Health Care Experiences and Health Outcomes of African-American Veterans

Despite overall improvement in the health status and longevity of the general U.S. population over the last several decades, Black Americans are less healthy, receive poorer care, suffer from more chronic diseases and have a shorter life span than White Americans. As a group, African-Americans are economically disadvantaged and have little or no health insurance. These factors play an important role in explaining Black-White differences in health status and in the quality of health care they receive. Veterans of the armed forces, however, are one population of African-Americans that typically have health insurance coverage, obtained through the Veteran’s Administration, VA. It is reasonable to expect, therefore, that African-American veterans have greater access to health care than their counterparts in the general population. Yet, health insurance coverage through the VA does not necessarily guarantee access to services and fair treatment by health care providers.

Prompted by concerns about the responsiveness of health care providers at the VA and their effect on their health status, the Tri-Ad Veterans, a Boston-based grassroots group of African-American veterans, asked the Institute on Urban Health Research at Northeastern University and the Boston Public Health Commission to look at the matter more closely. As a result, a pilot study, “Health Care Experiences and Health Status of African-American Veterans,” was conducted during the spring and summer of 2006.

A sample of African-American veterans was recruited via community outreach in the Roxbury, Dorchester, and Mattapan areas from May to June 2006. All study participants were age 18 or older, had identified themselves as African-American, had proof of veteran status, were fluent in English, and had received health care services in the past year. The sample of 141 African-American veterans was mostly male and middle-aged. Less than a quarter of the sample had completed college and/or gone on to graduate school. Only a fifth of the sample was currently married and almost 60% of the sample lived alone. Only eleven respondents (8%) were unemployed and looking for work for more than one year. Three of these eleven had a certified service-related disability. There were 37 other individuals reporting having a certified service-related disability (yielding a total of 40 or 28.4% of the sample). A majority of respondents reported having served in the Army and not having seen active combat.

Veterans responded to 82 items related to the following: demographics, health care experiences, satisfaction with health care services, opinions about discrimination in health care services, the quality of health services, physical and mental health status, and experiences of discrimination in other aspects of life.

While the majority of participants reported their health status to be good or excellent, a sizeable minority did not. The latter reported a high prevalence of high blood pressure, low back pain, arthritis, and diabetes, as well as mental health problems including depression and post traumatic stress disorder, many of which seemed unrelated to the older age of the sample.

On average, study participants were moderately satisfied with their health care. Yet, veterans indicated their dissatisfaction in areas related to their confidence in the diagnosis given by their doctors, access to medical specialists, time doctors spend with patients, and getting medical care in a timely manner. About 46% of African-American veterans had confidence in their providers and about 30% said they felt rushed when with the doctor. Further, over half reported not agreeing with a statement that black veterans were treated with the same respect by doctors and nurses as white veterans. Also, approximately 78% of the respondents did not disagree with a statement that they recalled having an experience of discrimination where they received health care services, suggesting that around 20% of our sample felt there was some discrimination by their health care provider.

An important finding was that African-American veterans disagreed with the statement that their health care providers asked about their spiritual needs. Many respondents expressed concern about this issue to researchers during survey administration. Future research needs to explore ways for VA providers to integrate chaplain services with medical care so African-American patients feel their spiritual needs are being met holistically by the medical team.

About two-thirds of the sample did not use the VA as their sole provider. This suggests that a large percentage of African-American veterans are using services outside a health care system that is available to them by function of their past military service. The decision to use non-VA services may have to do with different health benefits from current employment and/or dissatisfaction with VA services. However, only about 10% of the sample indicated a lack of willingness to use the VA in the future. Future research should examine reasons why Black veterans use services outside the VA system.

Uncovered were significant relationships between perceived discrimination from health care providers and satisfaction with care, and between quality of care and physical functioning. Urgently needed are a comprehensive review and monitoring of policies and provider behaviors that may wittingly or unwittingly contribute to unequal treatment in health care or perceptions of such treatment. Efforts to educate administrators and service providers about health disparities, unconscious clinician bias and cultural competency, including the assessment of spiritual beliefs, especially how these relate to patients’ illnesses, should be strengthened and made compulsory.

