Health Promotion International, 2016;31:270–279 doi: 10.1093/heapro/dau113 Advance Access Publication Date: 18 December 2014

Perceived vulnerability and HIV testing among youth in Cape Town, South Africa Department of Sociology, Memorial University, St. John’s, NL A1C 5S7, Canada *Corresponding author. E-mail: [email protected], [email protected]

Summary The importance of perceived vulnerability to risk-reducing behaviors, including HIV testing, is fairly established, especially among youth in sub-Saharan Africa. Yet, the majority of studies that examined this important relationship used cross-sectional data that inherently assume that perceived vulnerability does not change. While these studies have been useful, the assumption of perceived vulnerability as time invariant is a major flaw and has largely limited the practical usefulness of this variable in AIDS prevention and programing. Using longitudinal data and applying random-effects logit models, this study makes a major contribution to scholarship by examining if changes in perceived vulnerability associate with a change to test for HIV among 857 young people in Cape Town, South Africa. Results show that female youth who changed their risk perceptions were more likely to also change to test for HIV, but the effects were completely attenuated after controlling for theoretically relevant variables. No significant relationships were observed for males. Also, females who were virgins at wave 2 but had sex between waves were significantly more likely to have changed to test for HIV. Of most importance was that sexual behavior eliminated the effects of change in risk perceptions suggesting that a change in perception may have occurred as a result of changes in sexual behavior. AIDS prevention programs must pay particular attention to helping youth become aware of their vulnerability to HIV risks, especially as these have implications for risk-reducing behaviors, especially for females who are burdened. Key words: South Africa, risk perception, HIV testing, youth

INTRODUCTION Although improving steadily, South Africa continues to be one of many countries in sub-Sahara Africa severely affected by the AIDS pandemic. The national HIV prevalence among the general adult population has been estimated as ∼17.3%, translating into 5.6 million people living with the virus (UNAIDS, 2010; NDH, 2012). Significant geographic and demographic differences are also observed. Kwazulu-Natal continues to be heavily burdened while the Western Cape Province, with the city of Cape Town as its capital, is least affected. The risk of

infection is disproportionately higher among youth in South Africa, especially young women. For instance, prevalence was reported as 4.5 and 13.6% for young South African males and females aged 15–24, respectively (UNICEF, 2012). While practicing safer sex has been widely acknowledged as relevant to reducing HIV risks among youth in sub-Saharan Africa including South Africa (Hallett et al., 2006; Kirby, 2008; Muula, 2010), testing has emerged as more crucial. Testing informs individuals of their status and is an avenue through which HIV-positive patients may

© The Author 2014. Published by Oxford University Press. All rights reserved. For Permissions, please email: [email protected]

Downloaded from http://heapro.oxfordjournals.org/ at Orta Dogu Teknik University Library (ODTU) on May 21, 2016

Eric Y. Tenkorang*

Perceived vulnerability and HIV testing

BACKGROUND After several decades of research on health prevention and promotion, there is largely a consensus on what influences changes in health behaviors. This consensus partially results from the preponderance of theoretical approaches and models in the health behavior literature and the relative empirical support that these models have enjoyed

in the past. For instance, psychosocial and social learning models of health prevention and promotion, such as the Health Belief Model (HBM), the InformationMotivation-Behavioral Skills (IMB) model, among others, have all underscored the theoretical relevance of some proximal variables including knowledge and perceived vulnerability of contracting diseases as crucial to disease prevention. These models largely acknowledge individuals as rational actors whose decisions derive from a cost– benefit analysis of situations at hand. A number of assumptions underlie psychosocial and cognitive models employed in HIV prevention. It is assumed for instance that youth will engage in risk-reducing behaviors such as testing their serostatus when they perceive themselves at risk of contracting HIV, have knowledge about the disease and learn from the experiences of others infected with HIV (Sherr et al., 2007; Tenkorang et al., 2009; Tenkorang and Maticka-Tyndale, 2013). Consistent with this assumption, several studies have established linkages between risk perception and HIV testing behaviors (Peltzer et al., 2002; Worthington and Myers, 2003; Obermeyer and Osborn, 2007; Lynskey, 2008). Other studies have also documented associations between knowing a person who has died of HIV/AIDS and behavioral change (Rutenberg et al., 2003; Macintyre et al., 2004; Tenkorang and Maticka-Tyndale, 2013). While psychosocial/cognitive models have widely been used in understanding the spread of diseases and developing interventions, they have been criticized for their failure to accommodate cultural interpretations of disease (Kalipeni et al., 2007; Tenkorang et al., 2011). For instance, in settings where people believe disease is caused by agents over whom they have no control, cognitive models that focus on individual action may be ineffective. This has meant that researchers looked beyond cognitive predictors of diseases to include cultural, demographic, socioeconomic and life course variables (see Tenkorang and Maticka-Tyndale, 2013). The relevance of moving beyond psychosocial predictors to broader socio-cultural, demographic and life course factors in disease prevention is considered an important one especially when several studies have found these factors as influential to HIV testing behaviors (see Bwambale et al., 2008; Siziya et al., 2008; Tenkorang and Owusu, 2010). It is documented for instance that younger, wealthy and better educated individuals are more likely to test for their serostatus compared with older, poorer and less educated individuals (Machekano et al., 2000; Siziya et al., 2008; Luseno and Wechsberg, 2009; Macphail et al., 2009; Swenson et al., 2009; Wong et al., 2009). Consistent with previous literature, this study examines the effects of psychosocial factors (change in perceived vulnerability to HIV, knowledge

