Barriers to Cardiovascular Disease Preventive Behaviors among OEF/OIF/OND Women and Men Veterans

Abstract

Objective

OEF/OIF/OND Veterans have an elevated risk for developing cardiovascular disease (CVD), but research suggests that engagement in CVD preventive behaviors is low even among at-risk individuals. It is critical to understand barriers to prevention engagement among Veterans to inform the development of tailored interventions addressing barriers and reducing CVD incidence.

Methods

The Women Veterans Cohort Study (WVCS) survey of OEF/OIF/OND Veterans (586 women and 555 men) assessed patient, interpersonal, and systems level barriers to CVD risk prevention. Prevalence of barriers was determined, and chi-squares were conducted to examine sex differences. Multivariate logistic regressions were conducted to determine if sex differences remained when adjusting for demographic factors (age, marital status, education, employment status).

Results

Despite a low response rate (11.5%), endorsement of barriers was high for both women and men, with most (56.8%) not perceiving themselves to be at CVD risk. More men preferred making no lifestyle change (40.9% vs. 29.1%). More women endorsed lack of confidence (42.4% vs. 36.1%), stress (36.9% vs. 27.8%) and depression (36.9% vs. 27.8%), and inadequate social support (26% vs. 20.9%), along with the belief that their clinician does not perceive them as at risk (57.8% vs. 32%) and has not explained CVD preventive behaviors (19% vs. 12.3%). Multivariate analyses reduced statistical significance of sex differences.

Conclusions

Given the low response rate, testing of efforts – e.g., implementation science methods – to assess CVD risk reduction barriers in this population are needed, a task for which the Veterans Healthcare Administration is well suited.

Introduction

Cardiovascular disease (CVD) is the leading cause of death and disease burden globally, affecting an estimated 92.1 million American adults (Benjamin et al., 2017; Vasan & Benjamin, 2016). Although CVD associated mortality has decreased among women and men in the United States overall (Benjamin et al., 2017), mortality has increased among women aged 35–44 years (Ford & Capewell, 2007), the approximate age range for Veterans of Operation Enduring Freedom/Operation Iraqi Freedom/Operation New Dawn (OEF/OIF/OND), post-9/11 Veterans. This cohort has more than twice the women of any previous Veteran cohort, with women representing 18% of those who served (U.S. Congress, 2016). Therefore, understanding CVD risk among these Veterans and identifying sex differences in risk is an important public health issue.

Emerging literature is showing a high prevalence of CVD risk factors for OEF/OIF/OND Veterans. For example, weight gain after separation from the military is common (Haskell et al., 2017), with 65.8% of men and 46.7% of women being classified as overweight or obese (Rosenberger, Ning, Brandt, Allore, & Haskell, 2011). Smoking is also prevalent (Haskell et al., 2017), with nearly 26.2% of women and 42.6% of men classified as current smokers and 12.3% of women and 13.8% of men classified as former smokers (Volkman et al., 2015). Although the rates of traditional CVD risk factors among men Veterans are generally more elevated than among women Veterans, the rates of non-traditional factors associated with CVD risk are higher among women Veterans - e.g., the prevalence of depression approaching 50% and of PTSD over 20% among women (Haskell et al., 2010). During military service, these women are also at risk of encountering a range of traumatic experiences (Haskell et al., 2010), including military sexual trauma, which increases risk of PTSD, other mental health disorders, and chronic health problems (Kimerling, Gima, Smith, Street, & Frayne, 2007). Although research has not shown that reducing or treating depression and PTSD prevents development or modifies risk of CVD (Carney & Freedland, 2017), these factors have been linked to incident CVD (Cohen, Edmondson, & Kronish, 2015; Cohen, Marmar, Ren, Bertenthal, & Seal, 2009; Haskell et al., 2017) and within the OEF/OIF/OND cohort, to early incident hypertension (Burg et al., 2017). Given this constellation of traditional and non-traditional factors, engaging Veterans in behaviors to reduce CVD risk is critical. Yet, doing so requires that they and their healthcare providers appreciate that these efforts are warranted.

While research has shown that engagement in primary and secondary prevention activities can mitigate CVD risk, many individuals fail to engage in these activities. Furthermore, research has shown that there are persistent sex disparities in management of CVD risk among Veterans (Haskell et al., 2014; Vimalananda, Miller, Palnati, Christiansen, & Fincke, 2011; Whitehead, Czamogorski, Wright, Hayes, & Haskell, 2014). For example, despite an increased likelihood of having elevated low-density lipoprotein (LDL) cholesterol, women Veterans are less likely to receive treatment for elevated LDL, or have treatment modified, as compared to men Veterans (Haskell et al., 2014; Vimalananda et al., 2011). This suggests that women Veterans may encounter additional barriers to engagement in CVD preventive actions as compared to men Veterans.

