Content area
The aim of this review was to develop an overview of the sampling methods used in developed countries for behavioural surveillance among men who have sex with men. We abstracted information from peer-reviewed and non-peer-reviewed publications and, when needed, contacted first authors and surveillance authorities. Out of the 40 developed countries, 26 surveillance systems were identified in 23 countries. These surveillance systems made use of one or a combination of: venue-based sampling (n = 16); web-based sampling (n = 11); gay press/mail-out (n = 8); or respondent-driven sampling (n = 4). Differences in the sampling methods used were found by year of implementation and by region. More information is needed to better assess the strengths and limitations of the range of sampling methods. There was substantial variation in the way sampling methods were applied, most especially in venue-based sampling. To improve the ability to compare indicators across countries, key elements of the sampling methods should be standardized.[PUBLICATION ABSTRACT]
AIDS Behav (2010) 14:12521264 DOI 10.1007/s10461-010-9743-7
REVIEW PAPER
Sampling Methods Used in Developed Countries for Behavioural Surveillance Among Men who have Sex with Men
Dana Paquette John De Wit
Published online: 8 July 2010 Springer Science+Business Media, LLC 2010
Abstract The aim of this review was to develop an overview of the sampling methods used in developed countries for behavioural surveillance among men who have sex with men. We abstracted information from peer-reviewed and non-peer-reviewed publications and, when needed, contacted rst authors and surveillance authorities. Out of the 40 developed countries, 26 surveillance systems were identied in 23 countries. These surveillance systems made use of one or a combination of: venue-based sampling (n = 16); web-based sampling (n = 11); gay press/ mail-out (n = 8); or respondent-driven sampling (n = 4). Differences in the sampling methods used were found by year of implementation and by region. More information is needed to better assess the strengths and limitations of the range of sampling methods. There was substantial variation in the way sampling methods were applied, most especially in venue-based sampling. To improve the ability to compare indicators across countries, key elements of the sampling methods should be standardized.
Keywords HIV/AIDS Men who have sex with men
Sampling Behavioural surveillance
Biological and behavioural surveillance
Introduction
Many developed countries report a re-emerging epidemic of HIV among men who have sex with men (MSM). Increases in new diagnoses and in incidence have been noted in several countries [1]. The information produced by an effective HIV surveillance system is crucial to being able to respond to this rise in HIV infections with appropriate policy and interventions [2].
In the late 1990s, the UNAIDS/WHO Working Group of Global HIV/AIDS and STI Surveillance recommended the implementation of second generation surveillance. Second generation surveillance involves tailoring the surveillance system to the state of the epidemic and using multiple sources of data to better predict and track the course of the HIV epidemic [3]. A key component of second generation surveillance is the implementation and use of behavioural surveillance systems to gain insights into the extent to which behavioural changes are responsible for HIV trends [4, 5].
In developed countries, where the epidemic is generally concentrated or low-level, behavioural surveillance is recommended among population groups who are most at risk for HIV infection [3]. These groups, such as injecting drug users (IDU), MSM and commercial sex workers are often termed hard-to-reach or hidden populations because the behaviours they engage in are socially marginalized, stigmatized, illicit or illegal, making access and recruitment for behavioural surveys challenging tasks. As well, no sampling frame exists for these groups, which means that traditional sampling strategies such as random sampling or cluster sampling are often not feasible or cost-effective [6].
Several approaches to sampling hard-to-reach populations have been developed, and the main strategies include venue-based sampling, targeted sampling, respondent-driven sampling (RDS) and web-based sampling. Proponents
D. Paquette (&) J. De Wit
National Centre in HIV Social Research, Universityof New South Wales, Webster Building, UNSW 2052, Sydney, Australiae-mail: [email protected]
J. De WitDepartment of Social and Organizational Psychology, Utrecht University, Utrecht, The Netherlands
123
AIDS Behav (2010) 14:12521264 1253
of each of these sampling methods have touted their unique advantages, such as the ability of RDS to derive population estimates [6, 7], or the ability of web-based sampling to achieve large samples quickly [8, 9]. However, it remains unclear which strategy is best suited to sampling MSM in the context of behavioural surveillance in developed countries. As well, concerns have been raised regarding the variation in the sampling methods used for behavioural surveillance and the impact on the comparability of indicators across countries [10].
The aim of this review was to develop an overview of the sampling methods currently used in behavioural surveillance of MSM in developed countries. As well, we wished to describe patterns in the type of sampling methods used over time and by geographic region, and describe the challenges and benets associated with the use of various sampling methods.
Methods
Search Strategy
From November 2008 till April 2009, we searched for published manuscripts through Medline (19802008), Scopus and Google Scholar, using multiple combinations of keywords, which included: HIV, surveillance, behavioural surveillance, second generation surveil-lance; in combination with gay, homosexual, bisexual, men who have sex with men; and the names of each of the countries of interest and major/capital cities of the countries of interest. Published manuscripts were also found through reference lists of relevant publications. As well, we searched for non-peer-reviewed publications through websites of HIV surveillance authorities, Family Health International reports, EuroHIV surveillance reports and UNGASS Country Progress Reports.
Eligibility Criteria
We limited our search to countries in developed regions, as dened by the United Nations, Department of Economic and Social Affairs, Population Division [11], with a population of greater than 1,000,000.
Where it existed, this review sought to describe the countrys national behavioural surveillance system, implemented and maintained by, or on behalf of, the countrys public health and surveillance authorities. In countries where a national behavioural surveillance system for MSM did not exist, repeated cross-sectional surveys implemented in one or more cities of a country were described when present. These surveys were usually developed and maintained by community-based research organizations,
academic researchers or international agencies. We excluded repeated cross-sectional surveys of which the main purpose was to evaluate a specic prevention program or intervention. We also excluded research studies that were conducted on a one-time basis. If only one survey was conducted, but the report indicated that future rounds were planned and/or the survey was explicitly described as behavioural surveillance, it was included in the review.
The target population of the surveillance systems had to be MSM, and not other at-risk populations of which MSM may form a subgroup (e.g., sex workers, IDU).
Information Abstracted from Publications
From the articles and reports we abstracted the following information: sampling method; cities where surveys were conducted; the year behavioural surveillance was started; the intervals between survey rounds; eligibility criteria for surveillance participants; achieved sample size; sampling duration; whether biological samples were collected; whether participants were reimbursed, and if yes, by how much; any comments on the representativeness of the sample and the ease of operation. When the information in publications was incomplete, we contacted rst authors and/or the countrys surveillance authorities and asked them to provide further information and any other publications (peer-reviewed and non-peer-reviewed) that described their behavioural surveillance systems.