Downloaded from http://heapro.oxfordjournals.org/ at Orta Dogu Teknik University Library (ODTU) on May 21, 2016

be counseled to live healthily with the disease (Tenkorang and Maticka-Tyndale, 2013). Also, knowledge about one’s HIV serostatus may lead to sexual behavior change (Cribier et al., 1996; Allen et al., 2003; Wong et al., 2009). As part of reducing the HIV burden among youth and the general population, the South African government launched a national strategic plan aimed at maximizing opportunities for testing and screening for HIV and other comorbidities such as tuberculosis (NDH, 2012). This was after government had realized that even in spite of the high prevalence, HIV testing rates among youth remained low (Ramirez-Avila et al., 2010). Several reasons, including stigma, distance to testing facilities and concerns regarding accuracy and confidentiality of test results have often been cited as accounting for the low HIV testing rates among youth in sub-Saharan Africa and South Africa specifically (Kalichman and Simbayi, 2003; Meiberg et al., 2008; Negin et al., 2009; Kabiru et al., 2010). Psychosocial/cognitive models of HIV prevention also refer to the ‘biased’ risk perceptions or sense of invulnerability of youth as a major barrier to HIV testing (Tenkorang and Maticka-Tyndale, 2013). Perceived vulnerability is often obtained in surveys by asking respondents of their subjective perceptions of contracting HIV, whether they know someone living with HIV or if they are aware of any deaths as a result of AIDS. While previous studies have attempted establishing linkages between perceived vulnerability of disease and HIV testing behaviors, the majority of these studies used cross-sectional data that captured events at a snapshot and inherently assumes that perceived vulnerability does not change (Peltzer et al., 2002; Tenkorang and Owusu, 2010; Tenkorang and Maticka-Tyndale, 2013). Although these studies have been useful, it is argued here that the assumption of perceived vulnerability as time invariant is a major flaw that can limit its practical usefulness in AIDS prevention and programing. Also, the use of observational studies creates difficulties in drawing ‘causal’ connections between perceived vulnerability and HIV testing behaviors given the contemporaneous nature of the data. This paper adds to the extant literature by using longitudinal data to examine if changes in the perceived vulnerability of youth in Cape Town South Africa are associated with changes in HIV testing behaviors.

271

272

about HIV, HIV-related stigma, etc.), socio-economic and demographic predictors (age, racial background and school enrolment) and life course factors (sexual experience) on HIV testing behaviors among youth in Cape Town South Africa. Using longitudinal data, this paper offers a perspective different from previous studies that used cross-sectional data in exploring linkages between perceived vulnerability and HIV testing.

Data for this analysis come from the Cape Area Panel Survey (CAPS), a longitudinal study of adolescents in Cape Town, South Africa, that focuses on their sexual and reproductive health, school, work and household living conditions. The aim of this survey was to monitor the life transitions of young South Africans as they move through school, gain employment, leave home and begin sexual intercourse. The transitions made by these young adults are tracked over time, a significant departure from existing data sources providing similar information on young adults in South Africa. Currently, there are four waves of data collected with the first wave of data collection in August 2002. Specifically, the second (wave 2a) and wave 4 are employed for analysis. Wave 2a, which focused primarily on sex and HIV/AIDS, followed about 1360 respondents for re-interview from August 2003 to December 2003 after the first wave was collected in August 2002. Data for wave 4 that followed these respondents were collected in April 2006. Waves 2a and 4 were used mainly because questions on HIV testing were asked only for these waves. The analytic sample was restricted to 857 (male = 426 and females = 431) respondents who answered questions on HIV testing between both waves. The CAPS project design was approved by the ethical review boards at the University of Michigan and University of Cape Town. Written consent was obtained from all respondents, and for those below 18 years written consent was obtained from their parents (Lam et al., 2008).

Measures The outcome variable, ‘change in HIV testing behavior’ was created from questions that asked youth if they had ever tested for their HIV serostatus at both waves 2 and 4. The measure captured respondents who at wave 2 indicated that they had not tested but tested for their HIV serostatus at wave 4. This group is compared with those who maintained they had not tested for their HIV serostatus across waves 2 and 4. Two independent variables are used as proxy for perceived vulnerability. These include questions that asked youth their perceived

