Recent trends suggest that HIV in Asia is becoming a general epidemic, no longer seen as restricted among populations engaging in high-risk behavior, such as injection drug users (IDUs), men who have sex with men (MSM), and sex workers.1 In Asia and the Pacific, Thailand has the second highest rate of adult HIV prevalence2; however, the country is an exceptional example on an international level because of early and aggressive HIV prevention efforts. Unlike other Asian countries, the Thai government and its nongovernment organizations (NGO) counterparts have been successful in establishing policies and funding efforts that have made Thailand one of the few developing countries that have achieved the Millennium Development Goal 6: “to halt and reverse the spread of HIV/AIDS by 2015.”2 The national HIV prevention programs, in conjunction with an effective disease surveillance system, have resulted in Thailand being viewed as a unique success in the international battle against HIV and AIDS.
In one decade alone, a dramatic reduction in HIV incident cases was achieved and sustained. The increase of condom use, the decline of Thai men visiting sex workers, and the heightened awareness of HIV through effective education and information dispersal can all be attributed to effective government policies and public health efforts, resulting in the reversal of the spread of HIV from the explosive epidemic that it was in the early 1990s.3 Nevertheless, there is a possibility that HIV prevention strategies may be waning or less relevant given the changing nature of sexual behaviors from past to present decades. To ensure that the epidemic does not regain momentum, assessments should be made on the current state of the epidemic and prevention efforts; otherwise HIV transmission will continue to be a serious concern in Thailand.
Although there has been public praise for the work that Thailand has accomplished, there has been speculation about the efficacy of these interventions, specifically regarding how well they have succeeded in certain risk groups and why.4 An investigation into transmission rates responds to questions about efficacy because they provide an assessment that incorporates information on prevention efforts, the timing of effects of public health efforts in HIV, and transmission trends over time on a national level and in specific at-risk subpopulations. The transmission rate should not be mistaken for the transmissibility, which refers to the likelihood of contracting HIV based on the type of contact (ie, via blood, bodily fluids). Instead, the transmission rate is the quotient of new cases of HIV divided by the existing number of persons living with HIV for a specified period of time.5 In other words, the transmission rate is a measure that considers the potential of disease spread by accounting for the number of people capable of transmitting HIV.
Because antiretroviral therapy is prolonging the lives of HIV-infected individuals, the prevalence of the disease is increasing as is the opportunity for transmission to HIV-negative persons.5 Concurrently, an increase of HIV incidence is certainly a cause for concern but does not necessarily indicate failure of or shortcomings in HIV programs and policies.6 Although the estimates of the number of persons living with HIV in Thailand are expected to increase over time,3 and the number of annual new infections is important, the rate of HIV transmission is a more informative statistic in reviewing the success of HIV prevention as it incorporates the changes in disease incidence relative to prevalence, or the existent potential for transmission. Additionally, subgroup analyses provide a comparison of the changes in transmission rates in specific high-risk groups in comparison to the general population. The resultant product is a picture of the transmission dynamic between the core at-risk group and general population, which have been noticeably changing over time.
Transmission rates have been explored on a national level in the United States and in heterosexuals and MSM in the United Kingdom,6,7 but these measures have not yet been calculated in Thailand. Given the successes of Thailand's prevention programs in significantly reducing HIV infection, an analysis of transmission rates will provide a tangible measure that is in relation to the systematic programing implemented in Thailand. Furthermore, novel high-risk behavior subgroup analyses presented provide both the opportunity to explore which subgroups should targeted for intervention and to assess the utility and comprehension of this methodological tool.
National transmission rates and subgroup specific rates were derived from incidence and prevalence estimates generated by the Asian Epidemic Model (AEM). The AEM provided population size estimations of HIV-infected persons in Thailand and inputs of the “populations at-risk”, such as MSM and sex workers, to calculate transmission rates from 1985 to 2005. The AEM was created by the East-West Center8 in collaboration with the Thai Working Group on HIV/AIDS Projection to provide assessments for the Thai Ministry of Public Health.9,10 The AEM uses sentinel serosurveillance, behavioral surveillance, conscript seroprevalence, sexually transmitted infection, IDUs, MSM, and antiretroviral therapy data provided by the Thailand Ministry of Public Health, Department of Disease Control, and other research institutes and universities.
This macro simulation process model is a mathematical replication of the key processes driving HIV transmission. In southeast Asia, the drivers include transmission routes via sex work, intramarital and extramarital sex, MSM, and injection drug use. The projections are calculated annually for adults aged 15 years and older based on behavioral inputs with some adjustments. The AEM has been used in Thailand, Cambodia, Indonesia, Bangladesh, Vietnam, and China. The model has provided HIV statistics in Thailand to UNAIDS and the Thailand Department of Disease Control in the Thailand Ministry of Health.
