AIDS Education and Prevention, 16(2), 115–125, 2004 © 2004 The Guilford Press AWAD ET AL. BARRIERS TO HIV TESTING
DEVELOPMENT OF A MEASURE OF BARRIERS TO HIV TESTING AMONG INDIVIDUALS AT HIGH RISK Germine H. Awad, Lynda M. Sagrestano, Mark J. Kittleson, and Paul D. Sarvela
Rates of HIV antibody testing remain at approximately 45% of the general population. To more effectively design interventions to increase testing, comprehensive information is needed to understand the barriers to HIV testing. A measure of barriers to HIV testing was developed using the major barriers identified in the literature on barriers to health care utilization (Melnyk, 1988), and tested with a diverse group of individuals at high risk for HIV, including heterosexuals, men who have sex with men, injected drug users, and sex workers. An exploratory factor analysis indicated that the factor structure was replicated over 2 years of data collection. Three factors—Structural Barriers, Fatalism/Confidentiality Concerns, and Fear—emerged for both years. The reliabilities ranged from .75 to .87, indicating moderate to high internal consistency.
Health providers and public health organizations are consistently striving to increase the rate of HIV antibody testing for the general population, especially among those at high risk. Although people may believe there are several benefits associated with HIV testing (Rotheram-Borus et al., 2001), rates of HIV testing among the general population remain at approximately 43.8% to 45.6% in the United States (Centers for Disease Control and Prevention, [Centers for Disease Control and Prevention.], 2001a, 2003). A population-based survey of Canadians reported that approximately 34.9% of the population has been tested for HIV (Houston, Archibald, Strike, & Sutherland, 1998). One study found that even though individuals intended to be tested, when presented with the opportunity they declined (Wilson, Jaccard, & Minkoff, 1996). The purpose of the current study is to present a new measure of barriers to HIV testing that has been tested with a diverse group of individuals at high risk, including heterosexuals, men who have sex with men (MSM), injection drug users (IDUs), and sex workers. Although overall rates of testing are at approximately 45% (Centers for Disease Control and Prevention., 2001a, 2003), rates also vary based on several group characteristics. Data from the Centers for Disease Control and Prevention in 1995 to 1996 suggest that as a whole, 76% of individuals at high risk had been tested for HIV, with 68% of heterosexuals, 79% of MSM, and 82% of IDUs reporting testing, including testing through blood donation (Centers for Disease Control and Prevention, 1998). Germine H. Awad, Lynda M. Sagrestano, Mark J. Kittleson, and Paul D. Sarvela are with Southern Illinois University-Carbondale. Address correspondence to Germine Awad, M.A., Department of Psychology, Southern Illinois University, Carbondale, IL 62901-6502; e-mail:
[email protected]
115
116
AWAD ET AL.
Other studies have indicated that 53% of a sample of IDUs (Samet, Mulvey, Zaremba, & Plough, 1999) and 53% of homosexuals and bisexuals (Myers, Orr, Locker, & Jackson, 1993) report being tested. By race, the Centers for Disease Control and Prevention (2001a) reported 51.6% of Blacks, 39.5% of Hispanics, and 43.6% of Whites in the general population reported being tested. In contrast, only 24% of African American heterosexuals determined to be at high risk have been tested for HIV (Grinstead, Peterson, Faigeles, & Catania, 1997). In addition, there are some inconsistencies in the literature relating to rates of HIV testing for men and women. Some studies have reported that men are more likely than women to be tested (Samet et al., 1999), whereas others report higher rates of testing among women than men (Centers for Disease Control and Prevention., 2003; Simon, Weber, Ford, Cheng, & Kerndt, 1996), which may reflect high rates of testing among pregnant women. Several factors are associated with voluntary HIV testing. For women, having children, multiple sex partners, engaging in anal sex, and infrequent condom use significantly increased the likelihood that they would choose to be tested for HIV (Miller, Hennessey, Wendell, Webber, & Schoenbaum, 1996). For men, predictors of voluntary HIV testing included having sex with other men, a history of substance abuse, perceiving a partner to be at high risk for HIV, and being single (Stein, & Nyamathi, 2000; Houston et al., 1998). Among those with a recent history of substance abuse, 53% indicated that they have been tested for HIV. Survey respondents were more likely to get tested if they had a history of injection drug use, underwent prior detoxification and addiction treatment, had a history of sexually transmitted diseases (STDs), and engaged in sexual relations with multiple partners (Samet et al., 1999). Previous research has found that among lesbian/gay/bisexual youth, engaging in unprotected anal sex, perceiving low susceptibility to HIV, and perceiving greater barriers to HIV testing were associated with reduced likelihood of being tested for HIV (Maguen, Arimstead, & Kalichman, 2000). Gay and bisexual men from metropolitan areas, those not involved in an exclusive relationship, and those engaging in anal