INTRODUCTION

despite overall improvement in the health status and longevity of the general U.S. population over the last several decades, racial/ethnic disparities in health continue to persist and in some cases have increased (Institute of Medicine, 2003; Mills and Combs, 2002). Specifically, compared to White Americans, Black Americans are less healthy, receive poorer care, suffer from more chronic diseases and have a shorter life span (Spalter-Roth et al., 2005).

There is strong evidence that economic and social inequalities result in greater personal risk behaviors, more risky social and physical living environments (Borrell and Hatch, 2005), and policy (Kennedy, 2005) and provider biases (Spalter-Roth et al., 2005; Kawachi et al., 2005). As a group, African-Americans are economically disadvantaged and have little or no health insurance (Doty and Holmgren, 2006). In addition to having few economic resources for obtaining adequate health care, African-Americans may also have relatively little access to health education. Such inequalities play a significant role in explaining Black-White disparities in health status (LaVeist, 1992; Sampson, 2000) and in the quality of health care (Institute of Medicine, 2003). Perhaps what is less clear is how having health insurance and greater access to health care reduces these disparities and discrimination in health care.

Veterans of the armed forces are one population of African-Americans that typically has health insurance coverage (obtained through the Veteran’s Administration, VA). It is reasonable to expect, therefore, that African-American veterans have a greater access to health care than their counterparts in the general population. Yet, health insurance coverage through the VA may not necessarily guarantee actual access to services and fair treatment by health care providers. Institutionalized inequalities in access and quality of health care documented in the general population may be mirrored in the military health care system (Borowsky and Cowper, 1999; Mayberry et al., 2000). There is evidence that there are some race/ethnic-based differences in procedures and services provided to patients seen at the VA (Ferguson et al., 1998; Petersen et al., 2002; Goldstein et al., 2003). For example, Goldstein and associates (2003) found that hagiographies and ECGs were performed significantly less in Black than White stroke patients. It is important to note that veterans may not use the VA system for a variety of reasons, e.g., they may have better insurance through employment or they may not be located near a VA system.

Prompted by concerns about the responsiveness of health care providers at the VA and their effect on their health status, the Tri-Ad Veterans, a Boston-based grassroots group of African-American veterans, contacted the Institute on Urban Health Research at Northeastern University and the Boston Public Health Commission in order to request a study to look at the matter more closely. A study, “Health Care Experiences and Health Status of African-American Veterans,” was conducted during the spring and summer of 2006. The study focused on describing the following among a sample of African-American veterans in Boston: (1) their health status,

(2)

their perceptions regarding the quality of care they receive,

(3)

their perceived discrimination from health care providers, and (4) the relationships between their perceived discrimination and their experiences in health care and their health status.

To best understand the bases of the study, we present a targeted literature review highlighting key differences in indicators of health status among African-American veterans, dual utilization of health care services in the veteran population, and economic and social factors that contribute to racial discrimination and subsequent disparity in health care services. The literature review is followed by sections on the methodology of the study, results and a discussion of findings, and conclusions/recommendations.

Health Disparities in the General Population

Several indicators demonstrate disparities in the health of Americans based on their race. First, although average life expectancy among Americans has increased over the years (American Sociological Association, 2003), African-Americans continue to live an average of 7 fewer years than White Americans (National Center for Health Statistics, 1994) and experience a higher health-related death rate (Anderson and Smith, 2005). Second, African-Americans have disproportionately higher rates of a host of health conditions compared to White Americans. These include higher rates of cancer, heart disease, HIV and AIDS, diabetes and infectious diseases (Institute of Medicine, 2003). Not surprisingly, African-Americans also rate their health status lower than that of Whites (U.S. Department of Health and Human Services, 2004) and have higher rates of disability (Institute of Medicine, 2003; U.S. Department of Health and Human Services, 2006). Overall, more African-Americans than White Americans suffer from various health conditions and die prematurely (U.S. Department of Health and Human Services, 2006). African-American patients are more likely than White patients with identical psychiatric symptoms to be given more severe psychiatric diagnoses (Loring and Powell, 1988; Strakowski et al., 1993). Lastly, there are disparities in access to health care and in the experiences that patients report having in the health care system (Commonwealth Fund, 1999).

