Share this article on:

New strategies for HIV surveillance in resource-constrained settings: an overview

Diaz, Theresaa; De Cock, Kevinb; Brown, Timc; Ghys, Peter Dd; Boerma, J Tiese

doi: 10.1097/01.aids.0000172871.80723.3e
Original papers

Additional funding recently became available to help resource-constrained countries scale up their HIV treatment and prevention activities. This increased funding is accompanied by an increased demand for accountability from stakeholders. Many countries will need to make substantial improvements in their current HIV surveillance methods to monitor the collective national impact of these treatment and prevention initiatives. However, whereas most resource-constrained countries have monitored the prevalence of HIV, they have collected little information on other events in the HIV disease process, such as HIV incidence, rate of HIV drug resistance, number of deaths due to AIDS and only modest emphasis has been placed on AIDS reporting in generalized epidemics, resulting in severe underreporting. In addition, data on mortality trends are often not gathered. Furthermore, less than half of the countries with low-level/concentrated epidemics have tailored their surveillance systems to the local epidemic, behavioral surveillance is often not present, an integrated analysis of data is not widespread, and data are rarely used to inform policy. In January 2004, a conference was convened in Addis Ababa, Ethiopia, to examine new strategies for surveillance in resource-constrained countries, and their use in monitoring and evaluating HIV activities. This supplement summarizes the newest approaches and lessons learned for HIV/AIDS surveillance, based on presentations and discussions from that conference. This article provides an overview of HIV/AIDS surveillance in resource-constrained settings and discusses the history, current approaches, and future directions for HIV/AIDS surveillance in generalized and low-level/concentrated epidemics.

From the aCenters For Disease Control and Prevention, National Center for HIV, STD and TB Prevention, Global AIDS Program, Atlanta, GA, USA

bCenters For Disease Control and Prevention, National Center for HIV, STD and TB Prevention, Global AIDS Program, Nairobi, Kenya

cEast-West Center/Thai Red Cross Society Collaboration on HIV Modeling, Analysis and Policy, Bangkok, Thailand

dJoint United Nations Programme on HIV/AIDS, Geneva, Switzerland

eWorld Health Organization, Geneva, Switzerland.

Correspondence to Theresa Diaz, MD, MPH, Centers for Disease Control and Prevention, Global AIDS Program, MS E-30, 1600 Clifton Road, Atlanta, GA 30033, USA. E-mail: txd1@cdc.gov

Back to Top | Article Outline

Introduction

Comprehensive HIV/AIDS surveillance entails the collection, analysis and reporting of data at different points in the HIV disease process. In particular, these data are: the proportion of individuals who have become newly infected with HIV (i.e. HIV incidence); prevalent HIV infections; the number of HIV-infected individuals who newly develop AIDS (i.e. AIDS incidence); prevalent cases of AIDS; mortality from AIDS (Fig. 1), and also the monitoring of risk factors for acquiring HIV infection. Ideally, an HIV/AIDS surveillance system should monitor the levels and trends in the prevalence and distribution of HIV-related risk behaviors, the number of incident cases of HIV occurring in a year, the prevalence of HIV, the number of new AIDS cases occurring in a year, the level of HIV drug resistance, the prevalence of AIDS, and the number of deaths in a year attributable to HIV.

Fig. 1

Fig. 1

Most resource-constrained countries have monitored the prevalence of HIV infection but collected little information on the other events described above in the HIV disease process (Fig. 1). In 2000, the World Health Organization (WHO) and the Joint United Nations Programme on HIV/AIDS (UNAIDS) developed second-generation surveillance guidelines [1], which encouraged the expansion of surveillance activities, dependent on the type and stage of the HIV epidemic in a particular country to obtain a more complete picture of the epidemic. Epidemics are classified into generalized epidemics in which HIV is firmly established in the general population, and in which sexual networking in the general population is sufficient to sustain an epidemic independent of sub-populations at higher risk of infection [HIV seroprevalence in sentinel antenatal clinics (ANC) is at least 1% on a national basis and is used as a numerical proxy] [1]. In low level/concentrated epidemics HIV is not well established in the general population (HIV seroprevalence in sentinel ANC less than 1% is the suggested numerical proxy) [1].

Additional funding recently became available to help resource-constrained countries scale up their HIV treatment and prevention activities. These new funding initiatives include the Global Fund to Fight AIDS, Tuberculosis and Malaria, the US President's Emergency Plan for AIDS Relief, and the World Bank's Multi-Country AIDS Programme. In addition, WHO has set a goal of providing antiretroviral medications to 3 million individuals from resource-constrained countries by 2005 (the 3 by 5 initiative). These initiatives and funds should result in large increases in the provision of antiretroviral treatments, HIV testing, prevention of mother-to-child transmission programmes, and other prevention programmes. As more resources are invested in programming, surveillance systems will need to provide information on the number of individuals in need of these services, and also the overall impact of these activities.