We also planned to abstract information on any discussed strengths and limitations of the sampling methods and the way they were used in a specic surveillance system. However, beyond issues related to the representativeness of the resulting samples, very little comments were found regarding strengths and limitations of the sampling methods. We abstracted what was available, but did not analyse because this information was provided for only a small number of behavioural surveillance systems and sampling methods.
Results
Behavioural Surveillance Systems for MSM in Developed Countries
Table 1 summarizes the information we found regarding the existence of a HIV behavioural surveillance system of MSM, by country. Of the 40 countries included in his review, 23 had behavioural surveillance systems that were described in sufcient detail to be included in this review; six countries had behavioural surveillance systems, but were not included in this review since few details of the methods used could be found. Eight countries explicitly reported not having a behavioural surveillance system for MSM. For the
123
1254 AIDS Behav (2010) 14:12521264
Table 1 Behavioural surveillance and data sources by country
Countries Presence of HIV behavioural surveillance among MSM within country
Information source type Published information sources
Oceania
Australia Yes Journal articles, reports [1217]
New Zealand Yes Reports, website [1821]
Asia
Japan Noa Journal articles [22, 23]
Northern America
Canada Yes Conference abstract, website [24, 25]
USA Yes Journal article [2628]
Europe
Western Europe
Austria Noa Personal communication (Mar 10, 2009 e-mail from J. P. Klein to DP; unreferenced)
Belgium Yes Journal article, report, personal communication
(Feb 19, 2009 e-mail from V. Martens to DP; unreferenced) [29, 30]
Denmark Yes Journal article, personal communication (Jan 21, 2009 e-mail from S. A. Cowan to DP; unreferenced) [31]
Finland Noa Personal communication (Mar 11, 2009 e-mail from M. Salminen to DP; unreferenced)
France Yes Journal articles, reports, website [3236]
Germany Yes Journal article, personal communication (Jan 29, 2009 e-mail from U. Marcus to DP; unreferenced) [37, 38]
Greece Noa Report [39]
Ireland Yes Report [40]
Italy Yes Conference abstract, personal communication
(Nov 29, 2009 e-mail from M. Mirandola to DP; unreferenced) [41]
Netherlands Yes Factsheets, personal communication (Feb 26, 2009 e-mail from E. L. M. Op de Coul to DP; unreferenced. Mar 4, 2009 e-mail fromE. Roos to DP; unreferenced) [42]
Norway Yes, but insufcient information
Journal article [43]
Portugal None reportedb
Spain Yes Report, conference abstract, personal communication
(Jan 23, 2009 e-mail from C. Folch to DP; unreferenced) [44, 45]
Sweden Yes Report [46]
Switzerland Yes Journal articles, reports, personal communication
(Feb 4, 2009 e-mail from H. Balthasar to DP; unreferenced) [4750]
UK Yes Journal articles, reports, personal communication
(Jan 22, 2009 e-mail from D. Mercey to DP; unreferenced) [5158]
Central Europe
Albania Yes Report [59]
Bosnia and
Herzegovina
None reportedb
Bulgaria Yes, but insufcient information
Report [60]
Croatia Yes Journal article, personal communication (Jan 22, 2009 e-mail from I. Bozicevic to DP; unreferenced) [61]
Czech Republic Noa Personal communication (Jan 30, 2009 e-mail from I. Prochazka to DP; unreferenced)
Hungary Noa Personal communication (Mar 11, 2009 e-mail from M. Dudas to DP; unreferenced)
123
AIDS Behav (2010) 14:12521264 1255
Table 1 continued
Countries Presence of HIV behavioural surveillance among MSM within country
Information source type Published information sources
Macedonia,
F.Y.R.
Yes, but insufcient information
Report [62]
Personal communication (Mar 11, 2009 e-mail from K. Gajewska to DP; unreferenced)
Romania Noa Personal communication (Jan 12, 2009 e-mail from T. Kovacs to DP; unreferenced)
Serbia Yes Website [63]
Slovakia None reportedb
Slovenia Yes Journal article, personal communication
(Mar 3, 2009 e-mail from I. Klavs to DP; unreferenced) [64]
Poland Yes, but insufcient information
Eastern Europe
Belarus Yes, but insufcient information
Report [65]
Estonia Yes Report [66]
Latvia Noa Report [67] Lithuania Yes Conference abstract, report,personal communication
(Feb 11, 2009 e-mail from S. Caplinskas to DP; unreferenced) [68, 69]
Republic of Moldova
Yes, but insufcient information
Report [70]
Yes Reports [7173]
Ukraine Yes Journal article, report [74, 75]
a Indicates that the country does not have a behavioural surveillance system, as stated in documentation or in personal communication
b None reported: Indicates no information was found to indicate the presence or absence of a behavioural surveillance system
Russian
Federation
remaining three countries, no published information or local informants were found to clarify the presence or absence of a behavioural surveillance system among MSM.
Sampling Methods
The methods used in countries behavioural surveillance systems among MSM were categorized into four different types: venue-based sampling, web-based sampling, sampling by gay press/mail-out and RDS.
Venue-based sampling was dened as any sampling method where recruitment took place at one or more types of physical venues, including social venues (e.g., bars, clubs, gyms), events (e.g., gay pride events, gay lm festivals), sex on premises venues (e.g., saunas, bathhouses), public cruising areas (e.g., beats, public parks, bathrooms), and social groups and organizations. Sampling methods in which online recruitment efforts directed men to a questionnaire posted online were categorized as web-based sampling. The category gay press/mail-out included all surveillance systems in which questionnaires were
distributed through the gay press, and/or through agencies and organizations. In all gay press/mail-out surveys, the questionnaires were returned to a central organization by mail. Surveillance systems that were described as using RDS were categorized as RDS. However, these systems may not have adhered to all criteria that have been set out previously to qualify as RDS [76].
As shown in Table 2, 26 surveillance systems were identied within the 23 countries that reported having a HIV behavioural surveillance system among MSM (multiple surveillance systems were found in two of the countries). Within these 26 surveillance systems, 9 made use of multiple sampling methods and 17 made use of a single method. The sampling methods used were: venue-based sampling (n = 16); web-based sampling (n = 11); gay press/mail-out (n = 8); or RDS (n = 4).
In eight of the 26 surveillance systems, changes were made to the sampling method used. In six of these eight surveillance systems, a web-based sampling component was added to existing sampling methods and in two systems the sampling method was changed to web-based only.
