MILITARY MEDICINE, 173. 6:544, 2008

The Prevalence of Overweight and Obesity among U.S.Military Veterans

LCDR Nathaniel Almond. MC USN'; Leila Kahwati, MD MPHf; Unda Kinsinger, MD MPHf;Deborah Porterfield, MD

ABSTRACT Ovei^eight and obesity are increasingly contributing to disease burden among military populations. Thepurpose of this study was to calculate and examine Ihe prevalence of overweight and obesity among ilie veteranpopulation. Data were obtained from the 2(104 Behavioral Risk Facior Sur\eiilance Systctn. Overweight (body massindex 2:25 kg/m-) prevalence in veterans was ITt.y'A (SE. {)A%) for males and 53.6% (SE 1.7', ) for females. Obesity(body mass index 2:30 kg/m') prevalence in veterans was 25.3'7f (SE. 0.49f) for males and 21.2% (SE. \A^/() lorfemales. After adjusting for sociodemographics and health status, veterans were no more likely to be overweight (oddsratio, 1.05; 95% conlidence interval. 0.99-1.11) or obese (odds ratio 0.99: confidence interval. 0.93-1.05) thannonveterans. Despite previous participation in a culture and environment that selects for and enforces b(xly weightstandards, veterans have a high prevalence of overweight and obesity that is similar lo general population estimates.

INTRODUCTIONIncreases in the prevalence of overweight (body tnass index|BMI] >25 kg/m') and obesity (BMI >30 kg/nr) in theUnited States since I960 are well known.' ' Clinical exami-nation data from Ihe National Health and Nutrition Exami-nation Sui-vey (NHANES) found the prevalence of over-weight in U.S. adults increased from 45% in I960 to l%2 to66% in 2003 to 2004, while obesity prevalence increasedfrom 13% in 1960 to 1962 to 32% in 2003 to 2004.'^^Obesity and overweight are associated with increased mor-bidity and mortality as well as increased economic burden tosociety. The mortality attributed to obesity has been estimated tobe between 111.919 and 365,(XX) deaths annually.'-" Comorbidconditions associated with obesity include hypertension, dyslip-idemia. stroke, gallbladder disease, diabetes, coronaiy heart dis-ease, and osteoaiihritis. as well as breast, prostate, colorectal,gall bladder, and endometrial cancer.'" The economic cost ofobesity exceeds $90 billion dolkirs annually."

The epidemic of obesity significantly affects the military.First, the potential pool of recruits is decreased due to theincreasing proportion of young adults who do not meet mil-itary entry standards for weight, estimated at 13 to !8% ofU.S. men and 17 to 43% of U.S. women in the generalpopulation.'- Retention of active military personnel is alsodecreased secondary to the disease burden, with 1,419 per-

*NEPMU Five. Navat Staiion Sun Diego. 3233 Albacore Alley, SanDiego. C A 92136.

" VA National Cenlt r for Health Pritmoiion and Disease Prevenlion(NCP). 3022 Croasdaile Drive. Suite 2(K}. Durham. NC 27703.

:[;Deparlnienl of Soci;il Medicine. Scliool of ML'ilicine. University ofNorth Carolina al Chapel Hill. 121 MucNider. CB No. 7240. Ch ipcl Hill. NC

The views expres.sed in ihis article are those of ihe authors and do nofneces.sarily reHect the official policy or position ot" the Deparlnient of theNavy, Depanment of Defense. DepartmeiK of VA. or the U.S. governmeni.

This maniiscripi was received for review in July 2007. The revisedmanuscript wa.s accepted for publication in March 2(X)8,

soimel discharged in 2002 due to failing the body weightstandard.'^ Lastly, overweight and obesity add to health carecosts for the Department of Defense, whose total health carebudget is currently estimated at $36 billion with projectedcosts in 5 years to be $61 billion annually.'•*

The high prevalence of overweight and obesity has beenobserved in U.S. militiiry active duty populations." The prev-alence of overweight based on self-reported height andweight of the -~1.4 million''' active duty personnel in 2002was 62% in males, 32% in females, and 57% overall."' Thesenumbers are surprising given physical fitness and body fatstandards reqtiired of military service members. One expla-nation may be the inability of BMI, which is a proxy foradiposity, to distinguish "overfat" individuals irom very ath-letic and heavily muscled individuals in the BMI range of 25t(j 29.9 kg/tn-. The prevalence of obesity among active dutyduring this same time period was 10% in males, 4% infemales, and 9% overall.'^ These data, which show that mostof the overweight prevalence was in the BMI range of 25 to29.9 kg/m-. suggest that mu.scle mass might be contributingto the high number of overweight active duty members.