risks of contracting HIV and if they know of someone infected with HIV. These questions were asked from respondents at both waves of the survey. Response categories for ‘risk perception’ include ‘no risk’, ‘small risk’, ‘medium risk’ and ‘great risk’. These were transformed into dichotomous variables at both waves indicating youth who thought they had no risk perceptions and those who at least acknowledged they had some risks of contracting HIV. A new measure, ‘change in risk perceptions’ was created indicating youth who had not changed their risk perceptions across waves, those who had changed from ‘no risks’ in wave 2 to acknowledging they had ‘some risks’ at wave 4, those who thought had ‘some risks’ at wave 2 but changed to ‘no risks’ at wave 4 and finally, youth who maintained having ‘some risks’ of contracting HIV at both waves. The second measure of perceived vulnerability captured changes in respondents′ knowledge of someone infected with HIV. These include youth who reported they did not know of someone infected with HIV across both waves, those who changed from not knowing to knowing someone infected with HIV across both waves, those who changed from knowing to not knowing someone infected with HIV and those who maintained knowing persons with HIV across both waves. Given that sexual risks do inform perceived vulnerability and could be an important measure of the real/actual risks of youth, the analysis included a measure of sexual experience capturing the changes that occurred over time. Thus respondents who maintained being virgins across both waves were compared with those who were virgins at wave 2 but experienced sexual intercourse at wave 4, those who maintained they had experienced sex at both waves and those who for some reasons recanted at wave 4. Other theoretically relevant psychosocial, demographic and socio-economic variables are controlled in this study. Psychosocial variables include knowledge/myths about HIV created from a series of questions that asked respondents if they know HIV can be transmitted by sharing public toilets, sharing the same bath, sharing a bottle of water, kissing the lips, shaking hands, deep kissing, touching the genitals, using condoms during sex and having oral sex. Response categories include ‘yes = 2’, ‘may be = 1’ and ‘no = 0’ Using principal component analysis, the first five indicators loaded on the same latent construct called general myths about HIV transmission and the last 4 on another construct called myths related to sexual transmission of HIV. Reliability coefficients (Cronbach’s alpha) for both scales are 0.683 and 0.540, respectively. HIV-related stigma was also measured using questions that asked respondents whether they will still be friends with someone if they were HIV-positive, would drink from the same bottle as HIV-positive person, would buy fresh vegetables from

Downloaded from http://heapro.oxfordjournals.org/ at Orta Dogu Teknik University Library (ODTU) on May 21, 2016

DATA AND METHODS

E. Y. Tenkorang

273

Perceived vulnerability and HIV testing

Data analysis The dependent variable used in this study is dichotomous so logit models were used for analysis. The standard logit models are built on the assumption of independence of observations but the CAPS data have a hierarchical structure with participants nested within households and clusters, which could potentially bias the standard errors. To control for this dependence, I employed random effect models that enabled the estimation of the magnitude and significance of clustering both at the household and cluster levels (Pebley et al., 1996; Guo and Zhao, 2000; Raudenbush and Bryk, 2002). The extent of clustering in the models was measured using intra-class correlations (ICC). For standard logit models, the ICC at the household and cluster levels are estimated as τ 2α

π þ τ 2α þ τ 2β 3 2

and

τ 2β π þ τ α 2 þ τ 2β 3 2

where π3 is the variance at level 1 and τ 2α and τ 2β the variance at the household and cluster levels, respectively (Raudenbush and Bryk, 2002). The GLLAMM program available in STATA was used to build all models. 2

RESULTS Table 1 presents a univariate distribution of selected dependent and independent variables. On the average, male respondents are slightly older than female

Table 1: A univariate distribution of selected dependent and independent variables Variables

Male (N = 426), median/%

Change in HIV testing behaviors No–no 76.3 No–yes 23.7 Perceived vulnerability Change in risk perception No–no risk 20.2 Some–some risk 31.9 No–some risk 13.6 Some–no risk 34.3 Change in knowing someone infected with HIV No–no 52.2 Yes–yes 10.7 No–yes 8.7 Yes–no 28.4 Other psychosocial variables Median score general myths about −0.320 HIV Median score myths related to 0.052 sexual transmission Median score for stigma related to −0.241 HIV Median score for willingness to −0.157 disclose HIV status Demographic variables Average age of respondents 18.94 Race African 35.6 Coloreds 50.6 Whites 13.8 Religion Christians 62.7 Muslims 11.4 Others 12.0 None 13.8 Currently enrolled in school No 44.3 Yes 55.7 Change in sexual behavior No–no 20.0 Yes–yes 43.5 No–yes 26.1 Yes–no 10.3

Female (N = 431), median/%

56.4 43.6

25.0 26.3 18.0 30.6 47.5 12.9 11.7 27.5 −0.316 −0.187 −0.183 −0.157

18.31 37.1 42.9 20.0 67.1 10.0 8.0 14.9 31.7 67.4 27.4 35.5 28.7 27.4

respondents, the majority of respondents are Christians, Coloreds and are enrolled in school. The results show that compared with males, a higher proportion of females changed their HIV testing behaviors (23.7 versus 43.6%),

Downloaded from http://heapro.oxfordjournals.org/ at Orta Dogu Teknik University Library (ODTU) on May 21, 2016

HIV-positive shop-keeper and whether they think HIVpositive persons put others in their classes at risk. Response categories include ‘definitely no’, ‘probably no’, ‘probably yes’ and ‘definitely yes’. Higher values on the scale indicate rejecting stigma and lower values endorsing stigma about HIV. Reliability of the scale as measured by Cronbach′s alpha is 0.625. Willingness to disclose HIV-positive status, the third psychosocial variable, was measured with questions that asked if respondents were willing to disclose their HIV serostatus to their neighbors, a priest or church person, school teacher and to anyone. Response categories include ‘yes = 1’ and ‘no = 0’. Higher values on the scale indicate a greater willingness to disclose HIV serostatus while lower values show lower/unwillingness to disclose one′s status. Reliability coefficient as determined by Cronbach′s alpha is 0.524. Socio-demographic variables in the analysis include age of respondents measured in years, racial background with respondents identifying as ‘African’, ‘colored’ and ‘whites’ and whether respondents are currently in school studying given that schools are often used as platforms for disseminating HIV information.

