Antigua (2012): TRaC measuring condom use among female Spanish-speaking sex workers in Antigua. Round 3.

Background: According to the UNGASS 2010 Country Progress Report, at the end of December 2009 there were a reported eight hundred and fifteen (815) persons who tested positive for HIV, and another one hundred and twenty-six (126) persons with advance HIV. Of these persons, ninety-eight (98) received and continued Anti-Retro-viral Treatment (ART). The report indicates that the determinants of the epidemic include multiple sex partners, sex workers, sex tourism, unprotected sex and underlying psycho-social and economic factors. These psycho-social economic factors include an increase in migrant labour populations, mobility among native Antiguans, and transactional sex among the youth, all of which are driven by poverty or the pressure of a fashionable consumer lifestyle. With a conservative estimate, there are between 1300 and 1700 Spanish-speaking sex workers in Antigua, based on the results from the August 2008 Baseline and 2009 Round 2 PSI TRaC among Spanish-speaking Sex Workers in Antigua. These studies showed that 77% consistently use condoms with their clients in 2008, while 72% consistently use condoms with their clients in 2009. This means that over twenty percent of Sex Workers in each year engage in risky sexual behavior. Further, it was discovered that a very high percent (99.5%-2008 and 83%-2009) of sex workers in both years demonstrated incorrect condom use thereby exacerbating the risk of the transmission of HIV. Another significant improvement noted was the share of respondents who reported carrying a condom at the time of survey. In 2008, be 8% reported carrying a condom, but that percentage subsequently increased to 43% in 2010. The number of respondents who use female condoms with their clients also increased from 10% in the baseline study to 52% in follow-up. Taken together, these findings seem to suggest an overall decline in the engagement of risky sexual behaviors among Spanish-speaking Sex Workers which may have a positive impact on the risk of HIV transmission in Antigua/Barbuda. STUDY OBJECTIVES: This study is the third wave of the TRaC Survey. This survey aims to answer three fundamental questions related to segmentation, monitoring and evaluation of populations. 1. Segmentation. By each risk/behavior combination, which opportunity, ability, and motivation (OAM) constructs and population characteristics are significantly correlated with consistent condom use during the past thirty (30) days. Similar analysis will also be done with correct condom use if data allows it. 2. Monitoring. What are the levels and trends are evident in other key behavior, risk, OAM constructs, and exposure to PSI/SFH activities? (i.e. availability, knowledge, self efficacy, locus of control, threat, belief, outcome expectation, exposure) 3. Evaluation. Is exposure to PSI/SFH activities leading to changes in OAM, risk, and behavior? Specifically, the research objectives are to identify and describe the segments of the target population in terms of: a. Correct and Consistent Condom Use STUDY DESIGN/METHODOLOGY Round I of this study employed an RDS approach; however, this methodology was unsuccessful. Unfortunately, very few distributed coupons translated into use. This is because many sex workers harbor a high level of suspicion toward authorities, which is compounded with a fear of being deported. This may have caused many sex workers to believe that this coupon was a trick to detain and deport them. Measures were put in place to assure the females that this was not true. Once we assuaged their fears and confirmed that this was not a trick by members within their target population, we did see a minimal increase in the number of persons responding to the coupons. Nonetheless, most females who did contact interviewers to set up a time to meet never showed. Reasons proffered were they forgot, were no longer interested, or were busy. Based on this past experience, a time-location sampling approach was selected for Round II and Round III of the study. The total sample size will be 560 Spanish speaking sex workers. Data will be aggregated to create dashboard tables. PSI/SFH will identify the number of physical sites within St. John and neighboring communities where the target population can be found. Next, PSI/SFH will ascertain the size of the target population found at these identified sites at different days and time slots. This will allow PSI/SFH to determine the times and days where there is the maximum number of the target group members present. Based on data from the mapping exercise conducted in Round II of the study this information will be presented to key informants to verify that this is still true. Once the target population size at each location has been validated the list will be updated as necessary. Upon the successful completion of the updated list, all the sites and the number of persons estimated in the target population will be included in a table format (see Table 1). If there is cause to believe that the sex workers at different times are not homogeneous, then we will make distinctions between sites during weekdays and weekends. If there are no reasons to believe that at differing times there are different types of sex workers, then no distinction in day to day will be made . A random selection will then be made of the sites to schedule interviews. Care will be taken to ensure that selection is made during the busiest time periods when many persons of the target group are present. For instance if key informants note that 10 pm to 12 pm on weekdays at BAR ABC, are the busiest time periods, then that time slot will be selected to conduct the interviews. The target population as mentioned previously can come and go as they wish to these locations and therefore it is not possible to obtain a precise measure of the size of the target population. As a result, equal probability sampling will be used to select the clusters. The number of clusters we are required to select will be calculated based on the sample size divided by the average cluster size. For example, if the sample size is 200 and the number of persons in each cluster is 10, then 20 clusters are needed. In the event that the required numbers are not obtained at a particular location, then there will be a 10% increase in the number of persons in each cluster to account for any shortfalls. In each selected location, the interviewer will be required to obtain a count of the number of persons within the target group (this will be done by the interviewer informing the manager of the bar/brothel of the study and identifying the sex workers) and then calculate the interval. For instance, if the interviewer is provided with a count of 24 persons falling within the target population at a location and in each cluster the required number of persons is 12, then the interviewer would interview every other person fitting the target population criteria until the interviewer interviews 12 persons. In the event that the location has less than the required number of persons to interview, then all persons falling within the target group will be interviewed. As needed, additional clusters would then be selected to conduct interviews to make up the numbers