123
1256 AIDS Behav (2010) 14:12521264
Eligibilitycriteria
Sexwithanothermaninprevious
5years
Residentofcityorparticipates
regularlyincitysgaycommunity
Sexwithanothermaninprevious
5years
National(web-basedcomponent)Web-basedsamplingMenwhohadrecentlyparticipatedin
GAPSSwereineligibleforGOSS
Sexwithanothermanever
Minimumageof1618years,
dependingonprovincial/localethics
boardsandrequirements
Menortransgenderpeoplebornmale
Minimumageof18years
Residentsinalocallydened
geographicareaofinterest
Firsttimeparticipants
Gayorbisexualman
Web-basedsamplingLiveintheFrenchspeaking
communityofBelgium
Mail-out
Sexwithanothermaninpreviousyear
Notreported
200523yearsVenue-based
sampling
Currentsampling
method
Venue-based
sampling
Web-basedsampling
Gaypress
Venue-based
sampling
Intervalbetween
surveys
(dependingoncity)
19966months3years
2002yearsVenue-based
sampling
20033yearsVenue-based
sampling
25MSAsinfuturewavesAbletospeakEnglishorSpanish
199356yearsVenue-based
sampling
200014yearsVenue-based
sampling
(dependingon
Yearofrst
survey
DenmarkSexLifeSurveyCopenhagenandAarhus(venue-
basedsampling)
Sydney,Melbourne,Adelaide,
Canberra,Perthandregionalsites
inQueensland
Auckland(venue-based
component)
Winnipeg,Vancouver,Victoria
17metropolitanstatisticalareas
(MSAs)inrstwave:Atlanta,
GA;Baltimore,MD;Boston,
MA;Chicago,IL;Dallas,TX;
Denver,CO;FtLauderdale,FL;
NewarkNJ;NewYorkCity,NY;
Philadelphia,PA;SanDiego,CA;
SanFrancisco,CA;SanJuan,PR;
andWashington,DC
BarometreNet-GayWeb-basedsampling
EnquetePresseGayGayPress
GermanyGayMenandAIDS(GMA)National198724yearsWeb-basedsamplingMenwhohavesexwithmen
GayPress
samplingmethod)
Regional(Frenchspeaking
communityofBelgium)
National(Gaypressandmail-out,
web-basedsampling)
Table2Descriptionofbehaviouralsurveillancesystembycountry
CountryNameofsurveillancesystemCities/areaswheresurveysare
conducted
FranceBarometreGayNational198513years
NewZealandGayAucklandperiodicsexsurvey
(GAPSS)andGaymensonline
sexsurvey(GOSS)
USANationalHIVbehavioral
surveillanceformenwhohave
sexwithmen(NHBS-MSM)
NorthernAmerica
CanadaM-TrackMontreal,Ottawa,Toronto,
Oceania
AustraliaGaycommunityperiodicsurveys
(GCPS)
WesternEurope
BelgiumLesmodesdeviedesgaysetlesida
(GaymenlifestylesandAIDS)
123
AIDS Behav (2010) 14:12521264 1257
Menwhohadoraloranalsexwith
anothermaninpreviousyear
NetherlandsSchorerMonitorNational20001yearWeb-basedsamplingMenwhohavesexwithmen
SpainPartoftheIntegratedHIV/STI
IrelandAll-IrelandGayMensSexSurveyNational20031yearWeb-basedsamplingSexwithanothermaninpreviousyear
Mail-outResidentofIreland
ItalySIALON(Verona)Verona20082yearsVenue-based
sampling
CroatiaHIVsurveillancesysteminCroatiaZagreb20062yearsRDSAnalsexwithanothermaninprevious
year
Minimumageof18years
Eligibilitycriteria
Sexwithanotherman
Menwhohavesexwithmen
Web-basedsamplingResidentofSwitzerland
GayPress
UKLondonGymSurveyLondon19981yearVenue-based
sampling
Sexwithanothermaninpreviousyear
Notreported
Currentsampling
method
2006NotreportedRDSOral,analorpettingwithanother
manprevious6months
Ages1549years
Residentoftargetedtownsforprevious
6monthsormore
Notreported
Catalonia19932yearsVenue-based
sampling
UKGayMensSexSurveyNational19931yearWeb-basedsamplingSexwithanothermaninpreviousyear
Mail-outResidentofUnitedKingdom
Intervalbetween
surveys
Tirana2005NotreportedRDSSexwithanothermaninpreviousyear
199613years
Yearofrst
survey
Ljubljana20001yearVenue-based
sampling
National200412yearsWeb-basedsamplingNotreported
Mail-out
SwedenTheMSMSurvey2006National2006NotreportedWeb-basedsurveyNotreported
SwitzerlandGaySurveyNational198723yearsVenue-based
sampling
Venue-based
sampling
(dependingoncity)
London,Manchester,Brighton,
Glasgow,Edinburgh
SurveillanceSystemofCatelonia
(SIVES)
UKGayMensSexualHealthSurvey
(GMSHS)
SurveillanceStudy(Bio-BSS)
SerbiaNotreportedBelgrade,NoviSad,Nis,
Kragujevac,Sabac
SloveniaUnlinkedanonymousmonitoringof
HIVamongMSM
Table2continued
CountryNameofsurveillancesystemCities/areaswheresurveysare
conducted
CentralEurope
AlbaniaBehavioralandBiological
EasternEurope
EstoniaHIV/AIDS-relatedknowledgeand
behaviouramongMSMvisiting
gaywebsites
123
1258 AIDS Behav (2010) 14:12521264
2007NotreportedRDSOral/analsexortouchedanothermans
genitalsinprevious6months
Variations in Sampling Method
As seen in Table 3, there are a number of differences in the characteristics of sampling methods used for HIV behavioural surveillance in MSM. Not surprisingly, the median sample size was largest for web-based surveys and smallest for surveys using RDS. Web-based surveys also had the shortest median sampling duration at 4 weeks, followed by venue-based sampling at 9 weeks, and gay press/mail-out and RDS at 12 weeks. The gay press/mail-out sampling method tended to be the method implemented during the late 1980s and throughout the 1990s (median year of implementation = 1993), followed by venue-based (median year of implementation = 2000), which continues to be used for more recently implemented surveillance systems. Web-based surveillance and RDS had a median year of implementation of 2005 and 2006, respectively. With regards to region, venue-based sampling was used throughout all regions, while web-based and gay press/ mail-out were used almost exclusively by Western European countries and RDS by Central European countries.
Differences were also observed within sampling methods. Notably, aspects of venue-based sampling varied greatly, with much less variation in how web-based sampling, gay press/mail-out sampling and RDS surveys were conducted. Out of the 16 venue-based surveys, four compensated men for their participation in the survey, eight of these surveys included HIV testing and six were used in combination with other sampling methods.