intercourse were more likely to be tested, and homosexuals were more likely to be tested than bisexuals (Myers et al., 1993). For years, social scientists have investigated the link between attitudes and behavior (e.g., Ajzen & Fishbein, 1972). Although intuitively a discrepancy between attitudes and behavior is not expected, research in many domains indicates that attitudes often do not predict behavior (Ajzen & Fishbein, 1972). For example, a person may believe that being tested for HIV is important but not choose to get tested. Various factors, including barriers to HIV testing, may account for this discrepancy. The health belief model (HBM; Janz et al., 2002) is a conceptual framework that has examined barriers in the explanation of health related behaviors. The HBM consists of six main components including perceived susceptibility (belief concerning the likelihood of getting a condition), perceived severity (belief regarding the seriousness of the condition), perceived benefits (beliefs about the effectiveness of an action to reduce or eliminate the condition), perceived barriers (tangible and psychological costs associated with an advised action), cues to action (environmental prompts that activate a person’s willingness to take action), and self-efficacy (confidence in ability to take action). The model posits that for individuals to change a behavior they must first feel threatened by their current behavior and believe that change would bring about a valued result at minimal cost. They must also believe that they have the ability to overcome any barriers that might impede their progress. This process is sometimes accelerated by certain environmental cues that move individuals to action (Janz et al.,
BARRIERS TO HIV TESTING
117
2002). This model may help explain why some individuals choose to be tested, whereas others do not. Although there has been discussion of barriers to various types of health services in the literature (e.g., Melnyk, 1988), there is a dearth of literature that specifically examines the barriers to HIV testing. Melnyk (1988) reviewed several barriers associated with underutilization of health services including system barriers, time, distance, cost, availability, organization of services, and discrimination. System barriers refer to impediments due to some characteristic of the system that can be either observable or based on the perception of the individual seeking services. Time barriers include travel time to the health provider, waiting room time, speed of services rendered, and time conflicts with hours of operation. Distance barriers refer to proximity to the health provider. Cost may manifest itself in several ways including cost of service, less pay due to missing work, transportation costs, and inadequate insurance coverage. Unavailability of health services characterized by a low provider/consumer ratio and limited doctor office hours as well as organization of services (e.g., fragmentation of services, shortage of qualified personnel) also are considered barriers to health utilization. Finally, discrimination based on sex, race, social status, and age may serve as a barrier. It is reasonable to expect that similar barriers would emerge for HIV testing, specifically. To date, although several studies have touched on barriers to HIV testing, there are no scales specifically designed to measure attitudes toward and barriers to HIV testing. Boshamer and Bruce (1999) developed a 23-item scale that assessed perceptions of how friends might react to HIV testing, family’s concerns about the decision to get tested, other people’s reaction to HIV testing, and concerns about confidentiality. Scores on the scales were positively correlated with perceived knowledge about HIV and a greater likelihood of being tested. Although this measure assessed general attitudes toward HIV testing, it only touched on barriers to testing. Barriers that have emerged from other studies include overcrowded clinics and limited hours of operation (Stein & Nyamathi, 2000), concern about the trustworthiness of the staff (Harris, 1997), and lack of time (Simon et al., 1996). Myers and colleagues (1993) performed a factor analysis on 12 items aimed at identifying reasons why individuals fail to get tested for HIV. The factors that emerged included Desire for Anonymity, No Benefit and Denial, and Self-Perceived Health. Desire for anonymity was rated as the most important factor, which is not surprising given that one of the most cited reasons reported by individuals who refuse to take an HIV test is the concern about privacy (e.g., Kalichman, Kelly, Hunter, Murphy, & Tyler, 1993; Myers et al., 1993; Simon et al., 1996). Although previous studies have presented questionnaires that have tapped into barriers to HIV testing, none presents a complete picture of the barriers involved in the testing experience, nor has any been validated on diverse groups of individuals at high risk. The current study presents the development of a new measure of barriers to HIV testing that includes items addressing confidentiality concerns, structural barriers, and psychological barriers. Construct validity of the measure has been tested with a diverse group of individuals at high risk, including heterosexuals, MSM, IDUs, and sex workers.
