Five-year trend in healthcare access and patient‐reported health outcomes among women veterans

Abstract

Background:

This study aimed to describe the five‐year trend in healthcare access, health‐related quality of life (HRQoL), and health outcomes in women Veterans.

Methods:

A retrospective, pooled, cross‐sectional study design was employed. Five‐year trend was assessed using 2013 and 2017 Behavioral Risk Factor Surveillance System (BRFSS) data. Bivariate, multivariable logistic regression, and ordinary least squares regression analyses were conducted.

Results:

A total of 6561 women Veterans, aged 18 to 64 years (3534 in 2013 and 3027 in 2017) were included. Compared to 2013, more women Veterans in 2017 reported increased healthcare insurance and decreased cost as a barrier to care. Women Veterans with health insurance were less likely to report cost as a barrier to care. There were no statistically significant differences in HRQoL in bivariate or ordinary least squares regression analyses between BRFSS years.

Conclusion:

Federal policy and Veterans Health Administration expansion have had an impact on improving healthcare access to women Veterans. However, increased healthcare access alone does not translate into improved HRQoL or health outcomes for women Veterans. Future policies should not only focus on increasing healthcare access, but also improving health outcomes, especially HRQoL. The quality of the healthcare accessed must be a focus for future research and policy.

1 |. INTRODUCTION

Enlistment of women in the military has increased post‐9/11, during the Operations Iraqi Freedom, Enduring Freedom, and New Dawn.¹ Subsequently, women Veterans are one of the fastest growing segments of the United States veteran.² To date, approximately 9.8% of veterans are women.²

Women veterans returning from war require ongoing care for physiological and psychological health issues.^3,4 It is widely recognized that women Veterans’ health risks and needs are uniquely different from those of male Veterans.^5–7 For example, women veterans may experience gender‐specific health difficulties such as gynecological and urological symptomology, which may be worsened when comorbid with military sexual trauma and/or posttraumatic stress disorder (PTSD).⁸ Additionally, approximately 80% of post‐9/11 women veterans are aged 43 or younger,⁹ suggesting that reproductive health services are an important component of their healthcare.

The Veterans health administration (VHA), which provides care for Veterans was originally designed to deliver care to a male Veteran majority.^10,11 However, the number of women Veterans seeking care from the VHA has increased from 674,784 in 2011 to 836,191 in 2015.¹² Over the past 10 years, the VHA has expanded VHA healthcare access to improve health outcomes for women Veterans.¹³ The existing Women Veterans health program in the VHA was changed to a Strategic HealthCare Group in 2007, and in 2012 was restructured and renamed Women’s Health Services (WHS).¹⁴ The mission of WHS is to ensure timely, equitable, high‐quality healthcare delivery at VHA facilities. One of the priority areas for WHS is the provision of comprehensive primary care for women Veterans that includes reproductive health.¹⁵ Women’s health is also a research priority for VHA health services and research department.¹⁶

The Veterans’ Access, Choice, and Accountability Act of 2014 expanded access to healthcare via nonVHA healthcare facilities or providers for Veterans who have experienced greater than 30‐day appointment wait‐times or who live greater than 40 miles from a VHA facility.¹⁷ Finally, the Affordable Care Act, signed into law in 2010, made healthcare in the civilian sector available via health insurance expansion.¹⁸ This included, but is not limited to, provisions to prohibit denial of coverage based on pre‐existing conditions, access to obstetric/gynecologic care without prior authorization, reduced cost for preventative services, and extension of dependent coverage until age 26.

Despite the recognition and expansion of health services for women, women Veterans report poorer overall health,¹⁹ poorer functional status, and poorer mental health compared to civilian women.²⁰ Compared to their urban counterparts, rural women Veterans report significantly worse physical health functioning.²¹ From a 2010 population‐based survey of women Veterans, almost 1 in 5 women delayed healthcare or went without needed care in the prior 12 months; barriers included the inability to take off work, caregiver responsibilities, and transportation difficulties.²¹ Additionally, access, stigma, availability of gender‐specific care, and affordability have been cited as barriers.^22–24 For rural Veterans, barriers include frustrations with enrolling in and understanding the VHA healthcare system and the location of services.²⁵