Multiple studies have established overweight and obesityprevalence in the veteran population. The prevalence of over-weight for the over 5 million veterans who use outpatientVeterans Health Administration (VHA) facilities was 73%among men and 68% among women in 2(M)0. based onmeasured heights and weights in the medical record."* Obe-sity prevalence in this sample was 33% among men and 37%among women. Additionally, the overweight and obesityprevalence of VHA users was determined lo be 69 and 25%,respectively, using Behavioral Risk Factor Surveillance Sys-tem (BRFSS) 2000 data.'" Lastly, the prevalence of over-weight and obesity based on a 1996 cross-sectional survey ofVHA outpatient clinics found an overweight prevalence of75% overweight and an obesity prevalence of 34%.-*'

544 MILITARY MEDICINE, Vol. 173. June 2008

Prevalence of Overweight ami Obesity in U.S. Veterans

Tbe prevalence of overweight and obesity cannot be usedas an estimate for the entire veteran population because VHAusers are generally older, have poorer health status., and havelower income than Ihc overall veteran population.-' A recentstudy of BRFSS data from 2003 demonstrated that veteranswho did not use VHA facilities had an overall overweightprevalence of IVYc and an overall obesity prevalence of2A%p- Although this study was the first to address diseaseburden in veterans who did not use VHA facilities, it did notreport an overall prevalence of disease burden in the entireveteran population or report prevalence by gender.

In another study. 38- to 64-year-oid military health care-eligible individuals comprised of retirees and their spouseswere surveyed to detemiine their prevalence of overweight.^^This prevalence was 80% in males. 60% in females, and 70%overall. In this study, the majority of males were retirees,while the females were predominately spouses of the retirees.The limitations of not knowing whether respondents wereveterans or spouses of veterans and of sampling only respon-dents ages 38 to 64 prohibits drawing definitive conclusionsregarding all veterans, but it does suggest that a high percent-age of retired veterans are overweight.

Estimates of the previilence of overweight and obesity for theoverall U.S. military veteran population, which numbers 26.4million—or 12.7% of the U.S. population 18 or older^^—areunknown. Determination of overweight and obesity prevalencefor all U.S. veterans is needed for several reasons. First, knowl-edge of overweight and obesity prevalence has implications forprimary preventive efforts during the active duty periixi as ameans to prevent subsequent morbidity. Next, the prevalencecould ser\'e as a benchmark comparison for VHA and militaryretiree planners as well as other policy makers who are exam-ining smaller subsets of the veteran population. Lastly, as vet-erans may re-enlist or be recalled to duty, they constitute avaluable potential resource for the nation's defense and knowl-edge of their disease burden would inform military prepared-ness. The aims of this study were: (I) to detemiine the overallprevalence of overweight and obesity among all U.S. veteransby gender and age groups using self-reported BM! data from the2O(>4 BRFSS and (2) to compare the prevalence of overweightand obesity between the veteran and nonveteran population.

METHODS

Data Source and Study PopulationData used were from the 2(K)4 BRFSS, a monthly conducted.annually aggregated and reported, state-based, cross-sectional,random digit-dialed telephone survey, conducted by state healthdepiutments with assistance fmm the Behavioral SurveillanceBranch of the Centers for Disease Control and Prevention {furtherinformatiftii on BRFSS methcxiology exists at their web site:www.cdc.gov/brfss/technicaLinl\xlata/surveydata/2004.htm). TheBRFSS is the largest continuously conducted telephonehealth survey in the world. The BRFSS's annually revisedquestionnaire is designed to measure behavioral risk factors

In a representative sample of U.S. noninstitutionalized indi-viduals ages 18 and over. Information is collected by all 50state health departments using a disproportionate siratifiedsampling design. Data are weighted to represent populationestimates for each state. The BRFSS questionnaire includescore questions that are asked by every state, in addition tomodule questions that are used at each state's discretion.