274

Table 2: A bivariate analysis of selected dependent and independent variables Variables

Boys

Perceived vulnerability Change in risk perception No–no risk 1.00 Some–some risk 1.04 (0.058) No–some risk 0.988 (0.069) Some–no risk 0.962 (0.057) Change in knowing someone infected with No–no 1.00 Yes–yes 1.29 (0.064)*** No–yes 1.33 (0.066)*** Yes–no 1.09 (0.051) Other psychosocial variables General myths about 0.988 (0.021) HIV Myths related to sexual 0.970 (0.021) transmission Stigma related to HIV 0.967 (0.019) Willingness to disclose 0.978 (0.030) HIV status Demographic variables Age of respondents 1.03 (0.009)*** Race African 1.00 Coloreds 0.917 (0.044)** Whites 0.849 (0.097)* Religion Christians 1.00 Muslims 0.932 (0.073) Others 1.19 (0.064)*** None 1.13 (0.057)** Currently enrolled in school No 1.00 Yes 0.947 (0.044) Change in sexual behavior No–no 1.00 Yes–yes 1.18 (0.056)*** No–yes 1.14 (0.061)** Yes–no 1.01 (0.090)

Girls

1.00 1.32 1.18 1.10 HIV 1.00 1.21 1.31 1.02

(0.064)*** (0.069)** (0.065)

(0.068)*** (0.068)*** (0.059)

0.978 (0.028) 0.954 (0.023)** 0.963 (0.025) 0.997 (0.025)

1.06 (0.010)*** 1.00 0.841 (0.049)*** 0.611 (0.092)*** 1.00 0.877 (0.086) 0.934 (0.067) 0.997 (0.087) 1.00 1.24 (0.052)*** 1.00 1.88 (0.057)*** 1.54 (0.058)*** 1.31 (0.111)**

Note. Odds ratios are reported with robust standard errors in brackets. *P < 0.1; **P < 0.05; ***P < 0.01.

behavior. In all models, I estimate clustering at both the household and cluster levels indicated by the variance components and ICC. Multivariate results for females are somewhat consistent with bivariate findings. For instance, I find that change in risk perception is significantly associated with change in HIV testing behaviors except when demographic variables are controlled in Model 3. Further analysis show that it is the race variable that

Downloaded from http://heapro.oxfordjournals.org/ at Orta Dogu Teknik University Library (ODTU) on May 21, 2016

respectively. Also, compared with males, female youth were more likely to change perceptions about their vulnerability to HIV/AIDS. For instance, while about 18% of female respondents changed their risk perceptions from ‘no risks’ to ‘some risks’, ∼13.6% of male respondents did the same. A higher proportion of females compared with males changed from not knowing someone infected with HIV to knowing someone infected with the virus. The median scores indicate that youth in Cape Town, South Africa, reject general myths surrounding the transmission of HIV. However, male youth endorse myths related to sexual transmission compared with females. Both male and female youth endorse stigma related to HIV and are unwilling to disclose their serostatus to those close to them. Approximately 28.7% females broke their virginity across both waves, compared with 26.1% males. Bivariate analyses in Table 2 suggest that a change in perceived vulnerability is significantly associated with a change in the decision to test for one′s HIV serostatus. For instance, compared with female youth who thought they had no risks of contracting HIV across both waves, those who changed from ‘no’ risks to acknowledging they had ‘some’ risks also made a decision to change from not testing to testing their HIV serostatus. Similar results are obtained for female youth who maintained having some risks across both waves. It is interesting to find that both male and female youth who changed from not knowing someone infected with HIV to knowing a person infected with the virus also changed to test their HIV serostatus. This is also true for those who maintained knowing persons infected with HIV across waves. It is important to note that a change in sexual behavior is associated with a change in HIV testing behaviors for both boys and girls. For instance, females who had suddenly lost their virginity across waves were 54% more likely to change from not testing to testing their HIV serostatus. Demographic variables are significantly associated with HIV testing behaviors. Compared with Africans both Coloreds and Whites are significantly less likely to changing to test their HIV serostatus. Compared with those not enrolled in school, female youth enrolled are 24% more likely to have changed from not testing to testing their HIV serostatus. Similarly, older youth compared with younger ones are significantly more likely to have changed from not testing to testing for HIV. Multivariate models are presented separately for girls and boys in Tables 3 and 4, respectively. In all eight models are built, four each for males and females. The first model examines the relationship between change in perceived vulnerability and change in HIV testing behaviors. Model 2 adds other psychosocial variables; Model 3 adds demographic variables and Model 4 changes in sexual

E. Y. Tenkorang

275

Perceived vulnerability and HIV testing Table 3: Multivariate analysis of HIV testing behaviors among girls in Cape Town, South Africa Variables

Model 1

Model 3

Model 4

1.00 1.26 (0.068)*** 1.18 (0.072)** 1.12 (0.067)

1.00 1.19 (0.068)*** 1.13 (0.070)* 1.07 (0.066)

1.00 1.06 (0.066) 1.04 (0.068) 1.04 (0.063)

1.00 1.15 (0.074)** 1.24 (0.070)*** 1.01 (0.061)