A conference was convened in Addis Ababa, Ethiopia, in January 2004 to discuss recent developments and emerging strategies for surveillance in resource-constrained countries, and the relevance of these strategies for monitoring and evaluating the impact of this rapid expansion of programmes supported by international funding initiatives. This supplement summarizes the newest approaches and lessons learned for HIV/AIDS surveillance, based on presentations and discussions from that conference. This article provides an overview of HIV/AIDS surveillance in resource-constrained settings, and reports the history, approaches, and future directions for HIV/AIDS surveillance in generalized and low-level/concentrated epidemics.

Back to Top | Article Outline

Generalized epidemics

Issues and challenges

Although most resource-constrained countries with generalized epidemics have monitored the prevalence of HIV, they have also collected little information on the other events in the HIV process (Fig. 1), placed little emphasis on AIDS case reporting, which has resulted in severe underreporting, and have often failed to gather data on mortality trends. Behavioral surveillance is often not present, and integrated analysis of data is not widespread.

Back to Top | Article Outline

Evolution and current state

Initial attention in resource-constrained countries experiencing generalized HIV epidemics focused on AIDS case surveillance. At a meeting held in 1986 in Bangui, Central African Republic, participants agreed that widespread serological testing for HIV would not be possible, and that AIDS surveillance on the basis of a clinical case definition was appropriate. Clinicians had described the different manifestations of AIDS in Africa, emphasizing wasting, fever, and chronic diarrhea, and also less common findings such as pruriginous dermatosis or varicella zoster. A case of AIDS was defined as the presence of two major and one minor condition, or the diagnoses of disseminated Kaposi's sarcoma or cryptococcal meningitis alone [2]. Studies to validate this definition used hospitalized patients as the gold standard. The case definition was found to be specific as close to 90% of HIV-negative individuals did not fulfill its criteria; however, it lacked sensitivity because almost 50% of ill, HIV-infected individuals also did not meet the criteria [3]. The clinical nature of the case definition led to confusion about its purpose, which was surveillance, not clinical staging. Clinicians objected because the definition failed to capture many disease manifestations.

In addition, many resource-constrained countries were grossly underreporting AIDS cases. In 1997, WHO estimated that no more than 15% of cases were reported in these countries [4]. AIDS case surveillance was nonetheless useful for documenting the presence of AIDS in a particular country, providing descriptive epidemiology of the groups affected, the distribution of AIDS by age and sex, and identifying risk factors. However, AIDS case reporting proved inadequate for assessing the magnitude and trends of the epidemic.

The Centers for Disease Control and Prevention (CDC) AIDS surveillance case definition was revised in 1985, 1987, and 1993 [5], leading to a change in the case definition for resource-constrained countries. In 1994, WHO published revised recommendations for AIDS case surveillance. The Bangui definition was proposed for locations where HIV testing was not feasible, and an expanded definition was recommended for individuals with a positive HIV test and one of several conditions (i.e. weight loss, crypotoccocal meningitis, tuberculosis, Kaposi's sarcoma, neurological impairment, esophageal candidiasis, recurrent pneumonia, and invasive cervical cancer) [6]. Unfortunately, the publication of this expanded definition coincided with reduced interest in AIDS case surveillance in the 1990s, and it continues to have limited impact despite its potential to identify a larger proportion of individuals with AIDS [7].

Recognizing that AIDS case surveillance was a poor measure of recent HIV transmission, the CDC instituted a ‘family of serosurveys’ for HIV in 1987 for the United States, using the approach of unlinked anonymous testing (UAT) [8]. In UAT, blood drawn routinely for other purposes in healthcare settings is stripped of identifiers and tested later for HIV. UAT allows a more representative sampling of a study population than voluntary counseling and testing, for example, by eliminating selection bias associated with requesting consent. WHO guidelines on sentinel surveillance were first issued in 1989 recommending UAT for use in specific sentinel populations [9]. The sentinel population recommended for HIV surveillance in generalized epidemics is pregnant women in a sample of ANC. These women are considered representative of the general population. Data from this sentinel population form the basis for the generation of global, regional, and country estimates of numbers of HIV-infected individuals in countries with generalized epidemics. Walker and colleagues [10] summarized the steps applied to ANC data to generate estimates of national HIV infections [11]. Sentinel groups other than pregnant women include blood donors, sex workers, and sexually transmitted infection patients, tuberculosis patients, and hospitalized patients. Although, these sentinel groups are not representative of the general population, they can provide useful information on trends of the epidemic.

