Social support as a predictor of HIV testing in at-risk populations: a research note.
Abstract: This paper examines the relationship between social support and the probability of getting tested for human immunodeficiency virus (HIV) among at-risk adults in the United States. According to the literature, social support is one mechanism through which social capital is purported to work. Several studies have hypothesized that social capital influences public health, including HIV or Acquired Immune Deficiency Syndrome (AIDS) case rates and behaviors related to contracting HIV. In this analysis, I use social support as an individual level measure of social capital and examine its influence on a protective behavior, getting tested for HIV. The relationship of social capital and social support to behaviors related to HIV is relevant to the design and implementation of HIV prevention programs.
Subject: Social networks (Health aspects)
HIV testing (Usage)
HIV testing (Social aspects)
HIV infection (Diagnosis)
HIV infection (Prevention)
Author: Grosso, Ashley
Pub Date: 06/22/2010
Publication: Name: Journal of Health and Human Services Administration Publisher: Southern Public Administration Education Foundation, Inc. Audience: Academic Format: Magazine/Journal Subject: Government; Health Copyright: COPYRIGHT 2010 Southern Public Administration Education Foundation, Inc. ISSN: 1079-3739
Issue: Date: Summer, 2010 Source Volume: 33 Source Issue: 1
Topic: Event Code: 290 Public affairs
Geographic: Geographic Scope: United States Geographic Code: 1USA United States
Accession Number: 250033491
Full Text: BACKGROUND

Social capital is defined by political scientist Robert Putnam as "features of social organization, such as networks, norms and social trust, that facilitate coordination and cooperation for mutual benefit" (quoted in Carpiano, 2006, p. 168). The literature indicates that social and emotional support are important components of social capital. Berkman, Glass, Brissette and Seeman argue that social networks affect health behaviors through four pathways: provision of social support, social influence, social engagement and attachment, and access to resources (2000). They go on:

Social support is typically divided into subtypes which include emotional, instrumental, appraisal and informational support. Emotional support is related to the amount of 'love and caring, sympathy and understanding and/or esteem or value available from others (Berkman et al, 2000, p. 848).

According to Galea and Vlahov (2005), informal social ties affect social support. Social capital effects, including manifestations at the contextual level such as state level, are thought to offer social support on an ongoing basis (Galea and Vlahov, 2005, p. 348). Dressler (2004) defined social support as "the perceptions that help and assistance can be sought within one's network" (p. 27).

Previous studies have examined the relationship between social capital and HIV-related measures. For example, Campbell, Mzaidume, and Williams found that membership in certain types of groups, such as church groups and sports clubs, was associated with lower levels of HIV infection, while membership in other groups was associated with higher levels of infection in Carletonville, South Africa. The association between group memberships and sexual health also varied across different age and gender groupings. Similarly, Mushati, Mlilo, Campbell, Nyamukapa, and Gregson (2002) found in rural Zimbabwe that membership in certain forms of social group was positively associated with adoption of safer behaviors and HIV avoidance. My study will explore whether the relationship between social capital and HIV avoidance in the United States is similar to that discovered in these studies about rural communities in Africa.

Bhattacharya (2005) conducted a qualitative study that explored social capital resources and their influences upon HIV risk behaviors in a sample of 17 heterosexual Asian Indian immigrant men residing in New York City. Social capital's influence appeared to reduce acculturative stress and HIV risks, but individuals who lacked HIV transmission knowledge or sought peer solidarity through adherence to negative peer norms had elevated sexual risks for HIV. These findings suggest the promise of a social capital approach in the development of HIV prevention programs.

Previous studies have examined the relationship between social capital as an independent variable and AIDS case rates as a dependent variable. Holtgrave and Crosby (2003), using state- level data on social capital as measured by Putnam's Comprehensive Social Capital Index, found that higher rates of social capital were associated with lower AIDS case rates, and higher income inequality was associated with higher AIDS case rates, but social capital was the stronger predictor variable of the two.

Kim, Subramanian, Gortmaker, and Kawachi (2006) studied state and county level social capital in the United States in relation to obesity and physical inactivity. They argued that studies exploring the main associations and mediating pathways between social capital and disease would strengthen evidence for causal effects and thereby help to establish a stronger rationale for public health interventions. I will use some of the same data sources as this study, specifically the Behavioral Risk Factor Surveillance System (BRFSS) survey.