METHOD PARTICIPANTS Participants were drawn from two waves of data collection for the Illinois HIV Behavioral Surveillance Study. The first year included 827 individuals, and the second year included 1047 individuals. In Year 1, the sample was 63.6% male, 34.1% female;
118
AWAD ET AL.
38.3% White, 28.8% African American, and 27.2% Hispanic. With respect to sexual orientation, 18.3% self-identified as homosexual, 9.2% as bisexual, and 69.9% as heterosexual. In Year 2, the sample was 59.4% male, 40.4% female; 38.7% White, 36.6% African American, and 20.7% Hispanic. With respect to sexual orientation, 28.8% identified as homosexual, 55.5% as heterosexual, and 15.0% as bisexual. Using self-reported behavioral data, each participant was classified into risk categories. Due to multiple risk behaviors, individuals could be identified as belonging to more than one risk group (e.g., a gay man who injected drugs and engaged in sex work would be in at least three risk categories) and therefore between-group comparisons are not appropriate. Of the total sample, in the first year, 253 were MSM, 574 were heterosexuals, 278 were IDUs, and 194 engaged in sex work. In the 2nd year, 387 were MSM, 774 were heterosexuals, 391 were IDUs, and 415 engaged in sex work.
INSTRUMENTS A 26-page survey was developed based on the literature to assess the frequency and extent of sex and drug use behavior, risk reduction behaviors (e.g., HIV testing history, condom use, stage of change), risk-enhancing circumstances (e.g., homelessness), risk-reducing cofactors (e.g., self-efficacy, condom attitudes), service penetration of HIV prevention programs (e.g., barriers to service utilization), and group demographics. The measure was reviewed by a statewide evaluation committee and revised based on the committee’s recommendations. Based on the literature on barriers to service utilization in the mental health field (e.g., Melnyk, 1988), a series of 21 items were created, adapted specifically to barriers to HIV testing. In Year 1, participants who had not been tested in the preceding 2 years were asked to indicate whether each item represented a reason they did not get tested (“agree” or “disagree”). In Year 2, a subset of 13 of the items was included in the data collection. These items were selected by examining the likelihood of endorsement in Year 1 and dropping those that very few participants cited as barriers. Participants were asked how important each reason was to why they had not been tested. For each item, they first responded “yes” or “no,” and then if “yes,” they rated how important (“very or somewhat”). A 3-point scale was created with “no” coded as 1, “somewhat important” coded as 2, and “very important” coded as 3. The wording of the questionnaire changed slightly from Year 1 to Year 2, where Year 1 emphasized agreement and Year 2 emphasized importance.
