Mapping HIV/STI behavioural surveillance in Europe

The epidemics attributable to the Human Immunodeficiency Virus (HIV) and to other sexually transmitted infections (STI) remain a significant public health problem in Europe. The incidence of HIV infections is not decreasing, nor is it even stabilized in many populations, such as men who have sex with men (MSM), injecting drug users (IDU), or certain migrant populations, such as persons from Sub-Saharan Africa [1, 2]; furthermore, the reported incidence of several STI is increasing [3‐8].

Since the beginning of the HIV/AIDS epidemic, the importance of monitoring risk behaviour in various target populations has been recognized and different indicators have been proposed [9]. In the late 1980's, several European countries began to collect indicators of behaviour in various populations exposed to risk [10‐13], such as MSM [14‐17], IDU [18‐20], the general population [21, 22], youth [23] and sex workers (SW) [24, 25]; the concept of behavioural surveillance (BS) progressively emerged [26, 27].

The utility of such behavioural surveillance data in guiding programme development, especially in conjunction with biological surveillance data, became apparent during the 1990's. The Joint United Nations Programme on HIV/AIDS (UNAIDS)/World Health Organization (WHO) has defined second generation surveillance (SGS) [28, 29] as the «regular, systematic collection, analysis and interpretation of information for use in tracking and describing changes in the HIV/AIDS epidemic over time». This requires procedures to ensure a combined analysis of both behavioural surveillance (BS) data and data obtained from biological surveillance of HIV/AIDS and STI. BS allows for the monitoring of risks related to transmission of HIV and STI at the population level and provides a key source of information not only to understand the drivers of epidemics, but also for advocacy and for the planning and evaluation of prevention interventions [30, 31]. The type of BS to be conducted, in particular the populations to be included in the surveillance, depends of the type of HIV epidemic (generalised, concentrated, low level) [32]. Most European countries have a concentrated type of epidemic but some have a low-level type.

At the global level, behavioural indicators in various populations are included in the set of indicators defined to monitor commitment to the United Nations General Assembly Special Session on AIDS (UNGASS indicators) [33]. At the European level, the European Monitoring Centre for Drugs and Drug Addiction (EMCDDA) has proposed a set of core behavioural indicators for use with IDU[34].

The formalisation and organisation of BS in Europe is in progress [31, 35], but without any review of current practices having been conducted. There has been little consideration of behavioural surveillance systems in the scientific literature [36] and discussion on BS has mainly appeared only in journal supplements [37‐40]. In 2008, the European Centre for Disease Prevention and Control (ECDC) therefore commissioned an international team of experts to produce an in-depth analysis of BS related to HIV and STI in Europe and to develop a framework for the implementation of a key set of behavioural indicators. The specific objectives of the study were to provide an inventory of existing national BS systems and their functioning in Europe, including identification of the populations included, and the data collection methods and indicators used in each system. This article presents the resultant mapping of BS in Europe, based on a survey of EU/EFTA countries [41]. Surveillance is examined at the system level and in the following populations: general population, youth, IDU, MSM, persons living with HIV/AIDS (PLWHA), SW, STI clinic clients, migrants and ethnic minorities.

An international team of researchers was established with expertise in behavioural surveys in each of the 8 designated populations. The survey of all EU/EFTA countries was conducted from June to October 2008. Information was gathered on: the overall surveillance system, the specific populations surveyed and the topics covered and main indicators in use for each population. The data collection instrument consisted of 9 separate questionnaires sent to one identified contact person in each country. One questionnaire concerned the national behavioural and second generation surveillance system(s) as a whole, and there were an additional eight questionnaires covering each of the eight specific population groups. The questionnaires, provided in word format using dropdown lists, included both open and closed questions (see Additional files 1 and 2). In the system-related questionnaire, information was requested on:

the existence and definition of a national or regional BS system(s); indicating which populations were included and the national key behavioural indicators;

where in place, the functioning of the second generation system at national or regional level, including level of formalisation and documentation, interlinking of systems, responsibility, financing, analysis, dissemination and use of results, and sustainability.

In the questionnaires on specific populations, it was first asked whether a surveillance system was in place for this population and information was then requested on the existence of behavioural surveys (or other types of data collection used for surveillance) conducted since 1985. For each relevant survey (or other type of data collection) indicated, information was requested on:

methodology used, including year, sampling method, data collection method, and main topics covered;

main indicators followed;

related publications.

Twenty-eight of the 31 countries surveyed (90%) returned the questionnaire set; the remaining 3 countries did not respond within the prescribed deadline (Bulgaria, Romania, and Portugal; Bulgaria later sent a summary of its surveillance system in place since 2004, but these data could not be included in the current analysis). Most questionnaires were fully completed.