1.00 1.08 (0.074) 1.18 (0.072)*** 1.04 (0.060)

1.00 1.09 (0.070) 1.15 (0.067)** 1.06 (0.057)

0.993 (0.029) 0.964 (0.024)

1.02 (0.028) 0.978 (0.023)

1.01 (0.026) 0.979 (0.022)

0.985 (0.026) 0.990 (0.024)

1.02 (0.025) 1.02 (0.023)

1.01 (0.024) 1.03 (0.022)

1.04 (0.013)***

1.01 (0.013)

1.00 0.902 (0.061)* 0.671 (0.099)***

1.00 1.03 (0.064) 0.820 (0.101)**

1.00 0.907 (0.094) 0.908 (0.067) 0.964 (0.089)

1.00 0.885 (0.090) 0.946 (0.063) 0.964 (0.085)

1.00 1.07 (0.066)

1.00 1.05 (0.062)

0.206 (0.016)*** 0.0011 (0.006) 0.059 0.0003 −242.1594

1.00 1.70 (0.076)*** 1.43 (0.066)*** 1.37 (0.114)*** 0.178 (0.014)*** 0.0052 (0.007) 0.051 0.0015 −218.5019

0.230 (0.018)*** 0.0002 (0.007) 0.065 0.0005 273.7352

Note. Odds ratios are reported with robust standard errors in brackets. *P < 0.1; **P < 0.05; ***P < 0.01.

attenuated the statistical significance and effects of respondents who had changed their perceived risks from ‘no risks’ to ‘some risks’. Specifically, the majority of those who changed their risk perceptions from ‘no’ to ‘some’ risks were Africans who also had a higher likelihood of changing to test their HIV serostatus. When controlling

for change in sexual behavior, it is observed that the effect of risk perception is completely attenuated in Model 4. In particular, the effects of female youth who maintained they had some risks of contracting HIV across waves are completely lost. This suggests the risks involved in experiencing sexual intercourse may have also informed the risk

Downloaded from http://heapro.oxfordjournals.org/ at Orta Dogu Teknik University Library (ODTU) on May 21, 2016

Perceived vulnerability Change in risk perception No–no risk 1.00 Some–some risk 1.26 (0.064)*** No–some risk 1.16 (0.069)** Some–no risk 1.10 (0.064) Change in knowing someone infected with HIV No–no 1.00 Yes–yes 1.17 (0.068)** No–yes 1.27 (0.067)*** Yes–no 1.01 (0.059) Other psychosocial variables General myths about HIV Myths related to sexual transmission Stigma related to HIV Willingness to disclose HIV status Demographic variables Age of respondents Race African Coloreds Whites Religion Christians Muslims Others None Currently enrolled in school No Yes Change in sexual behavior No–no Yes–yes No–yes Yes–no Variance at household level 0.230 (0.017)*** Variance at community level 0.0008 (0.007) Intra-class correlations (household) 0.065 Intra-class correlations (Cluster) 0.0002 Log-likelihood ratios −300.5689

Model 2

276

E. Y. Tenkorang

Table 4: Multivariate analysis of HIV testing behaviors among boys in Cape Town, South Africa Variables

Model 1

Model 3

Model 4

1.00 1.09 (0.060) 0.993 (0.070) 1.03 (0.058)

1.00 1.08 (0.060) 1.01 (0.072) 0.997 (0.060)

1.00 1.08 (0.061) 0.997 (0.072) 1.01 (0.061)

1.00 1.30 (0.069)*** 1.34 (0.071)*** 1.09 (0.053)

1.00 1.25 (0.071)*** 1.27 (0.075)*** 1.08 (0.053)

1.00 1.25 (0.071)*** 1.26 (0.076)*** 1.08 (0.054)

1.01 (0.020) 0.991 (0.020)

1.01 (0.020) 0.993 (0.021)

1.01 (0.020) 0.992 (0.022)

0.983 (0.020) 0.992 (0.030)

0.986 (0.020) 1.02 (0.030)

0.988 (0.021) 1.02 (0.031)

1.02 (0.011)

1.02 (0.012)

1.00 1.01 (0.057) 0.929 (0.102)

1.00 1.02 (0.058) 0.930 (0.102)

1.00 0.966 (0.080) 1.13 (0.066) 1.07 (0.064)

1.00 0.974 (0.102) 1.13 (0.066) 1.06 (0.064)

1.00 0.992 (0.055)

1.00 0.996 (0.055)

0.171 (0.013)*** 0.0015 (0.005) 0.049 0.001 −206.2980

1.00 1.05 (0.065) 1.06 (0.065) 1.02 (0.097) 0.171 (0.013) 0.0011 (0.005) 0.049 0.001 204.8849

0.175 (0.013)*** 0.0023 (0.005) 0.050 0.001 −215.1604

Note. Odds ratios are reported with robust standard errors in brackets. ***P < 0.01.

perceptions for female youth in Cape Town, South Africa. Consistently, I find that respondents who had changed from not knowing someone infected with HIV to knowing a person with the virus also changed their decision to test for HIV. In fact, this remained the same for the male models. Male youth who maintained knowing someone with

HIV and those who had changed from not knowing to knowing person with HIV also changed their decision to test for the HIV serostatus. Estimates for variance components and ICC suggest that there are some unobservable factors that affect changes in HIV testing behaviors at the household level but not at the cluster level.