Other variations were also found in how venue-based sampling was applied, but are not shown in the table. For example, all the surveys in which the type of venue was reported (n = 14) recruited men from social venues (i.e., bars, clubs or gyms). However, 11 of the 14 also included one or more other types of venues, ranging from sex on premises venues, events, sexual health clinics, public cruising areas, community group meetings to sex shops. These surveys also varied in terms of where the questionnaires were completed and how they were returned. Six of the surveys asked men to complete the questionnaire in the venue where they were recruited; in ve of the surveys, men were asked to take the questionnaire away with them and return by mail or to the venue at a later date; two surveys provided the option of completing the survey in the venue or at a later follow-up time and place. This information was not provided for the remaining three surveys.
For the venue-based sampling surveys, differences were also found in whether the times and venues when/where men were recruited were randomly selected. Of the 14 surveys for which this information was provided, ve surveys applied a more systematic approach (e.g., venue-based application of time space sampling) and nine surveys made use of convenience venue-based sampling. For
Sexwithanothermaninpreviousyear
Ages1560years
Eligibilitycriteria
Homosexualorientation
Minimumageof18years
Menwhohavesexwithothermen
Ages15-49years
MoscowandStPetersburg2006NotreportedVenue-based
sampling
St.Petersburg2005NotreportedVenue-based
sampling
Currentsampling
method
Intervalbetween
surveys
Yearofrst
survey
Vilnius,Kaunas,Klaipeda20061yearVenue-based
sampling
Simferopol,Yalta,Dnipropetrovsk,
KriviyRig,Donezk,Ivano-
Frankivsk,Kyiv,Lugansk,
Mykolaiv,Odessa,Kherson,
Cherkasy
LithuaniaSexualbehaviourandHIV
prevalenceamongLithuanian
MSM
SentinelsurveysofHIVprevalence
andbehavioralrisksamongmen
havingsexwithmeninMoscow
andSt.Petersburg
2005Behavioralmonitoringsurvey
(FHI)
UkraineMonitoringbehaviourofmen
havingsexwithmenasa
componentofsecondgeneration
epidemiologicalsurveillance
Table2continued
CountryNameofsurveillancesystemCities/areaswheresurveysare
conducted
Russian
Federation
Russian
Federation
123
AIDS Behav (2010) 14:12521264 1259
Table 3 Details of sampling methods
a Sample size for most recent year or cycle of available data. Where data were collected in multiple cities, the sample sizes were totalled for all sites
b The opportunity to enter into a draw was not considered compensation. Small gifts, gift cards and money were considered compensation
Venue based n = 16
Web-based n = 11
Gay press/mail-out n = 8
RDS n = 4
Sample sizea
Range 7510,030 36115,058 2364,749 1991,764
Median 692 2,024 1,488 386
Sampling duration (weeks)
Range 0.1436 412 1216 1014
Median 9 4 12 12
Year initiated
Range 19922008 20002006 19852003 20052007
Median 2000 2005 1993 2006
Region
Oceania 2 1 0 0 Northern America 2 0 0 0
Western Europe 8 9 8 0
Central Europe 1 0 0 3
Eastern Europe 3 1 0 1
Sample characteristics
Participants compensatedb 4 0 0 4
Included HIV testing 8 0 0 3
Combined with other methods 6 8 8 0
the convenience venue-based surveys, the year of implementation ranged from 1993 to 2005 with a median year of 2000, while for those making use of a more systematic approach, the year of implementation tended to be more recent and ranged from 1996 to 2008, with a median of 2005. There was also a slight difference in the range of the duration (weeks) of sampling between convenience venue-based samples and those using a more systematic approach. For convenience samples, the number of weeks for data collection ranged from 0.14 to 18, with a median of 9 weeks, while for venue-based sampling using more systematic approaches, the number of weeks of data collection ranged from 4 to 36 weeks, with a median of 9 weeks.
In contrast, there seemed to be less variation in how the other sampling methods were conducted. None of the web-based surveys or gay press/mail-out surveys compensated participants or included HIV testing, and they were most often used in combination with surveys using other sampling methods. There was also little variation in the way RDS was conducted, with all participants being compensated, most included HIV testing and none of them being combined with surveys using other methods. However, while there may have been other variations in the way the above sampling methods were implemented, unlike for venue-based surveys, further information on these sampling methods was not provided in sufcient detail. Examples of such potential differences include the ways in which the web-based surveys were promoted (e.g., broadcast e-mails and/or banners that link to the questionnaire);
the number and types of publications used for gay press surveys; and the amount of formative research that was completed for RDS surveys.
Representativeness and Ease of Operation of Surveillance Systems
The most commonly noted concern regarding representativeness across all sampling methods was that the results were not generalizable to all MSM. This was a concern raised for 12 of the 16 venue-based surveys, 9 of the 11 web-based surveys, 6 of the 8 gay press/mail-out surveys and 1 out of the 4 RDS surveys. Comments were made that the results could only be generalized to men who attended venues, who accessed the Internet, read the gay press or who lived in a specic city. Response bias was an issue that was raised for four of the venue-based surveys, four of the gay press/mail-out surveys, one of the web-based surveys and one of the RDS surveys. As well, four of the venue-based surveys raised concerns about changes in the characteristics of individuals captured over time. This concern was not voiced for any of the other sampling methods.
Information on ease of operation was provided for only eight of the surveillance systems. Comments regarding ease of operation were provided for the surveillance system as a whole, and not by separate sampling methods. The surveillance systems for which this information was provided were: four with venue-based only surveys, two with venue-based, web-based and gay press/mail-out surveys,
123
1260 AIDS Behav (2010) 14:12521264
one with web-based and gay press surveys, and one using RDS.
The specic issues that were discussed encompassed the need to have good knowledge of the gay community and support of the community (n = 4; 2 venue-based only surveys, 1 using RDS and 1 with venue-based, web-based, gay press surveys); the importance of hiring appropriate staff and the need for staff training (n = 3; 2 venue-based only surveys and 1 using RDS); that the ongoing operation of the surveillance system had become relatively easy due to a number of years experience (n = 2; 1 venue-based only survey and 1 with web-based and gay press surveys); the need to conduct formative research throughout to ensure new venues are captured in the sampling frame (1 venue-based only survey); that the use of the Internet as a sampling method had facilitated data collection and analysis (1 with venue-based, web-based, mail-out surveys); and that the method used was cost-efcient (1 venue-based only survey).
Discussion
This review represents the rst published overview of the sampling methods used across all developed countries to reach MSM in behavioural surveillance surveys. The methods used in 23 countries are described, and variations over time, across regions and within the sampling methods themselves are highlighted.
Several patterns emerged in the sampling methods that were used. The surveillance systems that were rst implemented in the late 1980s and the 1990s tended to be either gay press/mail-out sampling or venue-based sampling. Newer surveillance systems or more recent changes to existing systems were more likely to be web-based sampling or more systematic approaches from which population estimates could potentially be calculated (i.e., RDS or venue-based applications of time space sampling).