Measures

Those who answered affirmatively to ever serving on activeduty in the U.S. Armed Forces, either in the regular militaryor in a National Guard or military reserve unit and whoindicated that they were retired, medically discharged, ordischarged from military services were considered to be"veterans" for this analysis. All other respondents were con-sidered nonveterans. This classilication is consistent with theU.S. census definition of a veteran. Self-reported height andweight were used to calculate BMI (weight in pounds/(heightin inches)-) x (703). Overweight and obesity prevalencedetermined by BMI classification was the primary outcomemeasure. BMI was classified as (I) normal or underweight(BMi <25 kg/nr). or (2} overweight but not obese. (BMI^25 kgMr but <30 kg/m-), or (3) obese (BMI >30 kg/m-),or (4) overweight (BMI ^25 kg/nr). These classllications foroverweight and obesity are according to the guidelines pub-lished by the National Heart. Lung, and Blood Institute.'*^

Other BRFSS variables were considered as possible con-founders of the relationship between veteran status and over-weight or obesity status. The variables included gender, age,race/ethnicity (Caucasian. African American, Hispanic, ormultiracial and not Hispanic), marital status, education (didnot graduate high school, graduated high school, attendedcollege or technical school, graduated college or technicalschool), annual household income (five levels), self-reportedhealth status (dichotomi/.ed as good to excellent and fair topoor), and smoking status (never, former, or current).

AnalysisThe analysis included only respondents who answered boththe veteran status questions and the self-reported height andweight questions (287,467 respondents of the total 303,882surveyed). STATA 9.0 (StataCorp LP. College Station.Texas) survey commands were used in all analyses to accountfor BRFSS's complex multistage cluster sampling surveydesign. Frequencies were calculated for various demographicsof both the veteran and nonveteran populations. Crude preva-lence of overweight but not obese (BMI ^25 kg/m- ai\d <30kg/m-), obese (BMI ^30 kg/m^), and overweight (BMI s25kg/m-) were detennined for veterans and nonveterans usingproportion estimation comniiuids for these weight categories.These resuhs were further stratified by age and gender.

To further examine the relationship between veteran statusand overweight and obesity, two logistic regression modelswere used to calculate adjusted prevalence odds ratios (ORs).In the first model, the dependent outcome was overweighl.

MILITARY MEDICINE, Vol. 173. June 20flS 545

Prevalence of Overweight and Obesity in U.S. Veterans

categorized as BMI ^25 kg/m- or <25 kg/m-. In the secondmodel, the dependent outcome was obesity, categorized asBMI >30 kg/m- or <3() kg/m-. Covariates that have beendemonstrated in previous studies to be related to either vet-eran status and/or to obesity were entered in the model, aswere covuriates that in exploratory analysis suggested possi-ble confounding. Interaction between veteran status and gen-der was assessed for inclusion in the models using the Waldtest. Variables that were not found to confound the relation-ship between veteran status and overweight and obesity wereremoved to arrive at a final model for the relationship be-tween (1) veteran status and overweight and (2) veteran statusand obesity. This study was determined to be exempt fromreview by the University of North Carolina School of PublicHealth Institutional Review Board.

RESULTSSubjects included 287,467 respondents who replied to height,weight, and veteran status questions; of these. 39.627 (13.8%)indicated that they were veterans. Table 1 displays populationestimates for various characteristics of veterans and nonvet-erans based on this sample. The veteran population wasprimarily male (94%), married (72%). older than 35 years(929!). and Caucasian (81%). More than 93% of veterans hadat icasl a high school education, and >43% of veteransreported incomes of >S50.000 dollars per year. Also, >80%rated their health as good, very good, or excellent. Moreveterans reported former smoking (43%) than never smoking(36%). The nonveteran population included a larger percent-age of females, non-Caucasians, and younger age groups thanthe veteran population.

Prevalence of Overweight and ObesityTable II describes prevalence of overweight but not obese,obese, and overweight in veterans and nonveterans by gender.Among veterans, prevalence of overweight but not obese washigher in males (48%) than females (32%); however, theprevalence of obesity was similar (25% in males, 21% intemales). Becau.se of the difficulty in distinguishing betweenbody fat and muscle tissue for BMIs in the 25 to 29.9 range,further investigation was limited to obesity alone.

The prevalence of obesity stratified by age and gender isdisplayed in Figure 1. The prevalence of obesity by age groupand gender followed a similar trend in both veterans andnonveterans with increasing prevalence as age increases up toage 2:65 where prevalence then declines. The prevalence ofobesity among male veteran ages 35 to 54 was slightly higherthan in nonveterans in this age group. Similar prevalence wasfound in all other age groups. The prevalence of obesityamong male veteran obesity prevalence was highest in theage group 45 to 54; among male nonveterans, the prevalenceof obesity was highest in the 55 to 64 age group.

