Estimating the size of the MSM populations for 38 European countries by calculating the survey-surveillance discrepancies (SSD) between self-reported new HIV diagnoses from the European MSM internet survey (EMIS) and surveillance-reported HIV diagnoses among MSM in 2009.
Estimating the size of the MSM populations for 38 European countries by calculating the survey-surveillance discrepancies (SSD) between self-reported new HIV diagnoses from the European MSM internet survey (EMIS) and surveillance-reported HIV diagnoses among MSM in 2009.
复制标题
DOI：
10.1186/1471-2458-13-919
复制
发表时间：
2013-10-03
影响因子：
4.5
通讯作者：
EMIS Network
中科院分区：
文献类型：
--
作者：
Marcus U;Hickson F;Weatherburn P;Schmidt AJ;EMIS Network
由于各种原因，包括MSM人群的规模未知，比较各国MSM中新诊断的艾滋病毒感染率具有挑战性。在本文中，我们提出了一种三角测量监测数据的方法，收集的数据在泛欧MSM互联网调查（EMIS）估计的规模，全国MSM人群和艾滋病毒被诊断的比率，其中通过计算调查监测的差异（SSD）作为衡量的选择偏差的调查参与者。2010年，第一个EMIS收集了欧洲38个国家18万多名男男性行为者自我报告的艾滋病毒诊断数据。将这些数据与国家艾滋病毒监测系统的数据进行比较，以探讨这两种方法中可能存在的抽样和报告偏倚。调查-监测差异（SSD）代表2009年诊断的调查成员（HIVsvy）与调查成员总数（Nsvy）的比率，除以2009年诊断的监测报告（HIVpop）与估计的男男性行为者总人口（Npop）的比率。由于家庭互联网接入的差异可能是调查选择偏差的一个关键组成部分，我们分析了在不同时间点进行连续MSM互联网调查的国家中，随着互联网接入水平的提高，家庭互联网接入与SSD之间的关系。使用经验定义的SSD来计算相应的MSM群体大小（Npop），使用公式Npop = HIVpop*Nsvy*SSD/HIVsvy。2003年和2010年之间的连续MSM互联网调查与不同水平的家庭互联网接入的调查-监测差异最好用一个潜在的方程来描述，高SSD在低互联网接入，下降到2左右的水平与广泛的访问。荷兰的SSD最低，为1.8，摩尔多瓦最高，为9.0。以2009年男男性行为者艾滋病毒诊断监测报告的现有最佳估计数（HIVpop）为例，男男性行为者的相对人口规模在15-64岁成年男性人口的0.03%至5.6%之间。最近诊断（2009年）的艾滋病毒在EMIS参与者和艾滋病毒在2009年诊断的男男性行为者之间的相关性是非常高的（R2 = 0.88）时，使用计算的MSM人口规模。在若干国家，Npop和HIVpop的低水平并不可靠。我们讨论和确定可能的测量误差的国家计算的MSM人口规模高于3%和低于1%的成年男性人口。在大多数情况下，监测系统中男男性行为者中新诊断的艾滋病毒数量似乎太低。在某些情况下，测量误差可能是由于EMIS样本量小。必须假定SSD是根据国家具体因素修改的。对基于社区的调查数据与监测数据进行比较表明，前者只有微小的抽样偏差，尽管各国参与率差异很大，但除少数国家外，不会严重扭曲国家间的可比性。互联网调查是对国家监测系统的有益补充，可以突出不足之处，在监测系统无法准确提供此类数据的国家，可以估计男男性行为者中新诊断感染的范围。
Comparison of rates of newly diagnosed HIV infections among MSM across countries is challenging for a variety of reasons, including the unknown size of MSM populations. In this paper we propose a method of triangulating surveillance data with data collected in a pan-European MSM Internet Survey (EMIS) to estimate the sizes of the national MSM populations and the rates at which HIV is being diagnosed amongst them by calculating survey-surveillance discrepancies (SSD) as a measure of selection biases of survey participants. In 2010, the first EMIS collected self-reported data on HIV diagnoses among more than 180,000 MSM in 38 countries of Europe. These data were compared with data from national HIV surveillance systems to explore possible sampling and reporting biases in the two approaches. The Survey-Surveillance Discrepancy (SSD) represents the ratio of survey members diagnosed in 2009 (HIVsvy) to total survey members (Nsvy), divided by the ratio of surveillance reports of diagnoses in 2009 (HIVpop) to the estimated total MSM population (Npop). As differences in household internet access may be a key component of survey selection biases, we analysed the relationship between household internet access and SSD in countries conducting consecutive MSM internet surveys at different time points with increasing levels of internet access. The empirically defined SSD was used to calculate the respective MSM population sizes (Npop), using the formula Npop = HIVpop*Nsvy*SSD/HIVsvy. Survey-surveillance discrepancies for consecutive MSM internet surveys between 2003 and 2010 with different levels of household internet access were best described by a potential equation, with high SSD at low internet access, declining to a level around 2 with broad access. The lowest SSD was calculated for the Netherlands with 1.8, the highest for Moldova with 9.0. Taking the best available estimate for surveillance reports of HIV diagnoses among MSM in 2009 (HIVpop), the relative MSM population sizes were between 0.03% and 5.6% of the adult male population aged 15–64. The correlation between recently diagnosed (2009) HIV in EMIS participants and HIV diagnosed among MSM in 2009 as reported in the national surveillance systems was very high (R2 = 0.88) when using the calculated MSM population size. Npop and HIVpop were unreliably low for several countries. We discuss and identify possible measurement errors for countries with calculated MSM population sizes above 3% and below 1% of the adult male population. In most cases the number of new HIV diagnoses in MSM in the surveillance system appears too low. In some cases, measurement errors may be due to small EMIS sample sizes. It must be assumed that the SSD is modified by country-specific factors. Comparison of community-based survey data with surveillance data suggests only minor sampling biases in the former that – except for a few countries - do not seriously distort inter-country comparability, despite large variations in participation rates across countries. Internet surveys are useful complements to national surveillance systems, highlighting deficiencies and allowing estimates of the range of newly diagnosed infections among MSM in countries where surveillance systems fail to accurately provide such data.
登录
查看更多内容
影响因子：
4.5
作者：
Marcus U;Schmidt AJ;Kollan C;Hamouda O
通讯作者：
Hamouda O
影响因子：
3.6
作者：
Evans, Alison Ruth;Wiggins, Richard D.;Elford, Jonathan
通讯作者：
Elford, Jonathan
影响因子：
2.6
作者：
Weatherburn, Peter;Schmidt, Axel J.;Marcus, Ulrich
通讯作者：
Marcus, Ulrich
DOI：
10.1590/s0102-311x2006000700002
发表时间：
2006-07-01
期刊：
Cadernos de Saúde Pública
影响因子：
--
作者：
Michaels, Stuart;Lhomond, Brigitte
通讯作者：
Lhomond, Brigitte
影响因子：
4.5
作者：
Marcus U;Schmidt AJ;Hamouda O;Bochow M
通讯作者：
Bochow M