Although some African countries conducted national HIV population-based seroprevalence surveys more than 15 years ago, clinic-based ANC sentinel surveillance became adopted as the norm because it was simple, flexible, sustainable, and capable of monitoring HIV trends. For many years countries have conducted nationally representative household surveys known as Demographic Health Surveys. These surveys have recently included HIV testing in some countries and these data have been compared with data from antenatal HIV surveillance results [11]. In Mali, Kenya, and Zambia estimates of HIV prevalence from ANC sentinel surveillance ranged from 19 to 29% higher than the Demographic Health survey population-based estimates. Whereas national general population-based HIV surveys may underestimate the true HIV prevalence because of participation bias, especially among men who have lower survey participation rates, results from the different African surveys suggest that the total number of HIV-infected individuals may be lower than that estimated by ANC sentinel surveillance, and more women are infected than men. The primary reason for this lower estimate is that ANC sentinel surveillance sites have traditionally been concentrated in urban areas, and urban areas have higher rates of HIV than rural areas. Population-based surveys that sample throughout a country will represent rural areas better than ANC sentinel surveillance does, resulting in a lower national HIV prevalence estimate.

Emphasis should be placed on the “front end” of the epidemic, as planning for prevention programmes requires good estimates of HIV incidence. As antiretroviral therapy becomes more widely available in resource-constrained countries, HIV disease and deaths should be more closely monitored. Planning for prevention programmes requires good estimates of HIV incidence. Soon serological tests for incident infections that distinguish recent from past infections with a single sample will be available for widespread use on non-B subtypes. Once these tests are available they can be applied to sentinel groups that are already sampled for HIV surveillance such as pregnant women and samples from national surveys [12].

As antiretroviral therapy becomes more widely available in resource-constrained countries, HIV disease and deaths should be more closely monitored. In the third decade of the epidemic, special studies of hospital patients, with HIV testing and application of the WHO expanded case definition for AIDS, could be useful in estimating the minimum incidence of AIDS in specific locations. Although not population-based, these data could form the basis of estimating requirements for care [13]. As antiretroviral therapy becomes more accessible, immunological testing of HIV-infected individuals in sentinel groups (e.g. CD4 cell counts) could give an estimate of the proportion of asymptomatic HIV-infected individuals who meet immunological criteria for treatment.

Data from tuberculosis case reporting could also provide useful insights. The life-time risk for tuberculosis in HIV-infected Africans approaches 50%, and tuberculosis incidence is closely linked to population HIV prevalence. Routine HIV testing of tuberculosis patients offers a valuable adjunct to the AIDS case reporting system. Furthermore, tuberculosis registers have greater completeness than AIDS case reporting in Africa and would be easier to strengthen.

Vital statistics in most African countries are too weak to monitor overall and AIDS-specific mortality rates reliably. However, special studies to assess the rates and proportions of deaths as a result of HIV, incorporating HIV testing of cadavers, and appropriate clinical information in defined geographical catchments [14], are a potentially useful approach. Valuable information can be obtained using selected demographic surveillance sites in defined populations where community workers collect information on all births, deaths, causes of death, and migration. There are demographic surveillance sites located in Asia and in sub-Saharan Africa, with the potential to monitor population-based mortality rates, especially when the HIV status of the population under surveillance is known. Tracking HIV-specific mortality in cohorts of treated individuals, analogous to trends described in the CDC HIV Outpatient Study or the Adult Spectrum of Disease Study [15], is another potentially useful method for tracking the survival of treated individuals and thus the impact of widespread antiretroviral therapy.

Countries with generalized epidemics also need behavioral surveillance data. Monitoring risk behaviors among the general adult population, youth, and high-risk populations may yield valuable information. In recent years, many data have been collected through population-based surveys on risk behaviors in the general adult population. These data have provided some insights into the age at first sexual encounter, the link between marriage patterns and sexual behavior, multiple partnerships, sexual mixing, and condom use [16–18]. However, many questions remain about the value of such data for trend analysis, comparisons between and within populations, evaluation of interventions, and influencing decision makers. Increasingly, behavioral data collection focuses on young people. Problems related to the measurement of sexual behavior in surveys observed among adult populations seem to multiply, given the greater stigma associated with socially unacceptable patterns of sexual behavior among young people. Young people, therefore, may be more likely to report more socially desirable behaviors. For example, young women have been found to underreport the number of sexual partners [19]. Much work is needed to improve such reporting. Finally, behavioral surveillance among risk populations, such as female sex workers (FSW), is a neglected component in generalized epidemics.