HIV incidence and prevalence estimates projected by the AEM are available on a national level, and for the following subpopulations: general population females, nonclient males, direct and indirect female sex workers (FSWs), male clients, male sex workers (MSWs), MSM, high-risk sharing IDUs, and low-sharing IDUs. The distinction between direct and indirect FSWs is based on the social factors surrounding sex work (ie, where services are exchanged) and the scientific relevance to the model (ie, higher frequency of partnering for direct, lower frequency for indirect sex workers).8 For the purposes of this investigation, the high and low-share IDUs groups were combined. This was done because of the lower total numbers of IDUs and because IDUs are targeted as a single group in terms of prevention and education. In addition to the actual incidence and prevalence from the baseline scenario based on the actual policy implementation, the AEM also provides a “no early intervention” projection scenario with substantially higher rates of cases, had there not been the intensive early intervention programs instated in the early 1990s. The AEM also incorporated data from a study in collaboration with the World Bank on behavior and transmission dynamics and antiretroviral supply and demand. Using information, such as the size of high-risk behavior populations, the proportion of men who visit sex workers, condom use, and needle sharing, the model produces HIV prevalence and incidence estimates in Thailand had there not been any prevention efforts implemented in the early 1990s.
In the AEM, the number of new infections on the national level is calculated from the behavioral inputs and the probability of transmissions. The number of persons living with HIV (prevalence) in a given year is calculated based on the cumulative number of new infections, less the AIDS and non-AIDS-related deaths. Within the at-risk subpopulations, the model adjusts for changes in behavior, which reflects the reality that there is translational movement between the groups. The annual transmission rate is calculated by dividing the incidence by the prevalence for the same year.
The transmission rate is commonly referred to as the incidence-to-prevalence ratio; however, because HIV incidence in the country in a year is a subset of prevalence, this investigation refers to the measure as a percent.
A transmission rate was calculated for every year from 1985 through 2005. The AEM provides incidence and prevalence projections from 1985 through 2025, but for the purposes of this investigation, rates were calculated through 2005, which is the last year that surveillance data and behavior inputs were available and applied. Data calculations were performed and graphics were created in Microsoft Excel 2000. The resultant data were reviewed in conjunction with a prevention program timeline that included specific details about duration, intensity, coverage, and content/methodology to qualitatively correlate the trends in transmission rates with prevention programing for vulnerable subpopulations and the overall Thai population.
The overall population in Thailand has seen a decline in HIV transmission rates, ranging from 100% at the beginning of the epidemic down to 2.9% in 2005 (Fig. 1). The largest declines occurred between 1988 and 1992, during which time the rate dropped to 20.3%; however, the transmission rates stabilized by 1997 and continued to decrease gradually from 5.3% to less than 3%. The turning point of the epidemic was in 1991, when HIV incidence reached its peak at over 150,000 new cases. From this year to the 2005, the number of new cases of HIV had been decreasing, whereas the number of persons living with HIV continued to increase, resulting in continued decreasing transmission rates.
The general female population rates followed the trends of the national population, with a sharper downward slope in early years to a steady decline to 3% in 2005 (Fig. 2). The nonclient male population had very low transmission rates of less than 1% over the entire study period. These rates contrast the trends in the high-risk behavior subgroups.
FSWs also showed a great decline in transmission rates; however, the transmission rate for direct sex workers was still at 15% in 2005. Both direct and indirect sex workers had higher transmission rates at every year, compared with the national population rates. In 1998, there is an unexpected drop in the transmission rates for both groups, but the rates then increased again the following year before leveling off.
Male clients had dramatic declines in transmission rate from 1989 to 1994, going from 84% to 15% in that time. By 1997, the rate decreased to 4% and remained steady through 2005. In 2005, MSM had transmission rates of 8%. After dramatic drops in the 1980s, the IDU transmission rate wavered between 10% and 12% in the next decade and then slowly leveled off to 7% by 2005.
MSW have unexpected results as their early year transmission rates exceeded 100%. From 1987 to 1989, MSWs transmission rates were 110%, 109%, and 104%. In this time period, there were a greater number of new cases of HIV counted in a year among MSWs than end-year prevalence counts. The transmission rate trend among MSWs did decrease, similar to the other populations, and reached a low of 31% in 1998; however, the rate increased in the following years, making MSWs have the largest transmission rate at 43%.