Due to smaller sample sizes, particularly in the older agecategories, the prevalence of obesity in female veterans by

TABLE I. Population Estimates of Selected Characteristics forVeterans and Nonveterans

Characteristic

GenderMaleFemale

Age (in groups)[8-2425-3435^1445-5455-64^65

Marital statusMarriedNot murried

RaceCaucasianAfrican AmericanHispanicMulliracial and

non-HispanicEducation

Did nol graduatehigh sL-hoiil

Graduated highschool

Attended college ortechnical

Graduated college ortechnical

.Annual householdincome

<$13.lH)US!5.<X)O-S25.OOOS25.(XX)-<$35.(HH)S35.O()O-<S5O.OO()$50.(XX) or more

Self-reported healthstatus

Good, very good, orexcellent

Fair or poorSmoking status

Never smokedFormer smokerCurrent .smoker

SampleNo."

287.467

287.467

286.767

285.031

284.043

252.127

286,600

286.618

Veterans'*Percent (% SE)

94.1(0.2)5.9 (0.2)

1.4(0.1)7.OfO.3)

12.1 (0.3)16.4(0.4)23.9 (0.4)39.0 (0.5)

71.5(0.4)28.5 (0.4)

81.2(0.5)8.6(0.3)5.5 (0.3)4.8 (0.3)

6.4 (0.2)

29.4 (0.4)

29.9 (0.5)

34.2(0.5)

6.8(0.3)16.2(0.4)14.0(0.4)19.3 (0.4)43.6(0.5)

80.2 (0.4)

19.8 (0.4)

36..1(O.5)43.2 (0.5)20.4 (0.4)

Nonveterans'Percent (% SE)

43.0 (0.2)57.0 (0.2)

14.9(0.2)19.8(0.2)21.5(0.2)19.0(0.2)11.3(0.1)l3.4fO.I)

57.0(0.2)43.0 (0.2)

69.0(0.2)10.0(0.1)14.9(0.2)6.2(0.1)

12.4(0.2)

29.9(0.2)

26.0(0.2)

31.7(0.2)

12.3(0.2)17.8(0.2)13.2(0.2)16.1 (0.2)40.7 (0.2)

84.5 (0.2)

15.5(0.2)

58.3 (0.2)20.6 (0.2)2M (0-2)

Data from the 2004 Behavioral Risk Factor Surveillance System."The sample population (N) varies according to numher of respondentswho answered each question." N = 39.627.' ^' = 247.840.

age groups was characterized by larger SEs. The prevalenceof obesity in female veterans ranged from a low of 6% in the18 to 24 age group to a high of 34% in the 55 to 64 age group.This range was wider than in nonveterans (15-30% in thesame age groups). Obesity prevalence among female veteransand nonveierans showed increasing trends with increasingage, just as was seen in males.

546 MILITARY MEDICINE, Vol. 173, June 2(H)8

Prevalence of Ovenveight ami Obesity in U.S. Veieranx

TABLE II.

MalesI A' = 114.805)Females[N = 172.662)Overall(A' = 287.467)

Prevalence of Overweight but

7o Overweighl hut tiotObese" (

Veteran

48.0(0.5)

32.4 (1.6)

47.1 (0.5)

'/c SE)

Nonveteran

42.8 (0.4)

29.4 (0.2)

35.2 (0.2)

TUM Obese, Obese, and

% Obese'

Veteran

25.3 (0.4)

21.2(1.4)

25.1 (0.4)

Ovei'wetghi by

{% SE)

Nonveteran

23.5 (0.3)

23.0 (0.2)

23.3(0.2)

Veteran Status and Gender

% Overweighr (% SE)

Veteran

73.3(0.4)

53.6(1.7)

72.2 (0.4)

Nonveteran

66.4(0.3)

52.4 (0.2)

58.410.2)

Data IVum 2004 Behavioral Ri.sk Faclor Surveillance System." BMI >25 kg/m- and <30 kg/m'."BMI >30kg/m-.• BMI ^25 kg/ml

A • male veterans |

Dmate non-veterans'

18-24 65+ Overall

B • female veterans

D female non-veterans

18-24 25-34 35-44 45-54

Age

55-64 65+ Oerall

FIGURE 1. Ohcsily prevalence (and SU) hy veicran slalus and gentler: data from the 2004 Behavioral Risk Factor Surveillance System. (A) Datii fur males.(B) Dalii lor females.

