelytoreport ps < 0.001; among deployed personnel, the reduction in risky
engaging in unprotected sex when they had been drinking than behavior was significant for four of the five behaviors [c2
whentheyhadnot(48%vs.45%),Cochran’sQ(1,N¼2094)¼13.95, (1,Ns¼1180e1225)(cid:3)7.58,ps<0.01],butdidnotreachsignificance
p < 0.001. In contrast, respondents were more likely to report forsuicideattempts[c2(1,N¼1225)¼2.94,p<0.10].
engaging in risky recreation or attempting suicide while not
drinking than while drinking (for risky recreation, 44% vs. 35%,
Cochran’sQ(1,N¼2098)¼89.25,p<0.001;forsuicideattempts,
5% vs. 3%, Cochran’s Q (1, N ¼ 2104) ¼ 15.14, p < 0.001). Despite
these differences, comparisons of deployers and non-deployers 1.8
yieldedidenticalpatternsofeffectsforthealcoholandno-alcohol
Total
items;therefore,alcoholandno-alcoholitemsforagiventypeof Not Deployed
rainsaklyybseesh.aviorwerecombinedintoasingleindexforallsubsequent Behaviors 1.5 Deployed
typTeasbolefr1ispkryovbiedheasvaiossro.cRieastpioonnsdaemntosnwghliofehtiamdeernegpaogretsdoinfdainffyeroennet Risky 1.2
teynpgeagoinfgriisnkyevebreyhaovthioerrwtyepreeoaftrissikgynibfiechaanvtiloyr.inAcsrceaansebdersieskenfoinr Number of 0.9
Traanbglein1g,tfhroemse2astsoocnieaatirolyns12w.eArneseuxbpslotarnattoiarly,wpritinhcoipdadlscroamtiopson(OeRntss) Mean
analysiswithvarimaxrotationperformedonlifetimeratesofthe 0.6
five risky behaviors revealed two underlying factors with eigen-
valuesgreaterthan1.Thefirstfactor,accountingfor39%ofthetotal
0.3
variance, consisted of unprotected sex, use of illegal drugs, and Civilian Pre-deployment Current
riskyrecreation;thesecondfactor,accountingfor21%ofthetotal Time Frame
variance,consistedofself-harmandsuicideattempts.Allloadings
Fig.1. Mean number of risky behaviors reported by time period. Note. For non-
ontheprimaryfactorweregreaterthan0.7,andnocross-loadings
deployedpersonnel,pre-deploymentandcurrenttimeframesareequivalentandare
weregreater than 0.3. This suggests that the five risky behaviors representedbythesamedatapoint.
C.J.Thomsenetal./JournalofPsychiatricResearch45(2011)1321e1331 1325
Table2 riskybehavioraftercontrollingfordemographicfactors,withORs
Percentageofrespondentsreportingeachtypeofhigh-riskbehaviorbytimeperiod greaterthan1.0indicatingthatriskybehaviorwasmorecommon
andcombatdeploymentstatus.
amongdeployers.
Typeofbehavior/Deploymentstatus Timeperiod AscanbeseeninthefirstcolumnofTable3,duringthecivilian
Civilian Pre-dep Post-dep Lifetime time frame, non-deployers were more likely than deployers to
reportengaginginriskybehavior.Thedifferencebetweenthetwo
High-riskrecreation 38 22 25 49
Deployed(n¼1225) 30 22 25 44 groupswassignificantforalltypesofriskybehaviorexceptsuicide
Non-deployed(n¼891) 50 22 e 56 attempts. Risky behavior while in the military but prior to
Unprotectedsex 37 29 29 52 deployment revealed a very different pattern. There were no
Deployed(n¼1225) 33 29 29 51 significantdifferencesbetweendeployersandnon-deployersinthe
Non-deployed(n¼891) 42 28 e 54 likelihoodofanytypeofriskybehaviorforthistimeperiod(seethe
Illegaldruguse 39 5 7 41 secondcolumnofTable3).Mostrelevanttoourprimaryhypothesis
Deployed(n¼1180) 33 6 7 36 are the comparisons between deployers and non-deployers in
Non-deployed(n¼874) 47 5 e 48 current rates of risky behavior. As shown in the third column of
Self-harm 13 6 10 19 Table 3, deployers were significantly more likely than non-
Deployed(n¼1220) 8 5 10 16 deployerstoreportcurrentengagementinriskyrecreation,illegal
Non-deployed(n¼889) 19 8 e 23 druguse,andself-harm;thetwogroupsdidnotdifferinreported
Suicideattempt 5 2 2 7 ratesofunprotectedsexorsuicideattempts.