Back to Top | Article Outline

Low level/concentrated epidemics

Issues and challenges

Countries with low-level/concentrated epidemics present unique challenges for surveillance [20]. The low prevalence, combined with the concentration of HIV in higher-risk populations, limits the usefulness of ANC surveillance used so effectively in generalized epidemics. Whereas both first and second-generation surveillance guidelines call for concentrating on higher-risk populations [1,9], these populations are often stigmatized or criminalized by the legal system, and those charged with conducting surveillance often have limited access to them [21].

Back to Top | Article Outline

Evolution and current state

Historically, countries with low level/concentrated epidemics relied on AIDS surveillance to monitor the epidemic. A variety of AIDS case definitions were used, making it difficult to interpret AIDS surveillance data from different countries. In Latin America in 1989, the Pan American Health Organization developed the Caracas case definition (revised in 1992), which consists of laboratory evidence of HIV and cumulative points assigned to a variety of conditions, and if the summation of these points exceeds 10 an individual is considered to have AIDS [22,23]. In Brazil, a more complex case definition was developed that included the CDC AIDS case definition and the Caracas definition, and more recently includes a CD4 T-cell count of less than 350 cells/mm3. In addition, many countries with concentrated epidemics use the European AIDS case definition developed in 1993 (i.e. the 1993 CDC AIDS case definition minus the CD4 T-cell count) and the CDC 1993 AIDS case definition. In addition to problems with the use of multiple AIDS case definitions, the completeness of case reporting is variable. In relatively wealthy countries (e.g. Barbados) case reporting is over 70%, whereas other countries have very low reporting rates; for example, it is estimated that only one out of seven AIDS cases were reported in Columbia [24].

Despite these problems, AIDS surveillance has proved useful for some countries that provide antiretroviral therapy. For example, Brazil has relied on AIDS surveillance systems to monitor its epidemics over time; decreasing rates of new AIDS cases have been demonstrated with the onset of antiretroviral therapies in 1996 [25,26]. However, the fact that AIDS surveillance monitors individuals in the late stage of the disease limits its utility for planning prevention programmes.

First-generation HIV surveillance guidelines, developed in 1989, were fairly limited in scope, with stated minimal objectives that included determining levels of infection and monitoring trends in key populations [9]. Many countries abided by such ‘minimalist’ objectives and did not collect the information they needed (e.g. risk behaviors and HIV prevalence among high-risk groups) to guide programmes to respond appropriately to their local epidemic situation. Concerns about the wider spread of HIV into the ‘general population’ often led to a prevention focus on larger lower-risk populations, instead of a more effective emphasis on those at higher risk [27,28].

Second-generation surveillance guidelines emphasize tailoring HIV surveillance systems to the country situation. In low-level/concentrated countries, tailoring means focused attention on those populations in which HIV is spreading or is likely to spread in the future. For most countries, these populations include FSW and their clients, men who have sex with men (MSM), and injecting drug users (IDU). In most cases, countries have tried to incorporate these populations into their surveillance systems, with varying levels of success. Walker et al. [29] reviewed the status of HIV serosurveillance systems in the late 1990s. For those parts of the world with low-level/concentrated epidemics, they found fully or partly functional systems in six out of 15 Latin American and Caribbean countries, 10 out of 17 Asian nations, 15 out of 28 eastern/central European/newly independent states, and no north African or middle Eastern countries.

A core component of second-generation surveillance systems is behavioral surveillance, which in low-level and concentrated epidemics provides direct information about where the epidemic potential is highest, where risk is occurring, and whether programmes are reducing risk [30]. Many gaps are evident in the behavioral components of second-generation surveillance systems. For example, early results from an assessment of behavioral surveillance systems currently being conducted by UNAIDS and WHO [31], with half of the countries reporting, only approximately half of Asian countries with behavioral surveillance systems include MSM and IDU.

Access to these groups remains a significant problem for many surveillance systems because of their stigmatized or illegal status, or the unwillingness of surveillance staff to work with them. Convenience samples have serious limitations, because the extent to which they represent the larger at-risk population is unknown. Other data gaps in low-level/concentrated countries include the proportion of adult men who visit sex workers and the sizes of the MSM, IDU, and FSW populations.

A basic tenet of second-generation surveillance, especially in low-level/concentrated epidemics is that epidemiological and behavioral data should be analysed in an integrated manner. To be meaningful, this analysis must include issues of prevention coverage and effectiveness. In the past 2–3 years, new tools for such integrated analysis have become available. Pisani et al. [28] presented a simple spreadsheet approach for determining the relative contributions of different populations to new infections based on fairly simple epidemiological and behavioral inputs, providing a basis for targeting prevention resources to reduce new infections. More advanced tools, such as the Asian Epidemic Model, have been developed and used for more comprehensive integrated analysis in the data-rich environments of Thailand and Cambodia [32–34]. Grassly et al. [35] applied similar models to explore the variation in the impact of IDU epidemics on sexual epidemics in countries such as Russia, China, and India.