While the VHA continues to expand services and collect data on the health status of women Veterans accessing the VHA facilities and clinics, nearly one‐third of women Veterans receive care outside the VHA.¹³ Systematic information regarding the prevalence or experience of women Veterans outside of the VHA systems is not available.²⁶ Failing to account for prevalence or experiences of women using nonVHA healthcare may “…bias conclusions about prevalence and incidence in utilization, diagnoses, and other veterans’ characteristics…”.²⁷ Previous studies have used the centers for disease control (CDC) behavioral risk factor surveillance system (BRFSS) data to study health indicators for military, Veteran, and civilian women,¹⁹ health and health behavior differences in United States military, Veteran, and civilian men,²⁸ and associations between health‐related quality of life (HRQoL) and financial barriers for women Veterans.²⁰ These studies have provided a snap‐shot status of healthcare and outcomes among women Veterans. It is not known whether the efforts by the VHA and the federal expansion of healthcare have translated into improved outcomes for women Veterans. No prior study has examined the trending of healthcare access, HRQoL, or health outcomes for women Veterans over a period of time. Analysis of such might provide insight into the effectiveness of ongoing efforts to increase access and improve care for women Veterans. Therefore, the primary objective of this study is to describe the five‐year trend in healthcare access, HRQoL, and health outcomes in women Veterans using the BRFSS data from 2013 and 2017. It is hypothesized that there will be a statistically significant improvement in healthcare access, HRQoL, and health outcomes for women Veterans in 2017 compared to 2013.

2 |. MATERIALS AND METHODS

2.1 |. Design

A retrospective, pooled, cross‐sectional study design was used.

2.2 |. Data

For the current study, we used 2013 and 2017 BRFSS data.^29,30 BRFSS is an annual survey of chronic health conditions, use of preventive care, and risk behavior of adults living in the United States. All 50 states collect data using a standardized core questionnaire, optional modules, and state added questions via telephone from eligible households.²⁹ Eligible households are defined as “a housing unit that has a separate entrance, where occupants eat separately from other persons on the property, and that is occupied by its members as their principal or secondary place of residence”,³¹ (p.5); eligible household members are all adults, aged 18 years and older, related, unrelated, boarders/roomers, and domestic workers who consider the household their home. Adult students living in college housing have been considered eligible respondents since 2011.

The BRFSS uses two samples: one for landline telephone respondents and one for cellular telephone respondents.³¹ A disproportionate stratified sampling is used for landline and cellular telephone respondents are randomly selected from a sampling frame of confirmed cellular area code and prefix combinations. Persons interviewed on cellular telephones are treated as one person households; for persons interviewed on landline telephones, an individual respondent is randomly selected from all adult members living in the household.

Stratification of landline telephone numbers is based on substate geographic regions. Cellular telephone samples were stratified only by state until 2013, at which time substate stratification was implemented (geographic specificity noted to be less reliable as compared to landline numbers).

From 2011 on, iterative proportional fitting or raking has been utilized as the weighting methodology for BRFSS data. To obtain a representative sample from each state, data are weighted to known proportions of age, sex, categories of ethnicity, geographic regions within states, marital status, education level, home ownership and type of phone ownership.³²

It has to be noted that BRFSS data are publicly available; therefore, they are not considered human subject research or subject to institutional review board review.

2.3 |. Analytical sample

For this study, inclusion criteria were (a) women Veterans, completed the interview (data collected for age, race, and sex—approximately halfway through the core BRFSS questionnaire), (b) interviewed in the year 2013 or 2017, and (c) aged 18 to 64. Women Veterans in this study are identified as those who answered yes to the question: “Have you ever served on active duty in the United States Armed Forces, either in the regular military or in a National Guard or military reserve unit? (Active duty does not include training for the Reserves or National Guard, but DOES include activation, for example, for the Persian Gulf War.)”.^33,34 The ages 18 to 64 were decided upon by the researchers due to the fact that one must be 18 years or older to enlist in the military and that persons after age 65 years old are entitled to health coverage through Medicare.

2.4 |. Measures

The theoretical framework that guided the selection of variables for this study was the behavioral health model.³⁵ The model is proposed as a way to “…discover conditions that either facilitate or impede (healthcare) utilization” (p. 4). The most current model demonstrates a dynamic relationship between the environment, population characteristics, and health behavior to predict health outcomes. Additionally, there are feedback loops that link health outcomes as subsequent predictors of health behavior and population characteristics. The variables presented below are discussed in accordance with the model components. The cross‐sectional, retrospective nature of our study prohibits evaluating the dynamic, predictive relationships between all factors, but instead we attempt to identify associations of the factors with health outcomes.

