Design-based inference in time-location sampling

The Development and the Assessment of Sampling Methods for Hard-to-Reach Populations in HIV Surveillance

Due to stigma or legal issues, populations with higher HIV risk are often hard to reach, which impedes accurate population estimation of HIV burden. To better sample hard-to-reach populations (HTRPs) for HIV surveillance, various sampling methods have been designed and/or used since HIV epidemic following the first reported AIDS cases in 1981. This paper describes the development and the assessment (i.e., validity and reproducibility) of approximately eight sampling methods (e.g., convenience sampling, snowball sampling, time location sampling, and respondent-driven sampling) for HTRPs in HIV surveillance, with a focus on respondent-driven sampling (RDS). Compared to other methods, RDS has been greatly assessed. However, current evidence is still inadequate for RDS to be considered the best option for sampling HTRPs. The field must continue to assess RDS and to develop new sampling approaches or modifications to existing approaches.

Adaptive Time-Location Sampling for COMPASS: A SARS-CoV-2 Prevalence Study in Fifteen Diverse Communities in the United States

BACKGROUND

COVID-19 has placed a disproportionate burden on underserved racial and ethnic groups, community members working in essential industries, those living in areas of high population density, and those reliant on in-person services such as transportation. The goal of this study was to estimate the cross-sectional prevalence of SARS-CoV-2 (active SARS-CoV-2 or prior SARS-CoV-2 infection) in children and adults attending public venues in 15 sociodemographically diverse communities in the United States and to develop a statistical design that could be rigorously implemented amidst unpredictable stay-at-home COVID-19 guidelines.

METHODS

We used time-location sampling with complex sampling involving stratification, clustering of units, and unequal probabilities of selection to recruit individuals from selected communities. We safely conducted informed consent, specimen collection, and face-to-face interviews outside of public venues immediately following recruitment.

RESULTS

We developed an innovative sampling design that adapted to constraints such as closure of venues, changing infection hotspots, and uncertain policies. We updated both the sampling frame and the selection probabilities over time using information acquired from prior weeks. We created site-specific survey weights that adjusted sampling probabilities for nonresponse and calibrated to county-level margins on age and sex at birth.

CONCLUSIONS

Although the study itself was specific to COVID-19, the strategies presented in this article could serve as a case study that can be adapted for performing population-level inferences in similar settings and could help inform rapid and effective responses to future global public health challenges.

Optimizing ancestral trait reconstruction of large HIV Subtype C datasets through multiple-trait subsampling

Large datasets along with sampling bias represent a challenge for phylodynamic reconstructions, particularly when the study data are obtained from various heterogeneous sources and/or through convenience sampling. In this study, we evaluate the presence of unbalanced sampled distribution by collection date, location, and risk group of human immunodeficiency virus Type 1 Subtype C using a comprehensive subsampling strategy and assess their impact on the reconstruction of the viral spatial and risk group dynamics using phylogenetic comparative methods. Our study shows that a most suitable dataset for ancestral trait reconstruction can be obtained through subsampling by all available traits, particularly using multigene datasets. We also demonstrate that sampling bias is inflated when considerable information for a given trait is unavailable or of poor quality, as we observed for the trait risk group. In conclusion, we suggest that, even if traits are not well recorded, including them deliberately optimizes the representativeness of the original dataset rather than completely excluding them. Therefore, we advise the inclusion of as many traits as possible with the aid of subsampling approaches in order to optimize the dataset for phylodynamic analysis while reducing the computational burden. This will benefit research communities investigating the evolutionary and spatio-temporal patterns of infectious diseases.

Geographically skewed recruitment and COVID-19 seroprevalence estimates: a cross-sectional serosurveillance study and mathematical modelling analysis

OBJECTIVES

Convenience sampling is an imperfect but important tool for seroprevalence studies. For COVID-19, local geographic variation in cases or vaccination can confound studies that rely on the geographically skewed recruitment inherent to convenience sampling. The objectives of this study were: (1) quantifying how geographically skewed recruitment influences SARS-CoV-2 seroprevalence estimates obtained via convenience sampling and (2) developing new methods that employ Global Positioning System (GPS)-derived foot traffic data to measure and minimise bias and uncertainty due to geographically skewed recruitment.

DESIGN

We used data from a local convenience-sampled seroprevalence study to map the geographic distribution of study participants' reported home locations and compared this to the geographic distribution of reported COVID-19 cases across the study catchment area. Using a numerical simulation, we quantified bias and uncertainty in SARS-CoV-2 seroprevalence estimates obtained using different geographically skewed recruitment scenarios. We employed GPS-derived foot traffic data to estimate the geographic distribution of participants for different recruitment locations and used this data to identify recruitment locations that minimise bias and uncertainty in resulting seroprevalence estimates.

