Veterans’ Preferences for Receiving Information about VA Services: Is Getting the Information You Want Related to Increased Healthcare Utilization?

Abstract

Although the current cohort of returning veterans has engaged more fully with care from the Department of Veterans Affairs (VA) than have veterans from previous eras, concern remains regarding low engagement with VA services,¹ particularly for specialty services for diagnoses that can most negatively impact quality of life. The current study used the framework of the Andersen model² to examine factors related to VA healthcare use in Operation Enduring Freedom/Operation Iraqi Freedom (OEF-OIF) veterans. Match between veterans’ preferences for source of information about VA programs and veterans’ actual sources of information about VA services was examined as an additional predictor of help seeking. The study included 1,161 veterans recruited from the southeast United States. Results suggested that veterans prefer to receive information from VA publications and the web, while they actually receive information from VA publications and other veterans. Logistic regression suggested that number of deployments, income, distance to VA, VA disability rating, self-rated health, and match between preferred source of information and actual source of information were significantly related to use of VA services since deployment. These results suggest that future outreach efforts should focus on targeting veterans’ health needs and preferences for care and source of information.

Access to and utilization of Department of Veterans Affairs (VA) Healthcare are issues that have garnered significant attention. Recent analyses have suggested that a majority (59%) of veterans serving in support of Operations Iraqi or Enduring Freedom (OEF-OIF) are enrolled in VA care, which represents an increase in use over past service eras³. For example, only about 20% of deployed Persian Gulf Veterans had enrolled in VA care by the year 2009.⁴ While recent policy changes have been enacted to improve access to services for veterans from all eras⁵, concerns remain regarding OEF-OIF veterans’ low engagement with VA services¹ as well as specialty services for diagnoses that can most negatively impact quality of life, such as post-traumatic stress disorder^6,7 (PTSD). For example, a 2004 study of returning veterans⁸ found that out of those veterans who screened positive for PTSD post-deployment, only 23% of OEF veterans and 29–40% of OIF veterans had received mental health services in the past year. The current study seeks to identify potential factors that might increase OEF-OIF veterans’ utilization of VA Healthcare, with a focus on how veterans prefer to receive information about VA programs.

Many studies have used the Andersen Model^2,9 to describe access to health care^10,11. This framework suggests that healthcare utilization is related to predisposing factors, such as demographic variables, enabling factors, such as income and health insurance, and need factors, including mental and physical health problems. A study of Vietnam veterans examined the Andersen Model in the prediction of any healthcare use over the prior 6 months, and indicated that although predisposing and enabling factors including age, income, and combat exposure were significantly related to mental and physical healthcare use, that the need variables were most strongly related to healthcare use¹¹.

A similar study based on the 2001 National Survey of Veterans found that older age, male gender, minority racial status, lower levels of education, unmarried status, combat exposure, lower income, lack of insurance, and overall greater symptom severity were all related to increased help seeking¹⁰. This study also found that need-based variables were the most significant predictors of utilization. Although the model has not been formally tested in OIF-OEF veterans, a recent study of mental health care utilization in returning veterans found that probable need for services, income, combat exposure, and perceptions of VA care were significantly related to use¹².

A majority of the research addressing healthcare utilization in returning veterans has focused on barriers to care such as cost, stigma, and health beliefs, with the intention of identifying targets for intervention for overcoming these barriers^1,13. Relatively recent research has begun to focus additionally on veterans’ preferences for care, with an eye towards creating tailored programs and outreach^14,15 that might draw patients towards needed care. In a qualitative study of veterans who had an active disability claim for PTSD, veterans reported that acceptance of diagnosis, availability of help, treatment encouraging beliefs, system facilitation, and social network encouragement were the major issues that factored into their treatment seeking¹⁴. In a national sample of OIF-OEF veterans, Sayer et al further found that veterans enrolled in VA care were highly interested in receiving information on VA services, particularly information on benefits (83%), occupational assistance (80%), self-help materials (75%), educational classes (62%) and individual therapy (61%). This research also suggested that veterans preferred information about VA services from the VA (57%), on the Internet (53%) or by mail (53%) or e-mail (43%).

These studies bring to light the importance of attention to veterans’ preferences for care and information, but to date, only limited research has examined the relationship between veterans’ healthcare preferences and actual help seeking behavior¹⁶. The current study uses the framework of the Andersen model to investigate the potential role of veterans’ preferences for receiving information about VA programs in relation to their use of VA programs. In particular, the current research examines the match between veterans’ preferred method of receiving information (e.g. television, website) and how they actually received this information. The researchers hypothesize that match between preference for and actual receipt of information about VA programs will be positively related to veterans’ use of available services, in addition to the factors predicted by the Andersen Model.

Methods

Participants

Participants included 1,161 OIF-OEF veterans recruited from the VA Veterans Integrated Service Network 6 covering North Carolina and parts of Virginia and West Virginia. Veterans had an average age of 39 years(standard deviation [SD] = 9.9), an average income of $54,500 (SD = 33,000), and an average of 1.6 deployments (SD = 0.9). For further descriptive statistics, see Table 1.