Deployed(n¼1225) 3 2 2 6
Non-deployed(n¼891) 7 3 e 9
3.4. Previousriskybehavior,combatdeployment,andcurrentrisky
pre-dep¼pre-deployment;post-dep¼post-deployment. behavior
The next set of analyses examined whether the impact of
The hypothesis that risky behavior would increase following deployment on each specific type of risky behavior was moder-
deployment was examined in two ways. First, we examined ated by prior engagement in that type of risky behavior. To
changes in high-risk behavior from pre-deployment to post- examine this issue,we conducted logisticregressionanalyses in
deployment among personnel who had combat deployed. Rele- which current risky behavior was predicted from demographic
vantpercentagesareprovidedinTable2.Bothriskyrecreationand factors (entered on the first step), prior engagement in the
self-harm increased significantly from pre- to post-deployment, behavior in question (entered on the second step), deployment
McNemarc2(1,Ns¼1220(cid:2)1225)¼8.28and31.51,respectively, status (entered on the third step), and the interaction between
ps <0.01.In contrast, changes inrates of unprotectedsex,illegal prior behavior and deployment status (entered on the fourth
druguse,andsuicideattemptsfrompre-topost-deploymentwere step). For those who had not deployed, previous behavior was
notsignificant,McNemarc2(1,Ns¼1180to1225)(cid:4)2.54,n.s. defined as civilian behavior; for those who had deployed, it
Asanadditionalwayofexaminingdeploymenteffectsonrisky included both civilian behavior and military behavior prior to
behavior, we conducted a series of logistic regression analyses deployment. Results of these analyses are provided in Table 4.
comparing the risky behavior of deployers and non-deployers Among the demographic variables, only being divorced, sepa-
during three different time frames: as a civilian; while in the rated,orwidowedemergedasaconsistentpredictorofallforms
military, pre-deployment; and currently. For deployers, current of risky behavior. Not surprisingly, the strongest predictor of
behavior was defined as post-deployment behavior; for non- current engagement in each specific type of risky behavior was
deployed personnel, current behavior was defined as pre- havingengagedinthattypeofbehaviorpreviously.Nonetheless,
deployment behavior (as there was no deployment period). In controlling for previous behavior did not alter the pattern of
each analysis, demographic variables (sex, age, rank, education,
service [USMC vs. USN], and marital status) were entered on the
first step, followed by deployment status. Table 3 provides the Table4
Results of logistic regression analyses predicting current high-risk behaviors
adjustedoddsratios fordeploymenteffectsoneachtypeofrisky
(adjustedoddsratios).
behaviorduringeachtimeperiod.TheseORsrepresenttheextent
Predictor Typeofhigh-riskbehavior
towhichdeployersandnon-deployersdifferedinthelikelihoodof
Risky Unprotected Drug Self-harm Suicide
recreation sex use attempt
Male 3.12*** 1.07 2.27 1.05 0.58
Olderage 0.95** 0.96** 0.95 0.94* 0.97
Table3
Higherrank 0.99 1.08 0.83 1.01 0.76
Adjustedoddsratios(ORs)representingdifferencesbetweendeployedandnon-
Highereducation 1.18 1.06 1.52* 0.96 1.23
deployedpersonnelinratesofhigh-riskbehaviors.
USN(vs.USMC) 0.48** 0.62* 1.28 0.84 0.79
Typeofbehavior Civilian Pre-deployment Currenta Singlea 1.16 1.58*** 1.14 1.02 0.78
Div/sep/widoweda 2.52** 2.55*** 2.59** 1.86* 3.90**
OR (CI95%) OR (CI95%) OR (CI95%) Priorbehavior 4.65*** 3.29*** 5.49*** 7.94*** 8.58***
High-riskrecreation 0.56 0.45,0.70 1.14 0.89,1.47 1.59 1.24,2.04 Combatdeployment 1.94*** 1.19 2.14** 1.92** 0.86
Unprotectedsex 0.68 0.55,0.85 1.03 0.82,1.29 1.15 0.91,1.46 Priorbehavior(cid:5) 2.54*** 1.27 1.39 2.27* 1.68
Illegaldruguse 0.72 0.58,0.90 1.42 0.88,2.29 1.86 1.18,2.95 deployment
Self-harm 0.62 0.45,0.85 0.69 0.46,1.05 1.52 1.06,2.20
ORsgreaterthan1.0indicateriskfactorsandORslessthan1.0indicateprotective
Suicideattempt 0.66 0.39,1.12 1.03 0.53,2.03 0.86 0.44,1.68
factorsforhigh-riskbehavior.DemographicvariableswereenteredonStep1,prior
ORsgreaterthan1.0(lessthan1.0)indicategreater(lower)likelihoodofhigh-risk high-riskbehavior(ofthesametypeasthebehaviorbeingpredicted)onStep2,
behavioramongpersonnelwhohadcombatdeployed,relativetothosewhohad combatdeploymentonStep3,andtheinteractiononStep4oftheanalyses.ORsare
not.StatisticallysignificantORs(p<0.05)areinboldfacetype. reported for the step on which the predictor entered the model. USN ¼ Navy;
a Currentbehaviorwasdefinedaspost-deploymentbehavioramongpersonnel USMC¼MarineCorps.