Similarly, access to quality health care seems to differ significantly by race even when patients have health insurance. Studies consistently indicate that compared to White Americans, African-Americans are less likely to have a regular source of care, to receive preventive health care, to be satisfied with the health care they receive, and to receive advanced treatments when these are needed (Collins et al., 2002).

Disparities in health status and access to care in Boston are not dissimilar to the patterns reported above. A recent report by the Boston Public Health Commission (2005) highlights the vast disparities in health status and health care access between Whites and African- Americans and other minority populations.

Health Disparities among Veterans

Some of the racial disparities in both health status and access to quality care noted above are mirrored among veterans. There appears to be no known information on the life expectancy of African-American veterans compared to White veterans. There is mixed data regarding mortality rates. For example, Selim and associates (2004) demonstrated that there were no differences in mortality rates in a sample of African-American and White veterans receiving outpatient VA care. Jha and colleagues (2001) found significantly lower rates of mortality among hospitalized African-Americans than among hospitalized White veterans. Kamalesh et al. (2005) also found that mortality from post myocardial infarctions was significantly lower in Black veterans with diabetes than White veterans with diabetes. Dominitz et al. (2002) found higher rates of mortality among Black veterans with squamous cell carcinoma than their White counterparts but no racial differences in mortality from adenocarcinoma. McGinnis et al. (2003) found higher death rates from HIV/AIDS among African-American veterans than among White veterans. There are several reports indicating greater morbidity and disability rates among African-American veterans compared to White veterans. For example, compared to their White counterparts, African-American veterans have higher rates of hospitalizations (McGinnis et al., 2003; Deswal et al., 2006), greater severity of stroke at time of hospital admission (Jones et al., 2000), greater complications post knee arthroplasty (Ibrahim et al., 2005), greater grafter failure post kidney transplantation (Chakkera et al., 2005), and poorer blood pressure control (Bosworth et al., 2006). Studies also report disproportionately higher rates of diabetes, high blood pressure, impaired movement and poor or fair health, as well as lower mental health functioning among African-American veterans compared to White veterans (Horner et al., 1991; Villa et al., 2003).

Despite these negative morbidity findings, Villa and colleagues (2003) did find that African-American veterans self reported higher levels of physical functioning than their White counterparts.

The data on access and quality of care are mixed with regards to the presence of racial disparities among veterans. For example, compared to White veterans, African-American veterans are less likely to receive certain needed medical procedures, such as diabetes treatment and monitoring (Heisler et al., 2003), knee arthroplasty (Jones et al., 2005), heart procedures (Ibrahim et al., 2005), and cardiac endarterectomy for treatment of carotid stenosis resulting from a stroke or transient ischemic attack (Oddone et al., 1993). Goldstein et al. (2003) noted no racial differences in the utilization of a variety of stroke-related procedures and services with exception of African-American veterans receiving significantly fewer cerebral angiographies and 24-hour ECGs. Deswal et al. (2006) found no significant racial differences in the quality of outpatient care of heart failure.

With regard to prescribed medications, fewer disparities seem to exist based on studies conducted to date. One study reported that African-American veterans were less likely than White veterans to receive thrombolytic therapy upon arrival for acute myocardial infarction (MI) (Petersen et al., 2002). Yet, research has also demonstrated that African-American patients in the VA are equally likely to receive beta-blocker therapy for acute MI, and somewhat more likely to receive angiotensin converting enzyme (ACE) inhibitor therapy at discharge (Petersen et al., 2002). Lipid-levels and receipt of lipid lowering therapy were found to be equal between White and African-American patients in the VA (Woodard et al., 2004). It was suggested that the equal access to pharmacy services may account for the lack of disparity. However, Safford et al. (2003) found that African-Americans were less likely to receive lipid-lowering therapy. In a study examining whether disparities existed in the use of aspirin therapy and smoking cessation therapy for the secondary prevention of coronary artery disease, African-American and White patients were equally likely to receive both interventions (Ambriz et al., 2004). In a subset of patients with hypercholesterolemia and a prior stroke, African-American patients were less likely to receive aspirin therapy (Ambriz et al., 2004). In research with female veterans with documented low bone mass, African-Americans were less likely than Whites to be prescribed calcium supplements and anti-resorptive medications (Wei et al., 2003). African-Americans also scored lower in satisfaction with their care provider (Harada et al., 2002).

