Determinants of self-perceived risk of HIV infection: population-based observations in Zambia

Background Perception of risk of HIV infection has been suggested to be an important area of study as it can be an assumed to be an indicator of one’s understanding of susceptibility to HIV infection and a precursor to behavioral change, which could determine future decision making regarding risk taking. Studies that have examined perception of HIV risk and its determinants still remain limited. Zambia is among the worst affected countries by the HIV pandemic in the sub-Sahara African region. The prevalence rate is 16.5% among adults aged 15-49 years and unprotected heterosexual intercourse is the main mode of transmission. The demographic health survey in 2002 showed that HIV prevalence is higher in the urban areas (23%) as compared to the rural areas (11%); and higher among women (17.8%) as compared to men (12.6%). Combating the spread of HIV in Zambia may require a deeper understanding of key issues that are driving the pandemic. The way people receive and process information on HIV and AIDS, how they perceive themselves at risk of HIV infection and what actions they take afterwards may be a key element to the effectiveness of the programmes that can be designed or improved to reduce the spread of the disease in the different sub-groups of populations in Zambia. In this study, we set out to investigate the levels of and factors that influence self perception of risk of HIV infection among adults in populations with high HIV prevalence levels. More specifically aimed to investigate the influence of socio-demographic factors, health status indicators and sexual behaviour on self perception of HIV risk. We also intended to assess the assumed relationship between self perception of HIV risk with worry of being HIV infected, the intention to seek Voluntary Counseling and Testing (VCT) and the actual use the VCT service when offered. Methods The data was from a population based survey conducted in 2003 in selected urban and rural areas in Zambia. The survey aimed among other things to generate knowledge on HIV prevalence, distribution of patterns and trends in HIV prevalence over time and impact of HIV on mortality and fertility levels. Information on socio-demography, health status, sexual behaviour, perception 6 of risk of HIV infection and VCT was collected from both men and women. Saliva samples were collected from consenting respondents for anonymous linked HIV testing. This study on perception of risk was limited to analyses of information collected from men and women aged 15-49 years who were not aware of their own HIV status. Logistic regression was used to assess the association between level of self-perceived HIV risk and HIV prevalence in rural and urban areas. Multiple linear regression models were used to examine factors associated with self perception of HIV risk, guided by a conceptual framework, which was developed based on previous research findings in this field. Background variables included age, sex, marital status, residence and level of education. Health status indicators included were self rated health, mental distress and HIV status, and the sexual behaviour indicators were number of sexual partners and STI experiences a year before the survey. Step-wise multiple linear regression modeling was used to assess the additive effects of the groups of variables. The association of self-perceived HIV risk with worry of being HIV infected, intention and use of VCT service as per the survey, were analyzed separately using bivariate correlations functions. Results The prevalence of HIV was 13.6% in the rural area and 18.0% in the urban areas, and only 13.6% of the respondents reported to know their HIV status. Fifty percent (50%) of the respondents rated themselves to be at no risk of HIV infection, while 17% to be at high/very high risk. The likelihood of being HIV infected for rural residents was about 1.8 times higher [95% CI 1.17 – 2.63] among those who perceived themselves at high/very high risk as compared to those who perceived themselves at no risk. The association was not significant among urban residents [AOR=1.3, 95%CI 0.97 – 1.78]. The overall positive predictive value of any risk was 20%, meaning that 20% of those who perceived themselves at risk were actually HIV infected. The results of the multiple linear regression analysis showed that perception of risk did not vary by residence except for urban men after adjusting for confounders in the model. However, there were striking differences in age and sex. Perception of risk increased with age among the young people below 30 years of age and dropped with age among those aged above 30 years. The younger men appeared to perceive themselves at higher risk as compared to the young women, 7 indicating a sharp contrast when compared with the age-sex distribution of HIV prevalence which showed that women were 4 times more likely to be infected than men. Being married was associated with a high perception of risk among women as compared to the single women whereas among married men, this association was negative and significant compared to the single men upon addition of other variables in the model. Level of education and mobility were positively associated with self-perceived HIV risk among men only; on the other hand, the effect of mobility was significant among both young men and women aged15-24 years. Health status indicators measured as self rated health, HIV status and mental distress were associated with self-perceived HIV risk, with mental distress having the strongest effect and HIV status having the weakest effect. However among young people, the effect of HIV status was only significant among young women. Sexual behaviour indicators differed also by gender. Having many sexual partners was significantly associated with high self-perceived HIV risk among men though weaker and insignificant among young men aged 15-24 years. However, having experienced STI symptoms was significantly associated with self perception of risk among both men and women. Separate analyses showed that self-perceived HIV risk was strongly associated with worry about being HIV infected, intention to seek VCT and actual use of VCT were associated with self perception of HIV risk for both men and women. Conclusion About 50% of the respondents perceived themselves at risk of HIV infection, whereas the HIV prevalence in the study population was 16.1%. The association between HIV status and perception of risk was relatively low with a positive predictive value of 20% i.e. only 20% of those who perceived themselves at any risk were actually HIV infected. However, these discrepancies in perception of risk and actual risk are not surprising due to the fact that infectiousness of HIV is very low: the probability of being infected by HIV per risk exposure is < 0.001. Perceiving oneself at risk of HIV infection was strongly associated with worry of being HIV infected, mental distress and the actual use of the HIV testing services offered to them as part of the survey, indicating that HIV is a big burden in this high HIV prevalence setting and perception of risk is important in health decision making. Perception of risk was also associated with deteriorating health and past risky sexual behaviour. Strong gender differences exist in the 8 population. There is need for the empowerment of women, so that they can have negotiating skills for safe sex. Improvement of education, especially access to basic education is very important for both men and women as it provides well-informed knowledge, confers skills necessary for assimilating health promotion information on HIV & AIDS which in turn is linked to risk reduction and having accurate and correct perception of risk of HIV infection. The study on self perception of risk is a complex process where individuals are to handle conflicting information