whohad previouslydeployed, and as pre-deployment militarybehavior among *p<0.05.**p<0.01.***p<0.001.
personnelwhohadnotpreviouslydeployed. a Thereferencegroupwasmarried/cohabiting.
1326 C.J.Thomsenetal./JournalofPsychiatricResearch45(2011)1321e1331
differences between deployers and non-deployers in current 25
Total
rdiespkyloybmeehnatviwora;s aasssoinciattehdewpirtehviionucrseaasneadlyrsaetses(oseferisTkayblreecr3e)-, nditions 20 NDoept lDoyepedloyed
ationalactivity,illegaldruguse,andself-harm,butwasunrelated Co
tboehrOaafvteiposrriomifnauprnyrepidnroitcetteriecnstgetdahrisgeehtx-hreaisnikndtbeaerthatcaetvmiioopnrt.sePodefrsdcueeinpcltiodaygeme.seonftdaenpdlopyreiorsr Psychiatric 15
atinodnnoofnp-rdeevpiolouysebreshraevpioorrtianrgeeparcohvitdyepdeionfFriigsk.y2.bSethaativsitoicraallsyasifgunnicf-- porting 10
Re
ibceahnatviinotrearancdtiosenlfe-hffaercmts(wseeereTaobblteai4n)e.dDfeoprlobyomthenritskwyasreacsrseoactiiaotneadl ntage 5
with increased likelihood of risky recreational behavior among Perce
respondents with a prior history of risky recreational behavior
0
(OR ¼ 2.55, CI95%: 1.82, 3.58), but not among those with no such Civilian Pre-Deployment Current
history (OR ¼ 1.19, CI95%: 0.77, 1.82). Similarly, deployment was Time Frame
associated with increased self-harm amongthosewho hadprevi-
ously engaged in self-harm (OR ¼ 2.61, CI95%: 1.35, 5.05), but not FNiogt.e.3.FoPrernceonnt-adgeeploofyerdesppoenrsdoennntsel,repproer-tdinegplopysmycehniattraicndcocnudrirteionntstibmyetimfraempeesriaorde.
amongthosewithnohistoryofself-harm(OR¼ 1.46,CI95%:0.89, equivalentandarerepresentedbythesamedatapoint.
2.40). The interaction was not statistically significant for unpro-
tectedsex,illegaldruguse,orsuicideattempts.Nonetheless,both
unprotectedsexanddruguseshowedasimilarpatternofeffects. 3.5. Combatdeploymentandpsychiatricproblems
Thatis,amongpersonnelwithnohistoryofengaginginunprotected
sex or using illegal drugs, current rates of that behavior did not Overall, 23% of respondents reported at least one psychiatric
significantly differ between deployers and non-deployers (for problematsomepointintheirlives,andthosewhohaddeployed
unprotectedsex,OR¼0.91,CI95%:0.64,1.30;fordruguse,OR¼1.74, (25%)weresignificantlymorelikelythanthosewhohadnot(20%)
CI95%:0.66,4.56).Incontrast,amongpersonnelwithapriorhistory to reporta lifetime psychiatric problem, c2 (1, N ¼ 2109) ¼ 5.70,
ofengagingintheriskybehavior,deployersdisplayedsignificantly p<0.05,F¼0.05.Fig.3depictsthepercentageofdeployedand
higher current rates of the behavior than non-deployers (for non-deployedrespondentsreportingpsychiatricproblemsduring
unprotectedsex,OR¼1.49,CI95%:1.07,2.08;fordruguse,OR¼2.29, each of the three time frames (i.e., civilian, pre-deployment, and
CI95%:1.34,3.90).Finally,forsuicideattempts,therewasnodiffer- current).Asbefore,fornon-deployedpersonnel,estimatesofpre-
encebetweendeployersandnon-deployersincurrentlikelihoodof deployment and current rates of risky behaviorare based on the
behaviorregardlessofwhethertheindividualreportedaprevious samedata,becausebothtimeframesareequivalentforthisgroup.
suicide attempt (OR ¼ 1.60, CI95%: 0.45, 5.68) or not (OR ¼ 0.65, Aswasthecaseforriskybehaviors(seeFig.1),Fig.3revealsacross-
CI95%:0.29,1.49). overinteraction;althoughnon-deployersinitiallyweremorelikely
Fig.2. Percentageofrespondentsreportingcurrentriskybehaviorsbydeploymentstatusandpreviousbehavior.Note.Analysescontrolledforsex,age,rank,education,branchof
service,andmaritalstatus.