With the unique experiences of military Veterans, the high representation of women among the OEF/OIF/OND Veteran cohort, and the apparent elevated CVD risk in part demonstrated by documented engagement in health risk behaviors, it is essential to understand the barriers to engagement in healthy, CVD risk lowering practices. Therefore, the purpose of this study was to examine barriers to CVD preventive behaviors among Veterans of OEF/OIF/OND, and to identify any differences in these barriers for women and men.

Method

The Women Veterans Cohort Study (WVCS) was established in 2007 (Haskell et al., 2011) with funding from the Veterans Health Administration (VHA) to identify sex specific medical and mental health outcomes of military service in the OEF/OIF/OND era. WVCS includes a geographically representative sample of over 1,000 women and men and an electronic health record sample of all OEF/OIF/OND Veterans registered with the VHA. The current study examines data from the survey only. Wave II of the WVCS survey was initiated in 2014, and included questions concerning engagement in, and barriers to engagement in, CVD risk lowering practices. The purpose of the current study was to examine these barriers endorsed by OEF/OIF/OND Veterans and to identify the barriers unique to women. The study was approved by the Institutional Review Board at VA Connecticut Healthcare System.

Participants

All OEF/OIF/OND women Veterans and a sub-sample of OEF/OIF/OND men Veterans who served between 2001–2014 and received care through the VA Healthcare Systems in New England, Indianapolis, IN, Los Angeles, CA, and Durham, NC (9,912) were eligible for participation in the study. Of these 9,912 eligible OEF/OIF/OND Veterans, 1,145 (586 women, 555 men, 4 participants did not report sex) Veterans completed the baseline and second wave of the WVCS survey between October 2015 and December 2016. This resulted in a response rate of 11.5%. The mean age was 43.86 (±10.92); 23.8% were single and 55.7% were married; 86.7% were white, 10% black/African American, and 8.2% Hispanic; and 31.5% had a bachelor’s/4-year college degree.

Procedure

The Department of Defense Manpower Data Center’s (DMDC) Contingency Tracking System provided the sampling frame for the study to the VHA. All 9,912 eligible OEF/OIF/OND women and men Veterans from the four participating VHA Centers were sent enrollment letters and the survey. Up to three mailings of the survey were sent to those who did not respond to the initial invitation. Veterans who were interested in participating consented to the study, and completed the survey via pen and paper, for which they received $20.

Demographics and Health Information

Demographic information (e.g., age, race, marital status) was collected through surveys. Participants also completed survey questions about health risk behaviors (e.g., smoking status), mental health symptoms (e.g., Patient Health Questionnaire-8 [PHQ-8], Generalized Anxiety Disorder-7 [GAD-7], PTSD Checklist-Military Version [PCL-M], and Insomnia Severity Index [ISI]), and mental health treatment (e.g., treatment for PTSD in the past 12 months), and receipt of treatment for CVD risk factors (e.g., hypertension, diabetes). Measures of mental health symptoms were selected because of their strong psychometric properties. Specifically, the PHQ-8 has shown strong reliability and validity (Kroenke & Spitzer, 2002). Similarly, the GAD-7 has demonstrated strong internal reliability (Cronbach α= .92), test-retest reliability (interclass correlation = 0.83), and construct validity (Spitzer, Kroenke, Williams, & Löwe, 2006). The PCL-M has also shown adequate internal consistency (Cronbach α= 0.75) as well as convergent validity (Wilkins, Lang, & Norman, 2011), while the ISI has shown adequate internal consistency (Cronbach α= 0.74) and content validity (Bastien, Vallières, & Morin, 2001). The phrase, ‘receiving treatment’ in reference to CVD risk factors and mental health treatment was not defined in the survey instructions, and therefore participants may have interpreted treatment to include both professionally assisted treatment and non-professional treatment.

Barriers to a Heart Healthy Lifestyle

A section of the survey included items assessing barriers to leading a heart healthy lifestyle (i.e., barriers to engaging in CVD preventive behaviors). Although several barriers may represent barriers to overall healthcare (e.g., too depressed, too stressed), this survey has previously been used in several studies examining national trends in U.S. women’s perceptions of barriers to engagement in CVD preventive behaviors (Mosca et al., 2013; Mosca, Mochari-Greenberger, Dolor, Newby, & Robb, 2010; Mosca et al., 2006). In addition, several barriers assessed CVD prevention specifically (e.g., I don’t perceive myself to be at risk for heart disease). Furthermore, the instructions to this section of the survey directed participants to respond to the barriers in relation to CVD (i.e., “The following is a list of things some people have said about living a heart healthy lifestyle.”).

Response options ranged from 1 (strongly disagree) to 4 (strongly agree). For analyses, response options were dichotomized into Disagree responses (strongly disagree and somewhat disagree responses) and Agree responses (somewhat agree and strongly agree responses). Barriers were categorized into: patient level – e.g. barriers such as stress and time that are within the patient’s level of control; interpersonal barriers – e.g., obligations that relate to another person; and systems level barriers – e.g., lack of money or healthcare insurance that are indicative of more systemic issues.