While the trend towards sampling methods that can produce population estimates (e.g., venue-based applications of time space sampling) suggests a perception that these sampling methods are superior to the traditionally used convenience samples, these more rigorous sampling methods may come at the expense of a surveillance systems simplicity, timeliness and stability (i.e., if a sampling method is too resource intensive, it may not be sustainable).
In addition to trends over time, there were also patterns found in the types of sampling method by region. For example, web-based sampling and gay press/mail-out were almost solely used by countries in Western Europe, while RDS was the sampling method of choice for Central European countries.
Clusters of the same sampling method found within certain time periods and geographic regions suggest that the
choice of sampling method may reect changing notions of best practice and/or what is being used in neighbouring countries, rather than being guided by local context, available resources and the countrys information needs. In order to select a sampling method that is most suited to the countrys local context, extensive ethnographic research should rst be conducted. This research should gather information on potential points of access that could be used for recruitment (e.g., presence of a gay scene, of gay venues and the gay press, access to and uptake of the Internet and the size of social networks among MSM). When making a decision on the most suitable sampling method, information gathered on the local context must be mediated by the availability of resources (e.g., funds and expertise), and where resources are not available, whether capacity can be built. For example, access to statistical and methodological expertise may be an important consideration if planning to implement a surveillance system making use of RDS or a venue-based application of time space sampling, in which the implementation and analysis may be more complex than other traditionally used methods.
Finally, the countrys information needs, which will depend on other existing sources of information, should also help guide the choice of sampling method. For example, if seroprevalence data are required, the capacity to collect biological samples or provide HIV testing to participants is most feasible through venue-based sampling or RDS, as compared to gay press/mail-out sampling and web-based sampling. While introducing HIV testing into behavioural surveillance may bias participation and increase logistical considerations [4], this added component can provide valuable information on trends in HIV prevalence, not otherwise available. The ability to collect biological samples may be a factor in the continued adoption of venue-based sampling, as evidenced in some of the more recently implemented surveillance systems.
Among the 26 surveillance systems identied through this review, roughly a third (n = 9) made use of multiple sampling methods. Where the resources are available, the use of different sampling methods can be valuable in capturing different segments of the MSM population, thereby providing more comprehensive information on the trends in risk behaviours. As well, the ability to compare multiple sources of data can contribute towards an understanding of the representativeness of the samples and identify any subgroups and their risk behaviours that are being systematically excluded from some of the sampling methods being used for behavioural surveillance. While it may not be feasible to employ multiple sampling methods on an ongoing basis, surveillance authorities may wish to consider the use of one main sampling method, supplemented by a survey that makes use of an alternate sampling method, which is conducted on a less regular basis (e.g., every 510 years).
123
AIDS Behav (2010) 14:12521264 1261
Unfortunately, very little information was available on how sampling methods were selected, making it difcult to determine whether decisions were made based on context, the availability of resources and information needs, or based on the popularity of the methods being used at the time or place of implementation. Indeed, despite a thorough search, details on many aspects of the sampling methods used in behavioural surveillance were often sparse. Recommendations for key data that should be reported for each RDS study were recently made [77]. Such standards for reporting are also needed for the other sampling methods used in behavioural surveillance systems. Given that differences in the way a sampling method is implemented can affect the quality of the surveillance data, such details should be routinely reported in journal articles and surveillance reports.
Despite the lack of available information, it was clear that not only were there differences in the sampling methods used over time and by region, there was also variation in the way the same sampling methods were applied across surveillance systems, most notably in the application of venue-based sampling. Whether participants were compensated, whether HIV testing was included, the number and types of venues used and the use of random sampling to select venues, days/ times or participants varied considerably across the surveys that used venue-based sampling.
While the choice of sampling method should be determined by local context, available resources and information needs, it may still be possible to standardize certain elements across sampling methods, such as eligibility criteria, as well as to develop standards of implementation within specic sampling methods. Efforts are currently being made to examine the issue of differences in sampling methods within member countries from the European Union and the European Free Trade Agreement (EU/ EFTA) [78]. Other developed countries would benet from being part of this process.
Almost all reports and publications dedicated some discussion to the representativeness of the samples resulting from the different sampling methods. Most often, this discussion focused on concerns that the sample was likely not representative of the MSM population as a whole, regardless of the type of sampling method used. To date, only a small number of studies have made use of random sampling to survey MSM [7981]. In the absence of data from such studies with which to compare, it is difcult to determine how the sampling methods compare in terms of representativeness. Further research is needed regarding how to evaluate the various sampling methods ability to approximate a random sample.
However, despite the emphasis placed on representativeness in most reports and publications reviewed, it is important to keep in mind that for surveillance purposes, it
is not necessary to capture representative samples of the whole MSM population, as long as the sample captures the relevant target group (e.g., those most likely to engage in risk behaviours), and stays relatively similar over time. This was recognized and discussed for some of the surveys, with concerns being raised about whether the participants recruited from venues are continuing to capture the relevant target group over time [18]. However, more research is needed to conrm whether such changes are occurring, and if yes, in what ways participants are changing, and the implications for behavioural surveillance.
Beyond the representativeness of the sample, there are other aspects of the sampling methods that are vital to the success of a surveillance system, such as the ease of operation, acceptability by participants and users of the data, timeliness in reporting and stability [82]. Little information was available in articles and surveillance reports on these aspects of the sampling methods. At a time when questions are being raised about the most appropriate sampling methods [7], these attributes need to be thoroughly examined and discussed in order to better assess which of the sampling methods, or combination of methods, is best suited for behavioural surveillance based on a countrys local context, needs and resources.
Of the 40 countries included in this review, eight countries (i.e., Japan, Austria, Finland, Greece, Czech Republic, Hungary, Romania and Latvia) explicitly reported not having a behavioural surveillance system among MSM. Countries without behavioural surveillance systems or that are in the process of building their surveillance systems will especially benet from more complete information regarding the relative strengths and limitations of the different sampling methods used for behavioural surveillance.
Efforts are currently also being made to harmonize behavioural indicators for MSM across EU/EFTA countries [10, 78]. However, the ability to compare indicators will remain limited as long as substantial differences continue to exist in the methods used to sample MSM. This paper contributes to this important discussion by providing an overview of the sampling methods used in developed countries and by suggesting areas for standardising elements of the sampling methods used to recruit MSM for behavioural surveillance surveys.
References
1. Sullivan PS, Hamouda O, Delpech V, Geduld JE, Prejean J, Semaille C, et al. Reemergence of the HIV epidemic among men who have sex with men in North America, Western Europe, and Australia, 19962005. Ann Epidemiol. 2009;19(6):42331.