Multivariate AnalysisTwo separate logistic regression models were used to exatn-ine the associLitions between (I) veteran status and over-weight and (2) veteran status and obesity. These modelsincluded veteran status as the independent outcome and eitheroverweight or obe.sity as the dependent outcome. Interactionhetvv'een gender and veteran status was assessed using theWald test in both the overweight and obesity models and nosignificant interactions were found. The following covariatesremained in both models: age. gender, marital status, race.education level, income level, smoking history, and self-ratedhealth status. Using these models, veterans were as likely as

nonveterans to be overweight (adjusted OR. 1.05: 957c con-fidence interval, 0.99, 1.11) or obese (adjusted OR, 0.99; 95%confidence interval. 0.93. 1.05).

DISCUSSIONThis study is the first to report the prevalence of overweightand obesity, based on self-reported height and weights, forthe entii'e U.S. military veteran population (regardless ofVHA user status} and is the first to examine disease burden ina nationally reptesentative veteran population by gender andage. Despite military entry standards for body weight andprevious participation in a culture that emphasized and en-

MILITARY MEDICINE, Vol. 173. June 2(K)S 547

Prevalence of Overweight and Obesity in U.S. Veterans

lurced physical titness, we found thai the prevalence ofoverweight and obesity is high in the veteran population andthai veterans were just as likely as nonveterans to be over-weight or obese. The military experience might be thought toconvey some long-lasting protective benefit against becomingoverweight or obese; however, this study does not supportthat hypothesis.

Previous findings of overweight and obesity prevalence inVHA users are interesting to contrast to this study.'" Both thisstudy and the 2000 study using measured heights and weightsof VHA users found 13% of males to be overweight or obese.However, a greater percentage of veteran males who used theVHA for medical care were obese (339r) than the overallveteran male population in our analysis (25%). This is con-sistent with previous observations that veteran VHA usershave n higher disease burden than veteran non-VHA users.''Female veterans in our analysis had inaikedly lower preva-lence of overweight (54%) and obesity (21%) than previouslyreported veteran females who used the VHA for medical care(68 and 33%. respectively), suggesting that female users ofVHA may be different from female veterans who do notobtain health care in the VHA.

This .study has several limitations. The BRFSS is a self-reported telephone survey; this type of sampling may notfully represent the general U.S. population. Also, the ten-dency to underreport weight and overreport height has beendocumented.'"^ Thus, self-reported BRFSS data tends to un-derestimate the prevalence of overweight and obesity com-pared to NHANES data, which uses measured heights andweights. In a study of BRFSS data from 1999 to 2000, thisunderestimation was 5.1% for overweight and 9.5% for obe-sity.-^ Understanding this underestimation is useful whencomparing data from BRFSS. NHANES, and other sources.

Additional research could further delineate the trajectoryof overweight and obesity and also suggest appropriate agegroups to approach with primary intervention programs. Ourstudy's finding of similar disease prevalence between veter-ans and nonveierans underscores the need for effective pri-mary preventive efforts for active duty members. The periodof active military service is a unique opportunity to studyprimary prevention programs in a coordinated health caresystem. Providing active .service members with life skillsduring their service could prevent futute obesity and comor-bidities associated with obesity, reducing the burden of dis-ease for the Depiu-tment of Veterans Affairs, the militaryhealth care system, and the overall U.S. medical system.Furthermore, it might ensure a more fit and militarily-readypool of individuals in the event of a national emergency thatresults in a recall to duty.

Baseline data on the prevalence of overweight and obesityafter discharge are essential to describing the overweight andobesity trajectory of veterans once they have transitionedfrom active service to the civilian sector. A longitudinal studyof active duty military members, titled the Millennium Co-hort study, is currently underway.-'* This study, which is

collecting self-reported height and weight at various time inter-vals, could further define the trajectory of disease burden in themilitiuy population. Understanding this trajectory may infonnnot only the development of military primary preventive effortsbut also nonmilitar)' behavioral mcxlification programs. Furtherresearch might also evaluate veteran comorbidities associatedwith ovei-weight and oliesity. including hyperlipidemia, hyper-cholesterolemia. and diabetes and the extent to which war-related injur>' or immobility, and mental health conditions im-pact the development of overweight and obesity.