Thailand is credited with swift action for HIV prevention from the early emergence of the epidemic. By 1984, there were already over 50 NGOs working on the HIV/AIDS epidemic in Thailand,11 and blood screening practices and surveillance pilot programs were being tested.4 However, the real push came when HIV was recognized as a national security issue. The first 3-year Program for Prevention and Control of AIDS launched in 1989, in line with substantial decreases in transmission rates on the national level in 1990, when HIV prevalence had doubled from the year before. During this time, transmission rates for the FSWs and male clients were still very high at 89% and 84%, respectively. By 1993, the rates of both groups decreased dramatically to 32% and 25%, in line with the 100% Condom Use Program, initially implemented in a small number of provinces in 1989 and subsequently put into action in all provinces in Thailand in 1992 after an endorsement from the National AIDS Committee for a national scale up from the pilot programme.12
The declines in transmission rates during the mid 1990s corresponds with the increasing number of NGOs (189 organizations by 1995)13 and their collaborations with government mandated provincial and regional committees. They are also attributed to the massive public information and education campaign in 1993, as transmission rates dropped to 10% on a national level in the following year. From this decade into the new millennium, the Thailand National HIV/AIDS Plan targets marginalized and key populations that drive the epidemic,14 which is supported by the lowering and stabilizing rates reaching 3% in 2005.
Last, in a comparison of this data with that of the World Bank study that predicts the state of the epidemic had there not been the focused prevention strategy started in 1989, it was determined that the transmission rates would presently be approximately 3 times higher had there not been early intervention (Fig. 3). HIV prevalence and incidence diverge in the hypothetical versus actual scenarios in 1990, and by 1992, the actual transmission rate was 20% versus the predicted 45%. The following year, the hypothetical rate was 29% and the actual rate was 10%. By 2005, there is still a 6% difference (9% v. 3%).
The current projections for HIV transmission rates in Thailand are promising at a national level, but variations in the rates for each subgroup highlight the need to focus on particular at-risk groups. In 2005, the US transmission rate was 5.2%,15 which is an absolute difference of 2.3% in comparison with Thailand's 2.9%. Thus Thailand's transmission rate on a national level was 44% lower than that of the United States. In Thailand, the transmission rate demonstrates that although there were over 500,000 existing cases of HIV, only 3% were new cases. The high prevalence is indicative of great potential for high transmission; however, prevention techniques must be in high use to have such a low transmission rate in a year.
HIV incidence was highest in 1991, and HIV prevalence continued to increase for several years. Despite the increasing potential for transmission, incidence actually declined during this period, which is reflected in the transmission rate. The national transmission rate decrease demonstrates the decline of new cases over time relative to the existing prevalent cases due to the increased prevention education and programing.
The general female population trends are closely related to the overall trends of the epidemic in the country as a whole. The majority of new cases are attributed to intimate partner transmission or other transmission modes, but not from IDUs or other high-risk behavior. However, low transmission rates of more recent years may be due to higher prevalence counts from sex workers who transition to the general female population.
The behavioral trends incorporated in the AEM provide evidence that the very low transmission rates of less than 1% among male nonclients over the entire study period may be due to the high HIV prevalence from the crossover of males from the client base to nonclient category. There is a high rate of movement of men from one behavior group to another. Therefore, the incident cases of men who are clients translate to prevalent cases of HIV in the nonclient subgroup in subsequent years, when these men no longer engage in high-risk behavior as clients.
For FSWs, in both the direct and indirect groups, there was an anomalous drop and then increase of transmission rate in 1998. This bump in the trend is likely due to the change in the mandated data collection on sex workers in Thailand. Until this year, a survey of “places of entertainment” was conducted annually, collecting reliable data on the numbers of sex workers at various establishments throughout all of the provinces. The AEM uses these data for reference on the population estimate; however, 1998 marked the transition year of the survey from a nationally standardized mandate to the responsibility of provincial and local organizations. The reporting of FSWs was more variable between the provinces, yet provided evidence that the population of sex workers in Thailand has decreased from this time to the present. Accordingly, the AEM fixes the number of FSWs living with and without HIV from this point on. The AEM also reflects the growing number of general females in its population estimates. Therefore, from 1998 on, HIV incidence among sex workers remains stable, whereas the prevalence decreases because of the shift of FSWs to the general female population, resulting in the dip in the transmission rate among FSWs in this year.