C.J.Thomsenetal./JournalofPsychiatricResearch45(2011)1321e1331 1327
Table5 Table6
Percentageofrespondentsreportingeachpsychiatricconditionbytimeperiodand Correlations(Phicoefficients)betweenpsychiatricconditionsandriskybehaviors.
combatdeploymentstatus.
Riskybehavior Condition
Psychiatriccondition/Deploymentstatus Timeperiod
Depression Anxiety PTSD
Civilian Pre-dep Post-dep Lifetime Riskyrecreation 0.17*** 0.16*** 0.10***
Depression 9 5 10 18 Unprotectedsex 0.14*** 0.14*** 0.08***
Deployed 6 4 10 17 Druguse 0.15*** 0.09*** 0.02
Non-deployed 14 6 e 18 Self-harm 0.30*** 0.26*** 0.09***
Suicideattempt 0.36*** 0.32*** 0.14***
Anxiety 6 4 9 14
Deployed 4 3 9 14 Correlationsarebetweenlifetimereportsofeachdiagnosisandofeachtypeofrisky
Non-deployed 8 5 e 13 behavior.PTSD¼post-traumaticstressdisorder.
***p<0.001.
PTSD 1 1 10 7
Deployed <1 1 10 11
Non-deployed 1 1 e 2
pre-deployment combined). Results revealed significant
pstrree-sdsedpis¼orpdreer-.dFeoprldoyepmloeynet;d,pnos¼t-d1e2p25¼;fpoorsnt-odne-pdleopylmoyeendt,.nPT¼SD89¼1.post-traumatic increases in the likelihood of both depression (McNemar c2 [1,
N ¼ 2116] ¼ 8.97, p < 0.01) and PTSD (McNemar c2 [1,
N¼2116]¼68.97,p<0.001).However,deployerswerenomore
thandeployerstoreportpsychiatricconditions,withinthecurrent
likely to report anxiety problems post- than pre-deployment,
time frame, deployers were more likely to report psychiatric
McNemarc2<1,n.s.
conditions.
Table5providesthepercentageofrespondentsreportingeach
psychiatric condition during each time frame, separately for 3.6. Dopsychiatricproblemsmediateassociationsbetween
deployedandnon-deployedpersonnel.Overall,depressionwasthe deploymentandriskybehavior?
mostcommonlyreportedproblem(17%),followedbyanxiety(14%),
andPTSD(7%).Deployersandnon-deployersdidnotdifferinlife- Thefindingthatpsychiatricconditionsshowsimilarpatternsof
timeratesofdepression(17%and18%,respectively),oranxiety(12% effectsasriskybehavior(asajointfunctionofdeploymentstatus
and14%,respectively),ps(cid:3)0.20.However,deployersweresignif- and time frame) is consistent with the possibility that mental
icantlymorelikelythannon-deployerstoreportPTSD(11%vs.2%), healthproblemsmediatetheassociationbetweendeploymentand
c2(1,N¼2089)¼62.28,p<0.001,F¼0.17. risky behavior. Also consistent with this possibility, psychiatric
The next set of analyses examined differences in psychiatric problems were significantly correlated with risky behaviors (see
problemsbetweendeployersandnon-deployersduringeachtime Table 6). Not surprisingly, correlations of psychiatric conditions
frame. As civilians, all three mental health conditions were more withovertlyself-destructivebehaviors(i.e.,self-harmandsuicide
commonamongnon-deployersthanamongdeployers,8.30(cid:4)c2(1, attempts) were generally stronger than correlations with risky
N¼2116)(cid:4)34.13,p<0.01,0.06(cid:4)F(cid:4)0.12.Differencesbetween behaviorsthatarenotovertlyself-destructive.