RESULTS

The geographic distribution of participants in convenience-sampled seroprevalence surveys can be strongly skewed towards individuals living near the study recruitment location. Uncertainty in seroprevalence estimates increased when neighbourhoods with higher disease burden or larger populations were undersampled. Failure to account for undersampling or oversampling across neighbourhoods also resulted in biased seroprevalence estimates. GPS-derived foot traffic data correlated with the geographic distribution of serosurveillance study participants.

CONCLUSIONS

Local geographic variation in seropositivity is an important concern in SARS-CoV-2 serosurveillance studies that rely on geographically skewed recruitment strategies. Using GPS-derived foot traffic data to select recruitment sites and recording participants' home locations can improve study design and interpretation.

ADAPTIVE TIME LOCATION SAMPLING FOR COMPASS, A SARS-COV-2 PREVALENCE STUDY IN FIFTEEN DIVERSE COMMUNITIES IN THE UNITED STATES

The COVPN 5002 (COMPASS) study aimed to estimate the prevalence of SARS-CoV-2 (active SARS-CoV-2 or prior SARS-CoV-2 infection) in children and adults attending public venues in 15 socio-demographically diverse communities in the United States. To protect against potential challenges in implementing traditional sampling strategies, time-location sampling (TLS) using complex sampling involving stratification, clustering of units, and unequal probabilities of selection was used to recruit individuals from neighborhoods in selected communities. The innovative design adapted to constraints such as closure of venues; changing infection hotspots; and uncertain policies. Recruitment of children and the elderly raised additional challenges in sample selection and implementation. To address these challenges, the TLS design adaptively updated both the sampling frame and the selection probabilities over time using information acquired from prior weeks. Although the study itself was specific to COVID-19, the strategies presented in this paper could serve as a case study that can be adapted for performing rigorous population-level inferences in similar settings and could help inform rapid and effective responses to future global public health challenges.

Simple anthropometric measures to predict visceral adipose tissue area in middle-aged Indonesian men

The diagnosing of central obesity requires ethnic-specific cut-offs of waist circumference (WC) and body mass index (BMI). This study aims to develop formulas to predict visceral adipose tissue (VAT) area based on WC and BMI to determine the cut-off points of central obesity in Indonesia. We conducted a cross-sectional study among 32 middle-aged Indonesian men. VAT area was measured using an abdominal CT scan, whereas WC and BMI were assessed through anthropometric measurements. Linear regression analysis was performed to define the formulas to predict VAT area using WC and BMI. Next, the optimal cut-off values of WC and BMI were determined using ROC curve analysis. Strong positive correlations were found between WC and VAT as well as BMI and VAT (r = 0.78; r = 0.67, p <0.001). The formula to predict VAT area from WC was -182.65 + (3.35 × WC), whereas the formula to predict VAT area from BMI was -57.22 + (6.95 × BMI). These formulas predicted WC of 88.5 cm and BMI of 23.9 kg/m2 as the optimal cut-off values for central obesity in middle-aged Indonesian men.

Awareness and Willingness to use Condoms and Preexposure Prophylaxis among Gay, Bisexual, and Other Cisgendered Men who Have sex with men in Slum Communities in Ghana. BSGH-004

Introduction: Research has begun to examine human immunodeficiency virus (HIV) prevention strategies within the Ghanaian context. Still, little is known about specific populations such as gay, bisexual, and other cisgender men who have sex with men (GBMSM) living in slum communities. We studied HIV prevention strategies such as condoms and preexposure prophylaxis (PrEP) in slum communities and the awareness and willingness to use these choices among GBMSM. This qualitative study examines HIV prevention strategies, specifically examining PrEP and condom use behaviors among GBMSM in Ghana. Methods: We conducted in-depth face-to-face interviews among 12 GBMSM from slums in Accra and Kumasi cities in Ghana. Data were analyzed through a summative content analysis with multiple reviewers to develop codes. Data were collected from participants in January 2022. Results: We found the fear and perceived risk of infection were motivators for consistent condom use, especially during anal sex. GBMSM living with HIV receiving antiretroviral therapy were more inclined to use condoms. We found motivations for using PrEP were influenced by the type of sexual activity and a history of negative HIV status. Also, the barriers to PrEP for GBMSM included limited access to healthcare facilities and the distance to these facilities. Conclusions: To improve condom and PrEP access and uptake, we recommend addressing structural barriers by increasing the number of health facilities and implementing targeted interventions to address the lack of information on HIV awareness and prevention. Involving peer educators may also effectively promote HIV prevention strategies, especially in communities with limited access to healthcare such as slums. Overcoming these access constraints could significantly enhance awareness and prevention of HIV, leading to improved health outcomes for GBMSM living in slum communities.