Movement towards simplification of data collection is noticeable in various populations, as exemplified by the increased use of internet-based surveys in MSM, service-based data collection in IDU, and in the use of general health surveys to collect data on sexual behaviour and drug use in the general population and in youth. These newer approaches offer advantages in terms of sustainability and cost provided that comparison over time can be maintained.

In MSM and IDU, where there is the longest experience of BS, the length and robustness of trends and the relatively low cost of non-probablistic methods of sampling (venue-, internet- or service-based) can explain the paucity of experience with RDS and the probable reluctance to adopt this recently promoted method for BS in these populations [43]. Favouring less costly data collection methods over RDS would seem a reasonable choice in the case of mature epidemics.

It is not easy to evaluate the adequacy of BS in European countries: in many cases the identification of populations with risk behaviour and the selection of populations to be included in a BS system has not been formally defined, or is incomplete. Several gaps may nevertheless be identified:

Some countries still have no data on MSM or IDU. Very few countries have collected data on sex workers and migrants, two populations difficult to reach and/or to identify reliably, and practically none has BS installed on a long-term basis. Few BS systems are in place in populations that are clearly identified as potentially at-risk of acquiring or transmitting infections and accessible through the health system, such as STI patients and PLWHA.

In several countries, the formalisation and functioning of the system and its use are less developed than might be expected, given the opportunities.

There are many challenges regarding the future of BS in Europe:

One of the challenges in the years to come will be to reinforce or maintain BS. Europe, particularly Western Europe, faces a decrease in the level of priority accorded to HIV (for example, in reaction to potential new epidemics, such as AH1N1), with a consequent decrease in the availability of funding for HIV prevention and research, including behavioural surveys. In certain Eastern European countries, international funding provided to initiate surveillance is decreasing. The necessity of maintaining BS may be challenged. Yet, there are many reasons to maintain BS in the Europe:

As mentioned above, the epidemiological situation regarding HIV and other STI is worsening in many countries. In addition, there is a continuous increase in the pool of HIV infected people.

A decrease in the use of protection and the development of alternative HIV/STI-related patterns of behaviour (risk reduction, stopping condom use under certain treatment conditions [44]) has begun to appear and deserves close monitoring.

There are concerns regarding access to prevention and care, especially for marginalised populations, such as certain categories of migrants [45] and IDU [46]; BS could be used for planning interventions and advocating for resources.

The sustainability of existing systems is not always guaranteed.

The cost of BS is one element threatening sustainability. Attempts have been made to reduce costs, while maintaining the availability of valid and reliable behavioural trends, and may serve as examples. These include: use of internet (MSM) and service-based surveillance (wherever possible); increased collaboration with existing PLWHA cohorts; insertion of sexuality modules in general health surveys replacing KAPB surveys; use of proxy data, such as information on SW obtained from clients in general population surveys.

The frequency of data collection should also be reconsidered. In several populations, such as the general population and youth, intervals between data collection may extend to 4-6 years, while in populations with higher HIV incidence, shorter intervals and more flexibility are needed. Qualitative studies may also be used to provide information on developments in specific populations and to facilitate decisions regarding the necessity of conducting surveys.

Our study has some limitations. The information delivered by the questionnaire was self-reported and the responses varied greatly as regards the level of detail included in the description of systems. The person to whom the questionnaire set was sent in each country was the ECDC contact person at national level for HIV biological surveillance; there is currently no specific ECDC contact person for behavioural surveillance. This person may have been unaware of the existence of surveys in some or all populations of interest, whether organised or not in a behavioural surveillance system, and may have given inadequate answers or may not have passed the population-specific questionnaires to the relevant national specialists. This would result in underreporting of existing surveys. However, it is very unlikely that any organised national behavioural surveillance would be unknown to the ECDC contact person. Furthermore, this survey was launched with a letter from the ECDC director, thus indicating a high level of importance. The report was also circulated to countries for validation, offering a further occasion for verification.

Few original questionnaires used for surveillance were sent by the countries. Nevertheless, most countries sent useful complementary information, such as references to documents or publications describing their systems.

The main challenge now faced is to build and maintain organised and functional behavioural and second generation surveillance systems at national level, with the following characteristics. First, there should be a regular assessment of the appropriateness and quality of the system. Second, a designated body should be set up, with clear responsibility for data coordination, collection, analysis and dissemination. This body should be qualified to make decisions regarding the simplification or extension of the system, on the basis of epidemiological and behavioural trends, cost, sustainability, quality and robustness of the system.