2.4.1 |. Dependent variables

Healthcare access

Three questions from the healthcare access core section were used as dependent variables.^33,34 The first measure, healthcare insurance, was measured by the question: “Do you have any kind of healthcare coverage, including health insurance, prepaid plans such as HMOs, government plans such as Medicare, or Indian Health Service?” The second, having a usual source of care (yes/no), was measured by the question: “Do you have one person you think of as your personal doctor or healthcare provider?” The third, cost as a barrier to care (yes/no), was measured by the question: “Was there a time in the past 12 months when you needed to see a doctor but could not because of cost?”

HRQOL measures

Four questions from the Health Status and Healthy Days core sections were used as measures of HRQoL.^33,34

1. Self‐reported general health was originally reported in a 5‐point scale. To satisfy an assumption of a linear relationship between item scores and the underlying health concept defined by their scales, this measure was transformed to a 0 to 100 scale (100, 84, 61, 25, 0) in accordance to results from empirical work.³⁶ There were nine women Veterans with missing data on this variable. For these women, we assigned the mean value (25).

2. Self‐reported physical health was measured by the number of days that physical health was reported as not good in the past 30 days, ranging from 0 to 30. There were 73 women Veterans with missing data on this variable. For these women, we assigned the mean value (5).

3. Self‐reported mental health was measured by the number of days that mental health was reported as not good in the past 30 days, ranging from 0 to 30. There were 82 women Veterans with missing data on this variable. For these women, we assigned the mean value (5).

4. Self‐reported functional status was measured by the number of days when poor physical or mental health was reported as not good in the past 30 days, ranging from 0 to 30. There were 34 women Veterans with missing data on this variable. For these women, we assigned the mean value (2).

After substituting the mean scores for missing data in all HRQoL measures, the overall mean score for each measure remained unchanged.

2.4.2 |. Key independent variable

BRFSS interview year

To analyze the five‐year trend in healthcare access and HRQoL measures, BRFSS interview year 2017 was compared to 2013.

2.4.3 |. Other independent variables

The independent variables have been sub‐grouped according to how they are viewed in the model. Predisposing characteristics include age and race. Age was grouped into four categories: (a) 18 to 29 years; (b) 30 to 39 years; (c) 40 to 49 years; and (d) 50 to 64 years. Race was categorized into five groups: (a) white; (b) African American; (c) Hispanic or Latina; and (d) other.

Enabling resource variables include marital status, education, employment, and household income. Marital status was categorized into two categories: (a) married and (b) not married. Education was categorized into: (a) less than high school diploma; (b) high school diploma or GED equivalent; and (c) above high school. Employment was categorized into: (a) employed; (b) not employed; and (c) retired. Household income was categorized into: (a) less than $10 000; (b) $10 000 to 25 000; (c) $25 000 to $75 000; and (d) above $75 000.

Need variables (chronic health conditions) include hypertension, heart disease, asthma, chronic obstructive pulmonary disease, kidney disease, and depression. Personal health practices (health behavior) variables included smoking status, alcohol consumption, physical activity, and body mass index (BMI). Smoking status was categorized into: (a) current; (b) former; and (c) nonsmoker. Alcohol consumption was categorized into: (a) heavy alcohol use; (b) moderate alcohol use; and (c) no alcohol use. Physical activity was categorized into: (a) yes, and (b) no. BMI was categorized into: (a) underweight/normal; (b) overweight; and (c) obese. External environment variables were categorized as: (a) northeast; (b) south; (c) midwest; and (d) west. In analysis related to HRQoL measures, healthcare access variables were included as independent variables.

2.5 |. Statistical analyses

Unadjusted group differences in women veterans’ characteristics by year of the survey (2013 and 2017) were analyzed with Rao‐Scott χ² tests. Statistically significant differences in healthcare access were also analyzed with Rao‐Scott χ² tests. Differences on average HRQoL between 2013 and 2017 were tested with t tests. Logistic regressions and ordinary least squares regressions were performed to evaluate the five‐year trend in healthcare access and HRQoL measures among women Veterans, after adjusting for predisposing characteristics, enabling resources, need, personal health practices, and external environment. Data were analyzed using the survey procedures in SAS software, Version 9.4,³⁷ which accounted for the disproportionate stratified sample sampling design of the BRFSS.