Discrimination in VA Care

The Institute of Medicine’s report “Unequal Treatment: Confronting Racial and Ethnic Disparities in Health Care” consistently demonstrated a decrease in the availability of health care procedures and treatment for racial and ethnic minorities and lower quality of service (Institute of Medicine, 2003). This was true despite insurance status, income, age, and severity of medical conditions. Bias, discrimination, and racism at the levels of the health care system, health care provider, and the provider-patient interaction are increasingly being examined as factors that influence health status. Though disparities have been examined, there is a limited amount of research regarding discrimination, health care, and health status related to African-American veterans in or outside the VA health system.

Discrimination may be inferred (Kreiger, 1999) indirectly from studies that examine whether known risk factors such as socioeconomic status, severity of illness, and insurance status explain differences in health status or receipt of care between racial groups served by the VA health system. The following are examples of such studies done in the VA health system. Whittle et al. (1993) suggested that clinical or social bias may have been a factor in their finding that Whites were more likely than Blacks to receive invasive cardiac procedures. Other researchers reported similar findings (Sedlis et al., 1997). Similarly, research with stroke patients in the VA health system demonstrated that clinician referral bias was an important explanation of racial variation even after controlling for known risk factors such as socioeconomic status and access to care (Oddone et al., 1993). Blow et al. (2004) suggested that clinician factors were a likely factor in accounting for ethnic disparities in diagnostic patterns of psychosis among veterans.

Studies are lacking that directly measure discrimination among veterans at the individual level. Kressin et al. (2004) conducted an examination of whether patients’ beliefs, such as evaluations of physicians’ degree of respect and caring, their trust in their physicians, and quality of cardiac care in the VA, were associated with racial disparities in cardiac catheterization rates. They found that Black patients were less likely to receive cardiac catheterization than Whites and were more likely to report experiences of racial and social class discrimination than Whites.

Another aspect of racism for veterans relates to their military experience. The military experience for veterans of color is unique. The stress of wartime service can be particularly exacerbated by the isolation and stress of discrimination and racism which contribute to post traumatic stress disorder (PTSD) in veteran populations (Loo, 1994; Loo et al., 2001; Johnson, 1994; Ruef et al., 2000). In particular, it has been found that African-American veterans experience a greater rate of PTSD than White veterans (Allen, 1986). Jones et al. (2000) discussed the effectiveness of incorporating support and coping strategies for dealing with racism into group therapy programs for African- American veterans. It may be important to consider accounting for this unique experience of discrimination as it relates to the receipt of health services in the VA and, ultimately, to health status. In addition, when psychiatrists treat African-Americans for PTSD they are much less likely to attribute the PTSD to combat than when treating PTSD in Whites (Murdoch et al., 2003). This is particularly damaging for African-American veterans because having PTSD that is not service-connected results in the denial of priority services at the VA (Murdoch et al., 2003).

Dual Use of VA and Non-VA Services

One of the concerns regarding discrimination in VA health care is that it may be pushing veterans of color to use services outside of the VA health system or, worse, may dissuade them from getting any health care at all. In a study of 577 veterans who reported VA primary care visits, 28% also reported non-VA primary care visits (Borowsky and Cowper, 1999). While the investigators did not examine race, veterans who were not satisfied with VA care and veterans who lived relatively close to a VA center (usually in urban areas) were significantly more likely to be dual users. The odds of dual use were reduced for those without other insurance and with less education.