However, substantial epidemiological and behavioral data gaps in key populations, and the inconsistent way in which these data have been collected over time, make integrated epidemiological and behavioral analysis challenging. Data on the effectiveness and coverage of current programmes is virtually non-existent. Although it is not currently collected in national surveillance systems, it is essential in conducting an analysis that relates behavior and epidemiology. It takes time and effort to seek out the information to fill data gaps, and integrated analysis requires skills from many disciplines. In practice, the national programme units charged with surveillance do not have the training, the time, or the manpower for this type of analysis. Accordingly, in most low-level/concentrated countries the closest that epidemiological and behavioral data have come to integration is to be presented at the same meetings. As monitoring and evaluation efforts are expanded, it is critical that integrated analysis efforts be supported, both through capacity building and the allocation of personnel with responsibility for such analysis. An integral part of this analysis should be the identification of major data gaps, improvements in surveillance systems to fill those gaps in the future, and the implementation of strategies to influence policies and programmes more effectively.

Back to Top | Article Outline

Data use for action

To be effective, public health surveillance should link data with action. Unfortunately, this is one area in which second-generation surveillance efforts have been more often linked with failure than with success. In a few countries, e.g. Thailand, Cambodia and Uganda, good use has been made of surveillance data to target prevention efforts, build high-level political support, and mobilize strong community responses. The epidemics in these countries have turned the corner with increased condom use, a decreased number of sex partners, and declining HIV prevalence in key populations [36–38]. However, this is not often the case. National surveillance data in Bangladesh, China, Indonesia, and the Philippines document ongoing low levels of consistent condom use between FSW and clients. The growing HIV prevalence among FSW seen in the surveillance data of many countries, e.g. China, Nepal, and Vietnam, further documents programme failures [39]. Similarly, years of worrying epidemiological and behavioral surveillance data among IDU documenting high HIV levels and continuing risk do not appear to be influencing national programmes to take HIV among IDU seriously. HIV prevalence among IDU remains high throughout Asia, and new outbreaks are reported among IDU in many parts of eastern Europe [40,41]. Clearly, in many places surveillance data are not adequately influencing programmes.

A number of factors contribute to this poor performance. Few integrated analyses have examined the contribution of specific populations, the links between specific populations and the overall epidemic in the country, sex and age dimensions of the epidemic or the impact of the epidemic on mortality, hospitalization, and the labor force. Integrated analyses are also needed to examine and understand behavioral trends and their impacts on HIV transmission in relation to underlying (e.g. socioeconomic changes, availability of intervention programmes to prevent HIV transmission) and biological (e.g. efficiency of transmission of HIV, duration of infectivity) determinates [42]. This leaves policymakers with the perception that the low-level/concentrated epidemic will remain small or, in generalized epidemics, that the impact is not as dramatic as it is in reality. It also leaves policymakers with the impression that contextual societal changes (e.g. economic, demographic) may do little to impact on the spread of HIV. This lack of analysis results from both limited staffing and limited capacity in national surveillance programmes as well as other institutional and political factors. For example, HIV surveillance may be conducted by the ministry of health whereas behavioral surveys may be conducted by other non-governmental organizations. These non-governmental organizations may be funded and directed by external agencies. These two groups may not be eager or willing to share data or work together to do integrated analyses. It has also produced static surveillance systems in many countries with embedded quality problems, ongoing access limitations, and other issues that are not systematically addressed. This lack of staffing, capacity and institutional and political barriers also leaves little time to advocate for appropriate responses on the basis of the data. In addition, in low-level/concentrated epidemics, HIV levels outside of at-risk populations are too low to motivate many policymakers who are unwilling to invest the political capital necessary to deal with highly stigmatized populations such as IDU, MSM, and FSW. As a consequence, policymakers remain complacent, programmes focus on the wrong populations, and prevention coverage remains too low to stop the spread of the epidemic.

Back to Top | Article Outline

The way forward

In both generalized and low-level/concentrated epidemics there is considerable room for improving surveillance systems. Major improvements would result from the application of existing strategies for surveillance. In addition, new techniques and approaches should be incorporated into existing surveillance systems to address the problems posed by the continued evolution of epidemics and the expanding prevention and care programmes in many countries. This supplement elaborates on these new techniques and approaches for selected surveillance topics.