Downloaded from http://heapro.oxfordjournals.org/ at Orta Dogu Teknik University Library (ODTU) on May 21, 2016

Perceived vulnerability Change in risk perception No–no risk 1.00 Some–some risk 1.06 (0.057) No–some risk 1.01 (0.067) Some–no risk 1.01 (0.056) Change in knowing someone infected with HIV No–no 1.00 Yes–yes 1.30 (0.065)*** No–yes 1.33 (0.067)*** Yes–no 0.091 (0.052) Other psychosocial variables General myths about HIV Myths related to sexual transmission Stigma related to HIV Willingness to disclose HIV status Demographic variables Age of respondents Race African Coloreds Whites Religion Christians Muslims Others None Currently enrolled in school No Yes Change in sexual behavior No–no Yes–yes No–yes Yes–no Variance at household level 0.178 (0.013)*** Variance at community level 0.0006 (0.005) Intra-class correlations (household) 0.051 Intra-class correlations (cluster) 0.0002 Log-likelihood ratios −233.3352

Model 2

Perceived vulnerability and HIV testing

DISCUSSION

a higher level of awareness among young South African females of the potential risks that HIV poses and the severity of the disease among them, especially for African females who are known to be heavily burdened. The findings also suggest that because youth perceive their sexual behaviors have implications for HIV transmission, they deem it relevant to know their HIV serostatus after becoming sexually active. Unlike risk perception, the effects of the other measure of perceived vulnerability (knowledge of someone infected with HIV) was statistically robust even after controlling for theoretically relevant covariates. It was interesting to find that for both males and females, knowledge of a person infected with HIV, in particular changing from not knowing to knowing someone infected with the virus led to a change in the decision to test for ones HIV serostatus. This finding provides support for the ‘experiential theory’ that posits that knowing someone living with HIV brings the disease closer to the awareness of people thus invoking behavioral changes including the decision to test for HIV. As HIV prevalence is relatively high in South Africa, it is possible that young people are exposed to the experiences of their own peers, family members and other relatives living with HIV, thus informing their decision to test for the disease. The study finds that some unobservable household level factors may be influential in explaining HIV testing behaviors among youth, in particular, female youth in Cape Town, South Africa. This finding is consistent with several others that claim that beyond individual level determinants, structural factors are relevant and must be considered in explaining risk-taking or riskreducing behaviors among youth in sub-Saharan Africa (Maticka-Tyndale and Tenkorang, 2010; Tenkorang and Maticka-Tyndale, 2013). Several policy questions emerge from this study. It is important that policy makers sensitize youth about their vulnerability to HIV risks. While it is encouraging that young South Africans in Cape Town realize the risks that accompany sexual intercourse, they should also be made aware of other potential sources of risks, especially those that derive from drug use. Engaging HIV-positive persons to speak with youth about their experiences living with the virus may be one of the several ways of raising the awareness of young people′s vulnerability to HIV risks. While findings for this study are useful, some limitations are worth-noting. Data used here are self-reported and subject to bias. It is widely known for instance that collecting information pertaining to HIV and sexual behavior is sensitive. This means respondents are more likely to provide answers that are socially desirable. In the case of HIV, some may deny testing for their serostatus just to avoid the stigma related to HIV or for fear of being tagged

Downloaded from http://heapro.oxfordjournals.org/ at Orta Dogu Teknik University Library (ODTU) on May 21, 2016

Although useful, the majority of studies establishing linkages between perceived vulnerability and HIV risks had mostly used cross-sectional data that assumed that perceived vulnerability does not change. I improve upon previous research by modeling change in perceived vulnerability and how such change affect changes in HIV testing behaviors of youth in Cape Town, South Africa. The results indicate that for females in particular, change in risk perceptions was associated with a change in HIV testing behaviors. Specifically, young women who had changed from not perceiving themselves at risk of contracting HIV to acknowledging they had some risks also changed from not testing to testing their HIV serostatus. Also, female youth who had perceived some vulnerability across waves were more likely to have changed to test for their HIV serostatus. Cognitive and socio-behavioral models that drive HIV prevention underscore perceived vulnerability as crucial to risk reduction including HIV testing. This is largely corroborated by previous research as demonstrated in these studies (Akwara et al., 2003; Macintyre et al., 2004; Tenkorang et al., 2010; Tenkorang and Owusu, 2010; Tenkorang and Maticka-Tyndale, 2013). Race attenuated the effects of risk perception but it was a change in the sexual behaviors of youth that completely mediated the effects of perceived vulnerability. Regarding race, it was further observed from a cross-classification analysis that young African girls were more likely to acknowledge a change in their risk perceptions compared with Whites and Coloreds. This may explain why they were more likely to have changed to test for their HIV serostatus. A number of interpretations are drawn from the above findings. First, that the results support the rationalchoice assumptions underlying cognitive models driving HIV prevention that individuals often engage in riskreducing behaviors such as testing their HIV serostatus when they are aware of their vulnerability to disease and believe it is beneficial to engage in such risk-reducing behaviors (Macintyre et al., 2001; Akwara et al., 2003; Macintyre et al., 2004; Tenkorang et al., 2009). Second, that for young South African females in Cape Town, experiencing sexual intercourse carries enormous HIV risks. This corroborates several other studies that argue that young people often interpret their perceived vulnerability of HIV risks mainly based on their sexual behaviors (Pebley et al., 1996; Akwara et al., 2003; Tenkorang et al., 2009), mostly downplaying other potential sources of risk. The fact that a higher proportion of females compared with males changed to acknowledge their susceptibility to HIV risks and also changed from not testing to testing for their HIV serostatus may be an indication of

277

278

HIV-positive (Plummer et al., 2004). Caution must be taken in generalizing the findings to South Africa, given that data were only collected in Cape Town located in the Western Cape Province. Nonetheless, the use of longitudinal data to examine changes in perceived vulnerability and HIV testing behaviors is unique and findings have practical policy implications.