3 |. RESULTS

The study sample consisted of 6561 women Veterans (N = 3534 in 2013 and N = 3027 in 2017) aged 18 to 64 years. Table 1 presents selected characteristics of the study sample by year. In both years, 2013 and 2017, the majority of the sample was white, nonHispanic (59.5% and 60.8% respectively). For both years, more than a third of women Veterans were in the age group 50 to 64 years. Marital status for both years in the sample was similar; in 2013, 56.8% were married and in 2017, 53.6% were married. For both years, greater than 80% of the sample had above high school education, greater than 60% was employed, and greater than 40% made above $25 000 per year. With regard to need variables, personal health practices, and the external environment, there were no significant differences between the study samples from 2013 and 2017, except for employment and region.

TABLE 1.

Selected characteristics of the study sample by year women veterans (age 18‐64 y) behavioral risk factor surveillance system, 2013 and 2017

	2013 (n = 3534)		2017 (n = 3027)
	N	Wt%	N	Wt%	χ2	P value	Sig
Predisposing characteristics
Race/Ethnicity					5.635	.131
White	2447	59.5	2068	60.8
African American	602	23.3	449	18.7
Latina	225	9.9	223	10.8
Other	260	7.2	287	9.8
Age, y					4.629	.201
18–29	302	14.0	257	15.4
30–39	620	23.8	490	21.5
40–49	835	26.0	635	22.6
50–64	1777	36.3	1645	40.5
Enabling resources
Marital status					2.879	.237
Married	1890	56.8	1640	53.6
Not married	1631	43.0	1380	45.9
Education
Less than high school	35	1.5	26	1.1
High school	530	16.7	446	15.6
Above high school	2966	81.8	2554	83.2
Employment					14.520	.002	**
Employed	2165	64.0	1912	66.7
Not employed	1017	29.4	757	24.4
Retired	344	6.4	343	8.3
Household Income					5.354	.253
Less than 10k	152	3.1	102	3.0
10k-25k	585	17.5	460	14.7
25k-75 k	1491	42.1	1188	41.6
Above 75k	1031	29.9	977	30.9
Need
Diabetes					0.088	.957
Yes	324	8.1	296	7.9
No	3204	91.7	2730	92.0
Hypertension					1.863	.394
Yes	1007	25.4	836	23.3
No	2524	74.6	2189	76.7
Heart disease					1.066	.587
Yes	171	3.7	156	4.5
No	3348	95.9	2859	95.2
Asthma					3.221	.200
Yes	600	16.6	456	13.9
No	2927	83.3	2563	86.0
COPD					0.849	.654
Yes	322	7.6	251	7.1
No	3199	92.0	2763	92.6
Chronic kidney disease					0.732	.693
Yes	76	2.5	95	3.0
No	3452	97.4	2929	96.9
Depression					3.614	.164
Yes	1154	31.2	1043	33.0
No	2371	68.6	1970	66.4
Personal health practices
Alcohol consumption					6.151	.104
Heavy alcohol use	215	5.3	205	8.0
Moderate alcohol use	1619	48.7	1394	48.8
No alcohol use	1648	44.4	1379	41.8
Smoking status					3.247	.355
Current smoker	782	20.5	640	21.3
Former	928	25.3	758	23.1
Nonsmoker	1807	53.5	1617	55.4
Physical activity					0.387	.824
Yes	2681	76.6	2256	75.4
No	843	23.0	764	24.2
Body mass index					3.196	.362
Underweight/normal	1221	34.2	969	37.0
Overweight	1076	32.5	951	29.9
Obese	1016	26.4	893	27.4
External environment
Region					10.719	.030	*
Northeast	398	10.8	396	13.2
South	770	17.8	760	19.0
Midwest	1526	52.8	1158	46.3
West	782	17.8	672	21.0

Note: Based on 3534 women (2013) and 3027 women (2017) who served in the armed forces, aged 18 to 64, who completed questionnaires during the year 2013 or 2017. Significant group differences were tested with χ² statistics.