Statistical Analyses

Prevalence of barriers to CVD preventive behaviors for the overall sample and separately for women and men Veterans was determined. Independent-samples t-tests were used to examine sex differences in age and mental health symptoms and chi-square analyses were conducted to examine differences in demographic characteristics, treatment for cardiovascular risk factors, health behaviors (e.g., smoking status), mental health treatment, and barriers to CVD preventive behaviors. Cramer’s V (_φc) was also calculated as an effect size. Cramer’s V is an effect size that ranges from .00 to 1.00. Values between .00-.10 show a very weak association, those between.10-.20 show a weak association, those between .20-.30 show a moderate association, and those ≥ .30 show a strong association (Marchant-Shapiro, 2015). Additionally, false discovery rate (FDR) p-values were calculated to adjust for multiple comparisons. Secondary analyses were also conducted to adjust for sex differences in demographic variables. Unadjusted and adjusted (age, marital status, education level, and employment status) logistic regression analyses were used to examine the effect of sex on barriers. All analyses were performed using SAS Enterprise Guide 6.1.

Missing Data

A few variables had missing data > 5% (see Tables 1–3). Cases with missing data were removed from analyses.

Table 1.

Sample Characteristics

	Overall Sample (N= 1145)	Women (n = 586)	Men (n = 555)	p
Age (M, SD)	43.86(SD = 10.92)	41.61(SD = 10.28)	46.21(SD = 11.08)	<0.0001*
Race				0.28
White	935(83.8)	466(79.5)	467(84.1)
Black/African American	109(9.8)	67(11.4)	42(7.6)
Asian	20(1.8)	12(2.1)	8(1.4)
American Indian and Alaska Native	10(0.9)	5(0.9)	4(0.7)
Native Hawaiian and Other Pacific Islander	3(0.3)	2(0.3)	1(0.2)
Other/Mixed	39(3.5)	23(3.9)	16(2.9)
Ethnicity				0.47
Hispanic	91(8.2%)	50(8.7)	40(7.5)
Non-Hispanic	1016(91.8)	523(91.3)	491(92.5)
Marital Status				<0.0001*
Married	638(55.7)	272(46.4)	364(65.6)
Divorced/Separated	231(20.2)	142(24.2)	88(15.9)
Widowed	4(0.4)	3(0.5)	1(0.2)
Single	272(23.8)	169(28.8)	102(18.4)
Education				<0.0001*
None or < HS	1(0.1)	1(0.2)	0(0)
HS or GED	234(20.7)	84(14.4)	149(27.2)
Associate’s/Junior/2-year college	271(23.9)	145(24.9)	124(22.7)
Bachelor’s/4-year college	357(31.5)	208(35.7)	149(27.2)
Graduate/Professional	269(23.8)	144(24.7)	125(22.9)
Employment				<0.0001*
Employed	694(60.7)	330(56.3)	362(65.5)
Unemployed	75(6.6)	37(6.3)	36(6.5)
Homemaker	25(2.2)	24(4.1)	1(0.2)
Student	113(9.9)	82(14)	31(5.6)
Retired	138(12.1)	58(9.9)	80(14.5)
Unable to work	98(8.6)	55(9.4)	43(7.8)
Treatment for Cardiovascular Risk Factors
Hypertension	226(25.1)	83(17.5)	142(33.6)	<0.0001*
Cerebrovascular Events	8(0.9)	6(1.3)	2(0.5)	0.23
Diabetes	77(8.7)	24(5.1)	53(13)	<0.0001*
Chronic Sleep Problems	417(45.6)	213(43.7)	203(48)	0.19
Exercisea
Strenuous Exercise (M, SD)	2.93(SD = 6.32)	2.90(SD = 8.26)	2.96(SD = 3.21)	0.88
Moderate Exercise (M, SD)	3.21(SD = 4.93)	3.18(SD = 5.77)	3.24(SD = 3.87)	0.85
Mild Exercise (M, SD)	3.42(SD = 3.83)	3.46(SD = 4.08)	3.39(SD = 3.55)	0.77
Smoking Status				0.13
Never	3(0.6)	0(0)	3(1.2)
Former	363(69.4)	176(67.7)	185(71.2)
Current	157(30)	84(32.3)	72(27.7)
Mental Health Symptoms
Patient Health Questionnaire-8 (M, SD)	7.45(SD = 6.28)	7.89(SD = 6.38)	6.98(SD = 6.14)	0.02*
Generalized Anxiety Disorder-7 (M, SD)	6.57(SD = 5.99)	6.94(SD = 5.98)	6.18(SD = 5.98)	0.04*
PTSD Checklist-Military Version (M, SD)	38.59(SD = 17.66)	38.79(SD = 17.70)	38.39(SD = 17.66)	0.71
Insomnia Severity Index (M, SD)	13.51(SD = 6.50)	13.47(SD = 6.58)	13.53(SD = 6.43)	0.87
Mental Health Treatment
Emotional Disorderb	494(54.1)	290(59.4)	203(48.1)	0.0006*
PTSD	416(45.5)	219(45.2)	197(46.2)	0.74
Drug/Alcohol	54(6.1)	23(4.9)	31(7.6)	0.10

^{^}Data are presented as n(%) unless otherwise indicated.