PROCEDURE Interviewer Identification and Training. In both years, interviewers for the statewide study were identified through state and local agencies providing HIV-related services and were typically individuals who worked for agencies in the region. A training program was developed and conducted with all interviewers. Care was taken that data collectors did not interview their own clients to avoid dual role relationships. Participant Recruitment. Staff from the state department of public health provided a list of subcontracting agencies throughout the state that assisted in the data-collection procedures. Subcontractors were asked to serve as a liaison to the targeted populations by matching interviewers with outreach workers who knew the population and could facilitate the interviewer in approaching individuals to participate. Interviews were conducted in many different locations, ranging from health departments and social service agencies to STD clinics, bars, parks, shelters, and needle exchanges. Prior to the interview, confidentiality was explained and informed consent was obtained. The survey was administered in a face-to-face format and took approx-
BARRIERS TO HIV TESTING
119
TABLE 1. Summary of Principal Axis Analysis with Oblimin Rotation for Year 1 Factors Item
Structural
Fatalism/ Confidentiality
Fear of Loss
h2
You don’t have transportation to get to the testing site.
.97
.04
–.15
.91
You don’t have enough time.
.59
.34
.00
.64
The testing site is too far away.
.49
.40
–.10
.53
You don’t know where to go for testing.
.45
–.15
.14
.21
.43
.27
.15
.44
There is no cure, so why get tested.
You don’t like the people who work at the testing site.
–.06
.92
–.01
.79
You can’t afford treatment, so why get tested?
–.06
.94
–.05
.86
You don’t want to know the results.
–.06
.79
.19
.70
People might recognize you at the testing site.
.30
.53
.11
.58
You are worried about confidentiality.
.25
.42
.23
.47
You are afraid of losing your health insurance.
–.09
–.00
.98
.91
You are afraid of losing your job.
–.03
.01
.92
.83
.15
.12
.54
.43
You are afraid of losing your partner.
Note. Bolded loadings represent those items that were theoretically consistent with the structural, fatalism/confidentiality, and fear of loss factors respectively.
imately 45 minutes to complete. Response cards were used to facilitate responses to survey questions with several choices. Following the interview, respondents were given coupons to a fast-food restaurant for their participation in the study (Year 1, $4; Year 2, $10). All data-collection procedures were approved by the university’s human subjects committee. When necessary, approval was also secured at the local level from individual agencies that required internal review.
RESULTS For each year’s data set, an initial principal component’s factor analysis was conducted to estimate the number of factors. Because only 13 of the original 21 items were included in the Year 2 data collection, only the corresponding 13 items from Year 1 were included in the analyses. For Year 1, three factors with eigenvalues greater than 1.0 surfaced, whereas four factors with eigenvalues greater than 1.0 emerged for Year 2. The scree plot also indicated between three and four factors for both years. Subsequently, a principal axis analysis using an oblimin rotation was performed with a three- and four-factor solution. An oblimin rotation was utilized because the factors were moderately correlated with one another (Tabachnick & Fidell, 1989). The three-factor solution was most interpretable for both Years 1 and 2. The factor solution from Year 1 accounted for 64% of the variance, whereas 50% of the variance was accounted for in Year 2. Tables 1 and 2 display the rotated factor structure coefficients for the items after oblimin rotation for Years 1 and 2, respectively. A cutoff of .30 was adopted for item inclusion on interpretable factors (Norman & Streiner, 1994). The factor structure was similar for both years with the exception of the fatalism/confidentiality item “I don’t want to know results,” which loaded .92 for Year 1 and .19 for Year 2. Due to its high loading in Year 1 and its theoretically suitability to the fatalism/confidentially construct, the item was kept in the Year 2 factor structure despite its low loading. In
120
AWAD ET AL. TABLE 2. Summary of Principal Axis Analysis with Oblimin Rotation for Year Factors Structural
Fatalism/ Confidentiality
Fear of Loss
I don’t have transportation to site.
.82
.10
–.07
.64
I don’t know where to go for testing.
.72
.09
.03
.44
Item
h2
I don’t have enough time.
.46
–.10
.05
.25
The testing site too far away.
.45
–.15
–.19
.42
.31
–.39
–.08
.43
I am worried about confidentiality.
I don’t like people at testing site.
–.06
–.86
.01
.67
I might be recognized at testing site.
.55
–.03
–.74
–.02
There is no cure so why get tested.