The military experience includes enforcement of a bodyweight standard and physical titness testing with mandatoryminimum physical fitness scores and decreased advancementopportunities for those who fail to meet them. The associationof increased abnormal eating behaviors with the current phys-ical fitness assessment cycle is well known.-'*'' This currentcyclical system, with predictable biannual weigh-ins, doesnot encourage year-round weight standards, nor does it pro-duce veterans who are any better equipped than their civiliancounterparts to maintain a healthy weight. Other reseiuchershave suggested more frequent assessment of weight andfitness to decrease unhealthy eating behavit)r resulting fromthe current system.-'' An alternative solution could be to alterthe assessment protocol to resemble the current military pwl-icy for drug screening.^' Weigh-ins could change from thecurrent cyclical process to become random and unannounced.

It may be unrealistic to expect veterans to switch easilyfrom the military's primarily external motivation to maintaina healthy weight to an internal motivation once the military'sexpectations are removed when the member transitions tocivilian life. A randomly enforced weight standard mightdecrease unhealthy eating behaviors and lead tt> lifelonghabits that a military member could continue after dischargefrom active duty. This approach could emphasize internalmotivation through individual concerns and responsibilityand would place emphasis on year-round maintenance of aconsistent healthy weight instead of on passing or failing aweight standard on a biannual, predictable schedule. Thisarea deserves further study.

REFERENCES1. Hediey AA, Ogden CL, Johnson CL, Carroll MD. Cunin LR, Regal

KM: Prevalence ol'overweighl and obesity among US children, adoles-cetils. and adults. iyy9-2(K)2. JAMA 2()O4; 291; 2847-50.

2. Ogden CL. Fry:ir CD. Carroll MD. Flegal KM; Mean btidy ucijiht.height, and body mu>.s index. United Stiiics 1%O-2(X)2. Adv Daia 2(XM:347; 1-17.

3. Kuczniarski RJ, Flegal KM, Campbell SM. Johnson CL: Increasing prev-alence of overweight among US adulLs: The Naiional Heallh and NutritionExamination Surveys, I960 to 1991. JAMA 1994; 272: 205-11.

4. CDC: Health. United States. 2003; Overweighl. Obesity, and HealthWeight among Persons 20 Years of Age and Over, According to Sex,Age. Race, and Hispanic Origin: United States. 1960-62. 1971-74.i976-S0, I9XS-94. and I999-2O(M). Available at wv^w.cdc.gov/nchs/dala/hus/tab!es/2(X)3/O3h usO68.pdf: accessed March 31. 2(W6.

5. Ogden CL. Carroll MD. Curtin LR. McDowell MA, Tabak CJ, FlegalKM: Prevalence of overweighl and obesity in the United States, 1999-2004. JAMA 2006; 295: 1549-55.

548 MILITARY MEDICINE. Vol. )7?, June 2008

Prevalence of Overweight ami Obesity in U.S. Veterans

6. Flegal KM, Cirauhard Bl. Williatnson DF. Gail MH: Excess deathsassociated with underweight, ovenveight, and obesity. .lAM A 2(X)S; 293:1861-7. 21.

7. Allison DB. Fontaine KR, Manson JE. Stevens J, Vanltallie TB: Anntialdeaths attributable to obesily in the United States. JAMA 1999: 282:l.'i3(l-8. 22.

8. Mokdad AH, Marks JS. Stroup DF. Gerberding JL; Correction: actualcauses i<f death in the United States. 2(KM1. JAMA 2(X)5: 293: 293-4. 23.

9. Mokdad AH. Marks JS. Stroup DF, Gerberding JL: Actual causes ofdeath in the United States. 20(N). JAMA 2004; 291: 1238-4.5.

U). Bray GA; Medical consequences of obesity. J Clin Endocrinol Metab 24.20(W: K9: 2583-9.

11. Finkel.-ilein EA. Fiebelknm IC. Wang G; Nalionul medical spendingattributable to overweight and obesity: how much, and who's paying? 25.Health Air (MillwotKl) 2003; W3(Suppl Web Bxclusive>l: 219-26.

12. Nolte R. Franckowiak SC. Crespo CJ. Andersen RE: U.S. militaryweight standards: what percentage of U.S. young adults meet the current 26.standards? Am J Med 2(«)2; I 13: 486-90.