^*Indicates significant differences.

^aData are presented as number of times on average in a week a participant engaged in exercise for more than 15 minutes. Strenuous exercise was defined as exercise that causes the heart to beat rapidly (e.g., running, vigorous swimming). Moderate exercise was defined as exercise that was not exhausting (e.g., fast walking, easy bicycling). Mild exercise was defined as minimal effort (e.g., yoga, archery, easy walking).

^bEmotional disorders (e.g., depression, anxiety).

Cases with missing data were removed from analyses: Mild Exercise missing = 6.3%; Insomnia Severity Index missing = 12.2%.

Table 3.

Models for the Effect of Sex on Barriers

	Unadjusted		Adjusted
	OR(CI)	p	OR(CI)	p
Patient Level Barriers
Perceive Self as Not at Risk	1.06(0.84–1.34)	0.62	1.00(0.78–1.30)	0.97
Lack of Confidence	1.29(1.02–1.64)	0.04*	1.35(1.04–1.76)	0.02*
Prefer Not to Change Lifestyle	0.57(0.45–0.74)	<.0001*	0.62(0.48–0.82)	0.0006*
Depression	1.44(1.12–1.85)	0.004*	1.27(0.95–1.69)	0.10
Fearful of Change	1.08(0.84–1.39)	0.55	0.98(0.74–1.29)	0.88
Stress	1.48(1.14–1.92)	0.004*	1.38(1.03–1.84)	0.03*
God/Higher Power	0.84(0.65–1.10)	0.20	1.10 (0.82–1.48)	0.52
Believe Lifestyle Change will Not Affect Risk	0.57(0.44–0.76)	<.0001*	0.61(0.45–0.83)	0.002*
Unsure of Preventive Behaviors to Take	1.17(0.88–1.57)	0.28	1.15(0.83–1.59)	0.41
Believe Changes are Too Complicated	1.11(0.82–1.49)	0.50	1.08(0.78–1.50)	0.65
Ill/Old	1.06(0.71–1.58)	0.79	1.21(0.78–1.89)	0.39
Lack Time	1.09(0.69–1.74)	0.70	1.17(0.70–1.97)	0.57
Confusion about Lifestyle Changes	0.83(0.52–1.31)	0.43	0.83(0.50–1.40)	0.49
Interpersonal Barriers
Family Obligations	0.83(0.66–1.05)	0.12	0.95(0.72–1.23)	0.68
Inadequate Social Support	1.37(1.03–1.81)	0.03*	1.32(0.98–1.80)	0.07
Systems Level Barriers
Media Confusion	0.89(0.70–1.14)	0.36	0.93(0.72–1.22)	0.61
Healthcare Professional Doesn’t Think I Need to Worry	1.38(1.08–1.76)	0.01*	1.30(0.99–1.70)	0.06
Lack of Money or Insurance Coverage	1.07(0.82–1.41)	0.60	0.89(0.65–1.20)	0.43
Healthcare Professional Doesn’t Explain Clearly What I should do	1.61(1.16–2.24)	0.005*	1.39(0.97–1.99)	0.07

Adjusted model is adjusted for sex, age, marital status, education level, and employment status.

^*Indicates significant findings.

Cases with missing data were removed from analyses: Too stressed barrier missing = 9.8% (unadjusted) and 12% (adjusted), Lack of time barrier missing = 23.5% (unadjusted) and 25.7% (adjusted), Healthcare professional doesn’t think I need to worry = 5.3% (adjusted), Confusion about lifestyle changes barrier missing = 7.9% (unadjusted) and 10.4% (adjusted), and Healthcare doesn’t explain clearly what I should do = 5% (adjusted).

Results

The participation rate in the study was low at 11.5% (1,145 participants/9,912 eligible participants). Table 1 shows sample characteristics. Women were significantly younger than men (41.61 years ± 10.28 vs. 46.21 years ± 11.08; p < 0.0001), with a higher representation of African Americans (11.8% vs. 8.2%) and a higher percentage reporting being single (28.8% vs. 18.4%). Women also reported higher educational attainment (35.7% vs. 27.2% with a bachelor’s/4-year college), but lower rates of employment (56.3% vs. 65.5%).