Prevalence estimates of trafficking in persons using statistical definitions: a cross-sectional high-risk community survey in Cape Town, South Africa

OBJECTIVES

Current research on trafficking in persons (TIP) relies heavily on legal and prosecutorial definitions. A public health approach has called for population-level assessment; however, identification of TIP victims lacks a standardised operational definition. This study applied the Prevalence Reduction Innovation Forum (PRIF) statistical definitions, developed by the US Department of State, to a community survey in Cape Town, South Africa.

DESIGNS

A high-risk sampling strategy was used. TIP screening questions from two instruments were matched with PRIF domain indicators to generate prevalence estimates. Sensitivity, specificity and receiver operating characteristics analyses were conducted to assess the performance of the two screeners.

SETTING

Cross-sectional survey conducted in Cape Town, South Africa, from January to October 2021.

PARTICIPANTS

South Africans and immigrants from other nations residing in Cape Town and its surrounding areas, aged 18 or older, who met the study inclusion criteria for a set of experiences that were identified as TIP risk factors.

PRIMARY AND SECONDARY OUTCOME MEASURES

Primary outcome measures were PRIF lifetime and past 12-month TIP positivity. Secondary outcome measures included individual and summary measures from the two screeners.

RESULTS

Our PRIF algorithm yielded a TIP lifetime prevalence rate of 17.0% and past 12-month rate of 2.9%. Summary measures from each TIP screener showed an excellent range of predictive utility. The summary screener measures yielded statistically significant differences among some demographic and background categories. Several screener items were shown less predictive of the PRIF statistical definition criteria than others.

CONCLUSIONS

Prevalence estimates of probable TIP were higher than those reported elsewhere. Our TIP screeners yielded an excellent range of predictive utility for the statistical definitions, promising the potential for wider applications in global and regional TIP research and policymaking. A more systematic sampling strategy is needed even if statistical definitions become widely used.

Protocol for a winter sentinel surveillance program of notifiable respiratory viruses in Queensland

BACKGROUND

With the reduction in access to polymerase chain reaction (PCR) testing and changes in testing guidelines in Australia, a reduced number of people are seeking testing for coronavirus disease (COVID-19), limiting the opportunity to monitor disease transmission. Knowledge of community transmission of COVID-19 and other respiratory viruses is essential to better predict subsequent surges in cases during the pandemic to alert health services, protect vulnerable populations and enhance public health measures. We describe a methodology for a testing-based sentinel surveillance program to monitor disease in the community for early signal detection of SARS-CoV-2 and other respiratory viruses.

METHODS/DESIGN

A longitudinal active testing-based sentinel surveillance program for respiratory viruses (including SARS-CoV-2, influenza A, influenza B and Respiratory Syncytial Virus) will be implemented in some regions of Queensland. Adults will be eligible for enrolment if they are part of specific community groups at increased risk of exposure and have not had a COVID-19 infection in the last 13 weeks. Recruitment via workplaces will occur in-person, via email and through online advertisement. Asymptomatic participants will be tested via PCR for SARS-CoV-2 infection by weekly self-collected nasal swabs. In addition, symptomatic participants will be asked to seek SARS-CoV-2 and additional respiratory virus PCR testing at nominated COVID-19 testing sites. SARS-CoV-2 and respiratory virus prevalence data will be analysed weekly and at the end of the study period.

DISCUSSION

Once implemented, this surveillance program will determine the weekly prevalence of COVID-19 and other respiratory viruses in the broader community by testing a representative sample of adults, with an aim to detect early changes in the baseline positivity rate. This information is essential to define the epidemiology of SARS-CoV-2 in the community in near-real time to inform public health control measures and prepare health services and other stakeholders for a rise in service demand.

Smartphone-based Respondent Driven Sampling (RDS): A methodological advance in surveying small or ‘hard-to-reach’ populations

Producing statistically robust profiles of small or ‘hard-to-reach’ populations has always been a challenge for researchers. Since surveying the wider population in order to capture a large enough sample of cases is usually too costly or impractical, researchers have been opting for ‘snowballing’ or ‘time-location sampling’. The former does not allow for claims to representativeness, and the latter struggles with under-coverage and estimating confidence intervals. Respondent Driven Sampling (RDS) is a method that combines snowballing sampling with an analytical algorithm that corrects for biases that arise in snowballing. For all its advantages, a major weakness of RDS has been around data collection. Traditionally done on-site, the process is costly and lengthy. When done online, it is cheaper and faster but under a serious threat from fraud, compromising data quality and validity of findings. This paper describes a real-life application of a RDS data collection system that maximizes fraud prevention while still benefiting from low cost and speedy data collection.