A 1992 study using data from the National Survey of Veterans examined the relationship between race/ethnicity and unmet health care needs for ambulatory care users of VA and non-VA facilities (Washington et al., 2002). The researchers demonstrated that Black and Hispanic veterans were most likely to report any VA use. American Indian/Eskimo, Hispanic, and Black veterans were 4.4, 2.5, and 1.9 times more likely, respectively, to report not being able to get needed care compared to White veterans. However, after these researchers adjusted their analyses for VA ambulatory care use, they found less disparity between American Indian/Eskimo or Hispanic veterans and White veterans and no disparities between Black and White veterans.

A similar investigation examined woman veterans’ reports of dual use of VA and non-VA providers based upon the VA health care provider’s gender, provision of routine gynecological care by the VA provider, use of VA women’s clinic, and overall satisfaction with VA care (Bean-Mayberry et al., 2004). The study demonstrated that provision of routine gynecological care by VA providers and use of VA women’s clinics was associated with a significantly lower likelihood of dual use. On the other hand, dissatisfaction with care and higher income were significantly associated with an increased likelihood of dual use. Thus, there appear to be two different reasons for using health care services outside the VA: having another presumably better health care insurance plan and dissatisfaction with VA services.

Causes of Health Disparities

There are a number of factors which alone or in combination are responsible for the levels of inequality in health status affecting racial and ethnic minorities. Beyond socio-economic and environmental factors -- ethnic and racial minorities tend to live in disadvantaged areas with high incidences of environmental pollutants like lead paint which harm human development (Jacobs et al., 2002), there are other barriers to access and quality of care different groups receive.

There may be several reasons for disparities in access to health care for veterans, such as (a) discontinued care which means fewer doctor visits which leads to poorer outcomes and emergency room visits (Gary et al., 2004); (b) lack of financial resources which disproportionably affects minorities (Geronimus, 2000); (c) structural barriers including transportation issues and inability to schedule appointment times during convenient hours; (d) scarcity of providers in low-income areas (Lillie-Blanton et al., 2001); (e) literacy issues which make it hard to understand doctor’s orders and medication information; and (f) lack of diversity and cultural awareness in the medical workforce (Chen et al., 2005).

Possible reasons for the disparities in quality of health care include (a) communication problems between the health care provider and the patient that can lead to incorrect diagnosis or improper usage of medications and (b) provider discrimination (Stewart, 1995; Institute of Medicine (2003)). Provider discrimination occurs when a provider of care treats people of different groups according to stereotypes or differently than other patients (Dedier et al., 1999).

The literature reviewed provides the underpinnings for the relationships between aspects of health care and health outcomes we set out to explore in this study of African-American veterans in the Boston area. The literature review points to the need to explore three key research questions: (1) what are African-American veterans’ perceptions of the health care services they receive, discrimination in health care and the status of their own physical and mental health?, (2) what are the relationships between use and/or non-use of VA services and patient satisfaction, quality of care, health care discrimination, and health status?, and (3) what factors, including perceived discrimination, predict satisfaction with health care, quality of care, and physical and mental health status among African-American veterans?

METHODS

Study Design

The study employed a cross-sectional design with a sample of African-American veterans recruited via community outreach. We utilized convenience sampling methods to obtain completed surveys during twelve recruitment sessions at six different locations throughout the Roxbury, Dorchester, and Mattapan area from May to June 2006. Outreach was conducted by Tri-Ad Veterans, a grassroots organization of African-American veterans. Criteria for participation included: age 18 or older, self identified as African-American, having proof of veteran status, English fluency, and having received health care services in the past year. The study was approved by the Northeastern University Institutional Review Board and funded by the Boston Public Health Commission and Institute on Urban Health Research at Northeastern University.

Study Procedures

The study was advertised on multiple occasions through a local television station (weekly program on BNN, “Speaking About”), a local radio station (W.I.L.D., FM and WRBB), and a local newspaper (Bay State Banner). Flyers with recruitment sessions were posted at the different locations where veterans were likely to congregate such as at the American Legion Posts. At least one Northeastern University (NU) trained researcher and one representative of the Tri-Ad Veterans group were present at each of the recruitment sessions. The Tri-Ad Veterans representative assisted in identifying potential respondents, handing out survey materials, verifying the veteran status of the respondents, distributing and collecting survey materials, and documenting the distribution of the $20 research incentives. The Northeastern University researcher assisted in obtaining informed consents from subjects, distributing and collecting surveys and consent forms, and documenting the distribution of the $20 research incentives.