CVD Risk Factors

Fewer women reported receiving treatment for hypertension (17.5% vs. 33.6%, p < 0.0001) and diabetes (5.1% vs. 13%, p < 0.0001) in the past year vs. men. Women however, had higher depression scores (M = 7.89 ± 6.38 vs. M = 6.98 ± 6.14, p = 0.02) and anxiety scores (M = 6.94 ± 5.98 vs. M = 6.18 ± 5.98, p = 0.04) as compared to men. Furthermore, over half of the total sample reported mild to severe depression (PHQ-8 score >5) and/or mild to severe anxiety (GAD-7 score ≥5). Women represented 53.6% of the sample reporting mild to severe depression and 54.4% of the sample reporting mild to severe anxiety. Over half of the total sample also scored in the diagnostic range on the PCL-M (scores ≥33) with women representing 51.3% of the sample scoring in the diagnostic range. Additionally, more women reported that they received treatment for emotional disorders (25.3% vs. 17.7%, p = 0.0006) as compared to men. See Table 1 for additional CVD risk factors.

Barriers to Engagement in CVD Preventive Behaviors

Table 2 provides the prevalence of barriers to CVD preventive behaviors for the overall cohort, and separately for women and men.

Table 2.

Sex Differences in Barriers

	Overall Sample (N= 1145)	Women (n = 586)	Men (n = 555)	p	FDR p	φc
Patient Level Barriers
Perceive Self as Not at Risk	645(56.5)	335(57.3)	308(55.8)	0.62	0.69	0.02
Lack of Confidence	452(39.6)	248(42.5)	202(36.5)	0.04*	0.09	0.06
Prefer Not to Change Lifestyle	398(34.9)	167(28.6)	228(41.2)	<0.0001*	0.0002*	−0.13
Depression	370(32.5)	212(36.4)	157(28.4)	0.004*	0.02*	0.09
Fearful of Change	353(30.9)	186(31.9)	167(30.2)	0.55	0.69	0.02
Stress	334(32.2)	190(36.5)	144(28.1)	0.004*	0.02*	0.09
God/Higher Power	303(26.9)	147(25.3)	156(28.7)	0.20	0.38	−0.04
Believe Lifestyle Change will Not Affect Risk	273(23.9)	111(19)	161(29.1)	<0.0001*	0.0007*	−0.12
Unsure of Preventive Behaviors to Take	229(20.l)	125(21.4)	103(18.8)	0.28	0.49	0.03
Believe Changes are Too Complicated	217(19.l)	116(19.8)	100(18.3)	0.50	0.68	0.02
Ill/Old	106(9.4)	56(9.6)	50(9.1)	0.79	0.79	0.01
Lack Time	79(9)	39(9.4)	40(8.7)	0.70	0.74	0.01
Confusion about Lifestyle Changes	79(7.5)	38(6.9)	41(8.2)	0.43	0.62	−0.03
Interpersonal Barriers
Family Obligations	624(55.1)	310(53)	313(57.6)	0.12	0.25	−0.05
Inadequate Social Support	261(23.2)	150(25.9)	110(20.3)	0.03*	0.08	0.07
Systems Level Barriers
Media Confusion	405(35.7)	200(34.4)	203(37)	0.36	0.57	−0.03
Healthcare Professional Doesn’t Think I Need to Worry	398(35.6)	226(39.3)	172(32)	0.01*	0.03*	0.08
Lack of Money or Insurance Coverage	287(25.4)	151(26)	134(24.6)	0.60	0.69	0.02
Healthcare Professional Doesn’t Explain Clearly What I should do	176(15.7)	108(18.8)	68(12.6)	0.004*	0.02*	0.09

^{^}Data are presented as n(%) unless otherwise indicated.

^*Indicates significant differences.

FDR refers to a false discovery rate.

φ_c refers to Cramer’s V effect sizes.

Cases with missing data were removed from analyses: Too stressed barrier missing = 9.8%, Lack of time barrier missing = 23.5%, and Confusion about lifestyle changes barrier missing = 7.9%.

Patient Level Barriers

Most barriers were endorsed by one fifth to over one half of the overall cohort. The most frequently endorsed patient level barrier was the perception of not being at risk for heart disease, with 57.3% of women and 55.8% of men endorsing this belief (p = 0.62). Additional barriers were endorsed by both women and men Veterans: 31.9% of women and 30.2% of men endorsed being fearful of change, 19.8% of women and 18.3% of men endorsed the belief that lifestyle changes for cardiovascular health are too complicated, 21.4% of women and 18.8% of men endorsed feeling unsure of the preventive behaviors to engage in, and 25.3% of women and 28.7% of men endorsed the belief that God or a higher power determines their health. More men endorsed a preference for not changing their lifestyle (41.2% vs. 28.6%, p < 0.0001), and the belief that lifestyle changes will not affect their CVD risk (29.1% vs. 19%, p = <0.0001). In contrast, more women endorsed lack of confidence in the ability to make necessary lifestyle changes (42.4% vs 36.5%, p = 0.04), and both stress (36.5% vs. 28.1%, p = 0.004) and depression (36.4% vs. 28.4%, p = 0.004) as barriers.