Improving Underestimation of HIV Prevalence in Surveys Using Time-Location Sampling

We sought to find a method that improves HIV estimates obtained through time-location sampling (TLS) used to recruit most-at-risk populations (MARPs). The calibration on residuals (CARES) method attributes weights to TLS sampled individuals depending on the percentile to which their logistic regression residues belong. Using a real country database, provided by EMIS-2010, with 9591 men who have sex with men (MSM) and an HIV prevalence of 12.1%, we simulated three populations (termed "pseudo-populations") with different levels of HIV. From each pseudo-population, 1000 TLS samples were drawn, and the HIV prevalence estimated by the TLS method and by the CARES method were recorded and compared with the HIV prevalence of the 9591 men. Results showed that the CARES method improves estimates given by the TLS method by getting closer to the real HIV prevalence.

GPS-estimated foot traffic data and venue selection for COVID-19 serosurveillance studies

The initial phase of the COVID-19 pandemic in the US was marked by limited diagnostic testing, resulting in the need for seroprevalence studies to estimate cumulative incidence and define epidemic dynamics. In lieu of systematic representational surveillance, venue-based sampling was often used to rapidly estimate a community’s seroprevalence. However, biases and uncertainty due to site selection and use of convenience samples are poorly understood. Using data from a SARS-CoV-2 serosurveillance study we performed in Somerville, Massachusetts, we found that the uncertainty in seroprevalence estimates depends on how well sampling intensity matches the known or expected geographic distribution of seropositive individuals in the study area. We use GPS-estimated foot traffic to measure and account for these sources of bias. Our results demonstrated that study-site selection informed by mobility patterns can markedly improve seroprevalence estimates. Such data should be used in the design and interpretation of venue-based serosurveillance studies.

Intercepted journeys: Associations between migration and mobility experiences and depressive symptoms among substance using migrants at the Mexico-Guatemala border

Substance use and depressive psychiatric symptoms have been associated with migration and mobility. The Mexico-Guatemala border is a key transit point for internal, regional, and international migration flows. However, there is limited knowledge of the role of substance use, migration, and mobility on mental health among migrants at this border. Our paper explores the association of migration and mobility patterns with possible major depressive symptoms among migrants at this key geographic region. We recruited 392 substance-using migrants using modified time-location sampling. Crude and adjusted logistic regression models were developed. We found that 12% of the sample had possible major depressive symptoms. After adjusting for relevant covariates, including gender, income, and perceived homelessness, we found that recent rural-urban and short-term migrants had higher odds of possible major depressive symptoms, whereas international migrants had lower odds. Findings of this paper suggest that although migrants experience hardship and uncertainty, they may respond with complex and nuanced forms of coping and planning.

Migration experiences, life conditions, and drug use practices of Russian-speaking drug users who live in Paris: a mixed-method analysis from the ANRS-Coquelicot study

Background

After the collapse of the Soviet Union at the beginning of the 1990s, people who inject drugs spiked in Eastern Europe. Facing local repression and an array of factors encouraging emigration, some users have migrated to France. This population now make up to a third of the patient list of some harm reduction services in Paris. This article aims to present original data on the sociodemographic profiles of these users, on their migration trajectory, their life conditions, and on the evolution of their drug use practices since arriving in Paris.

Methods

Data were collected as part of the ANRS-Coquelicot Survey, an HIV and HCV seroprevalence study among French-speaking people who use drugs. A sub-sample of Russian-speaking drug users who had relocated from Eastern Europe to live in Paris completed a quantitative questionnaire (N = 150) and a qualitative semi-structured interview (N = 20). The survey aimed to describe participants’ backgrounds, and a thematic analysis of interviews was conducted to explore participants’ migration histories, their life conditions in Paris, and their drug use practices before and after arriving in France.

Results

This study highlights the great vulnerability of the participating population, often following a loss of social status after migrating to France. Another important finding is that participants had better access to harm reduction tools and reduced their risk of exposure to HIV and HCV infections linked to needle sharing. Although 60% said they had already shared a syringe in their lifetime (49.9% of them in their home country), the proportions shrank to 13.9% after they arrived in France and to 9.3% in the month before the study, a proportion that is lower than among French-speaking people who use drugs.

Conclusions

Our main findings on the profiles and behaviors of the study population lead us to make two recommendations: to offer stronger global care to these users in Paris and to reform drug policy in their home countries by integrating it into a public health approach.