McDougal et al. (pp. S25–S30) describe new serological and nucleic acid tools for incidence measurement, offering an opportunity to detect and monitor new HIV infections directly, a valuable addition to existing epidemiological techniques. These new laboratory techniques may prove a valuable supplement to the traditional monitoring of risk behaviors. In the near future, such monitoring of behaviors and new infections will grow in importance, with the potential increase in risk behaviors associated with the wider introduction of antiretroviral medications.

Martin et al. (pp. S59–S65) address the critical issue of quality assurance in HIV serological testing for HIV surveillance. In the past, several countries have overestimated prevalence, because of absent or underperforming quality assurance programmes for surveillance.

Magnani et al. (pp. S67–S72) describe new respondent-driven sampling techniques for surveillance among hidden populations. These new sampling techniques may prove valuable for calibrating existing time–location and convenience surveillance samples to determine whether they are representative, and adequately capture the higher-risk components of at-risk populations. Although there are many issues in implementing these approaches, especially in developing country settings where they may ultimately be too costly and time consuming for routine surveillance, they provide another valuable window on the larger picture of risk in low-level/concentrated countries.

Calleja et al. (pp. S9–S17) consider nationally representative household surveys that have been recommended for the calibration of ANC-based surveillance in countries with generalized epidemics. The quality of past household surveys has varied considerably. This paper specifically addresses logistical issues in the conduct of these surveys that influence the quality of the surveys, and hence their value in calibrating ANC-based systems.

Hladik et al. (pp. S19–S24) take a critical look at the future of facility-based sentinel surveillance in generalized epidemics. With increased coverage of HIV testing services, data generated by HIV testing and counseling systems is an increasingly important source of information on the epidemic. In particular, HIV testing conducted as part of prevention of mother-to-child transmission programmes is expected to generate larger denominators than have traditionally been available from sentinel surveillance among pregnant women. It will be critically important to identify and address any bias inherent in this new source of data.

Zaba et al. (pp. S39–S52) address an issue at the heart of second-generation surveillance [1]: linking HIV seroprevalence to behavioral data. They describe advantages and limitations of linking at the individual, community, and national levels, both in low-level/concentrated and generalized epidemics. They examine the biases of the measurement of sexual behaviors, our ability to monitor sexual behaviors consistently over time, and the interpretation of sexual behavior trends in relation to the social context, individual and community behaviors and HIV viral dynamics.

Diaz et al. (pp. S31–S37) focus on the new challenges posed in many resource-constrained settings where access to treatment is increasing. Many of these countries had not invested in systems to measure and track HIV/AIDS morbidity, because these data were of little use in programming. With expanding treatment systems, there is a need for improving systems that monitor HIV/AIDS morbidity and the impact of antiretroviral drugs.

Pervilhac et al. (pp. S53–S58) address the final stage of the surveillance cycle – the use of surveillance data. They review recent innovative experiences, available tools, approaches in the use of HIV/AIDS surveillance data, and also organizational issues related to the link between surveillance and policy.

Unless these new tools are coupled with improved ‘traditional’ epidemiological and behavioral surveillance techniques and better analysis, they will be no more effective in moving responses than current approaches. Traditional systems still have serious quality issues; sampling is sometimes inconsistent, applications of the protocols in the field are spotty, and problems are sometimes seen in reporting data to the central level. Difficulties still exist in accessing the key populations driving the epidemic in many countries. Furthermore, analysis capacity is largely non-existent, meaning that the data are not being appropriately used, and the links to policy and programmes remain weak and insufficient. Addressing these existing problems, while adding new technologies, and addressing expanded monitoring and evaluation demands and expanded care and treatment programmes, presents serious challenges to national surveillance programmes.

Surveillance must become a political priority. Currently, surveillance has few real proponents, and risks becoming lost under more general headings of monitoring and evaluation and strategic information. The international community should stand ready to help national surveillance programmes through expanded technical support, capacity-building activities, and financial backing.

Back to Top | Article Outline

Acknowledgements

The authors would like to thank Sadhna Patel, with the CDC Global AIDS Program for her hard work in coordinating this conference. In addition, they would like to thank the Ministry of Health of Ethiopia and the CDC Global AIDS Program in Ethiopia for hosting the conference and the European Commission Second Generation Surveillance for their contribution to the meeting.