High turnover of MSWs to other male subgroups accounts for the greater than 100% MSW transmission rates from 1987 to 1989. It is a reflection of probable short service time when infection occurs and a resulting quick transition to no longer being an MSW by the end of the year. Therefore, the MSW is counted as a new case in 1 year, but by the end of the year, he has already moved out of the sex worker category to the nonclient male population. The quotient of incidence divided by prevalence would never exceed 100% on a national level. However, due to the behavior models incorporated in the AEM, there is the exceptional situation where disease prevalence in a subgroup is less than the number of new cases characterized by a high-risk behavior.
MSWs still present the largest concern, with the highest transmission rate of all the subgroups in 2005. At 43%, it is over 10 times the national rate. In the years leading up to 1998, there is evidence of increasing condom use within this population. Although there was high turnover resulting in a reduction of MSWs living with HIV, the transmission rate decline is actually due to the incidence decreasing at a faster rate than the prevalence. This trend, along with the stabilized rate of condom use after 1998, results in incidence decreasing at a slower rate than prevalence, and thus higher transmission rates from 1998 to 2005. Because this is a transient group and the smallest subgroup, it is more difficult to track and provide outreach and education.
MSM have a lower transmission rate than MSWs, but this rate is still higher than the overall population. With a high HIV prevalence of 42,000 persons in 2005, there is potential for increased transmission within this group if more prevention efforts are not made. The Thai government has acknowledged where improvements may be made in this subgroup as the 2008 10th National Economic and Social Development Plan addresses the need for more provision and easy-access to condoms, lubricants, and education targeting MSM.16 This will be imperative as HIV prevalence in this group continues to increase, which is indicative of high potential of transmission in future years.
It is more difficult to correlate prevention programing with the smaller at-risk groups, such as IDUs and MSM. Although most of the literature highlights prevention efforts and surveillance in the larger urban areas, where many IDUs and MSM live, these particular groups are hidden segments of the larger overall population making it difficult to assess direct impacts. There are reports of how prevention is promoted in drug treatment clinics in the late 1980s in Bangkok, but the drop from 81% in 1988 to 26% in 1 year for IDUs may be due to saturation in the subgroup in addition to decrease in prevalence due to death. After the initial years of the epidemic, there were few remaining susceptible persons to infect in the network due to the nature of the IDUs network. It is characterized with high probability of transmissibility, high frequency of injection (2-3 times per day), and disconnect from other networks. The continued trend among IDUs is that many of the persons living with HIV have stopped participating in this high-risk behavior and have thus moved to another subpopulation (size of 58,000 in 1993 to 16,000 in 2005). Therefore education is important not just to those who engage in high-risk behavior, but to all of their potential partners in the general public. Information on HIV prevention is essential to those who live with HIV because they are the main transmission risk but also to those whose partners currently or formerly engaged in high-risk behavior.
The limitations of this investigation are in line with any research looking at population level data and the accuracy of HIV incidence and prevalence; however, the AEM has been celebrated as a realistic picture of the epidemic.8 Between the development of the model and the ample availability of epidemiological and behavioral data in Thailand, the AEM projections are referenced by the Thai Ministry of Public Health and UNAIDS for reports on the epidemic. That being said, the model outputs are only as reliable as the surveillance and behavior data inputs,8 therefore, careful attention should still be paid to the early years of surveillance. In addition, when reviewing the transmission rate trends for the individual high-risk groups, analysis should take into account the fluctuations in disease prevalence due to the ability for persons to move from one category to another. With these ongoing shifts in high-risk behavior populations, future surveillance efforts should include more tracking in behavior to fill in data gaps.10 Cost-effectiveness is an area of research that should be examined in the future. An analysis of program budgets, more details on the number of recipients of the programs, and the cost measured per program or per person can provide further detail on the effects of prevention programing on HIV transmission. At the time of this assessment, the prevention allocation in the National HIV/AIDS Program budget was not available for all of the years in this study.
It should be noted that transmission rate calculations in the early years are likely to be more variable than presented because surveillance techniques were still being developed and standardized and therefore, susceptible to reporting bias. Therefore, although the greatest declines in transmission rates occur at the end of the decade in the 1980s, the decreases after this time period present the effects of concerted programing efforts toward the stabilization of HIV transmission.
The investigation into HIV transmission rates in specific high-risk behavior groups does open up a discussion regarding the interpretation of this methodological tool. Although past use of the transmission rate in the United States was related to the probability that an HIV+ individual would transmit HIV to others, there is a notable variation in its interpretation for subpopulations that merits further investigation. The annual transmission rate compares the number of new infections in one year to the number of HIV-positive persons. On a national population level, it is an indication of the likelihood that the HIV+ individual will transmit HIV to others. The mixing of individuals from different high-risk groups requires careful interpretation of the transmission rate.