Post-Traumatic Stress Disorder in Homeless Migrant Mothers of the Paris Region Shelters

Migrant women are disproportionately more likely to experience traumatic events in their country of origin, during migration and after arriving in the host country. Homeless women are more likely to be exposed to multiple victimizations in childhood (emotional or physical maltreatment) and in adulthood (sexual abuse, street victimization). This study’s objective was to describe the factors associated with the likelihood of post-traumatic stress disorder (PTSD) among homeless migrant mothers in the Paris region. Face-to-face interviews were conducted by bilingual psychologists and interviewers in a representative sample of homeless families in the Paris region. PTSD was ascertained using the Mini International Neuropsychiatric Interview (MINI) (n = 691 mothers). We studied PTSD in mothers using weighted Poisson regression. Homeless migrant mothers had high levels of PTSD (18.9%) in the 12 months preceding the study. In multivariate analysis, PTSD was associated with departure from the country of origin because of violence (PR = 1.45 95% CI 1.03; 2.04), depression in the preceding 12 months (PR = 1.82 95% CI 1.20; 2.76), and residential instability (PR = 1.93 95% CI 1.27; 2.93). Homeless migrant mothers have high levels of traumatic events and PTSD. Improvements in screening for depression and PTSD and access to appropriate medical care are essential for this vulnerable group.

Diphtheria-Tetanus-Polio, Measles-Mumps-Rubella, and Hepatitis B Vaccination Coverage and Associated Factors among Homeless Children in the Paris Region in 2013: Results from the ENFAMS Survey

Background: The number of homeless families has increased considerably since the 1990s in France. We aimed to estimate the homeless children vaccination coverage (VC) for diphtheria, tetanus, polio, measles-mumps-rubella and hepatitis B and identify factors associated with insufficient VC according to birthplace. Methods: A cross-sectional survey was conducted among homeless shelter families in the greater Paris area. A nurse conducted face-to-face interviews and collected vaccination records. We analyzed factors associated with insufficient VC, stratified by birthplace and vaccine, using robust Poisson regression. Results: The study included 214 children born in France and 236 born outside France. VC in French-born homeless children was high (>90% at 24 months for most vaccinations) and similar to levels observed in the general population, whereas VC in those born outside France was low (<50% at 24 months for all vaccines). Factors significantly associated with insufficient VC among children born outside France were age, parents with French-language difficulties, and changing residence at least twice in the previous year. Children in contact with the healthcare system at least once in the previous year had significantly higher VC, irrespective of vaccine and birthplace. Conclusion: Special attention should be paid to homeless children born outside France, with recent European and French recommendations confirming the need for catch-up vaccination in children with undocumented VC.

Estimating prevalence from dried blood spots without using biological cut-offs: application of a novel approach to hepatitis C virus in drug users in France (ANRS-Coquelicot survey)

Seroprevalence estimation using cross-sectional serosurveys can be challenging due to inadequate or unknown biological cut-off limits of detection. In recent years, diagnostic assay cut-offs, fixed assay cut-offs and more flexible approaches as mixture modelling have been proposed to classify biological quantitative measurements into a positive or negative status. Our objective was to estimate the prevalence of anti-HCV antibodies among drug users (DU) in France in 2011 using a biological test performed on dried blood spots (DBS) collected during a cross-sectional serosurvey. However, in 2011, we did not have a cut-off value for DBS. We could not use the values for serum or plasma, knowing that the DBS value was not necessarily the same. Accordingly, we used a method which consisted of applying a two-component mixture model with age-dependent mixing proportions using penalised splines. The component densities were assumed to be log-normally distributed and were estimated in a Bayesian framework. Anti-HCV prevalence among DU was estimated at 43.3% in France and increased with age. Our method allowed us to provide estimates of age-dependent prevalence using DBS without having a specified biological cut-off value.

Prevalence of hepatitis C infection, screening and associated factors among men who have sex with men attending gay venues: a cross-sectional survey (PREVAGAY), France, 2015

Background

Over the last 20 years, Hepatitis C virus (HCV) infection prevalence has dramatically increased among HIV-infected men who have sex with men (MSM) in many countries worldwide. It is suspected that this increase is primarily driven by sexual behaviours linked to blood exposure. Monitoring these behaviours is crucial to understand the drivers of the epidemic. This study assessed the prevalence of chronic HCV infection among MSM attending gay venues and associated chronic HCV risk factors. HCV screening and associated factors were described.

Methods

The cross-sectional survey PREVAGAY, based on time-location sampling, was conducted in 2015 among MSM attending gay venues in 5 French metropolitan cities. A self-administered questionnaire was completed and capillary whole blood on dried blood spots (DBS) collected. Possible factors associated with chronic HCV prevalence and with HCV screening in the previous year were investigated using Poisson regression.

Results

Chronic HCV infection prevalence from DBS analysis was 0.7% [IC95%: 0.3–1.5] in the study’s 2645 participants and was 3.0% [1.5–5.8] in HIV-positive MSM. It was significantly higher in those who reported the following: (lifetime) slamming (with or without the sharing of injection equipment); (during the previous year) fisting and chemsex, unprotected anal intercourse with casual partners, using gay websites and/or of mobile-based GPS applications, and having more than 10 sexual partners.