Table 1.

Descriptive Statistics

Variable	N	%
Male	959	82.7
Female	196	16.9
Ethnicity
White	818	70.5
African-American	198	17.1
Latino	78	6.7
Other	79	7.1
Married	816	70.3
Unmarried	331	28.5
Dependent Children	727	64.5
No Dependents	412	35.5
Currently Employed	859	74.1
Currently Unemployed	301	25.9
VA Disability Rating
None	612	52.8
0 – 40%	332	28.5
≥ 50%	182	15.79
Current Health Rating
Excellent	135	11.6
Very Good	304	26.2
Good	443	38.2
Fair	205	17.7
Poor	47	4.1
Military Component
Reserves	219	18.9
National Guard	220	19.0
Active Duty	689	59.4
Officer	299	25.7
Enlisted	850	73.3
Military Branch
Army	567	48.9
Navy	249	21.5
Air Force	165	14.2
Marines	154	13.3
Match between preferred and actual info source
0	609	52.5
1	205	17.7
2	180	15.5
3	142	12.2
4+	24	2.0
Used VA Services since Deployment
No	756	65.2
Yes	404	34.8

Recruitment and Procedures

A list of potential separated active duty personnel, National Guard members and Reservists who had deployed to Iraq or Afghanistan was obtained from the Defense Manpower Data Center. Out of approximately 70,000 eligible veterans in the Veterans Integrated Service Network 6 area, 5,000 were randomly sampled. Out of the 5,000 surveys, 1.4% were uncompleted due to extenuating circumstances, 18.5% were undeliverable, and 80% were successfully delivered. Of the 4,004 surveys that were delivered, 1161 were completed and returned for a final cooperation rate of 29%.

A survey contractor was used throughout the study to enhance privacy protections, so no identifying information was obtained by study researchers. A Dillman Total Design Method¹⁷ was used, and participants were first sent an alert letter 1 week in advance of the study informing them of the nature and purpose of the research project. They then received the survey package which included the survey as well as a cover letter with study contact information, informed consent and confidentiality information. A follow-up packet was mailed if the veteran had not responded within 5 weeks, and veterans were considered non-responders if they did not respond within 5 weeks of the second survey mailing. This study was approved by the Durham VA Medical Center Institutional Review Board and the VA Office of Management and Budget.

Measures

The complete survey consisted of questions regarding demographic information, military history, rating of physical and mental health, and health care service utilization. The selection of items was informed by qualitative needs assessment utilizing six 10–12 person focus groups¹⁸. In this study, the outcome variable of interest was binary response to the question, “Since your last deployment, have you used VA for healthcare services?” Also examined were predisposing factors including gender, ethnicity, marital status, and number of deployments in addition to the enabling variable of income. Need for services was represented by level of VA disability, rated on a scale from 0% to 100% disability in increments of 10%, and on self-reported rating of health ranging from Excellent to Poor.

At the conclusion of the survey, veterans were asked to respond to the following question, “People learn about VA programs from different sources. Did you receive any information about VA programs from the following sources?” and asked to indicate yes/no if they had received information from: newspapers, TV, radio, VA information or publications, other government publications, other veterans, discharge counseling, Posdeployment Health Reassessment (PDHRA), veterans’ organizations, website, friends/family/co-workers, chaplain/religious leader, command/superiors, family readiness groups, combat stress control/behavioral health contacts, toll-free hotline, Military OneSource, or other. They were then asked, “In the last column above, please indicate how you would prefer to receive information in the future.” For the purposes of this study, a variable was computed that represented a count of the number of matches between the format in which veterans received information about VA programs vs. the format in which they would prefer to receive such information.

Data Analysis

Individual items were initially examined to determine veterans’ preferences for information source as well as the number of matches between their preferences for and actual receipt of information. A chi-square test was used to examine the relationship between match and service utilization, and a multivariate logistic regression was used to predict service utilization with demographics, need variables, and match factors included as covariates.

Results

Results were first analyzed by examining the relative frequencies of how veterans actually received information about VA programs versus how veterans would prefer to receive information about VA programs (See Table II). Frequency data indicated that most veterans received information through VA publications, and most veterans also preferred this method of communication. However, relatively few veterans indicated a preference for receiving information from other veterans and from discharge counseling, although the number receiving information in these formats was quite high.

Table 2.

Frequency Data Comparing Veterans’ Receipt of Information vs. Preferences for Information Source

Source	% of Veterans that received info this way	% of Veterans that would prefer to receive info this way
VA Publications	74	37
Other Veterans	69	6
Counseling at Discharge	59	11
Internet	57	33

Preliminary analysis of the relationship between information source match and VA use revealed a significant chi-square, X²(1) = 3.47, p < .05 (1 sided), and indicated that 207 veterans in this study had used VA services since deployment and had at least 1 match between how they preferred to receive information and how they actually received information. An exploratory logistic regression was then performed with gender, racial status, age, marital status, number of deployments, information source match, employment status, income, distance to closest VA hospital, self-reported health level, and VA disability level as predictors of VA health care use. Although the overall model was significant X²(6, N=1070) = 245.88, p < .001, a number of the predictors were not, including gender, age, race, marital status, and employment status (p = .43 – .89).