.20
–.38
–.12
.33
I can’t afford treatment so why get tested.
.48
–.35
.06
.50
.18
–.19
–.21
.22
–.08
.05
–1.0
.91
I don’t want to know results. I am afraid of losing my health insurance. I am afraid of losing my job.
.04
.04
–.89
.79
I am afraid of losing my partner.
.05
–.27
–.38
.35
Note. Bolded loadings represent those items that were theoretically consistent with the structural, fatalism/confidentiality, and fear of loss factors respectively.
Year 2, the item “I can’t afford treatment so why get tested” loaded higher on structural barriers (.48) than fatalism/confidentiality (-.35) but we chose to place it within the fatalism/confidentiality construct due to its theoretical suitability to the construct, as it was above our cutoff for both factors and therefore could be included in either. The remaining items for both years loaded on their respective scales in the expected fashion. Examination of the factors for each year indicated that the items loading on each factor were similar for the two data sets. Scale scores were created by averaging the items that loaded on their respective factors. The first factor, Structural Barriers, consisted of five items (e.g., “I don’t have transportation to get to the testing site”), and had an alpha of .79 for Year 1, and .75 for Year 2. The second factor, Fatalism/Confidentiality Concerns, also consisted of five items (e.g., “I am worried about confidentiality”; “There is no cure, so why get tested”), and had an alpha of .87 for Year 1, and .77 for Year 2. The third factor, Fears, consisted of three items (e.g., “I am afraid of losing my partner”), and had an alpha of .81 for Year 1, and .79 for Year 2. Means, standard deviations, and reliabilities for each factor for the full sample and each risk group for Years 1 and 2 are presented in Table 3. Pairwise t-tests were conducted on the three-factor subscales. Because individuals can be categorized in several risk groups (e.g., a women may be an IDU and a sex worker) groups were not compared with one another. Instead, scores on the three-factor scales were compared with one another within each risk group. Specifically, within risk group, difference scores were computed comparing each of the factors (i.e., fatalism/ confidentiality with fear, fatalism/confidentiality with structural, structural with fear). Mean difference scores among the factor subscales are presented in Table 4. Results suggest that fatalism/confidentiality concerns emerged as the most important concern for heterosexuals and IDU in both Years 1 and 2. There were no differences among the barriers for MSM in Year 1, but differences did emerge in Year 2. For
BARRIERS TO HIV TESTING
121
TABLE 3. Means, Standard Deviation, and Internal Validity Estimates for Structural, Fatalism/Confidentiality, and Fear Barriers for the Full Sample and by Risk Group
M Fatalism/Confidentiality Full Sample Heterosexuals MSM IDUs Sex Workers Fear of Loss Full Sample Heterosexuals MSM IDUs Sex Workers Structural Full Sample Heterosexuals MSM IDUs Sex Workers
Year 1 SD
n
M
Year 2 SD
n
1.33 1.38 1.23 1.40 1.59
.37 .38 .33 .42 .37
158 114 48 40 23
.87 — — — —
1.54 1.55 1.44 1.57 1.67
.62 .66 .58 .63 .65
326 132 91 97 72
.77 — — — —
1.22 1.23 1.20 1.19 1.31
.35 .36 .34 .29 .30
179 131 50 50 33
.81 — — — —
1.33 1.26 1.35 1.36 1.30
.61 .56 .65 .63 .54
321 129 88 93 70
.79 — — — —
1.23 1.26 1.19 1.20 1.45
.32 .33 .29 .27 .33
176 129 51 46 32
.79 — — — —
1.31 1.30 1.21 1.32 1.33
.47 .48 .39 .52 .53
324 131 90 97 71
.75 — — — —
Note. MSM = men who have sex with men; IDUs = injection drug users. Sample sizes represent valid cases.