At each recruitment session, the NU researcher and community representative provided potential respondents with a consent form and survey. The NU researcher explained the study and briefly reviewed key points in the consent form. Respondents were given instructions to complete and sign the consent form if they agreed to the study conditions and then completed the survey and returned it to the researcher. The survey took respondents on average 20-25 minutes to complete.

Survey Instrument

The survey was developed by NU researchers in consultation with Tri-Ad Veterans and tested for face and content validity among a small group of veterans. It was comprised of 82 items related to the following: demographics, health care experiences, satisfaction with health care services, perception of discrimination in health care services, perception of the quality of health services, physical and mental health status, and experiences of discrimination in other aspects of life. The survey used a variety of scales and different question formats including close-ended and open-ended items.

The first section (items 1-10) involved questions about the respondent’s age, gender, ethnicity, educational level, marital status, if respondent lives alone and employment status. These items were asked since they are often associated with changes in physical and mental health status and because of the possibility that such factors had to be controlled for when looking at relationships between perceived discrimination and health status. There were four items in this section about the respondent’s military service: (1) the branch of the military served, (2) participation or not in active combat, (3) if respondent had a service-connected disability, and (4) if the respondent had been certified for this disability and what the service rating was. The latter two items were asked as possible control variables for the effects of service-connected disability on health status.

The second section (items 11-21) asked respondents about their access to and overall satisfaction with health services. Items on access to health care services asked if the respondent had someone to take them to the doctor, if they had used health services in the past three years, if they had been seen by a provider (VA, non-VA, etc.) in the past 12 months, if they planned to use the VA in the future, and if they had any form of health and prescription insurance outside of VA coverage. The item asking about the receipt of services from a VA and/or non-VA provider in the past 12 months served as a key study variable in the bivariate and multivariate analyses. This variable was subsequently dichotomized into two groups based on significant differences uncovered in the data analyses: those who received care from a combination of VA and non-VA providers and those who received care from either VA-only providers or non-VA only providers. Most of these items were based on the Veterans SF-36 questionnaire (Kazis et al., 2004b). Another item asked respondents to indicate if they had ever decided not to disclose their veteran status in a non-VA setting and why (a separate open-ended follow-up item). This latter item aimed to identify the extent that the veteran label affected the veteran’s comfort in using non-VA services. Open-ended questions in this section asked respondents why they were not currently using VA services and/or not planning to go to a VA facility in the future. Respondents were asked to rate their satisfaction with the health care services they received from their provider and to rate the quality of those services. Several items in the next section of the survey focused on specific aspects of the health services respondents had received.

The third section (items 22-39) directly utilized items from the short-form of the Patient Satisfaction Questionnaire (PSQ)III developed by the Rand Corporation (Marshall and Hays, 1994). This short form, known as PSQ-18, contains 18 items tapping each of the seven dimensions of satisfaction with medical care measured by the PSQ-III: general satisfaction, technical quality, interpersonal manner, communication, financial aspects, time spent with doctor, and accessibility and convenience. Respondents utilized a Likert scale ranging from strongly agree (1) to strongly disagree (5). Since several items were positively stated, these items were recoded during data analysis. Thus, the higher the PSQ-18 score, the more satisfied respondents were with their health care services. The PSQ-18 subscale scores are substantially correlated with their counterparts in the larger, 50-item PSQ-III and possess adequate internal consistency and reliability (Marshall and Hays, 1994). Moreover, both the magnitude of the correlation coefficients and the overall pattern of correlations among PSQ-18 subscales are highly similar to those observed for the PSQ-III. The Rand Corporation suggests the use of the PSQ-18 in situations where the need for brevity precludes administration of the full-length PSQ-III. To spare our respondents the burden of a long survey, we chose to utilize this short version of the PSQ-III. The PSQ-18 was shown to have excellent reliability in this sample (Cronbach’s alpha= 0.937).

