In sub-Saharan Africa, 90–95 percent of HIV infections are transmitted through heterosexual sex. As a result, encouraging changes in heterosexual behavior is a large part of the HIV prevention effort in that region. However, research has shown that, on average, Africans have not changed their sexual behavior very much in response to HIV. This is particularly surprising in light of large changes in behavior among another high-risk group—gay men in the United States.
In the new study "HIV and Sexual Behavior Change:Why Not Africa?" Emily Oster, Becker Fellow for the Gary S. Becker Center on Chicago Price Theory, founded by Richard O. Ryan, MBA ´66 at the University of Chicago Graduate School of Business, analyzes this apparent lack of behavioral response among Africans. Most prior estimates of behavioral response have focused on very limited and specialized populations. Oster provides the first estimates of behavioral response among the general population in a large number of African countries.
In theory, it seems reasonable that people would adopt safer sexual practices, including monogamy, in response to HIV risks. However, Oster notes that the "cost" of HIV infection and premature death depends on an individual�s life expectancy without HIV and their future income. In other words, individuals with higher future value of life should have a greater response to HIV.
HIV infection usually results in death approximately ten
years from the time of infection.
"If someone knows they will die for certain in ten years
even without HIV, the incentive to avoid risky behavior that
will expose them to HIV is minimal," says Oster.
Oster explores whether this theoretical relationship holds true in practice. Her results suggest a strong correlation between income, life expectancy, and behavior change. Individuals with higher income and longer expected future life span are more likely to respond to HIV risk by lowering their number of sexual partners. This may explain some of the difference in behavioral response between Africa and the United States. And, because mortality threats and poverty remain fundamental barriers to HIV prevention in Africa, interventions designed to decrease mortality risks, such as malaria, could have significant effects on HIV prevention.
"While economists won’t be surprised at these results, for epidemiologists and those in the public health field this may be a slightly different way to look at the world," says
Oster. "The idea that people wouldn't protect themselves from a risk of death because their life is not as valuable runs contrary to intuition; yet it seems to hold true in
the data."
An Investment in Health
In the study, Oster approaches health as an investment. Just as firms make choices about investing in new equipment or new plants, individuals make choices every day about investments in their health. Choices about smoking, exercising, and eating
well can all be viewed as investments in long-term health and longevity. Similarly, in a world where HIV is common, the choice to avoid risky sexual behavior can be viewed as an investment in future survival.
Oster makes a similar argument about income: individuals who expect to be wealthier in the future have more incentive to invest in their future health and survival.
Estimating Responses to HIV
For Oster’s study, regional HIV rates were calculated based on data on pregnant women from the U.S. Census HIV/AIDS Surveillance Database. Oster combined this data with population-based measures of HIV prevalence from the Demographic
and Health Surveys (DHS), which focus on fertility, contraception, child health, and sexual behavior in African countries fromthemid-1980s onward. The analysis focuses
on Benin, Burkina Faso, Ethiopia, Ghana, Kenya,Malawi, Mali,Namibia, and Zimbabwe.
Oster observed two variables: 1) reporting for more than one sexual partner in the last year; and 2) the total number of partners in the last year.
Approximately 2 percent of women reported having more than one partner, while approximately 11 percent of men reported having more than one partner.
Oster also needed measures of income and non-HIV life expectancy in order to test the predictions of her theoretical model of health investments. Durable-good ownership and whether the household has electricity served as primary measures of income. To estimate life expectancy, Oster used information on patterns of mortality in Africa in the absence of HIV.
The first set of results show that, on average, there is very little individual response to HIV in Africa. This is consistent with previous research, which has shown that sexual behavior in Africa remains largely unchanged as a result of HIV.
Oster then analyzed the relationship between behavior change and income and life expectancy. She found that individuals who have more income and more expected years of life without HIV change their sexual behavior more by reducing their number of sexual partners.
Oster also analyzed the effect documenting risks on behavioral response, including the interaction between malaria prevalence and the HIV rate. In areas with high rates of malaria, people already face a high risk of death and may have little incentive to change their sexual behavior. This is reflected in the data: individuals who live in areas with high malaria prevalence have lower response to HIV.
Comparing the United States and Africa
The lack of behavioral response in Africa is often contrasted with the extensive behavioral response of gay men in the United States since themid-1980s.Oster suggests that one part of the explanation may be that gay men in the United States are wealthier and have longer life expectancy than individuals in Africa, and thus have a higher future value of life.
To determine the magnitude of behavioral differences across groups, Oster estimated future utility losses from HIV infection for individuals. Given data on individual income, future survival time, and HIV transmission rates, Oster estimated the "price" per sexual partner for a given individual: the dollar value of the future income lost from HIV infection, multiplied by the probability of infection with a given sexual partner.
Data on gay men came from the MultiCenter AIDS Cohort Study, conducted from1984 through the late 1990s. Oster focused on the period before 1990, when effective treatment became available. The main variable was reporting two or more sexual partners in the last two months, parallel to the variable for reporting multiple partners for the sample of African men. On average, 60 percent of the respondents
reported having multiple sexual partners. Income estimates were based on census data. The 1980 census-based Life Tables for the United States were used for calculating
expected survival.
Oster found that a $10,000 increase in the "price" of a sexual partner in Africa decreased the probability of having multiple sexual partners by 3.2 percentage points. Among men in the United States, this decrease was around 3.5 percentage points. The results suggest a similar level of responsiveness for these two groups. However, the major difference between the groups is that the price per partner is much higher in the
United States than in Africa. This suggests that if Africans were as rich and as long-lived as people in the United States, they might have similar behavior change.
To evaluate the effect of different policy responses to the HIV epidemic, Oster developed a simulation model to track how the epidemic and sexual behavior evolves over time.
Based on the simulation model, Oster finds that behavioral changes make relatively little difference in the magnitude of HIV rates predicted by the model. Even predicting 35 years into the future of the epidemic, the difference in prevalence is only 0.5 percentage points, which reflects the small degree of estimated behavior change.
Oster used the simulation model to evaluate three types of interventions: 1)HIV drug treatment; 2) education; and 3) an additional $1,000 in annual income. The model introduces an intervention at year 25. Educational intervention and additional
income would decrease HIV rates, but both effects are extremely small. Since there is some concern about increases in risky behavior when drug treatment is introduced or made more widely available, an alternative may be treatment in conjunction with emphasis on behavior change.
Prevention Strategies
While existing work on HIV in Africa focuses on cultural barriers to changing behavior, Oster suggests that standard economic theory may provide significant insights.
"In many sub-Saharan African countries, the incentive to change sexual behavior in response to HIV risks is very low," says Oster. "Policymakers must recognize that the issue is not entirely amateur of cultural differences. There is an important
link between response to HIV and other mortality risks."
"HIV and Sexual Behavior Change: Why Not Africa?" Emily Oster.