MSM in Year 2, both Fatalism/ Confidentiality Concerns and Fear were rated as more important than structural barriers. For sex workers in Year 1, fatalism emerged as the most important concern followed by structural barriers. In Year 2, the most important barrier for sex workers was fatalism/confidentiality. Comparing across the three barrier subscales, for the full sample, as well as for heterosexuals, sex workers, and IDUs, Fatalism/Confidentiality Concerns were rated as more important than Fear or Structural Barriers in all the barriers subscales with the exception of MSM in Year 1. Analyses were also conducted based on ethnic group membership and sex, and no significant group differences emerged.
DISCUSSION Although there have been scales created to identify barriers to HIV testing, most have targeted only a few types of barriers rather than a more comprehensive set of barriers to HIV testing (e.g., Myers et al., 1993; Simon et al., 1996). The current study presents the development of a Barriers to HIV Testing scale designed to include the major barriers identified in the literature on barriers to health care utilization (Melnyk, 1988). An exploratory factor analysis indicated that the factor structure was replicated over 2 years of data collection, despite differences in the wording and response categories of the instructions from Year 1 to Year 2. Three factors—Structural Barriers, Fatalism/Confidentiality Concerns, and Fear—emerged for both years. The reliabilities ranged from .75 to .87, indicating moderate to high internal consistency. A unique and important contribution of this study is the diverse sample used. The participants in the current study were a nonstudent community sample of individuals at high risk for HIV and consisted of groups from culturally diverse backgrounds.
122
Fatalism/Confidentiality—Fear
.230****
SD
.479
.583
.561
SD
.300
.394
.371
.241****
.104*
–.134**
M
MSM (n = 88)
–.044
.030
–.045
M
MSM (n = 44)
Note. Mean values represent difference scores. *p < .10; **p < .05; ***p < .01; ****p < .001.
.196****
Fatalism/Confidentiality—Structural
Structural—Fear
Fatalism/Confidentiality—Fear
M –.030
Difference Scores
Heterosexuals (n = 132)
.108****
.131****
Structural—Fear
Fatalism/Confidentiality—Structural
M .027
Difference Scores
Heterosexuals (n = 104)
.413
.536
.499
SD
Year 2
.234
.288
.284
SD
Year 1
High–Risk Groups
.248****
–.193***
–.040
M
IDU (n = 97)
.198***
.128
–.062
M
IDU (n = 30)
.431
.625
.640
SD
.382
.525
.397
SD
TABLE 4. Difference Scores among Structural, Fatalism/Confidentiality, and Fear Barriers by Risk Group
.316****
.331****
.025
M
Sex Workers (n = 71)
.092
.283***
.210***
M
Sex Workers (n = 20) SD
.478
.528
.501
SD
.322
.320
.351
BARRIERS TO HIV TESTING
123
Whereas most of the studies presented in the literature utilized student samples and/or individuals from White middle-class backgrounds, this measure was normed on individuals at high risk for HIV with diverse behavioral risk profiles, including heterosexuals, IDUs, MSM, and sex workers. Across risk groups, Fatalism/Confidentiality Concerns emerged as the most important barrier to testing for all risk groups, indicating further effort is needed to address confidentiality and fatalism concerns in a more thorough manner. Because of the stigma associated with HIV and AIDS, confidentiality becomes a crucial issue for HIV testing sites. A standard policy concerning confidential HIV testing includes recording the name of a person who tests positive and confidentially reporting those names to public health officials (Centers for Disease Control and Prevention, 2001b). Lab technicians and state health department personnel usually have access to test results. Results are usually only sent to a doctor when a release form is signed. The guarantee of confidentiality may not ameliorate concerns of privacy for individuals seeking HIV testing. The fear of health providers breaking confidentiality is exacerbated when reports of specific instances of breaches are made public. In one instance, an HIV patient’s children were made aware of her positive HIV status from a child of a health care provider (Moyer, 1999). Health care providers and individuals seeking HIV testing may have different notions of what constitutes confidentiality, and test seekers may be more conservative in their definitions of confidentiality. Alternatives