In high-risk behavior groups, the interpretation is dependent on the risk group and their partners. Therefore, the same interpretation would not be used, as the allocation of persons in the various groups is by the risk behavior, but not indicative of from whom HIV was acquired. For example, although MSM and IDUs HIV transmission is likely occur within the subgroup, sex workers acquire HIV from their male clients and vice versa, a phenomenon known as the feedback loop. Consequently, the originally published interpretation of transmission rates must be modified for these subgroups, with concentration on the analysis of the relationship of new cases of HIV to existing cases, not on likelihood of transmission within the subgroups. To translate the same interpretation, it will be necessary to understand the mapping of interactions between male clients, sex workers, and their partners and to calculate a composite rate for the groups involved. For the purposes of this investigation, the national level, MSM, and IDUs groups could also be understood using the original definition, but the remaining subpopulations are still reviewed individually to examine the relationship between incidence and prevalence over time. A future investigation may include an analysis of transmission rates combining subgroups that interact. For example, combining sex workers and male client incidence and prevalence data would provide a transmission rate that may be analyzed with the original interpretation.
Finally, an argument may be made that causation cannot be proved between population data, prevention programing, and the change in HIV transmission rates; however, there is compelling information that the timing of prevention efforts correlates with the fall of transmission rates in Thailand. Although other phenomena in addition to prevention programing, such as the effects of antiretroviral therapy, cannot be ruled out, the evidence here strongly supports the notion that prevention programs preceded transmission rate declines in a way that suggests these programs were effective.
Future investigations in transmission rates will be able to provide insight in the changing epidemic, from spread in high-risk behavior groups to the general public, including young persons. In the following years, more information on new prevention and education efforts, such as further efforts in female condom use promotion and work in intimate partner relationships, should be reflected in hopeful decreased rates in the populations targeted. The current population of young people does not fall into the high-risk groups, but they are vulnerable because we see that HIV is not a disease of only sex workers and IDUs. It should be noted that this group was not among the audience of the notable prevention efforts heavily powered in the 1990s.
In the future, if data are available by age cohort, transmission rates may provide evidence for a revisiting of the education and prevention campaigns in the 1990s. It may also be helpful to breakdown the current years' transmission rates to shorter time periods of less than a year to better understand the higher transient groups, such as MSWs. Additionally, future developments in the methods and interpretation of the HIV transmission rate should take into account the nature of human interactions that play an important role in the spread of HIV. Investigators must understand behavior, the notion of high turnover, and external factors that influence transmission in certain high-risk subpopulations. The transmission rate is best analyzed in conjunction with the factors that influence both incidence and prevalence in all groups of interest.
This analysis shows decreasing trends in transmission rates on a national level, and in high-risk subpopulations in Thailand, consistent with the timing of strategic prevention programing in the country. Continued monitoring will be necessary, so future revisions to the AEM or updates with more recent data will warrant a revisiting of the transmission rates. When data are available, focus should be made on a comparison of the transmission rates through the present with newer prevention program development, such as the recent funding provided by the Global Fund for the Thai Drug User's Network and the formation of the Thai Women and AIDS Task Force,17 to see the more subtle changes in the past 5 years.
The national transmission rate provides an outlook on trends in Thailand which is informative as HIV is no longer considered a disease of sex workers, MSM, and IDUs, but instead a disease that has the potential to spread to the general population through intimate partner transmission. There is variability in trends of transmission rates between the different high-risk behavior groups and the general population, with sex workers having the highest rates. However, although the smaller subpopulations have lower rates, there may need to be more focus on groups with less surveillance, such as IDUs and MSM. The results of this analysis may guide decision makers in how to focus future prevention efforts, and may provide justification in the participation of organizations and communities, such as those involved in Positive Prevention18 and its evolution to “Positive Health, Dignity, and Prevention” for persons living with HIV.19 In Thailand, transmission rates have shown steady decline and stability in recent years, confirming the success15 of lowering numbers of new cases of HIV despite the ever-growing prevalence of people living with HIV.
The authors would like to thank Mr. Najib Assifi, Ms. Laura Vinha, Ms. Vanda Asapahu, Ms. Nadia van der Linde, and the United Nations Population Fund Thailand Country Office and Asia Pacific Regional Office teams; Mr. Patrick Brenny of the UNAIDS Thailand Country Office; Dr. Trace Kershaw, Dr. Mayur Desai, and Dr. Kaveh Khoshnood of the Yale School of Public Health; Dr. Kristopher Fennie of the Yale School of Nursing; Ms. Grace Choi of New York University; and the East-West Center and Policy Research and Development Institute Foundation for resources, data, and input on this investigation.