Thus overall, over half of women and men Veterans did not believe themselves to be at CVD risk, and between 6.9% and 42.5% of women endorsed some patient level barrier to engaging in CVD risk lowering behaviors.

Interpersonal Barriers

More women endorsed inadequate social support (25.9% vs. 20.3%, p = 0.05) as a barrier as compared to men, while family obligations as a barrier were equivalently high among both women and men (53% vs. 57.6%).

Systems Level Barriers

More women than men endorsed the belief that their clinician does not perceive them as at risk for CVD (39.3% vs. 32%, p = 0.01), and a lack of discussion by their clinician about how to lead a heart healthy lifestyle as compared to men (18.8% vs. 12.6%, p = 0.004). Women and men endorsed equivalently high rates of other systems level barriers, with 26% of women reporting lack of money or insurance coverage (vs. 24.6% for men) and 34.4% reporting confusion in media information as barriers (vs. 37% for men).

Secondary Analyses

After logistic regressions adjusting for sex, age, marital status, education level, and employment status, sex differences for several barriers to engagement in CVD preventive behaviors remained. Statistical differences between men and women in depression, inadequate social support, the belief that their clinician does not perceive them as at risk for CVD, and a lack of discussion by their clinician about how to lead a heart healthy lifestyle as barriers, were eliminated. Unadjusted and adjusted models are reported in Table 3.

Discussion

WVCS is the first study to examine barriers to CVD preventive behaviors among OEF/OIF/OND women and men Veterans. This Veteran cohort has the highest representation of women, approximately double the percent of women as compared to previous Veteran cohorts (U.S. Congress, 2016). Over half of both men and women Veterans who completed the WVCS survey did not perceive themselves to be at risk for CVD, and between approximately one-fifth and one-half endorsed most patient level, interpersonal, and/or systems level barriers. These barriers included being fearful of making lifestyle changes, believing that changes are too complicated to make or being unsure of what changes to make, having too many competing family obligations, and lack of financial resources. These findings are concerning in light of data demonstrating increasing CVD risk after military separation - e.g., obesity, continued tobacco use, high prevalence of depression and PTSD (Burg et al., 2017; Haskell et al., 2017; Haskell et al., 2010; Rosenberger et al., 2011; Volkman et al., 2015), early incident hypertension (Burg et al., 2017), and increasing multiple major risk factor prevalence with each decade of life (Vimalananda et al., 2013). This group is of an age where CVD related mortality has increased in recent years (Ford & Capewell, 2007). Yet findings from the current study indicate that many OEF/OIF/OND Veterans appear unaware of their CVD risk and that multiple barriers to engagement in CVD preventive behaviors exist.

Sex differences in barriers were also examined, with women overall endorsing more barriers than men, even though endorsement by both men and women was high. For example, while more men indicated that they preferred not to make lifestyle changes, more women did not believe that change would affect their risk. More women also endorsed a lack of confidence in making change, and both stress and depression as barriers to change, along with inadequate social support. They also reported a much higher prevalence of chronic sleep problems, and higher scores on measures of anxiety and depression. Yet, secondary analyses controlling for a range of demographic variables, including age, marital status, education, and employment, reduced the statistical significance of sex differences in many barriers, including depression, and inadequate social support. This suggests that other demographic factors may also play a role in barriers to engagement in CVD preventive behaviors. It is important to note however, that women were also significantly younger, almost twice as likely to be either divorced or single as compared to men, and less likely to be employed, in part due to being a student or unable to work. Thus, the effect of adjusting for these variables may in part describe a pathway by which women experience the endorsed barriers - e.g., due to social constraints that are accompanied by psychosocial stress and lack of resources.

It is important to evaluate these findings in the context of the low response rate, which may affect both the findings and generalizability of results. The low participation rate is lower than typical response rates observed in other studies examining CVD risk. Studies examining CVD awareness and barriers typically report response rates to online surveys of between 97%−100% (Mosca et al., 2006; Mosca et al., 2010). Yet, response rates to phone interviews for the same surveys report lower response rates of between 22%−27% (Mosca et al., 2010; Mosca et al., 2013). Furthermore, although the response rate in the present study is lower than response rates observed in other studies examining CVD risk, there is evidence that response rates among Veterans are often lower than among the general population. For example, one study of Veterans from the OEF/OIF era reported a 26.3% response rate (Coughlin et al., 2011). A similar survey among Veterans reported a response rate of 22% (National Academies of Sciences, Engineering, and Medicine, 2018).