Abbreviations: COPD, chronic obstructive pulmonary disorder; Sig, significance; Wt, weighted.

^***P < .001;

^**.001 ≤ P < .01;

^*.01 ≤ P < .05

Bivariate analyses of the study sample by year for health access measures are presented in Table 2. A statistically significant trend was observed in two measures of healthcare access (health insurance and cost as a barrier to healthcare). A higher percentage of women Veterans reported having health insurance in 2017 compared to 2013 (93.0% vs 88.4%; χ² 9.329; P = .002). A lower percentage of women Veterans reported cost as a barrier to healthcare in 2017 than in 2013 (13.2% vs 17.9%; χ² 6.879; P = .009). No statistically significant changes were observed in the third healthcare access measure (usual source of care).

TABLE 2.

Five‐year trend in healthcare access and outcomes women veterans (age 18‐64 y) behavioral risk factor surveillance system, 2013 and 2017

	2013		2017
	N	Wt %	N	Wt %	χ2	P value	Sig
Healthcare access
Health insurance					9.329	.002	**
Yes	3168	88.4	2835	93.0
No	358	11.6	187	7.0
Cost prevented care					6.879	.009	**
Yes	530	17.9	375	13.2
No	2999	82.1	2645	86.8
Usual source of care					0.098	.754
Yes	2970	82.7	2579	83.3
No	557	17.3	439	16.7

	Mean	95%CI	Mean	95%CI	P value	Sig
Health-related quality of life measures
General health	69.15	(67.45, 70.84)	68.78	(66.49, 71.06)	.779
Poor physical health days	4.61	(4.05, 5.16)	4.78	(4.19, 5.37)	.686
Poor mental health days	4.92	(4.38, 5.46)	5.45	(4.84, 6.07)	.215
Poor MH or PH limited activities	2.14	(1.77, 2.50)	2.49	(2.08, 2.89)	.217

Note: Based on 3534 women (2013) and 3027 women (2017) who served in the armed forces, aged 18 to 64, who completed questionnaires during the year 2013 or 2017. General Health Status was transformed and ranged from 0 to 100. Poor physical health days, poor mental health days and limited activities ranged from 0 to 30 d. Significant group differences were tested with χ²statistics and t statistics.

Abbreviations: CI, confidence interval; MH, mental health; PH, physical health; Sig, significance; Wt, weighted.

^***P < .001;

^**.001 ≤ P < .01;

^*.01 ≤ P < .05

In multivariable logistic regression analysis for healthcare access, adjusting for predisposing characteristics, enabling resources, need, personal health practices, and external environment, a similar trend was observed (Table 3). Women Veterans in 2017 were 1.79 times as likely as those in 2013 to report having health insurance (95% confidence interval [CI] = 1.26, 2.54; 001 < P < .0). Women Veterans in 2017 were less likely to report cost as a barrier to care compared with those in 2013 (adjusted odds ratio [AOR] = 0.72; 95%CI = 0.054, 0.96; .01 ≤ P < .05) without controlling for health insurance. When health insurance was added to the model, no statistically significant difference remained.

TABLE 3.

Adjusted odds ratios and 95% confidence intervals of time from logistic regression on healthcare access and regression coefficients and standard errors from OLS regression on HRQoL measures women veterans aged 18 to 64 y behavioral risk factor surveillance system, 2013 and 2017

Healthcare access measures
	AOR	95%CI		Sig
Health insurance = yes
2017	1.79	(1.26, 2.54)		**
2013 (reference)
Cost barrier to care = yes
Model 1 (without health insurance)
2017	0.72	(0.54, 0.96)		*
2013 (reference)
Model 2 (With health insurance)
2017	0.79	(0.58, 1.08)
2013 (reference)
Usual source of care = yes
Model 1 (without health insurance)
2017	0.99	(0.77, 1.29)
2013 (reference)
Model 2 (with health insurance)
2017	0.92	(0 71 1 90)
2013 (reference)
Health-related quality of life measures
	Beta	Standard error	P value	Sig
Health status
2017	−0.809	1.147	.481
2013 (reference)
Poor physical health days
2017	0.174	0.333	.601
2013 (reference)
Poor mental health days
2017	0.434	0.352	.219
2013 (reference)
Days poor MH or PH limlited activities
2017	0.395	0.232	.089
2013 (reference)

Note: Based on 3534 women (2013) and 3027 women (2017) who served in the armed forces, aged 18 to 64, who completed questionnaires during the year 2013 or 2017. Asterisks represent differences in dependent variables in 2017 as compared to 2013.