Akwara A. P., Madise J. N., Hinde A. (2003) Perception of risk of HIV/AIDS and sexual behaviour in Kenya. Journal of Biosocial Science, 35, 385–411. Allen S., Meinzen-Derr J., Kautzman M., Zulu I., Trask S., Fideli U., et al. (2003) Sexual behavior of HIV discordant couples after HIV counseling and testing. AIDS, 17, 733–740. Bwambale F. M., Ssali S. N., Byaruhanga S., Kalyango J. N., Karamagi C. A. S. (2008) Voluntary HIV counselling and testing among men in rural western Uganda: implications for HIV prevention. BMC Public Health, 8, 263. doi: 10.1186/1471-2458-8-263. Cribier B., Chmitt M. P., Coz C. L., Grosshans E. (1996) Changes in sexual behavior of patients attending an HIV testing centre: a prospective study 1988–1994. Genitourinary Medicine, 72, 37–42. Guo G., Zhao H. (2000) Multilevel modeling for binary data. Annual Review of Sociology, 26, 441–462. Hallett T. B., Aberle-Grasse J., Bello G., Boulos L. M., Cayemittes M. P., Cheluget B., et al. (2006) Declines in HIV prevalence can be associated with changing sexual behavior in Uganda, urban Kenya, Zimbabwe, and urban Haiti. Sexually Transmitted Infections, 82, (Suppl.1), i1–i8. Kabiru C. W., Luke N., Izugbara C. O., Zulu E. M. (2010) The correlates of HIV testing and impacts on sexual behavior: evidence from a life history study of young people in Kisumu, Kenya. BMC Public Health, 10, 412–424. Kalichman S., Simbayi L. C. (2003) HIV testing attitudes, AIDS stigma, and voluntary HIV counseling and testing in a black township in Cape Town, South Africa. Sexually Transmitted Infections, 79, 442–447. Kalipeni E., Oppong J., Assata Z. (2007) HIV/AIDS, gender, agency and empowerment issues in Africa. Social Science and Medicine, 64, 1015–1028. Kirby D. (2008) Changes in sexual behaviour leading to the decline in the prevalence of HIV in Uganda: confirmation from multiple sources of evidence. Sexually Transmitted Infections, 84, (Suppl. 2), ii35–ii41. Lam D., Cally A., Nicola B., Anne C., Murray L., Menendez A., et al. (2008) The Cape Area Panel Study: A Very Short Introduction to the Integrated Waves 1-2-3-4 Data. The University of Cape Town, Cape Town, South Africa. Luseno W. K., Wechsberg W. M. (2009) Correlates of HIV testing among South African women with high sexual and substance-use risk behaviors. AIDS Care, 21, 178–184. Lynskey V. (2008) Risk perception and the stigma of HIV/AIDS: why routine testing will change how Americans view the

disease. Interdisciplinary Journal of Health, Ethics, and Policy, 7, 7–11. Machekano R., McFarland W., Hudes E. S., Bassett M. T., Mbizvo M. T., Katzenstein D. (2000) Correlates of HIV test results seeking and utilization of partner counseling services in a cohort of male factory workers in Zimbabwe. AIDS & Behavior, 4, 63–70. Macintyre K., Brown L., Sosler S. (2001) It’s not what you know, but who you knew: examining the relationship between behaviour change and AIDS mortality in Africa. AIDS Education and Prevention, 13, 160–174. Macintyre K., Rutenberg N., Brown L., Karim A. (2004) Understanding perceptions of HIV risk among adolescents in KwaZulu-Natal. AIDS and Behavior, 8, 237–250. MacPhail C., Pettifor A., Moyo C., Rees H. (2009) Factors associated with HIV testing among sexually active South African youth aged 15–24 years. AIDS Care, 21, 456–467. Maticka-Tyndale E., Tenkorang E. Y. (2010) A multi-level model of condom use among male and female upper primary school students in Nyanza, Kenya. Social Science and Medicine, 71, 616–625. Meiberg A. E., Bos A. E. R., Onya H. E., Schaalma H. P. (2008) Fear of stigmatization as barrier to voluntary HIV counseling and testing in South Africa. East African Journal of Public Health, 5, 49–54. Muula A. S. (2010) HIV prevalence and associated behavioral factors in Lesotho, Kenya, Malawi and Uganda. Croatian Medical Journal, 51, 271–273. National Department of Health. (2012) The 2011 National Antenatal Sentinel HIV & Syphilis Prevalence Survey in South Africa. Department of Health, Republic of South Africa, 2012. http://www.journaids.org/images/uploads/2011_ nat_antenatal_sentinel_HIV_and_syphilis_prevalence_survey. pdf (last accessed 7 January 2014). Negin J., Wariero J., Mutuo P., Jan S., Pronyk P. (2009) Feasibility, acceptability and cost of home-based HIV testing in rural Kenya. Tropical Medicine and International Health 14, 849–855. Obermeyer C. M., Osborn M. (2007) The utilization of testing and counseling for HIV: a review of social and behavioral evidence. American Journal of Public Health, 97, 1762–1774. Pebley A. R., Goldman N., Rodriguez G. (1996) Prenatal and delivery care childhood immunization in Guatemala: do family and community matter? Demography, 33, 231–247. Peltzer K., Mpofu E., Baguma P., Lawal B. (2002) Attitudes toward HIV-antibody testing among university students in four African countries. International Journal for the Advancement of Counseling, 24, 193–203. Plummer M. L., Ross D. A., Wight D., Changalucha J., Mshana G., Wamoyi J., et al. (2004) “A bit more truthful”: The validity of adolescent sexual behavior data collected in rural northern Tanzania using five methods. Sexually Transmitted Infections, 80, ii49–ii56. Ramirez-Avila L., Nixon K., Noubary F., Giddy J., Losina E., Walensky R. P., et al. (2010) Routine HIV testing in adolescents and young adults presenting to an outpatient clinic in