METHODOLOGY

This study examines the relationship between social support, one component of social capital, and getting tested for HIV. The unit of analysis is individuals. The dataset is from the 2007 BRFSS survey, sponsored by the Centers for Disease Control. This is a nationally representative telephone survey (N=430,912). I limit the analysis to the segment of the population (N=17,793) who reported having at least one risk factor for Hepatitis B infection (including injecting street drugs by needle, having sex with a man who has had sex with another man, having two or more sex partners in the past year, and receiving clotting factor concentrate for hemophilia). Since the modes of transmission for HIV are similar to those listed, I treat this variable as a measure of risk for HIV infection. In the following discussion, the variable names from the 2007 BRFSS Codebook Report are in parentheses. I recoded as necessary to drop refused or unsure responses or create dichotomous variables. I used Stata 10.0 for the analysis.

The independent variable for this model is social support (EMTSUPRT), as measured by how often a respondent gets the amount of social and emotional support needed. The dependent variable is whether a respondent has ever been tested for HIV (HIVTST5). I limit the regression to those respondents reporting HIV risk factors (HEPBRSN2). The control variables are age (AGE), gender (SEX), ethnicity (HISPANC2), race (_PRACE), marital status (MARITAL), education (EDUCA), and income (INCOME2). A measure of mental health (MENTHLTH) is also included because people with more emotional problems might be expected to have less social capital. I hypothesize that people reporting at least one risk factor for HIV infection who have higher social capital will be more likely to report having ever been tested for HIV.

FINDINGS

A logistic regression showed that for the at-risk respondents, a one-unit increase in social support based on a five-point Likert scale question produced about a four percent increase in the odds ratio for getting tested for HIV. Marital status and some racial categories (American Indian/Alaska Native, Other, and No preferred race) were not significantly related to whether a respondent had ever been tested for HIV. Older respondents were slightly less likely than younger respondents to report ever having been tested. Males had a ten percent lower odds ratio compared to females, probably because pregnant women often receive HIV tests as part of prenatal medical visits. Hispanic/Latino respondents had a thirteen percent lower odds ratio for getting tested than respondents who did not report this ethnic background. Native Hawaiian/Pacific Islander respondents had a fifty-five percent lower odds ratio for having been tested than all other racial categories. Asian respondents had a forty-three percent lower odds ratio than all other racial categories. African Americans had an eighty-six percent higher odds ratio than other racial groups for getting tested for HIV. A one-unit increase in income on an eight-point ordinal scale produced an eight percent decrease in the odds ratio for ever having an HIV test. An increase in education level based on a six-point ordinal scale was associated with a twenty-eight percent increase in the odds ratio for getting tested. Finally, each additional day a respondent's mental health was not good in month before the survey led to a one percent increase in the odds ratio for getting tested for HIV.

CONCLUSION

Previous research has shown a relationship between social capital and HIV. This study has shown that the more often an individual at risk for HIV infection gets the social and emotional support needed, the more likely he or she is to report having been tested for HIV. The results of this study seem to suggest the promise of a social capital approach to designing and implementing HIV prevention programs. Providing social and emotional support to populations at risk (such as offering support groups or services like childcare or transportation) may decrease risk behaviors or promote health protective behaviors.

This study extends the previous research on social capital and HIV in several ways. While previous studies examined group membership as a measure of social capital, this study looked at social support, another dimension of the social capital concept through which health behaviors are affected. Earlier research showed the relationship between social capital and AIDS case rates at the state-level, while this study demonstrated the impact of social capital on a behavior related to HIV at the individual level. Finally, previous research looked at AIDS case rates and risk behaviors for HIV infection, while this research focused on a health-promoting behavior, getting tested for HIV.

One limitation of this research is potential lack of full model specification. HIV is a complex phenomenon, and not all of the factors influencing whether an individual will seek an HIV test are included in the BRFSS. Another limitation is measurement. The BRFSS question about risks combines very different populations, such as men who have sex with men and injection drug users, so we cannot know whether one segment of this combined group is responsible for higher levels of HIV testing.

Further studies should examine the relationship of social capital to other health protective behaviors, such as seeking care once an individual receives a positive result on an HIV test. Future research should also examine the relationship between HIV testing and social capital at other levels, such as the state or municipality, to determine where and what type of interventions incorporating the concept of social capital would be most effective in preventing the spread of HIV.

ACKNOWLEDGEMENTS

I am grateful to Yahong Zhang, Gregg Van Ryzin, Frank Thompson, Daniel Smith, and Suzanne Piotrowski for their helpful comments on earlier drafts of this paper.