13. Institute of Medicine (U.S.) Subcommittee tin Military Weighl Manage-ment: Weight Managetnent: State of the Science and Opponunities for 27.Military Programs. Washinglon. DC, Naiional Academies Press. 2<K)3.

14. Matec7.un J: General Remarks by Deputy Navy Surgeon General. Pre-sented at the 45th Occupational Hoallh and Preventive Medicine Con-ference Opening Ceremonies. March 20. 2006. Hampton. VA. 2X.

15. Defense Almanac. Washington, DC. Deparlment of Defense, 2001.16. Bray R. Janice M. Vinous AA, Pemberton MR: 2002 Departmetil of

Defense Survey of Health Related Behaviors among Military Personnel. 29.Available al ww\«..tricare.osd.mii/inain/neu's/DoDSurvey.htm#3.2.6; ac-cessed October i. 2(K)5.

17. Kress AM. Peterson MR. Hartzoll MC: Ass<x;iation between obesity and 30.depressive symptoms among U.S. military aclive duly service personnel.2002. J Psychosom Res 200ft; 60: 263-71. 31.

18. Das SR. Kinsinger LS. Yancy WS Jr. et al: Obesily prevalence amongveterans at Veterans Affairs medical facilities. Am J Prev Med 2005: 28:291-4. 32.

19. Wang A. Kinsinger LS. Kahwaii LC. et al: Obesity and weight controlpractices in 2000 among veterans using VA facilities. Obes Res 2(H)5;13; 1405-11. 33.

20. Nowicki EM. Billington CJ. Levine AS. Hoover H, Must A, Naumova

E: Overweight, obesity, and associated disease burden in Ihe VeteransAfl'airs ambulatory care population. Milit Med 2003; 168: 252-6.Agha Z. Lolgren RP. VanRuiswyk JV. Layde PM; Are patients alVeteran.s Affairs medical centers sicker: a comparative analysis of healthstatus and medical resource use. Arch Iniern Med 20(H): i60: 32.^2-7.Nelson KM: The burden of obesity among a national probability sampleof veterans. J Gen Intern Med 2(H)6; 21: 915-9.Kress .AM. Hart/el MC. Peterson MR: Btirdenol disease associated withoverweight and obesity among U.S. military retirees and their depen-denls. aged 3S-64, 2(X)3. Prev Med 2005; 41: 63-9.Richardstin C. Waldrop J; United States Bureau of the Census: Veterans.2000. Washington. DC. U,.S. Dept. of Commerce Economics and Sta-tislics Adtniiiistration. U.S. Census Bureau. 2003.Naiionai Institutes of Heallh: Clinical Guideiincs on the Idcntihcatmn.B\aluation. and Treatment of Overweight and Obesity in Adults-ThcEvidence Repart. Obes Res 1998; ftlSuppl 21: 5IS-209S.Nawaz H. Chan W, Abdulralintan M. Larson D. Kat/ DL: Self-reportedweight and height: implications for obesity research. Atn J Prc\ Mod2(K11;2O: 294-8 .

Yun S. Zhu BP. Black W. Brownson RC: A comparison of naiiunalesiimytes of obesily prevalence from the behavioral risk factor surveil-lance syslem and the National Heallh and Nutrition ENatnination Survey.Int J Obes (Lond] 2(H)6; 30: 164-70.Gray GC. Chcsbrough KB, Ryan MA. ei ;il: The Millennium CohonStudy: a 21-year prospective cohort study ol 140.000 military pej^sonnel.Milil Med 2(X)2; 167: 483-8.Carltun JR. Manos GH. Van Sl>ke JA: Anxiety and abnormal eatingbehaviors associated with cyclical readiness testing in a naval hospitalactive duty population. Milit Med 2(10.'i: 170: 663-7.McNully PA: Prevalence and contribuiing factors of eating disorderbehaviors in active ditty Navy men. Milii Med 1997; 162; 753-(S.McNiilty PA; Prevalence and contributing factors of eating disorderbehaviors in a population of female Navy nurses. Milit Med 1997; 162:703-6.

McNulty PA: Prevalence and contributing factors of eating disorderbehaviors in active duty service women in the Army. Navy. Air Force,and Marines- Milit Med 2001: 166: 53-S.Military Personnel Drug Abuse Testing Program. DoD Directive lOIO.LWashington. DC. Department uf Defense. December 9. 1994.

MILITARY MEDICINB. Vol. 173. Jtine 2008 549