2. Rehle T, Lazzari S, Dallabetta G, Asamoah-Odei E. Second-generation HIV surveillance: better data for decision-making. Bull World Health Organ. 2004;82(2):1217.
123
1262 AIDS Behav (2010) 14:12521264
3. UNAIDS/WHO Working Group of Global HIV/AIDS, STI Surveillance. Guidelines for second generation HIV surveillance. Geneva: UNAIDS; 2000. p. 40. (Report No.: WHO/CDS/CSR/ EDC/2000.5).
4. Zaba B, Slaymaker E, Urassa M, Boerma JT. The role of behavioral data in HIV surveillance. AIDS. 2005;19(Suppl 2): S3952.
5. Garnett GP, Garcia-Calleja JM, Rehle T, Gregson S. Behavioural data as an adjunct to HIV surveillance data. Sex Transm Infect. 2006;82(Suppl 1):i5762.
6. Heckathorn DD. Respondent-driven sampling: a new approach to the study of hidden populations. Social Prob. 1997;44(2):17499.
7. Magnani R, Sabin K, Saidel T, Heckathorn D. Review of sampling hard-to-reach and hidden populations for HIV surveillance. AIDS. 2005;19(Suppl 2):S6772.
8. Zhang D, Bi P, Hiller JE, Lv F. Web-based HIV/AIDS behavioral surveillance among men who have sex with men: potential and challenges. Int J Infectious Dis. 2008;12(2):12631.
9. Chiasson MA, Parsons JT, Tesoriero JM, Carballo-Dieguez A, Hirsheld S, Remien RH. HIV behavioral research online. J Urban Health. 2006;83(1):7385.
10. Editorial Team. Workshop on European behavioural indications for men who have sex with men. Euro Surveill. 2008;13(15).
11. United Nations Statistics Division. Composition of macro geographical (continental) regions, geographical sub-regions, and selected economic and other groupings. New York: United Nations Statistics Division; 2009 [updated 2009 Apr 1; cited 2009 Apr 10]. http://unstats.un.org/unsd/methods/m49/m49regin.htm
Web End =http://unstats.un.org/unsd/methods/m49/m49regin.htm .
12. Frankland A, Zablotska I, Prestage G, OConnor S, Martin P, Imrie J. Gay community periodic survey: Queensland 2007. Sydney: National Centre in HIV Social Research, University of New South Wales; 2007. p. 31. (Report No.: GCPS Report 1/2008).
13. Frankland A, Zablotska I, Prestage G, Spence G. Gay community periodic survey: Adelaide 2007. Sydney: National Centre in HIV Social Research, University of New South Wales; 2007. p. 33. (Report No.: GCPS Report 2/2008).
14. Zablotska I, Frankland A, Prestage G, Down I, Ryan D. Gay community periodic survey: Sydney 2008. Sydney: National Centre in HIV Social Research, University of New South Wales; 2008. p. 13.
15. Zablotska I, Brown G, Frankland A, Prestage G, Kippax S, Langdon T. Gay community periodic survey: Perth 2006. Sydney: National Centre in HIV Social Research, University of New South Wales; 2006. p. 51.
16. Zablotska I, Prestage G, Chong S, Schamburg K, Mills D, Blattman T, et al. Gay community periodic survey: Canberra 2006. Sydney: National Centre in HIV Social Research, University of New South Wales; 2006. p. 34.
17. Van de Ven P, Crawford J. Change in sexual practice among Australian men who have sex with men, 19921996. AIDS. 1998;12(Suppl B):S66.
18. Saxton P, Dickson N, Hughes T. Findings from the Gay Auckland periodic sex survey. Auckland: New Zealand AIDS Foundation; 2006. p. 104.
19. New Zealand AIDS Foundation. 2006 Ofine and Online Sex Surveys [Internet]. Auckland: New Zealand AIDS Foundation; [cited 2008 Nov 12]. http://www.nzaf.org.nz/files/gapss%20goss%20summary%20sheet.pdf
Web End =http://www.nzaf.org.nz/les/gapss%20 http://www.nzaf.org.nz/files/gapss%20goss%20summary%20sheet.pdf
Web End =goss%20summary%20sheet.pdf .
20. Saxton P, Dickson N, Hughes T. GAPSS 2004: Findings from the Gay Auckland periodic sex survey. Auckland: New Zealand AIDS Foundation; 2004. p. 88.
21. Saxton P, Dickson N, Hughes T, Paul C, Saxton P, Dickson N, Hughes T, Paul C. GAPSS 2002: ndings from the Gay Auckland periodic sex survey. Auckland: New Zealand AIDS Foundation; 2002. p. 60.
22. Inoue Y, Yamazaki Y, Kihara M, Wakabayashi C, Seki Y, Ichikawa S. The intent and practice of condom use among HIV-positive men who have sex with men in Japan. AIDS Patient Care STDs. 2006;20(11):792802.
23. Kihara M, Ono-Kihara M, Feldman MD, Ichikawa S, Hashimoto S, Eboshida A, et al. HIV/AIDS surveillance in Japan, 1984 2000. J Acquir Immune Dec Syndr. 2003;32(Suppl 1):S5562.
24. Public Health Agency of Canada. HIV/AIDS: M-Track survey [Internet]. Ottawa: Public Health Agency of Canada [updated 2008 Mar 6; cited 2008 Nov 12]. http://www.phac-aspc.gc.ca/aids-sida/about/mtrack-eng.php
Web End =http://www.phac-aspc.gc.ca/ http://www.phac-aspc.gc.ca/aids-sida/about/mtrack-eng.php
Web End =aids-sida/about/mtrack-eng.php .
25. Paquette D, Kropp R, Cule S, Wylie J, Husbands W, Myers T, et al. HIV prevalence and associated risk behaviours among men who have sex with men (MSM) in Canada. In: XVII International AIDS Conference. Mexico City, 2008 [abstract TUPE0252].
26. Sanchez T, Finlayson T, Drake A, Behel S, Cribbin M, Dinenno E, et al. Human immunodeciency virus (HIV) risk, prevention, and testing behaviorsUnited States, National HIV Behavioral Surveillance System: men who have sex with men, November 2003April 2005. MMWR CDC Surveill Summ. 2006;55(6):116. (Erratum in: MMWR Morb Mortal Wkly Rep. 2006;55(27): 752).27. MacKellar DA, Gallagher KM, Finlayson T, Sanchez T, Lansky A, Sullivan PS. Surveillance of HIV risk and prevention behaviors of men who have sex with mena national application of venue-based, timespace sampling. Public Health Rep. 2007; 122(Suppl 1):3947.