Only 41.3% [38.2–44.5] of the participants reported HCV screening during the previous year. Screening was significantly more frequent in MSM under 30 years of age, those who were HIV-positive, those vaccinated against hepatitis B and meningococcus C, and those who reported the following (during the previous year): more than 10 sexual partners, at least one sexually transmitted infection and fisting.

Conclusion

Chronic HCV infection prevalence in MSM attending gay venues was significantly higher in HIV-positive MSM and in those with risky sexual behaviours. Reflecting current screening recommendations for specific populations, previous HCV screening was more frequent in HIV-positive individuals and those with risky sexual behaviours. Nevertheless, HCV screening coverage needs to be improved in these populations. Comprehensive medical management, which combines screening and linkage to care with prevention strategies, is essential to control HCV among MSM.

Estimating Sizes of Key Populations at the National Level: Considerations for Study Design and Analysis

Background:

National estimates of the sizes of key populations, including female sex workers, men who have sex with men, and transgender women are critical to inform national and international responses to the HIV pandemic. However, epidemiologic studies typically provide size estimates for only limited high priority geographic areas. This article illustrates a two-stage approach to obtain a national key population size estimate in the Dominican Republic using available estimates and publicly available contextual information.

Methods:

Available estimates of key population size in priority areas were augmented with targeted additional data collection in other areas. To combine information from data collected at each stage, we used statistical methods for handling missing data, including inverse probability weights, multiple imputation, and augmented inverse probability weights.

Results:

Using the augmented inverse probability weighting approach, which provides some protection against parametric model misspecification, we estimated that 3.7% (95% CI = 2.9, 4.7) of the total population of women in the Dominican Republic between the ages of 15 and 49 years were engaged in sex work, 1.2% (95% CI = 1.1, 1.3) of men aged 15–49 had sex with other men, and 0.19% (95% CI = 0.17, 0.21) of people assigned the male sex at birth were transgender.

Conclusions:

Viewing the size estimation of key populations as a missing data problem provides a framework for articulating and evaluating the assumptions necessary to obtain a national size estimate. In addition, this paradigm allows use of methods for missing data familiar to epidemiologists.

"Starfish Sampling": a Novel, Hybrid Approach to Recruiting Hidden Populations

We sought to leverage the strengths of time location sampling (TLS) and respondent-driven sampling (RDS) for surveys of hidden populations by combing elements of both methods in a new approach we call "starfish sampling." Starfish sampling entails random selection of venue-day-time units from a mapping of the locations where the population can be found, combined with short chains of peer referrals from their social networks at the venue or presenting to the study site later. Using the population of transmen in San Francisco as a case example, we recruited 122 eligible participants using starfish sampling: 79 at randomly selected venues, 11 on dating applications, and 32 by referral. Starfish sampling produced one of the largest community-recruited samples specifically for transmen to date. Starfish sampling is a flexibility method to recruit and sample hidden populations for whom conventional TLS and RDS may not work in theory or practice.

Hair concentration of trace elements and growth in homeless children aged <6years: Results from the ENFAMS study

BACKGROUND

Growth is an important indicator of health in early childhood. This is a critical developmental period, during which a number of factors, including exposure to metals, might play a role in later physical and metabolic functions.

OBJECTIVE

To study the association between exposure to arsenic (As), cadmium (Cd), mercury (Hg), lead (Pb) and selenium (Se), and physical growth of children from homeless families aged <6years.

METHODS

This study was based on data of the cross-sectional survey (ENFAMS), which was conducted by the Observatoire du Samu Social on a random sample of homeless sheltered families in the Paris region during winter 2013. Families with children under 6years (N=324) were interviewed in 17 languages using face-to-face questionnaires. A nurse took anthropometric measures and collected hair samples where As, Cd, Hg, Pb and Se levels were measured. We calculated weight-for-age Z-score (WAZ), height-for-age Z-score (HAZ) and BMI-for-age Z-score (BMIZ) of children, using the 2006 WHO Child Growth Standards as a reference. Associations between ln-transformed metal exposures and growth outcomes were tested by multivariable linear regression models with adjustment for potential confounders (including maternal anthropometrical and socio-demographical characteristics, gestational age, child birthweight, breastfeeding, food insecurity of the child). Due to missing data (1.6% to 14.2% depending on the variables), we used multiple imputation by chained equations.

RESULTS

A strong positive correlation was found between Pb and Cd levels (r=0.65; p<0.001). Positive associations between Se level and HAZ (β=0.61; p=0.05) and between Cd and BMIZ (β=0.21; p=0.03) and negative associations between As and HAZ (β=-0.18; p=0.05) were no more significant after multiple imputation. A weak negative trend was observed between Cd and HAZ (β=-0.14; p=0.14), while positive trends were found between Se and both WAZ (β=0.55; p=0.10) and HAZ (β=0.51; p=0.06) after multiple imputation.