Back to Top | Article Outline

References

1. World Health Organization and Joint United Nations Programme on HIV/AIDS. Second generation surveillance of HIV. WHO/CDS/EDC/200.5, UNAIDS/00/03; 2000.
2. World Health Organization. Workshop on AIDS in Central Africa. WHO/CDS/AIDS 85.1; 1986.
3. De Cock KM, Colebunders R, Francis H, Nzilambi N, Laga M, Ryder AW, et al. Evaluation of the WHO clinical case definition for AIDS in rural Zaire. AIDS 1988; 12:219–221.
4. World Health Organization. Global AIDS surveillance – Part 1. Wkly Epi Rec (WHO) 1997; 72:359.
5. Centers for Disease Control and Prevention. 1993 Revised classification system for HIV infection and expanded surveillance case definition for AIDS among adolescents and adults. MMWR RR 1992; 41:1–17.
6. World Health Organization. WHO case definitions for AIDS surveillance in adults and adolescents. Wkly Epi Rec (WHO) 1994; 69:273–280.
7. Greenberg AE, Coulibaly IM, Kadio A, Coulibaly D, Kassim S, Sassan-Morokro M, et al. Impact of the 1994 expanded World Health Organization AIDS case definition on AIDS surveillance in university hospitals and tuberculosis centers in Cote d’Ivoire. AIDS 1997; 11:1867–1872.
8. Allen DM, Lee NC, Schulz SL, Pappaioanou M, Dondero TJ Jr, Onorato IM. Determining HIV seroprevalence among women in women's health clinics. Public Health Rep 1990; 105:130–134.
9. World Health Organization/Global Program on AIDS. Field guidelines for HIV sentinel surveillance: a manual for national AIDS control programmes. Geneva: WHO; 1989.
10. Walker N, Stanecki KA, Brown T, Stover J, Lazzari S, Garcia-Calleja JM, et al. Methods and procedures for estimating HIV/AIDS and its impact: the UNAIDS/WHO estimates for the end of 2001. AIDS 2003; 17:2215–2225.
11. Boerma JT, Ghys PD, Walker N. Estimates of HIV-1 prevalence from national population-based surveys as a new gold standard. Lancet 2003; 362:1929–1931.
12. Dobbs T, Kennedy S, Pau CP, McDougal JS, Parekh BS. Performance characteristics of the immunoglobulin G–capture BED–enzyme immunoassay, an assay to detect recent human immunodeficiency virus type 1 seroconversion. J Clin Microbiol 2004; 42:2623–2628.
13. De Cock KM, Porter A, Odehouri K, Barrere B, Moreau J, Diaby L, et al. Rapid emergence of AIDS in Abidjan, Ivory Coast. Lancet 1989; 2:408–411.
14. De Cock KM, Barrere B, Diaby L, Lafontaine F, Gnaore E, Porter A, et al. AIDS: the leading cause of adult death in the West African city of Abidjan, Cote d’Ivoire. Science 1990; 249:793–796.
15. Jones JL, Hanson DL, Chu SY, Fleming PL, Hu DJ, Ward JW. Surveillance of AIDS-defining conditions in the United States. Adult/Adolescent Spectrum of HIV Disease Project Group. AIDS 1994; 8:1489–1493.
16. Boerma JT, Gregson S, Nyamukapa C, Urassa M. Understanding the uneven spread of HIV within Africa: comparative study of biologic, behavioral, and contextual factors in rural populations in Tanzania and Zimbabwe. Sex Transm Dis 2003; 30:779–787.
17. Ferry B, Carael M, Buve A, Auvert B, Laourou M, Kanhonou L, et al. Comparison of key parameters of sexual behavior in four American urban populations with different levels of HIV infection. AIDS 2001; 15(suppl. 4):S41–S50.
18. Bessinger R, Akwara P, Halperin D. Sexual behaviors, HIV and fertility trends: a comparative analysis of sex countries; Phase I of the ABC study. Measure Evaluation/USAID; (www.cpc.unc.edu/measure/publications/special/special.html); August 2003.
19. Nnko S, Boerma JT, Urassa M, Mwaluko G, Zaba B. Secretive females or swaggering males? An assessment of the quality of sexual partnership reporting in rural Tanzania. Soc Sci Med 2004; 59:299–310.
20. Pisani E, Lazzari S, Walker N, Schwartlander B. HIV surveillance: a global perspective. J Acquir Immune Defic Syndr 2003; 32(suppl. 1):S3–S11.
21. Schwartlander B, Ghys PD, Pisani E, Kiessling S, Lazzari S, Carael M, Kaldor JM. HIV surveillance in hard-to-reach populations. AIDS 2001; 15(suppl. 3):S1–S3.