The low response rate observed in the current study may be attributable in part to the original consent process, which required Veterans to consent in person and to the length of the survey (72 pages), and to the type and amount of incentive. Indeed, previous research has found survey fatigue among Veterans to affect response rates, as Veterans are frequently invited to participate in health surveys (Coughlin et al., 2011). Furthermore, Veterans have previously expressed a concern that government benefits can be affected by personal information revealed in surveys, which may also contribute to low response rates (Coughlin et al., 2011). Strategies for improving the low response rate among Veterans include, providing multiple options for completing survey research (e.g., mail survey, web-based survey, phone survey) (Coughlin et al., 2011). Additionally, eliminating the requirement for in-person consent in similar survey studies, may also increase participation. Finally, the length of the survey and survey fatigue among Veterans must also be addressed. Shortening the length of the survey as well as the number of surveys that Veterans are invited to participate in may also improve response rates.

Veteran survey research typically underrepresents minority Veterans and those with fewer years of education (Coughlin et al., 2011). Thus, it is possible that endorsement of barriers may be lower than would be observed in a more representative sample of Veterans, given that minority groups and those with lower educational attainment are more likely to encounter barriers related to health-care access and are at greater risk for disease burden (Liao et al., 2011; Williams, Priest, & Anderson, 2016). Alternatively, given that younger Veterans are also underrepresented in Veteran survey research (Coughlin et al., 2011), it is possible that younger Veterans may encounter fewer barriers and be at a decreased risk for CVD. Yet, it is important to note, the present findings are consistent with previous research showing multiple barriers to leading a heart healthy lifestyle among non-Veteran women (Mochari-Greenberger, Mills, Simpson, & Mosca, 2010; Mosca et al., 2013; Mosca, McGillen, & Rubenfire, 1998; Mosca et al., 2010; Mosca et al., 2006). Furthermore, as previously discussed, young Veterans in particular are at increased risk of CVD.

Efforts to understand the underlying issues have identified a number of barriers that contribute to low rates of engagement in CVD risk lowering efforts (Mochari-Greenberger, Mills, Simpson, & Mosca, 2010). Among women in particular, patient level barriers have included being stressed, lacking time, being fearful of change, feeling that change is too complicated, and having the belief that God or a higher power determines one’s health (Mochari-Greenberger et al., 2010; Mochari, Ferris, Adigopula, Henry, & Mosca, 2007; Mosca et al., 2006). Women also report a number of interpersonal barriers such as family obligations (Mosca et al., 2010; Mosca et al., 2006), and systems level barriers such as lack of money or insurance coverage (Mochari-Greenberger et al., 2010; Mosca et al., 2006), and confusion in media sources about what change is needed (Mosca et al., 2010). Women are also found to differ from men in what they consider the most important barrier to change, with women endorsing low self-esteem - e.g., low confidence - as their top barrier and men rating lack of time (Mosca, McGillen, & Rubenfire, 1998). Overall, our current research supports previous findings that women report higher rates of patient level barriers to engagement than men.

The impact of the barriers endorsed by women may be further compounded by the behavior of the clinicians serving them. Approximately 40% of women (vs. 32% of men) endorsed the belief that their clinician does not perceive them to be at CVD risk, and almost 20% (vs. 12% of men) indicated that their clinician does not engage them in discussions of how to lead a heart healthy lifestyle. While sex differences on these barriers became non-significant when adjusting for demographic variables, the higher endorsement for these items by women suggests that there may be systems level differences in how CVD risk is perceived for women and men Veterans. Indeed, prior research has shown both that women Veterans who have elevated LDL-cholesterol are less likely than their male counterparts to receive lipid lowering therapy (Haskell et al., 2014), and that those with elevated blood pressure are less likely to receive a diagnosis of hypertension or anti-hypertensive treatment (Burg et al., 2017). It is arguable that respondents to the WVCS Wave II Survey are indeed at low risk for CVD, and thus the perception of not being at risk that they share with their clinician may be accurate, though the prevalence of factors that contribute to incident CVD risk, including hypertension, chronic sleep problems, tobacco use, PTSD, and depression are notably high. Indeed, there is a confluence of factors in OEF/OIF/OND women Veterans that includes stress, depression, PTSD, financial and family issues, and an increasing prevalence of multiple major CVD risk factors with each decade of life from age 35 on (Vimalananda et al., 2013). Thus, prudence argues that clinicians should be discussing CVD risk with these women Veterans (Vimalananda et al., 2013).