CONCLUSION

Overall, our results found no strong association between exposure to metals and physical growth of homeless children but we observed some trends that were consistent with previous studies. More research is required studying these associations longitudinally, along with higher sample sizes, for better understanding the sources of exposure in homeless population and the potential effects on growth.

Defining and surveying key populations at risk of HIV infection: Towards a unified approach to eligibility criteria for respondent-driven sampling HIV biobehavioral surveys

Substantial resources are invested in human immunodeficiency virus biobehavioral surveys using respondent-driven sampling for measuring progress towards the UNAIDS 90-90-90 goals and to obtain other essential data on key populations. Survey data are used to meet country needs as well those of development partners, whose data needs may sometimes diverge. Surveys using differing eligibility criteria impede comparisons across surveys. With scant literature and guidelines on how to approach eligibility criteria, diverse criteria are used within and across countries. We conducted a review of peer-reviewed human immunodeficiency virus respondent-driven sampling biobehavioral survey literature published through December 2013. We describe eligibility criteria of 137 articles representing 214 surveys. Reporting on age, risk behavior, and reference period of risk behavior was nearly universal; however, reporting on gender, geography, and language was less common. Multiple definitions were used for each criterion, making comparisons challenging. We provide a framework for how to approach defining eligibility to improve consistency and comparability across surveys.

Time location sampling in men who have sex with men in the HIV context: the importance of taking into account sampling weights and frequency of venue attendance

Sex between men is the most frequent mode of HIV transmission in industrialised countries. Monitoring risk behaviours among men who have sex with men (MSM) is crucial, especially to understand the drivers of the epidemic. A cross-sectional survey (PREVAGAY), based on time-location sampling, was conducted in 2015 among MSM attending gay venues in 5 metropolitan cities in France. We applied the generalised weight share method (GWSM) to estimate HIV seroprevalence for the first time in this population, taking into account the frequency of venue attendance (FVA). Our objectives were to describe the implementation of the sampling design and to demonstrate the importance of taking into account sampling weights, including FVA by comparing results obtained by GWSM and by other methods which use sample weights not including FVA or no weight. We found a global prevalence of 14.3% (95% CI (12.0-16.9)) using GWSM and an unweighted prevalence of 16.4% (95% CI (14.9-17.8)). Variance in HIV prevalence estimates in each city was lower when we did not take into account either the sampling weights or the FVA. We also highlighted an association of FVA and serological status in the most of investigated cities.

Self-reported injection practices among people who use drugs in French prisons: Public health implications (ANRS-Coquelicot survey 2011-2013)

INTRODUCTION AND AIMS

The aims of this study were to describe the prevalences of injection practices and needle/syringe sharing in people who use drugs in French prisons, and to investigate associated factors.

DESIGN AND METHODS

Using the ANRS-Coquelicot survey (2011-2013), a random sample of 1718 people who used drugs in free society was included. Information regarding a history of incarceration, drug-injection practices inside prison and needle/syringe sharing was collected during interviews.

RESULTS

In our sample, 65.5% reported a history of injection and 57.4% had been incarcerated at least once. Among those who reported both of these conditions, 14% reported injection practices inside prison, 40.5% of whom had shared needles/syringes. In the multivariable model, the following variables were associated with injection practices inside prison: being a Russian-speaking detainee, having spent more time in prison, and having started to inject before 1996 and especially before 1987. Being Russian speaking was also associated with needle/syringe sharing in prison.

DISCUSSION AND CONCLUSIONS

The prevalences of injection practices and needle/syringe sharing in prisons are alarmingly high. Effective interventions to prevent the transmission of infectious diseases among people who use drugs in the prison setting are essential. The implementation of international recommendations on the principle of equivalence between prisons and the community is still very limited in most countries, and should be complemented with tailored interventions for the most vulnerable prison populations, especially Russian-speaking detainees.

Age- and time-dependent prevalence and incidence of hepatitis C virus infection in drug users in France, 2004-2011: model-based estimation from two national cross-sectional serosurveys

Hepatitis C virus (HCV) infection is a public health issue worldwide. Injecting drug use remains the major mode of transmission in developed countries. Monitoring the HCV transmission dynamic over time is crucial, especially to assess the effect of harm reduction measures in drug users (DU). Our objective was to estimate the prevalence and incidence of HCV infection in DU in France using data from a repeated cross-sectional survey conducted in 2004 and 2011. Age- and time-dependent HCV prevalence was estimated through logistic regression models adjusted for HIV serostatus or injecting practices. HCV incidence was estimated from a mathematical model linking prevalence and incidence. HCV prevalence decreased from 58·2% [95% confidence interval (CI) 49·7-66·8] in 2004 to 43·2% (95% CI 38·8-47·7) in 2011. HCV incidence decreased from 7·9/100 person-years (95% CI 6·4-9·4) in 2004 to 4·4/100 person-years (95% CI 3·3-5·9) in 2011. HCV prevalence and incidence were significantly associated with age, calendar time, HIV serostatus and injecting practices. In 2011, the highest estimated incidence was in active injecting DU (11·2/100 person-years). Given the forthcoming objective of generalizing access to new direct antiviral agents for HCV infection, our results contribute to decision-making and policy development regarding treatment scale-up and disease prevention in the DU population.