deployers and non-deployers during military service but before To directly examine the hypothesis that psychiatric problems
deployment were similar to civilian differences, but smaller in mediate the associationbetweendeployment andrisky behavior,
magnitude. Within this time frame, non-deployers were signifi- we conducted additional logistic regression analyses paralleling
cantlymorelikelythandeployerstoreportanxiety(p<0.01),and thosedescribedinSection3.4,eachpredictingadifferentspecific
the difference in depression approached significance (p ¼ 0.05). type of current risky or self-destructive behavior. As before, to
However, rates of PTSD did not significantly differ (p > 0.4; see control for possible demographic differences between groups,
Table 5). The pattern of differences between deployers and non- demographicswereenteredonthefirststep.Onthesecondstep,
deployersduringthecurrenttimeframewasacompletereversal weenteredpriorengagementinthattypeofriskybehavior,aswell
ofthepatternobservedduringcivilianlife(and,toalesserdegree, as psychiatric condition (one variable representing prior psychi-
military/pre-deployment life); that is, deployers reported signifi- atric problems and one representing current problems). On the
cantly higher rates of all three conditions than non-deployers thirdandfourthsteps,weentereddeploymentandtheinteraction
during the current period, 14.70 (cid:4) c2 (1, N ¼ 2116) (cid:4) 78.23, of deployment (cid:5) prior behavior, respectively. To the extent that
p<0.01,0.08(cid:4)F(cid:4)0.19(seeTable5). psychiatric problems mediate the effects of deployment on risky
McNemar change tests were used to examine whether behavior,wewouldexpecteffectsofdeploymentonriskybehavior
deployers were more likely to report each condition following tobeweakerwhenhistoryofpsychiatricconditionsiscontrolled
deployment than before deployment (i.e., civilian and military/ thanwhenitisnot.
Table7
Resultsoflogisticregressionanalysespredictingcurrenthigh-riskbehaviors(adjustedoddsratios).
Predictora Typeofhigh-riskbehavior
Risky Unprotected Drug Self-harm Suicide
recreation sex use attempt
Priorbehavior 4.70*** 3.17*** 5.32*** 7.19*** 6.29***
Priorpsychiatriccondition 1.16 1.39* 1.16 1.11 0.73
Currentpsychiatriccondition 4.79*** 2.80*** 2.48*** 4.29*** 14.00***
Combatdeployment 1.64*** 1.06 1.90** 1.52* 0.50
Priorbehavior(cid:5)deployment 2.50*** 1.18 1.33 2.09* 2.09
*p<0.05.**p<0.01.***p<0.001.
a Demographicvariableswereenteredonthefirststepandarenotreportedhere(seeTable4);priorbehaviorandpriorandcurrentpsychiatricconditionwereenteredon
Step2;combatdeploymentwasenteredonStep3;theinteractionwasenteredonStep4.ORsgreaterthan1.0indicateriskfactorsandORslessthan1.0indicateprotective
factorsforhigh-riskbehavior.ORsarereportedforthesteponwhichthepredictorenteredthemodel.USN¼Navy;USMC¼MarineCorps.
1328 C.J.Thomsenetal./JournalofPsychiatricResearch45(2011)1321e1331
ResultsoftheseanalysesareprovidedinTable7.Becausethey engaginginunprotectedsexwithsomeoneotherthanone’sregular
havebeenprovidedpreviously(seeTable4)andareunalteredin partner and participating in recreational activities that carried
thepresentmodel,resultsfor thedemographic variablesentered a risk of severe injury or death; each of these was reported by
onthefirststeparenottabled.AcomparisonofTables4and7with approximately half of participants. Illegal drug use also was
respect to the effects of prior behavior, combat deployment, and common,reportedbynearlytwoinfiveparticipants.Self-injurious
their interaction reveals that controlling for prior and current behaviorwasreported byone in five,andattempted suicidewas
psychiatricconditionshadlittleimpactonthepatternofsignificant reportedbyoneinfourteenparticipants.
effects: (1) prior behavior remained a strong and significant Compared to the USMC population as a whole, the present
predictorofeverytypeofcurrentbehavior,evenaftercontrolling samplewassomewhatyounger,lowerinrank,andmorelikelyto
for psychiatric problems; (2) combat deployment remained beenlisted(vs.officers).Thispatternofnon-representativenessis
asignificantpredictorofthesamethreeoutcomesasbefore(risky commoninresearchsamplingmilitaryunitsduringtimesofhigh
recreation,druguse,andself-harm);and(3)theinteractionofprior operational tempo (e.g., see Hogeetal., 2004), and likely reflects
behavior and deployment remained significant for the same two the fact that higher-ranking personnel are more likely than their
outcomes as before (risky recreation and self-harm). Thus, there lower-ranking counterparts to be occupied with other activities
waslittleevidencethatpsychiatricproblemsmediateddeployment duringscheduledresearchtimes.Giventhatriskybehaviorismost
effectsonriskybehavior.Nonetheless,havingapsychiatriccondi- common among young people, the present results may over-
tionduringthecurrenttimeframewasstronglyassociatedwithall estimateratesofriskybehaviorintheUSMCasawhole.However,
fiveoutcomes.Incontrast,havingapreviouspsychiatriccondition thisbiasmaybecounterbalancedbythelikelihoodthatpersonnel
was a significant predictor of only one current outcome (unpro- whoengageinthehighestlevelsofriskybehaviorandthosewith
tectedsex). the most severe psychiatric problems may be unavailable for
participationduetoattrition,disciplinaryaction,orinjury,which
4.Discussion would reduce the apparent prevalence of risky behavior in the
sample,relativetoitstrueprevalenceintheUSMC.