The VA launches a healthcare system wide American Heart Association (AHA), ‘Go Red Challenge’ each year, encouraging each healthcare site to focus on raising awareness of heart disease in women Veterans, urging these Veterans toward active engagement in their heart health. Reaching OEF/OIF/OND women Veterans through this initiative should be a key focus of this annual program. The AHA Go Red initiative and the associated “Life’s Simple 7” metric - quitting smoking, maintaining ideal BMI, healthy diet, physical activity, and blood pressure, cholesterol, and blood glucose at guideline recommendations - could be used to assess and promote ideal cardiovascular health for these Veterans (Lloyd-Jones et al., 2010). At the same time, it is essential to determine whether the prevalent endorsement of barriers demonstrated in this sample is representative of the larger OEF/OIF/OND Veteran cohort. The importance of this issue is in part a reflection of the low response rate on the WVCS survey. The ‘Go Red Challenge’ provides a useful context for accomplishing this VHA wide assessment.

More women also endorsed depression, stress, and lack of confidence as barriers, and these are not addressed by ‘Go Red’ or ‘Life’s Simple 7’. Furthermore, prior research has shown that women Veterans are not satisfied with the CVD risk related care they receive through the VHA (Goldstein, Stechuchak, et al., 2017) and prefer gender-specific care - e.g., weight reduction programs (Goldstein, Oddone, et al., 2017). These findings together suggest that, in concert with a comprehensive, system-wide assessment of barriers, there is a need for developing and testing tailored interventions to address the specific needs of women Veterans. Individually tailored interventions, including motivational interviewing tailored to the CVD related risks and barriers of a given Veteran regardless of sex could be useful to acknowledge a given barrier and increase problem solving related to specific barriers. Indeed, motivational interviewing is the focus of a VA-wide initiative (Cucciare et al., 2012), and as previous research has suggested may be a promising intervention for patients with CVD risk factors because of the focus on behavior change (Thompson et al., 2011). Behavioral/lifestyle interventions as recommended by the U.S. Preventive Services Task Force (USPSTF; Moyer, 2012) provide other examples of effective individually tailored interventions that may be used to promote CVD risk reduction (Lin et al., 2014) =. It is important to note however, that neither approach addresses interpersonal or systems level barriers. Given the low response rate, it is critical for future research to test new risk stratification and surveillance strategies for OEF/OIF/OND Veterans, which will enable the VA to better understand the risks, barriers, and needs for this cohort of Veterans and be better prepared to develop interventions and initiatives to address these barriers.

Strengths and Limitations

The current study included a large and geographically representative sample of OEF/OIF/OND Veterans, with equal representation by women and men. Therefore, sex differences that emerged can inform a tailored approach toward development, testing, and implementation of interventions to promote engagement in CVD preventive behaviors. Yet, while the current study provides important information, it is not without limitations. As noted, the response rate was low, and the sample not random, demonstrating an imperative to assess CVD risk reduction barriers more broadly in this at-risk Veteran group. Given the non-random sample and the low response rate it is possible that these results will not generalize to other women or men Veterans or to other populations. Indeed, previous research has shown that Veteran survey research may underrepresent minority and younger Veterans, and those with fewer years of education (Coughlin et al., 2011). Additionally, primary outcomes – e.g., barriers – were based on the survey data, which is self-report and may be influenced by recall bias and social desirability. The data are also cross-sectional, which precludes the ability to examine causal relationships. Furthermore, although there were sex differences in several barriers, effect sizes were small, though the prevalence of endorsed barriers by both women and men Veterans is troubling, in light of the unfolding CVD risk being reported in the literature.

Conclusions

In this sample of OEF/OIF/OND women and men Veterans it was found that this Veteran group does not perceive themselves as at risk for CVD, and a large percentage endorse a number of patient level, interpersonal, and systems level barriers to making CVD risk lowering behavior changes. In addition to this high level of endorsement regardless of sex, a number of sex differences in barriers were found, with secondary analyses suggesting a contribution of demographic factors related to age, employment, and partner status. Understanding these sex differences in barriers is particularly important because of persistent sex disparities in cardiovascular risk management that exist both within and outside the VHA system (Haskell et al., 2014; Hyun et al., 2017; Vimalananda et al., 2011; Whitehead et al., 2014). Furthermore, research shows that CVD risk factors for OEF/OIF/OND Veterans become apparent at a relatively young age, that multiple major risk factors occur in early middle age and that, particularly for women, non-traditional factors such as depression, stress, and PTSD that are associated with incident CVD are common (Burg et al., 2017; Haskell et al., 2017; Haskell et al., 2010; Rosenberger et al., 2011; Vimalananda et al., 2013; Volkman et al., 2015). Efforts to reach these Veterans about CVD risk and preventive behaviors, whether through the ‘Go Red’ initiative or related programs should be emphasized. These initiatives that are ongoing with the VHA could also provide an important context for determining whether the findings for this sample are representative of the larger OEF/OIF/OND Veteran cohort. In addition to these risk-behavior focused initiatives, women Veterans may benefit from support systems that are dedicated exclusively to their unique needs and barriers (Goldstein et al., 2018; Mattocks et al., 2012). Efforts to assist Veterans with caregiving responsibilities may also help to reduce family obligations as a barrier (Mattocks et al., 2012).