REFERENCES

Berkman, L.F., Glass, T., Brissette, I., & Seeman, T.E. (2000). From social integration to health: Durkheim in the new millennium. Social Science & Medicine, 51, 843-857.

Bhattacharya, G. (2005). Social capital and HIV risks among acculturating Asian Indian men in New York City. AIDS Education and Prevention, 17(6), 555-567.

Campbell, C., Mzaidume, Y., & Williams, B. (2002). The role of social capital in promoting or hindering HIV prevention: A case study of a South African mining community. In Proceedings of the XIV International AIDS Conference. Bologna, Italy: Monduzzi Editore.

Carpiano, R.M. (2006). Toward a neighborhood resource-based theory of social control for health: Can Bourdieu and sociology help? Social Science Medicine, 62, 165-175.

Centers for Disease Control and Prevention. (2007). Behavioral Risk Factors Surveillance System. Atlanta, GA: Department of Health and Human Services.

Dressler, W.W. (2004). Culture and the risk of disease. British Medical Bulletin, 69, 21-31.

Galea, S., & Vlahov, D. Urban health: Evidence, challenges, and directions. (2005). American Review of Public Health, 26, 341-364.

Holtgrave, D.R., & Crosby, R.A. (2003). Social capital, poverty, and income inequality as predictors of Gonorrhoea, Syphilis, Chlamydia and AIDS case rates in the United States. Sexually Transmitted Infections, 79(1), 62-64.

Kim, D., Subramanian, S., Gortmaker, S.L., & Kawachi, I. (2006). U.S. state- and county-level social capital in relation to obesity and physical inactivity: A multilevel, multivariable analysis. Social Science & Medicine, 63(4), 1045-1059.

Mushati, P., Mlilo, M., Campbell, C., Nyamukapa, C., & Gregson, S. (2002). Social capital and HIV avoidance in men and women in rural Zimbabwe. In Proceedings of the XIV International AIDS Conference. Bologna, Italy: Monduzzi Editore.

ASHLEY GROSSO

Rutgers University, Newark
Table 1. At-Risk Population Reports of Social Support, HIV
Testing, Mental Health, and Demographic Characteristics:
Descriptive Statistics (N=17,793)

                                                        Standard
Variable                    Minimum   Maximum   Mean    Deviation

How often do you get the
social and emotional
support you need?           1         5         3.8     1.1

Have you ever been tested
for HIV?                    0         1         0.67    0.47

Number of days mental
health not good in past
month                       0         30        6.7     10.1

Age                         18        99        42.6    14.7

Education                   1         6         4.8     1.1

Married/member of
unmarried couple            0         1         0.3     0.46

Income                      1         8         5.2     2.3

Male                        0         1         0.54    0.5

Hispanic/Latino             0         1         0.08    0.28

African American            0         1         0.14    0.35

Asian                       0         1         0.01    0.11

Native Hawaiian/Pacific
Islander                    0         1         0.007   0.08

American Indian/Alaska
Native                      0         1         0.03    0.17

Other                       0         1         0.04    0.21

No preferred race           0         1         0.004   0.06

Note: For all dichotomous
variables, Yes=1, No=0

Table 2. Logistic Regression Model of the Association of HIV Testing
with the Explanatory Factors

                                                          P >
                                       Standard           [absolute
Tested              Odds Ratio         Error      z       value of z]

Social Support      1.04               0.02        2.10   0.036

Mental Health       1.01               0.002       6.68   0.000

Age                 0.99               0.002      -1.95   0.051

Education           1.28               0.03       12.49   0.000

Marital Status      1.06               0.04        1.51   0.131

Income              0.92               0.01       -7.93   0.000

Gender              0.9                0.03       -2.62   0.009

Hispanic/Latino     0.87               0.07       -1.70   0.089

African American    1.86               0.11       10.08   0.000

Asian               0.57               0.09       -3.53   0.000
Native
Hawaiian/Pacific
Islander            0.45               0.09       -3.79   0.000

American
Indian/Alaska       0.97               0.11       -0.3    0.765
Native

Other               1.08               0.11        0.76   0.445

No preferred race   1.06               0.32        0.20   0.844

N=13883             Pseudo

P<0.0000            [R.sup.2]=0.0231
Gale Copyright: Copyright 2010 Gale, Cengage Learning. All rights reserved.