Abbreviations: AOR, adjusted odds ratio; CI, confidence interval; HRQoL, health‐related quality of life; MH, mental health; OLS, ordinary least squares; PH, physical health; Sig, significance.

^***P < .001;

^**.001 ≤ P < .01;

^*.01 ≤ P < .05.

While not a primary objective our study, it must be noted that women Veterans with health insurance were less likely to report cost as a barrier to care compared to those without health insurance (AOR = 0.17; 95%CI = 0.12, 0.26; P < .001). Additionally, those with cost as a barrier to care were less likely to report a usual source of care than those without cost as a barrier (AOR = 0.47; 95%CI = 0.34, 0.66; P < .001).

With regard to HRQoL measures (general health, number of poor physical health days, number of poor mental health days, and the number of days poor mental or physical health prevented usual activities) there were no statically significant differences observed between the year 2013 and the year 2017 in bivariate analyses (Table 2). Additionally, no statistically significant differences were found for all HRQoL measures in ordinary least squares regression models between years (Table 3).

4 |. DISCUSSION

There were relatively few differences with regard to predisposing characteristics, enabling resources, health behaviors, need, and external environment variables between women Veterans in 2013 and 2017. Women Veterans were more likely to be insured in 2017 than in 2013, which was congruent with our hypothesis. Reduction in cost as a barrier to care and an increase in usual source of care were observed, but not statistically significant in logistic regressions with health insurance. However, our analyses revealed that women without health insurance were more likely to report cost as a barrier and less likely to have a usual source of care. Taken together, these findings suggest that expansion of healthcare insurance may be a large factor in healthcare access, particularly in terms of cost barriers and obtaining a usual source of care.

It is important to mention that, while well below the 12.7% national average of uninsured persons in 2017,³⁸ there remained 7.0% of women Veterans from interview year 2017 who were uninsured. More concerning is the fact that though only 7.0% reported being uninsured, still yet 13.2% reported cost as a barrier to care and, 16.7% did not have a usual source of care. While it may be unrealistic to achieve a zero percent uninsured, 0% cost as a barrier, and 100% with a usual source of care, it is certainly something to which policy makers ought to aspire.

It was surprising that HRQoL outcomes were relatively unchanged regardless of improved healthcare access. Thus, the expansion of access/coverage does not translate into improved health outcomes or HRQoL. In 2017, women Veterans reported poor health greater than 15% of a month. As poor HRQoL is a valid indicator of service needs, policy makers need to assess unmet needs among women Veterans and develop targeted efforts to meet the needs. Future studies are necessary to determine the barriers and facilitators of HRQoL.

Poor HRQoL may present a significant economic impact. Absenteeism can result in decreased productivity, increased costs, increased safety issues, and poor morale among coworkers.³⁹ For women Veterans in 2017, the average number of days (in a 30‐day period), for which usual activities (self‐care, work, or recreation) were prevented by poor mental or physical health, was 2.49 days. With regard to this population, aged 18 to 64 years—prime working age, 2.49 missed days of employment/month may translate into a significant economic impact (for both the employee and the employer).

This study has important implications for both policy and practice. It highlights that increased healthcare access alone does not translate into improved HRQoL or health outcomes for women Veterans. Therefore, future policies should not only focus on increasing healthcare access, but also improving health outcomes, especially HRQoL. To do so, challenges that women Veterans face, such as rural/urban disparities in health and access to healthcare as well as other barriers to receiving VHA services, should be considered.^21,40 These challenges do still exist in improving women Veterans’ health despite significant progress made over the last years.⁴¹ For example, HRQoL of women Veterans can be influenced by social, emotional, physical, and other factors.⁴² A comprehensive study by Murphy and Hans⁴³ reveals the need for improvements in every area of population‐health including healthcare, employment, social, and physical environment. Hence, future studies need to include a comprehensive assessment of women Veterans’ health from a population‐health perspective. Such studies need to collect data on various determinants of health including social services, issues related to child care, children’s education, housing, employment, rurality of residence, availability of telehealth services in rural areas, support networks, knowledge gaps regarding VHA services, and caregiving responsibilities.