The model was respecified using only the significant predictors, and the results of this analysis are presented in Table III. Predisposing factors (i.e. number of deployments), enabling factors (i.e. income, distance to VA), need factors (i.e. health rating and VA disability), and matched preferences for receiving information were each associated with VA healthcare utilization. Odds ratios indicated that an increase in match in information source was related to a 13% increase in the likelihood of help seeking. Health status was the strongest predictor of VA healthcare use, with a decrease in health status increasing the likelihood of seeking help by 45%

Table 3.

Logistic Regression Predicting VA Healthcare Utilization

	−2 Log Likelihood		Nagelkerke R2		X2
Overall Model	1146.34		28.2		10.00
Variable	B	Wald	Sig	Exp(B)	95% CI
Match	.13	4.63	.03	1.13	1.01–1.27
Health	.37	20.03	.00	1.45	1.23–1.70
VA Disability	.18	72.00	.00	1.20	1.15–1.25
Income	−.23	54.72	.00	.79	.74–.84
# Deployments	−.27	8.97	.00	.76	.64–.91
Distance to VA	−.20	7.65	.01	.82	.71–.94

Discussion

The results of this study both confirm and extend the application of the Andersen Model in a VA healthcare population. These results suggest that predisposing, enabling, and need factors, including number of deployments, income, distance to VA, health rating, and disability rating, are relevant predictors of VA service utilization. Notably, preferences for care also emerged as a key variable in the prediction of VA help seeking. Results suggest a strong relationship between service utilization and match between how veterans receive information about VA services versus how they would prefer to receive information, such that an increase in match of information source is related to an increase in service use. Although need, represented by disability and self-evaluation of health, was the strongest predictor of service use, these findings suggest that veterans’ preferences are an important component that should be considered in the modeling of help seeking behavior.

As stated, the variables representing need for care, including decreased rating of one’s health and increased level of disability, both predicted increases in help seeking as expected based on previous research¹². The enabling factors in this study, including income and distance to closest VA, were also significant predictors of service use. Further distance to VA was related to decreased help seeking, confirming that access to care has an important relationship to use of care. Income was predictive of help seeking, but in an unexpected direction. Although income is typically seen as enabling the help seeking process², in this study higher income led to less VA healthcare use. This is likely due to the higher rates of VA healthcare utilization by enlisted personnel and those without private insurance^3,18.

The predisposing variable of exposure to combat was negatively related to service use in this study, which was surprising, given that combat exposure has been related to increased service use in previous research¹⁰. This discrepancy may be due to sampling. Whereas veterans in previous studies had a wide range of combat exposure, including many who had seen no combat, all veterans in the current study were deployed in the service of OIF/OEF. In light of this sample, it is possible that combat exposure is related to decreased service use due to increased barriers commonly associated with immersion in military culture such as reluctance to acknowledge problems and ask for help. Other predisposing factors, including gender, age, race, marital status, and employment status, were not related to VA healthcare use in this study. Results from past studies have been variable regarding the significance of similar demographic variables. However, this result is generally consistent with the finding that more treatment-related variables, such as need and preferences for care, have stronger relationships with service utilization than do predisposing variables.

Overall, this study demonstrated that the Andersen Model is a relevant conceptualization of help-seeking behavior in the veteran population, with particular contributions from veteran characteristics reflecting need for and access to care. However, this study also suggests an extension of this model to include veterans’ preferences for information about VA programs. Several recent articles^14,15 have begun to call for attention to veterans’ preferences in our attempts to increase VA service utilization, and the current study again confirms that preferences are indeed a relevant factor in veterans’ use of care. Although many veterans are receiving information in a way in which they prefer, often from VA publications, results from this sample indicate that many veterans are receiving information in ways that they would not prefer (such as from other veterans), and that many veterans would like to receive more information from the web. It appears that one potential way to bridge the gap between veterans and the wide range of VA healthcare services available to them may be to tailor outreach and advertising campaigns to the preferences for receiving information as well as the most pressing health concerns of returning veterans.

The study is limited by a restricted response rate, and consequently, results should be interpreted cautiously. As with preferences for information about care, it appears that veterans in this era have relatively low response rates for paper-and-pencil techniques¹⁹ and may respond better to research queries online. Despite this limitation, sample characteristics closely map on to those of larger studies of OIF/OEF veterans^1,20 and appear to match closely with figure representing veterans of this era. Future studies might focus on these issues at a national level, and could likely improve cooperation rates by employing data collection techniques that might appeal more to returning veterans. In conjunction with other recent studies addressing the critical role of veterans’ preferences for care, this study indicates the need for further research on the preferences for type of care and source of information about care in returning veterans of this era. Targeted marketing and outreach tailored to the current cohort’s preferences may play a key role in engaging this group of veterans in VA care.