The fourth section (items 40-44, 49-57, 59) included several questions about the respondent’s quality of medical care. Respondents used a Likert scale ranging from strongly agree

(5) to strongly disagree (1) where a higher score reflected better quality of care. These questions came from previously published items. Items 40-44 were taken from the Evaluation of VA Care scale reported by Kressin and colleagues (2002). Other items, 49-57 and 59, were adapted from those used in the 2001 Commonwealth Fund’s Minority Health Survey (Collins et al., 2002). The reliability analyses indicated this quality of care index to have excellent reliability (Cronbach’s alpha=0.931).

The fifth section, (items 45-48 & 58) asked close-ended and open-ended items about perceived discrimination in health care. These items were adapted from the Perceptions of Racism Scale (Green, 1995). We developed our perceived health care discrimination measures using three of these items: items 45, 47, and 58. We selected these three items since items 45 and 47 focused specifically on racial discrimination among physicians and nurses, and item 58 was a global discrimination measure about whether respondents personally experienced discrimination from their providers. The reliability analysis showed this three-item perceived health care discrimination index to have good reliability (Cronbach’s alpha = 0.794).

A high score on this three-item perceived discrimination index is interpreted as reflecting less discrimination and a low score as indicating more perceived discrimination. Items 45, 47, and 58 were combined into one measure of health care discrimination for the primary study analyses, but additional analyses were conducted to see if there were significant differences between a 3-item and 1-item health care discrimination measure. Since items 46 and 48 asked about perceived gender discrimination of providers, these were not included in the health care discrimination measures. Many respondents commented that they were not able to make those assessments since they were male and were unaware of how providers treated women.

The sixth section (items 61-71) included items from both the SF-12 version 2 survey and a larger Veterans SF-36 Survey that assessed physical and mental status over different periods of time (Ware et al., 2005; Kazis et al., 2004b). The SF-12 version 2 made a few modifications from the version 1 published in 1996 (Ware et al., 1996). These modifications included changes in item wording, instructions, response categories, and format recommended by the developers of SF-36.

The Veterans SF-36 Survey was an extension of the Medical Outcomes Study (MOS) and assessed health-related quality of life but with a specific focus on health issues affecting the veteran population, in particular, functional status of the respondents and their general well-being.

Items 61-66 were taken directly from the SF-12 version 2 (Ware et al., 2005). Item 67 was fairly close to the item in SF-12 version 2 with the minor exception that the response choice of “a good bit of the time” was omitted in the SF-12 version 2 but retained in the current survey. Since this response choice was not chosen by many respondents, we combined “a good bit of the time” with “most of the time” response choices. This small change enabled us to use the SF-12 version 2 scoring algorithms. The Cronbach’s alpha for the physical component summary and the mental component summary were 0.89 and 0.86 for the SF-12 version 2 (Ware et al., 2005). Items 69-71 were taken directly from the Veterans SF-36 Survey. Item 68 was a new item that was added in the present survey to control for the possibility that the responses to the rest of the survey may have been affected by recent life changes. Higher physical and mental component scores reflected individuals with higher physical and mental functioning/status.

The seventh and final section (items 72-82) was developed from nine items used by Krieger and colleagues (2005). Questions 72 and 73 specifically were developed from an Experience of Discrimination (EOD) tool developed by Krieger and colleagues (Krieger, 1990; Krieger and Sidney, 1996). These latter items asked about the frequency of lifetime experiences of unfair treatment and racial or gender discrimination. Using Krieger’s scoring approach (Krieger et al., 2005) for the situation version of the EOD, we added the number of situations in which a participant reported experiencing racial discrimination. These authors reported that the EOD situation scale’s reliability was high, a Cronbach’s alpha of 0.88, and a test-retest reliability correlation of 0.89. Thus, it appears that this specific measure of discrimination is fairly reliable and valid. A higher number on this measure reflected greater experiences of lifetime discrimination.