28. Lansky A, MacKellar D, Gallagher KM, Lin LS, Sullivan PS, Onorato IM. Evaluation of the new National HIV behavioral surveillance system (NHBS) is imperative. Sex Transm Dis. 2006;33(4):2723. (author reply 45).
29. Martens V, Huynen P. Knowledge and behaviour of men who have sex with men towards HIV/AIDS in the French Community of Belgium. Arch Public Health. 2006;64(1):1326.
30. Martens V. Modes de vie et comportements des gays face au SIDA. Rapport de lenquete 2004 pour lassociation Ex aequo. Observatoires du SIDA et des sexualites (FUSL);2005, 35 p.
31. Cowan SA. HIV and risk behaviour among men who have sex with men in Denmarkthe 2006 Sex Life Survey. Euro Surveill. 2008;13(48):pii:19050.
32. Leobon A, Frigault L-R, Levy J. Results of the Net Gai barometre 2004 survey: Internet use among French homosexual men. Montreal: World Cong Sexol; 2005. (abstract 706.2).
33. Michel AO, Alexandre A, Monvoisin D, Velter A, Desenclos JC, Semaille C. Barometre Gay 2002, France: partnership for a large survey and feedback to respondents. Bangkok: International Conference on AIDS; 2004. (abstract MoPeD3969).
34. Velter A, Michel A, Semaille C. Rapport Barometre Gay 2002. Saint-Maurice: Institut de veille sanitaire; 2005. p. 56.
35. Leobon A, Frigault L-R. Frequent and systematic unprotected anal intercourse among using the Internet to meet other ment for sexual purposes in France: results form the Gay Net Barometer 2006 Survey. AIDS Care. 2008;20(4):47884.
36. Adam P, Hauet E, Caron C. Recrudescence des prises de risque et des MST parmi les gays. Resultats preliminaires de lEnquete Presse Gay 2000. Saint-Maurice: Institut de veille sanitaire; 2001.
37. Bochow M. Are younger German gay men more at risk from HIV? Results of a national survey in the gay press in Germany. Cult Health Sex. 2000;2(2):18395.
38. Marcus U, Voss L, Kollan C, Hamouda O. HIV incidence increasing in MSM in Germany: factors inuencing infection dynamics. Euro Surveil. 2006;11(9):15760.
39. Ministry of Health and Social Solidarity [Greece]. Greece: UNGASS country progress report 2008 (Reporting Period: JanuaryDecember 2007). Athens: Ministry of Health and Social Solidarity [Greece]; 2008. p. 17.
123
AIDS Behav (2010) 14:12521264 1263
40. Devine P, Hickson F, McNamee H, Quinlan M. Real lives: ndings from the All Irelance Gay Mens Sex Surveys, 2003 and 2004. The Rainbow Project and The Gay Mens Health Project; 2006, 73 p.
41. Mirandola M, Foschia J-P, Baldassari E, Gios L, Ramarli D, Fontana R. Sialon: behavioural survey questionnaire development and survey procedures piloting among men who have sex with men (MSM) in Veneto, Italy. In: AIDS 2008XVII International AIDS Conference. Mexico, 2008 [abstract CDC0199].
42. Hospers HJ, Dorer TT, Zuilhof W. Fact sheets: Schorer Monitor 2008. [Internet]. Amsterdam: Schorer; [cited 2009 Jan 24]. http://www.schorer.nl/48/schorer-monitor/schorer-monitor-2008/
Web End =http://www.schorer.nl/48/schorer-monitor/schorer-monitor-2008/ .
43. Blystad H, Nilsen, Aavitsland P. Increase in reported HIV infections among MSM in Oslo, Norway. Euro Surveill. 2004;8(11):pii = 2405. [Internet cited 2009 Jan 23]. http://www.eurosurveillance.org/ViewArticle.aspx?ArticleId=2405
Web End =http:// http://www.eurosurveillance.org/ViewArticle.aspx?ArticleId=2405
Web End =www.eurosurveillance.org/ViewArticle.aspx?ArticleId=2405 .
44. Rodes A, Maguire A, Vall M, Tort X, Rabella N. Changes in HIV risk related behaviour among gay men in Barcelona, Spain. In: International Conference on AIDS. Vancouver, 1996 [abstract Tu.C.2416].
45. Centre for Sexually Transmitted Infection and AIDS Epidemio-logical Studies of Catalonia (CEEISCAT). Sistema integrat de vigilancia epidemiologica de Sida/VIH/ITS a Catalunya (SIVES) Biennial report. Barcelona: Generalitat de Catalunya, Departement de Salut 2008; 2008, 89 p.
46. National Board of Health and Welfare [Sweden]. Sweden: country progress report. Stockholm: National Board of Health and Welfare [Sweden]; 2008. p. 34.
47. Dubois-Arber F, Jeannin A, Moreau-Gruet F, Benninghoff F, Spencer B. 15 years of behavioural surveillance in Switzerland. In: International Conference on AIDS. Barcelona, 2002 [abstract WePeC6181].
48. Balthasar H, Jeannin A, Dubois-Arber F. Surveillance of HIV/ AIDSRelated behaviors among men who have sex with men Switzerland, 19922004. Surveillance des comportements face au VIH/sida chez les hommes ayant des rapports sexuels avec des hommesSuisse, 19922004. Int J Public Health. 2007; 52(1):2738.
49. Balthasar H, Jeannin A, Dubois-Arber F. First anal intercourse and condom use among men who have sex with men in Switzerland. Arch Sex Behav. 2009;13(2):10008.
50. Balthasar H, Jeannin A, Dubois-Arber F. Les comportements face au VIH/SIDA des hommes qui ont des relations sexuelles avec des hommes. Resultats de Gaysurvey 2007. Lausanne: Institut universitaire de medecine sociale et preventive Lausanne; 2008. p. 21.
51. Elford J, Bolding G, Maguire M, Sherr L. Sexual risk behaviour among gay men in a relationship. AIDS. 1999;13(11):140711.
52. Elford J, Bolding G, Davis M, Sherr L, Hart G. Trends in sexual behaviour among London homosexual men 19982003: implications for HIV prevention and sexual health promotion. Sex Transm Infect. 2004;80(6):4514.
53. Elford J, Bolding G, Davis M, Sherr L, Hart G. The internet and HIV study: design and methods. BMC Public Health. 2004;4:39.
54. Weatherburn P, Hickson F, Reid D, Jessup K, Hammond G. Multiple chances ndings from the United Kingdom Gay Mens sex survey 2006. London: Sigma Research; 2008. 48 p.
55. Nardone A, Dodds JP, Mercey DE, Johnson AM. Active surveillance of sexual behaviour among homosexual men in London. Commun Dis Public Health. 1998;1(3):197201.