to confidential testing include anonymous testing sites that utilize a coding system in which individuals are provided with an identifier code instead of using their name (Centers for Disease Control and Prevention, 2001b). This system is designed to reduce the number of clinic workers that would have access to the results. In anonymous testing centers only the individuals getting tested can disclose their results. Although anonymous testing may alleviate most of the concerns test takers may have, unfortunately it is not available in all states (Centers for Disease Control and Prevention, 2001b). Another alternative to confidential testing includes the use of hometesting kits that can be purchased at drugstores. Home test users prick their fingers and place their blood sample on a card that is mailed to a licensed laboratory (Centers for Disease Control and Prevention, 2001b). Purchasers of the kit are given an identification number to use when phoning in for their results. Home testing is a viable alternative for those who are extremely concerned about their privacy. Fear of loss was another barrier that was noted by participants in the study. Fear of loss included fear of losing a partner, job, and/or health insurance. The aforementioned methods to improve confidentiality and anonymity may help alleviate the fear associated with losing a job and health insurance, as it would be the individual’s choice if they wanted to inform their place of employment or health insurance company of their HIV status. However, a positive test followed by illness and treatment could still lead to such losses, and issues of confidentiality in testing cannot alleviate the stigma or fear of stigma experienced by HIV-positive individuals. In terms of losing a partner, counseling programs should be more readily available for those coping with the possibility of losing their partner after disclosing a positive HIV status. Issues related to fear associated with testing should be emphasized for MSM, given that fear of loss was just as important as issues of confidentiality for this risk group. There were several limitations in the current study. First, although the current study utilized a diverse sample, cross validation is needed to assess whether the factor structure can be replicated with other populations (e.g., students samples, samples in other regions). Second, the trichotomous and dichotomous scale options have led to
124
AWAD ET AL.
decreased variability in the responses. Although the data in the current study was not significantly skewed (i.e., an adequate amount of variance was found; Gorsuch, 1983), and the validity of the factor analysis was not compromised, it should be noted that it is usually better to have a measure with more variability. Therefore, future studies should incorporate response options that allow for greater variability. In future studies a 5-point Likert-type scale, ranging from 1 (“not at all important”) to 5 (“very important”), should be used to increase response options for participants. Third, future studies should examine other aspects of validity, such as predictive validity, to determine whether this measure will help predict the probability that an individual will volunteer for HIV testing. Fourth, the data were only collected on those who had not been tested in the past 2 years; however, people who were recently tested likely also had barriers to testing that were not examined in the current study. Future research should include both those who have and those who have not been recently tested to fully understand the barriers to HIV testing. The HBM suggests that individuals are less likely to move to action if the barriers to testing outweigh the benefits (Becker, 1974; Janz et al., 2002; Kalichman, 1998). Therefore, AIDS educators, counselors, and clinics should emphasize the benefits of HIV testing and work to alleviate the barriers associated with testing. Because diverse groups may experience different barriers, tailoring your intervention may be successful in increasing rates of HIV testing for individuals from different groups, including people of color, MSM, sex workers, and IDUs (Kalichman et al., 1993). This newly developed measure may enable health providers and clinic operators to better assess and address the possible barriers associated with HIV testing in their communities.