Downloaded from http://heapro.oxfordjournals.org/ at Orta Dogu Teknik University Library (ODTU) on May 21, 2016

REFERENCES

E. Y. Tenkorang

Perceived vulnerability and HIV testing

Tenkorang E. Y., Rajulton F., Maticka-Tyndale E. (2009) Perceived risks of HIV/AIDS and first sexual intercourse among youth in Cape Town, South Africa. AIDS & Behavior, 13, 234–245. Tenkorang E. Y., Adjei J., Gyimah S. O. (2010) Perceptions of HIV/ AIDS risk and sexual risk-taking of young people in Ghana. Canadian Journal of Development Studies, 31, 329–347. Tenkorang E. Y., Obeng-Gyimah S., Maticka-Tyndale E., Adjei J. (2011) Superstition, Witchcraft and AIDS prevention in subSaharan Africa: The case of Ghana. Culture, Health & Sexuality 13, 1001–1014. UNICEF. (2012) Preventing HIV Infection among Adolescents and Young People, 2012. http://www.unicef.org/esaro/ 5482_HIV_prevention.html (last accessed 7 January 2014). United Nations Program on HIV/AIDS. (2010) Global Report: UNAIDS Report on the Global AIDS Epidemic. http://www. unaids.org/documents/20101123_GlobalReport_em.pdf (last accessed 10 May 2014). Wong L. H., Van Rooyden H., Modiba P., Richter L., Gray G., McIntyre J. A., et al. (2009) Test and tell: correlates and consequences of testing and disclosure of HIV status in South Africa (HPTN 043 Project Accept). Journal of Acquired Immune Deficiency Syndrome, 50, 215–222. Worthington C., Myers T. (2003) Factors underlying anxiety in HIV testing: risk perceptions, stigma, and the patient– provider power dynamic. Qualitative Health Research, 13, 636–655.

Downloaded from http://heapro.oxfordjournals.org/ at Orta Dogu Teknik University Library (ODTU) on May 21, 2016

Durban, South Africa. Presented at the 3rd Pediatric Workshop, 15–16 July 2010, Rome, Italy. Raudenbush S. W., Bryk A. S. (2002) Hierarchical Linear Models: Applications and Data Analysis Methods. Sage Publications, Newbury Park, CA. Rutenberg N., Kaufman C. E., Macintyre K., Brown L., Karim A. (2003) Pregnant or positive: adolescent childbearing and HIV risk in Kwazulu-Natal, South Africa. Reproductive Health Matters, 11, 122–133. Sherr L., Lopman B., Kakowa M., Dube S., Chawira G., Nyamukapa C., et al. (2007) Voluntary counseling and testing: uptake, impact on sexual behavior, and HIV in a rural Zimbabwean cohort. AIDS, 21, 851–860. Siziya S., Muula A. S., Rudatsikira E., Mataya R. H. (2008) Correlates of HIV testing among women in Malawi: results from the 2006 Multiple Indicator Cluster Survey. Tropical Medicine and International Health, 13, 1351–1356. Swenson R. R., Rizzo C. J., Brown L. K., Payne N., DiClemente R. J., Salazar L. F., et al. (2009) Prevalence and correlates of HIV testing among sexually active African American adolescents in four US cities. Sexually Transmitted Diseases, 36, 584–591. Tenkorang E. Y., Owusu G. (2010) Correlates of HIV/AIDS testing in Ghana: some evidence from the Demographic and Health Surveys. AIDS Care, 22, 296–307. Tenkorang E. Y., Maticka-Tyndale E. (2013) Individual-andschool-level correlates of HIV testing among secondary school students in Kenya. Studies in Family Planning, 44, 169–187.

279

Perceived vulnerability and HIV testing among youth in Cape Town, South Africa.

The importance of perceived vulnerability to risk-reducing behaviors, including HIV testing, is fairly established, especially among youth in sub-Saha...
168KB Sizes 2 Downloads 6 Views