The present study is the first, to our knowledge, to compare We found that rates of all risky behaviors were considerably
ratesofabroadrangeofriskyorself-destructivebehaviorsacross higherduringcivilianlifeanddeclinedsignificantlyuponentering
deploying and non-deploying military personnel and, among themilitary.Giventhatmostoftheriskybehaviorsconsideredhere
deployers, before versus after deployment. More specifically, we are more common at younger ages (Briere and Gil, 1998;
examinedtheeffectsofdeploymentonratesoffivetypesofriskyor Zuckerman, 2007), this decline may be partially attributable to
self-destructivebehaviorthatseemquitedisparateonthesurface. maturationeffects.Italsoislikelythatdifferencesbetweenmilitary
Relationships between some of these behaviors have been and civilian environments contribute to reductions in risky
demonstratedinthepast(e.g.,MuehlenkampandGutierrez,2007; behavior over time. For example, there is less tolerance of some
Zuckerman,2007).However,wecouldlocatenopublishedresearch types of risky behaviors in the military than in civilian environ-
thathasexaminedassociationsamongsuchabroadconstellationof ments. With respect to illegal drug use in particular, aggressive
risky behaviors. Although all five behaviors were significantly militarypolicieshaveledtodramaticreductionsoverthepastthree
associated, factor-analytic results suggested that they constitute decades(seeBrayetal.,2006;Brayetal.,2003).Itisthereforenot
two distinct clusters,one representing deliberateself-harm (self- surprisingthatillegaldruguseshowedthebiggestdeclinesfrom
harm and suicide attempts), and the other representing risky civilian to military life, from 39% to 5%. Similararguments might
behaviors that may or may not be motivated by self-destructive apply to reductions in other risky behaviors upon entering the
impulses(unprotectedsex,riskyrecreationalactivities,andillegal military. Of course, the militaryand civilianworlds may differ in
druguse). Nonetheless,becauseeachtype ofbehaviorisconcep- numerous other respects, as well, which also may influence the
tually and practically distinct, we analyzed the effects of combat likelihood of risky behavior. For example, self-harm and suicide
deployment on individual behaviors rather than on these two attempts might decrease upon entering the military because the
underlyingfactors. militaryprovidesanincreasedsenseofpurpose,greaterstructure,
Insomecases,demographiccorrelatesofdifferenttypesofrisky ormoresocialsupportthantheaverageciviliancontext.
behaviorweresimilar.Forexample,marriedpersonnelandthosein Ascivilians,non-deployersweresignificantlymorelikelythan
the Navy (vs. Marine Corps) were generally at lower risk for deployerstoreportfouroffiveriskybehaviors(theexceptionbeing
engaginginalltypesofriskybehavior.Inothercases,associations suicideattempts)andallthreepsychiatricconditions.Thesefind-
with demographic factors varied across different types of risky ings are consistent with Haley’s (1998) notion of the “healthy-
behavior. For example, although higher ranking personnel, those warrioreffect”.Thatis,inanatmosphereinwhichdeploymentis
whowereolder,andthosewithmoreeducationgenerallywereless thenorm,thosewhodonotdeployarelikelytohavephysicalor
likely to report all types of risky behaviors, these characteristics psychological problems that prevent them from doing so. (For
wereunrelatedtothelikelihoodofunprotectedsex.Patternsofsex additional evidence supporting this effect see Bell et al., 2010.)
differences also varied across the risky behaviors examined. However, our results indicate that pre-deployment differences
Consistent with previous research, men were more likely than betweendeployersandnon-deployersinbothriskybehaviorand
womentoreportriskyactivitiesassociatedwithsensationseeking psychiatric problems were most pronounced during the civilian
(riskyrecreationalactivities,unprotectedsex,andillegaldruguse; timeframe.Paradoxically,then,thepresentdatasuggestthat,with
Zuckerman, 2007), but were less likely than women to report respect to risky behavior and psychiatric problems, the “healthy-
attemptedsuicide(Zhangetal.,2005).Therewasnosexdifference warrioreffect”maybesubstantiallyattenuatedduringmilitarylife,
inthelikelihoodofself-harm.Althoughsomestudieshavereported relativetocivilianlife.