22. Pan American Health Organization. Working Group on AIDS Case Definition. Bull Pan Am Health Org 1989; 10:9–11.
23. Weniger BG, Quinhoes EP, Sereno AB, de Perez MA, Krebs JW, Ismael C, et al. A simplified surveillance case definition of AIDS derived from empirical clinical data. The Clinical AIDS Study Group and the Working Group on AIDS case definition. J Acquire Immune Defic Syndr 1992; 5:1212–1223.
24. Pan American Health Organization. HIV and AIDS in the Americas an epidemic with many faces. Washington, DC: Pan American Health Organization; 2001. ISBN 92 75 12360 8.
25. Pio Marins JR,de Fátima Jamal L, Chen S, S. Hudes E, Barbasa A, Berti de Azevedo Barras M, et al. Sobrevivência atual dos pacientes com aids no Brasil. Evidência dos resultados de um esforço nacional (Survival of AIDS patients in Brazil: Evidence of the results of a national effort). Boletim Epidemiológico Aids – Ano XV no. 02. Year XV, no. 02 Oct 2001–March 2002.
26. Casseb J, Pereira LC, Silva GL, Medeiros LA. Decreasing mortality and morbidity in adult AIDS patients from 1995 to 1997 in São Paulo, Brazil. AIDS Patient Care STDs 1999; 13:213–214.
27. Ainsworth M, Teokul W. Breaking the silence: setting realistic priorities for AIDS control in less-developed countries. Lancet 2000; 356:55–60.
28. Pisani E, Garnett GP, Grassly NC, et al. Back to basics in HIV prevention: focus on exposure. BMJ 2003; 326:1384–1387.
29. Walker N, Garcia-Calleja JM, Heaton L, Asamoah-Odei E, Poumerol G, Lazzari S, et al. Epidemiological analysis of the quality of HIV sero-surveillance in the world: how well do we track the epidemic? AIDS 2001; 15:1545–1554.
30. Brown T. Behavioral surveillance: current perspectives, and its role in catalyzing action. J Acquir Immune Defic Syndr 2003; 32(suppl. 1):S12–S17.
31. Pervilhac C, Zaniewski E, Garica-Calleja JM, Stanecki K, Walder N. Assessing the quality of HIV behavioral surveillance systems. In: XVth International AIDS Conference. Bangkok, Thailand, 2004 [Abstract MoPeC3621].
32. Saidel TJ, Des Jarlais D, Peerapatanapokin W, Dorabjee J, Singh S, Brown T. Potential impact of HIV among IDUs on heterosexual transmission in Asian settings: scenarios from the Asian Epidemic Model. Int J Drug Policy 2003; 14:63–74.
33. Cambodian Working Group on HIV/AIDS Projection. Projections for HIV/AIDS in Cambodia: 2000–2010. Phnom Penh: National Center for HIV/AIDS, Dermatology and STD; 2002.
34. Thai Working Group on HIV/AIDS in Thailand. 2000–2020. Bangkok: Division of AIDS, Ministry of Public Health; 2001.
35. Grassly NC, Lowndes C, Rhodes T, Judd A, Renton A, Garnett GP. Modeling emerging HIV epidemics: the role of injecting drug use and sexual transmission in the Russian Federation, China and India. Int J Drug Policy 2003; 14:25–43.
36. Phoolcharoen W, Ungchusak K, Sittitrai W, Brown T. Thailand: lessons from a strong national response to HIV/AIDS. AIDS 1998; 12(suppl. B):S123–S135.
37. Phalla T, Leng HB, Mills S, Bennett A, Weinrawee P, Gorbach P, Chin J. HIV and STD epidemiology, risk behaviors, and prevention and care response in Cambodia. AIDS 1998; 12(suppl. B):S11–S18.
38. Stoneburner RL, Low-Beer D. Population-level HIV declines and behavioral risk avoidance in Uganda. Science 2004; 304:714–718.
39. Monitoring the Pandemic Network. AIDS in Asia: face the facts. MAP report; ISBN 974-92253-4-1. 2004; 22. Monitoring the AIDS Pandenic (MAP) Network. www.mapnetwork.org/reports/aids_in_aisa.html
40. Ghys PD, Saidel T, Hoang TV, Savtchenko I, Erasilova I, Mashologu YS, et al. Growing in silence: selected regions and countries with expanding HIV/AIDS epidemics. AIDS 2004; 18:1–6.
41. Joint United Nations Programme on HIV/AIDS, (UNAIDS). Report on the global HIV/AIDS epidemic. Geneva: UNAIDS; 2004. pp. 5–6.
42. Boerma TJ, Weir SS. Integrating demographic and epidemiological approaches to research on HIV/AIDS. The proximate-determinants framework. J Infect Dis 2005; 191 (suppl. 1): 561–567.
Keywords:

AIDS; HIV; surveillance

© 2005 Lippincott Williams & Wilkins, Inc.