REFERENCES Ajzen, I., & Fishbein, M. (1972). Attitudes and normative beliefs as factors influencing behavioral intentions. Journal of Personality and Social Psychology, 21, 1-9. Becker, M. (1974). The health belief model and personal health behavior. Thorofare, NJ: Slack. Boshamer, C. B., & Bruce, K. E. (1999). A scale to measure attitudes about HIV-antibody testing: Development and psychometric validation. AIDS Education and Prevention, 11, 400-413. Centers for Disease Control and Prevention. (1998). HIV testing among populations at risk for HIV infection—Nine states. Morbidity and Mortality Weekly Reports, 47, 1086-1091. Centers for Disease Control and Prevention. (2001a). HIV testing among racial and ethnic minorities—United States, 1999. Morbidity and Mortality Weekly Reports, 50, 1054-1058. Centers for Disease Control and Prevention. (2001b). HIV testing policies, 2000. Retrieved July 2003, from http://hardtruth.qti.net/hiv_testing_policies.htm
Centers for Disease Control and Prevention. (2003). HIV testing—United States, 2001. Morbidity and Mortality Weekly Reports, 52, 540-542. Gorsuch, R. L. (1983). Factor analysis (2nd ed.). Hillsdale, NJ: Erlbaum. Grinstead, O. A., Peterson, J. L., Faigeles, B., & Catania, J. A. (1997). Antibody testing and condom use among heterosexual African Americans at risk for HIV infection: The national AIDS behavioral surveys. American Journal of Public Health, 87, 857-859. Harris, R. M. (1997). To be or not be tested for the AIDS virus. Nursing Research, 46, 293-294. Houston, S., Archibald, C. P., Strike, C., & Sutherland, D. (1998). Factors associated with HIV testing among Canadians: Results of a population-based survey. International Journal of STD and AIDS, 9, 341-346. Janz, N. K., Champion, V. L., & Strecher, V. J. (2002). The health belief model. In K. Glanz, B. K. Rimer, & F. M. Lewis (Eds.), Health behavior and health education: Theory, research, & practice (3rd ed., pp. 45-66). San Francisco, CA: Jossey-Bass.
BARRIERS TO HIV TESTING Kalichman, S. C., Kelly, J. A., Hunter, T. L., Murphy, D. A., & Tyler, R. (1993). Culturally tailored HIV-AIDS risk-reduction messages targeted to African American urban women: Impact on risk sensitization and risk reduction. Journal of Consulting and Clinical Psychology, 61, 291-295. Maguen, S., Armistead, L. P., & Kalichman, S. (2000). Predictors of HIV Antibody testing among gay, lesbian, and bisexual youth. Journal of Adolescent Health, 26, 252-257. Melnyk, K. A. M. (1988). Barriers: A critical review of recent literature. Nursing Research, 37, 196-201. Miller, K.S., Hennessy, M., Wendell, D. A., Webber, M. P., & Schoenbaum, E. E. (1996). Behavioral risks for HIV infection associated with HIV-testing decisions. AIDS Education and Prevention, 8, 394-402. Moyer, P. (1999). Confidentiality in HIV testing remains controversial. WebMD Medical News. Retrieved July 2003, from http:// cont e n t . h e a l t h . m s n . c o m / c o n t e n t / a r t icle/19/1728_50477.htm Myers, T., Orr, K.W., Locker, D., & Jackson, E. A. (1993). Factors affecting gay and bisexual men’s decisions and intentions to seek HIV testing. American Journal of Public Health, 83, 702-704.
125 Norman, G. R., & D. L. Streiner. (1994). Biostatistics: The bare essentials. St. Louis, MO: Mosby. Rotheram-Borus, M. J., Mann, T., Newman, P. A., Grusky, O., Frerichs, R. R., Wight, R. G., & Kuklinski, M. (2001). A street intercept survey to assess HIV-testing attitudes and behaviors. AIDS Education and Prevention, 13, 229-238. Samet, J. H., Mulvey, K. P., Zaremba, N., & Plough, A. (1999). HIV testing in substance abusers. American Journal of Drug and Alcohol Abuse, 25, 269-280. Simon, P.A., Weber, M., Ford, W.L., Cheng, F., & Kerndt, P.R. (1996). Reasons for HIV-antibody testing refusal in heterosexual sexually transmitted disease clinical population. AIDS, 10, 1549-1553. Stein, J. A., & Nyamathi, A. (2000). Gender differences in behavioral and psychosocial predictors of HIV testing and return for test results in a high-risk population. AIDS Care, 12, 343-356. Tabachnick, B.G., & Fidell, L.S. (1996). Using multivariate statistics (3rd ed). New York: HarperCollins. Wilson, T. E., Jaccard, J., & Minkoff, H. (1996). HIV-antibody testing: Beliefs affecting the consistency between women’s behavioral intentions and behavior. Journal of Applied Social Psychology, 1996, 1734-1748.