56. Dodds JP, Johnson AM, Parry JV, Mercey DE. A tale of three cities: persisting high HIV prevalence, risk behaviour and undiagnosed infection in community samples of men who have sex with men. Sex Transm Infect. 2007;83(5):3926.
57. Elford J, Bolding G, Sherr L, Hart G. No evidence of an increase in serosorting with casual partners among HIV-negative gay men in London, 19982005. AIDS. 2007;21(2):2435.
58. Williamson LM, Hart GJ. HIV prevalence and undiagnosed infection among a community sample of gay men in Scotland. J Acquir Immune Dec Syndr. 2007;45(2):22430.
59. Family Health International. Albaniabehavioral and biological surveillance study report. Arlington: Family Health International; 2006. p. 142.
60. Ministry of Health [Republic of Bulgaria]. Republic of Bulgaria: country progress report (Reporting Period: January 2006 December 2007). Soa: Ministry of Health [Republic of Bulgaria]; 2008. p. 30.
61. Bozicevic I, Rode OD, Lepej SZ, Johnston LG, Stulhofer A, Dominkovic Z, et al. Prevalence of sexually transmitted infections among men who have sex with men in Zagreb, Croatia. AIDS Behav. 2009;13(2):3039.
62. Ministry of Health [Republic of Macedonia]. Republic of Macedonia: UNGASS country progress report (Reporting Period: January 2006December 2007). Skopje: Ministry of Health [Republic of Macedonia]; 2008. p. 39.
63. Meskovic D. BSS among MSM population in Serbia in 2006, 2007: ppt presentation [Internet] [cited 2008 Nov 21]. http://www.integration-projects.org/euromsmhiv/presentations/A8%5BDANIEL_MESCOVIC%5DBSSSerbia.ppt#256,1,Slide1
Web End =http:// http://www.integration-projects.org/euromsmhiv/presentations/A8%5BDANIEL_MESCOVIC%5DBSSSerbia.ppt#256,1,Slide1
Web End =www.integration-projects.org/euromsmhiv/presentations/ http://www.integration-projects.org/euromsmhiv/presentations/A8%5BDANIEL_MESCOVIC%5DBSSSerbia.ppt#256,1,Slide1
Web End =A8%5BDANIEL_MESCOVIC%5DBSSSerbia.ppt#256,1,Slide1 .
64. Klavs I, Poljak M. Unlinked anonymous monitoring of human immunodeciency virus prevalence in high- and low-risk groups in Slovenia, 19932002. Croat Med J. 2003;44(5):5459.
65. Minister of Health [Republic of Belarus]. Republic of Belarus: National report on the implementation of the declaration of commitment on HIV/AIDS (Reporting Period: January 2006 December 2007). Minsk: Minister of Health [Republic of Bela-rus]; 2008. p. 27.
66. Estonian Ministry of Social Affairs. Estonia: country report on national response to HIV/AIDS epidemic 2007. Tallinn: Estonian Ministry of Social Affairs; 2008. p. 26.
67. Ministry of Health [Latvia]. Latvia: UNGASS country progress report (Reporting Period: January 2006December 2007). Riga: Ministry of Health [Latvia]; 2008. p. 35.
68. Ministry of Health [Lithuania]. Lithuanian country report. Vilnius: Ministry of Health [Lithuania]; 2007. p. 13.
69. Strujeva O, Caplinskas S, Gricius A. Sexual behaviour and HIV prevalence among Lithuanian men who have sex with men (MSM). In: 4th IAS Conference on HIV Pathogenesis, Treatment and Prevention. Sydney, 2007 [abstract CDC077].
70. National Centre on Health Management [Republic of Moldova]. Republic of Moldova: report on progress (Reporting Period: January 2006December 2007). Chisinau: National Centre on Health Management [Republic of Moldova]; 2008. p. 84.
71. Smolskaya TT, Yakovleva AA, Kasumov VK, Gheorgitsa SI. HIV sentinel surveillance in high-risk groups in Azerbaijan, the Republic of Moldova and the Russian Federation. Copenhagen: World Health Organization; 2004. p. 174. (Report No.: EUR/03/5057956).
72. World Health Organization. HIV prevalence and risks among men having sex with men in Moscow and Saint Petersburg. Geneva: WHO; 2007. p. 84.
73. Family Health International. Behavioral Monitoring Survey: Russia 2005. Arlington: Family Health International; 2007. p. 128.
74. Kruglov YV, Kobyshcha YV, Salyuk T, Varetska O, Shakarishvili A, Saldanha VP. The most severe HIV epidemic in Europe: Ukraines national HIV prevalence estimates for 2007. Sex Transm Infect. 2008;84(Suppl 1):i3741.
75. Balakireva OM, Bondar TV, Kasyanchuk MG, Kis ZR, Leschinskiy EV,Sheremet-SheremetyevSP. Monitoringbehaviour of menhaving sex with men: analytical report on results of the research. Kyiv, Ukraine: ICF International HIV/AIDS Alliance in Ukraine; 2008.
76. Johnston LG, Malekinejad M, Kendall C, Iuppa IM, Rutherford GW. Implementation challenges to using respondent-driven sampling methodology for HIV biological and behavioral
123
1264 AIDS Behav (2010) 14:12521264
surveillance: eld experiences in international settings. AIDS Behav. 2008;12(Suppl 1):S13141.77. Malekinejad M, Johnston LG, Kendall C, Kerr LRFS, Rifkin MR, Rutherford GW. Using respondent-driven sampling methodology for HIV biological and behavioral surveillance in international settings: a systematic review. AIDS Behav. 2008;12(Suppl 1): S13141.
78. Institute for Social, Preventive Medicine, Centre Hospitalier Universitaire Vaudois, University of Lausanne Switzerland. ECDC technical report: Mapping of HIV/STI behavioural surveillance in Europe. Stockholm: European Centre for Disease Control; 2009. p. 164.
79. Brewer DD, Golden MR, Handseld HH. Unsafe sexual behavior and correlates of risk in a probability sample of men who have
sex with men in the era of highly active antiretroviral therapy. Sex Transm Dis. 2006;33(4):2505.80. Xia Q, Osmond DH, Tholandi M, Pollack LM, Zhou W, Ruiz JD, et al. HIV prevalence and sexual risk behaviors among men who have sex with men: Results from a statewide population-based survey in California. J Acquir Immune Dec Syndr. 2006;41(2): 23845.
81. Catania JA, Osmond D, Stall RD, Pollack L, Paul JP, Blower S, et al. The continuing HIV epidemic among men who have sex with men. Am J Public Health. 2001;91(6):90714.
82. CDC. Updated guidelines for evaluating public health surveil-lance systems. MMWR Morb Mortal Wkly Rep. 2001;50(13): 129.
123
Springer Science+Business Media, LLC 2010