thatwomenaremorelikelythanmentoengageinself-harm(e.g., Both within-subjects and between-group analyses revealed
Whitlocketal.,2006),others(e.g.,BriereandGil,1998),including significant effects of deployment in increasing the likelihood of
apreviousstudyofmilitaryrecruits(Klonskyetal.,2003),havenot. self-harm and risky recreation. In addition, between-groups (but
In the present sample, nearly three-quarters of respondents notwithin-subjects)analysesindicatedsignificantincreasesinthe
reportedthattheyhadpreviouslyengagedinatleastonetypeof oddsofillegaldruguseamongdeployersrelativetonon-deployers,
riskybehavior.Themostcommonlyreportedriskybehaviorswere even aftercontrolling forpriordrug use. The magnitude of these
C.J.Thomsenetal./JournalofPsychiatricResearch45(2011)1321e1331 1329
changes was small relative to the changes that were observed Therefore,although23%mayoverestimate thenumberofpersonnel
betweencivilianandmilitary(pre-deployment)life(seeFig.1).Itis whohavebeendiagnosedwithoneofthesethreepsychiatricprob-
importanttonote,however,thatanyincreaseinriskybehaviorover lems, it may underestimate the number of personnel who suffer
timeoccursinoppositiontothegeneraldeclineinriskybehavior symptomsassociatedwithoneofthesedisorders.
that accompanies the aging process (Briere and Gil, 1998; Preliminary analyses yielded evidence consistent with a medi-
Zuckerman,2007).Becauseriskybehaviorisexpectedtodecrease ating role of psychiatric conditions in the association between
overtime,anyincreaseeevenofapparentlysmallmagnitudeeis deploymentandriskybehavior.First,asforriskybehavior,deploy-
noteworthy. mentwasassociatedwithincreasesinpsychiatricproblems.Specif-
Ofthetwotypesofovertlyself-destructivebehaviorexamined, ically,duringthecurrenttimeframe,deployersweremorelikelythan
self-harm was significantly impacted by deployment whereas non-deployerstoreportallthreepsychiatricconditions,andamong
suicide attempts were not. Perhaps effects of deployment on deployers,thelikelihoodofdepressionandPTSD(butnotanxiety)
suicide-related thoughts and behaviors would have been more increasedsignificantlyfrompre-topost-deployment(cf.Hogeetal.,
readilyapparent if we had assessed milder (and more prevalent) 2004).Inaddition,psychiatricproblemswerepositivelyassociated
forms such as suicide-related ideation. Similarly, the inconsistent withriskybehavior.However,directtestsdidnotsupportamediating
effectsofdeploymentonillegaldrugusemayreflectthefactthat roleforpsychiatricproblems.Psychiatricconditionse particularly
this is an uncommon behavior, at least during military life. thosewithinthecurrenttimeframeewereassociatedwithincreased
However, reduced statistical powerresulting fromlowbase rates likelihoodofriskybehavior,aboveandbeyondtheeffectsofprior
cannotaccountforthenullfindingregardingdeploymenteffectson risky behavior. Interestingly, this was true for both overtly self-
engaginginunprotectedsexwithsomeoneotherthanone’sregular destructivebehaviorsandbehaviorsthatwerelessobviouslyinten-
partner,asthiswasthemostprevalenttypeofriskybehaviorinthe ded to harm the self. However, controlling for prior and current
present sample. It is not clear why risky sexual behavior, unlike psychiatric conditions did not substantially reduce the impact on
risky recreational behavior, was not influenced by deployment. currentriskybehaviorofeitherthemaineffectofdeploymentorthe
However,onemightspeculatethatmotivationsunrelatedtorisk- deploymentbypriorbehaviorinteraction.