Another important area that should be studied is the state of knowledge about care of Veterans in nonVHA settings. Unfortunately, BRFSS data do not contain information on care by setting for us to provide evidence specific to nonVHA settings. However, based on published literature we can speculate that there is a knowledge gap in this area. For example, Fredricks and Nakazawa⁴⁴ surveyed physicians (the majority were in primary care or internal medicine) to assess civilian physician knowledge of Veterans’ issues. The authors reported that despite providing care for a high percentage of Veterans, the physician participants recognized the need for more training on Veteran‐specific healthcare issues. A report by Tanielian et al⁴⁵ indicated that only one in three providers were familiar with military or Veteran‐specific health conditions. Limited knowledge and understanding of military service and its impact on health was identified as one of the barriers to provide healthcare to Veterans by nonVHA providers in a qualitative study where ten primary care providers were interviewed.⁴⁶ In that study, other barriers were identified including limited knowledge of resources available to Veterans, as well as the lack of coordination between non‐VHA and VHA systems. Therefore, future research should address these barriers and focus on the development of interventions that can improve Veterans’ healthcare in nonVHA settings. Further, well‐planned implementation of interventions or policies enhancing Veterans’ healthcare outside VHA systems, including time for sufficient development of community service networks, should be considered because, without such, the quality of services provided may be inadequate.⁴⁷

To our knowledge, this article represents the first trend study of women Veterans on HRQoL and health access. Additionally, while most studies on women Veterans have utilized data from VHA databases or utilized Veteran samples selected from Veterans who use the VHA healthcare system, BRFSS is a population‐based survey and the results from this study may be more inclusive of Veterans who do not utilize VHA services. Thus, it is an important contribution to the knowledge about the status of women Veterans’ healthcare.

However, this study is not without limitations. A retrospective, cross‐sectional design allows for associations only, thus prohibiting the ability to identify causality. As a telephone survey, those persons without a telephone, such as impoverished or homeless Veterans (who are particularly vulnerable to healthcare access disparities), are underrepresented. Further, the nature of the BRFSS questions, single item, self‐report, may be vulnerable to bias. Additionally, data related to Veteran‐specific health conditions, such as PTSD, anxiety, suicidal ideation, suicide attempts, and traumatic brain injury, were unavailable and prevented a more thorough investigation of chronic health conditions affecting Veterans. Finally, while it is likely that population‐based survey data are more inclusive of Veterans who do not utilize VHA services, we were unable to identify those who exclusively use VHA services, exclusively use nonVHA services, or those who are dual users.

5 |. CONCLUSION

This five‐year trend study offers a significant contribution to the existing literature on women Veterans’ health in that it provides a more comprehensive evaluation of women Veterans beyond those who use VHA services exclusively. It is evident from the study that federal policy and VHA expansion have had an impact on improving healthcare access to women Veterans. Thus, this study highlights the importance of healthcare expansion in terms of health insurance as a means to reduce cost as a barrier to care and increase the number of persons who report a usual source of care. However, it is evident that improving healthcare access alone does not improve HRQoL and health outcomes for women Veterans. Future research should include more complex research designs to identify other variables, including the quality of care delivered, that may affect health outcomes and HRQoL for women Veterans.

AUTHOR BIOGRAPHIES

Billie S. Vance is a Clinical Associate Professor and Ph.D. student at West Virginia University. She is conducting research aimed at improving women Veterans’ health and the quality of care delivered to women veterans from non‐VHA care settings.

Khalid Alhussain, Pharm D, is a Ph.D. student in the Department of Pharmaceutical Systems and Policy at West Virginia University. He is also a Teaching Assistant in the College of Clinical Pharmacy at King Faisal University. Alhussain’s areas of research interests include health outcomes research, comparative effectiveness research, and patient‐reported outcomes.

Usha Sambamoorthi, Ph.D., is a Professor and Interim‐Chairperson of the Department of Pharmaceutical Systems and Policy at West Virginia University (WVU) School of Pharmacy. She is also the Director of the Graduate Program in Health Services and Outcomes Research. Dr. Sambamoorthi’s research focuses on improving population health through examination of healthcare access, quality, and outcomes using “real‐world” large data.