Hepatitis C virus and HIV seroprevalences, sociodemographic characteristics, behaviors and access to syringes among drug users, a comparison of geographical areas in France, ANRS-Coquelicot 2011 survey

BACKGROUND

People who use drugs (PWUDs) are at a high risk for hepatitis C virus (HCV) and human immunodeficiency virus (HIV), but they have different characteristics depending on the local context. In France, seroprevalence, sociodemographic, and behavior information have only been studied at a national level rather than at a local level. The aim of this study was to describe and examine profile and drug use practice differences in seven French cities and departments and to assess whether these differences can explain HCV and HIV seroprevalence variations between French geographical areas.

METHODS

Data were collected from the cross-sectional ANRS-Coquelicot survey conducted for the second time in 2011 among drug users having injected or snorted drugs at least once in their life. Professional interviewers administrated a face-to-face questionnaire in six different areas in France: Paris, Marseille, Bordeaux, Lille, Strasbourg and the Seine-Saint-Denis department (Paris suburbs). Participants were asked to self-collect a fingerpick blood sample in order to search for the presence of anti-HIV and anti-HCV antibodies and to estimate seroprevalence in PWUDs.

RESULTS

Overall, HCV and HIV seroprevalence was 44% [95% CI: 39.6-47.9] and 10% [95% CI: 7.5-12.6] respectively. The highest HCV seroprevalence was 56% in Marseille and the lowest was 24% in Bordeaux and for HIV the highest was 18% in Seine-Saint-Denis and the lowest was 0% in Lille. The population's age differed between areas and could mostly explain HCV seroprevalence variation but not exclusively. Profiles and practices, different in each area, can also explain this variation. In multivariate analysis, HCV seroprevalence was lower in Bordeaux (prevalence ratio [PR]=0.64), Strasbourg (PR=0.76), and Seine-Saint-Denis (PR=0.8) than in Paris. Nearly one-third of injectors declared having had difficulties to obtain syringes in the 6 previous months, but disparities existed between areas.

CONCLUSION

HCV risk exposure in PWUDs remains high in France and varies between different areas. Innovative harm reduction strategies including educative programs about safe injecting and supervised consumption rooms need to be developed.

A Venue-Based Survey of Malaria, Anemia and Mobility Patterns among Migrant Farm Workers in Amhara Region, Ethiopia

BACKGROUND

Mobile populations present unique challenges to malaria control and elimination efforts. Each year, a large number of individuals travel to northwest Amhara Region, Ethiopia to seek seasonal employment on large-scale farms. Agricultural areas typically report the heaviest malaria burden within Amhara thereby placing migrants at high risk of infection. Yet little is known about these seasonal migrants and their malaria-related risk factors.

METHODS AND FINDINGS

In July 2013, a venue-based survey of 605 migrant laborers 18 years or older was conducted in two districts of North Gondar zone, Amhara. The study population was predominantly male (97.7%) and young (mean age 22.8 years). Plasmodium prevalence by rapid diagnostic test (RDT) was 12.0%; One quarter (28.3%) of individuals were anemic (hemoglobin <13 g/dl). Nearly all participants (95.6%) originated from within Amhara Region, with half (51.6%) coming from within North Gondar zone. Around half (51.2%) slept in temporary shelters, while 20.5% regularly slept outside. Only 11.9% of participants had access to a long lasting insecticidal net (LLIN). Reported net use the previous night was 8.8% overall but 74.6% among those with LLIN access. Nearly one-third (30.1%) reported having fever within the past two weeks, of whom 31.3% sought care. Cost and distance were the main reported barriers to seeking care. LLIN access (odds ratio [OR] = 0.30, P = 0.04) and malaria knowledge (OR = 0.50, P = 0.02) were significantly associated with reduced Plasmodium infection among migrants, with a similar but non-significant trend observed for reported net use the previous night (OR = 0.16, P = 0.14).

CONCLUSIONS

High prevalence of malaria and anemia were observed among a young population that originated from relatively proximate areas. Low access to care and low IRS and LLIN coverage likely place migrant workers at significant risk of malaria in this area and their return home may facilitate parasite transport to other areas. Strategies specifically tailored to migrant farm workers are needed to support malaria control and elimination activities in Ethiopia.