taking (e.g., sexual attraction, desire for intimacy) play a larger Asnotedpreviously,ourmeasuresofpsychiatricconditionsare
role in sexual behavior than in the other types of risky behavior imprecise.Similarly,ourmeasuresofriskybehaviorsareeachbased
examined. Also, unprotected sex is only one type of risky sexual on only one or two items. Reliability and precision would be
behavior. In a review of characteristics predictive of sexual risk- enhancedbyincludingmultipleitemstoassesseachconstruct.In
taking, Hoyle et al. (2000) found that unprotected sex was more addition, there is causal ambiguity in interpreting associations
weaklyrelatedtosensationseekingthanothermeasuresofsexual between current risky behaviors and current psychiatric condi-
risk-taking,suchasnumberofpartnersandriskyencounters(e.g., tions. Although this association could indicate that psychiatric
sex with strangers). Perhaps other types of risky sexual behavior issuesareimplicatedincausingriskybehavior,itisequallypossible
aremorestronglyinfluencedbydeployment. that engaging in risky behaviors is a marker of psychiatric prob-
One important finding of the present research is that not all lems. That is, people who report engaging in risky, or especially
individuals were equally likely to demonstrate increases in risky self-destructive, behaviors may be most likely to be labeled by
behavior following deployment. As predicted,deploymenteffects medicalormentalhealthprofessionalsashavingpsychiatricissues.
onriskybehaviorswerelargelyrestrictedtoindividualswhohad More generally, causal conclusions must be temperedbythe fact
engagedinriskybehaviorsinthepast.Infact,deploymentdidnot thatthepresentresearchiscross-sectional.Althoughourinclusion
haveasignificanteffectonanyofthefiveriskybehaviorsexamined of both within-person and between-group comparisons and the
amongindividualswithnopriorhistoryofengaginginthattypeof replication of effects across both types of comparisons enhances
riskybehavior.Incontrast,deploymentsignificantlyincreasedfour thevalidityofthepresentconclusions,onlylongitudinalresearch
of thefiveriskybehaviors(theexceptionbeingsuicideattempts) canyieldfirmcausalconclusions.
amongindividualswhohadengagedinthatbehaviorinthepast. Inaddition,itisimportanttoconsiderthepossibilitythatresponse
Thus, deployment does not appear to introduce new risk-taking biasesinfluencedthepresentresults.Thecurrentstudyreliedupon
behaviorsamongthosewhohadnotengagedinthempreviously, retrospective accounts of risky behaviors as well as psychiatric
butrathertoincreaseengagementinriskybehaviorsamongthose conditions.Themostobviousretrospectivebiaseunderreportingdue
whohadalreadyengagedinthem.Thisfindingisconsistentwith to forgetting e cannot account for the present results, given that
the proposition that previous engagement in risky behavior may reportsofriskybehaviorswerehighestfortheearliestperiodexam-
serveasamarkerforproblemsorvulnerabilitiesthatarelikelyto ined(i.e.,asacivilian).Anotherpossibilityisthatpsychiatriccondi-
beexacerbatedbydeployment. tions may distort memory, such that individuals suffering from
Inadditiontoexaminingeffectsofdeploymentonriskyandself- depressionorPTSDmorereadilyrecalltraumaticevents.However,
destructivebehaviors,thisstudyconsideredapotentialmediatorof findingsregardingthishypothesishavebeeninconsistent(Fergusson
these associations: psychiatric problems, including depression, et al., 2000; Wilson et al., 2008); moreover, generalizing this
anxiety,andPTSD.Nearlyoneinfour(23%)ofrespondentsreported hypothesis to encompass the recall of risky behaviors is not
thattheyhadbeentoldbyamedicalormentalhealthprofessional straightforward.Italsoispossible thatrespondentsunderreported
that they had one of these disorders. Depression was the most theirriskybehaviorsandpsychiatricproblemsduetoconcernsthat
commonlyreportedproblem(17%),followedbyanxiety(14%),and such disclosuresmight negatively impacttheircareers. We donot
PTSD(7%).Despitethefactthatrespondentsmayunderreportnega- thinkthis is likely, however.All personnelwere assuredthattheir
tiveeventsduetosocialdesirabilityconcerns,wesuspectthatthis responseswereanonymousandsurveysweremailedtouniversity
figuremayoverestimatethepercentageofpersonnelwithpsychiatric researchers for processing, rather than being collected by military
disorders.Thismaybethecasebecausepeopleeincludinghealth personnel.Previousresearch(Olsonetal.,2004)hasdemonstrated
careprofessionalseoftenusethetermsdepression,anxiety,andPTSD thatself-reportsofnegativeevents(specifically,childhoodabuse)are
inanontechnicalway,inreferencetosymptomsthatdonotmeet atmuchhigherlevelsundertheseconditionsthaninofficialsurveys
diagnosticcriteria.Atthesametime,however,peoplewhoexperience thatwouldbecomepartoftherespondent’smilitaryrecord.
significantsymptomsofthesedisordersdonotnecessarilycommu- The present results suggest that post-deployment increases in
nicatethosesymptomstoamedicalormentalhealthprofessional. psychiatricproblemscannotaccountforincreasesinriskybehavior