Data Analysis

Data analysis was conducted using SPSS 14.0. We conducted descriptive statistics (means, standard deviations, and frequencies) on demographic variables: age, gender race ethnicity, educational background, marital status, lived alone, was looking for employment, was working, branch of military served, whether veteran saw combat, and had a certified service-related disability. Descriptive statistics were collected for health care access and other related variables: used health care within the last 3 years, if seen by a VA and/or non-VA provider, indicated future use of VA health care system, had a form of private or government health insurance, had help to go to the doctor if ill, overall satisfaction with health care provider, and overall perceived quality of health care from provider. Descriptive statistics were also provided for the following outcome measures: patient satisfaction, perceived quality of care, perceived health care discrimination, physical and mental health status. Independent sample t-tests were performed to compare between two groups of users, nonusers, and dual users of the VA. Analysis of Variance (ANOVA) was performed when comparing among all three groups of users of VA and/or non-VA providers. Stepwise hierarchical regressions were conducted using those study variables for which there were significant correlations between background variables and study outcomes. These regressions were conducted to identify independent predictors of perceived quality of care, patient satisfaction, and physical and mental health status even after controlling for age, educational level, working or not, had certified disability or not, had help to go to the doctor if ill, use of VA or non-VA only providers, lifetime discrimination, and perceived health care discrimination. For all significance testing, a p-value of greater than or equal to 0.05 was set as the a priori level to achieve statistical significance. Reliability of scales and indices was assessed using the Cronbach alpha.

Responses to the open-ended questions were recorded and indexed. These responses ranged from a couple of words to a couple of sentences and they provided us with the veteran’s own words in describing utilization of health care services, access to health care services and status of their own health.

RESULTS

Sample Profile

The demographic characteristics of the study participants are described in Table 1. Participants were largely male and, on average, middle-age. All self identified as Black/African-American and close to 10% also identified as part of another racial/ethnic group. Less than a quarter of the sample had completed college and/or gone on to graduate school. Approximately one fifth were currently married and almost 60% of the sample lived alone. About 60% of the sample was not working. Twenty-five respondents (17%) reported looking for work. There were 37 other individuals reporting having a certified service-related disability (yielding a total of 40 or 28.4% of the sample). A majority of respondents reported having served in the Army and not having seen active combat.

Table 2 shows data related to the nature and extent of the sample population’s access to health care. With regard to where participants received health care in the last three years and the last 12 months, the sample was almost evenly divided into those who obtained care in the VA only, those who obtained care from non-VA providers only and those who received care from both VA and non-VA providers. With regard

to considering the VA as a place to get health care in the future, about half of the sample indicated willingness to go to the VA as a primary source of care and almost a third indicated they were only willing to use the VA as a back-up to non-VA care. A little more than 10% of the sample indicated they would not use the VA in the future. More than half of the sample had some additional type of health insurance other than with the VA but did not have additional prescription insurance outside of that with the VA or TRICARE. Approximately 79% of those who did not go to the VA in the past 12 months had either private or government insurance (outside of VA coverage or TRICARE). Most respondents reported having help to go to the doctor if they were ill. Finally, over half of the respondents were “extremely”, “quite” or “moderately” satisfied with their health care provider and over two-thirds believed they received “good” or “excellent” care from their health care provider.

Perceptions of health care and discrimination in health care, and physical and mental health status among African-American veterans Patient satisfaction. Table 3 presents descriptive results of all responses to the 18 items of the Patient Satisfaction Questionnaire. A majority of respondents “agreed” or “strongly agreed” with many statements indicating satisfaction in many aspects of health care. For example, most agreed that their doctors had some good communication skills as evidenced by a majority noting that their doctors were good about explaining reasons for medical tests, being friendly and courteous, and not ignoring them. About half the respondents felt the doctors spent plenty of time with them and disagreed that their providers were in a hurry when treating them. Most respondents felt their doctor’s office had everything needed to provide complete medical care, and the medical care they had received was near perfect. Yet a majority of respondents felt that doctors made them wonder if their diagnosis was correct. Despite these concerns about accurate diagnoses, respon