Programme science: a route to effective coverage and population-level impact for HIV and sexually transmitted infection prevention.

Improvements in context-specific programming are essential to address HIV and other sexually transmitted and blood-borne infection epidemics globally. A programme science approach emphasises the need for context-specific evidence and knowledge, generated on an ongoing basis, to inform timely and appropriate programmatic decisions. We aim to accelerate and improve the use of embedded research, inquiry, and learning to optimise population-level impact of public health programmes and to introduce an effective programme coverage framework as one tool to facilitate this goal. The framework was developed in partnership with public health experts in HIV and sexually transmitted and blood-borne infections through several workshops and meetings. The framework is a practice-based tool that centres on the use of data from iterative cycles of programme-embedded research and learning, as well as routine programme monitoring, to refine the strategy and implementation of a programme. This programme science approach aims to reduce programme coverage gaps, to optimise impact at the population level, and to achieve effective coverage. This framework should facilitate the generation of programme-embedded research and learning agendas to inform resource allocation, optimise population-level impact, and achieve equitable and effective programme coverage.

Assessing awareness and use of HIV self-testing kits after the introduction of a community-based HIV self-testing programme among men who have sex with men in Kenya.

Men who have sex with men (MSM) bear a disproportionate burden of new HIV infections in Kenya, while experiencing discrimination, leading to suboptimal levels of HIV care. HIV self-testing (HIVST) is a tool to increase HIV screening and earlier diagnosis; however, questions remain regarding how best to scale-up HIVST to MSM in Kenya. The main objective of this study was to examine changes in knowledge and use of HIVST after implementation of a community-led HIVST project. Participants were MSM recruited from Kisumu, Mombasa, and Kiambu counties. Data were collected from two rounds (Round 1: 2019; Round 2: 2020) of serial cross-sectional integrated biological and behavioural assessments (IBBA), pre-, and post-project implementation. Two main outcomes were measured: 1) whether the respondent had ever heard of HIVST; and 2) whether they had ever used HIVST kits. Changes in outcomes between IBBA rounds were examined using modified multivariable Poisson regression models; adjusted prevalence ratios (aPR) and 95% confidence intervals (95% CI) are reported. A total of 2,328 respondents were included in main analyses. The proportion of respondents who had heard of HIVST increased from 75% in Round 1 to 94% in Round 2 (aPR: 1.2, 95% CI: 1.2-1.3), while those reporting using an HIVST kit increased from 20% to 53% (aPR: 2.3, 95% CI: 2.0-2.6). Higher levels of education and HIV programme awareness were associated with both outcomes. Awareness and use of HIVST kits increased after implementation of a community-led HIVST implementation project, demonstrating the importance of integration with existing community groups.

Risk factors of HIV and variation in access to clean needles among people who inject drugs in Pakistan.

We identified key risk factors for HIV among people who inject drugs (PWID) in Pakistan and explored access to free clean needles. Multivariable logistic regression was used to investigate associations between HIV prevalence and demographic, behavioral, and socio-economic characteristics of PWID. Data came from the Government of Pakistan's Integrated Biological and Behavioral Surveillance (IBBS) Round 5 (2016-17; 14 cities). A secondary analysis investigated associations with reported access to clean needles. Unweighted HIV prevalence among 4,062 PWID (99% male) was 21.0%. Longer injecting duration (Odds ratio [OR] 1.06 [95% confidence interval: 1.02-1.10]; per year), higher injecting frequency (OR 1.67 [1.30-2.13]; per unit increase), and injecting heroin (OR 1.90 [1.11-3.25]) were positively associated with HIV prevalence. There was no association between using a used syringe at last injection and HIV. Having>10 years of education had lower odds of HIV than being illiterate (OR 0.58 [0.35-0.95]). Having a regular sexual partner (OR 0.74 [0.57-0.97]) or paying for sex with the opposite sex (OR = 0.62 [0.45-0.85]) had lower odds of HIV than not. Conversely, PWID paying a man/hijra for sex had higher odds of HIV (OR 1.20 [1.00-1.43]). Receipt of clean needles varied by city of residence (0-97% coverage), whilst PWID with knowledge of HIV service delivery programs had higher odds of receiving clean needles (OR 4.58 [3.50-5.99]). Injecting behaviors were associated with HIV prevalence among PWID, though risks related to paying for sex remain complicated. Geographical variation in access to clean needles suggests potential benefits of more widely spread public health services.

Key MessagesWhat is already known on this topicThe HIV epidemic in Pakistan is concentrated among key populations including people who inject drugs.What this study addsInjecting practices, sexual behaviors, and socio-economic factors are associated with HIV prevalence among people who inject drugs. Access to harm reduction services is varied in Pakistan.How this study might affect research, practice, or policyAccess to clean free needles, as well as service delivery programs, with a broad geographical reach remain important to curb the HIV epidemic among people who inject drugs in Pakistan.

Healthcare utilization among persons living with HIV in Manitoba, Canada, prior to HIV diagnosis: A case-control analysis.

BACKGROUND: Understanding care patterns of persons living with HIV prior to diagnosis can inform prevention opportunities, earlier diagnosis, and engagement strategies. We examined healthcare utilization among HIV-positive individuals and compared them to HIV-negative controls. METHODS: Data were from a retrospective cohort from Manitoba, Canada. Participants included individuals living with HIV presenting to care between 2007 and 2011, and HIV-negative controls, matched (1:5) by age, sex, and region. Data from population-based administrative databases included physician visits, hospitalizations, drug dispensation, and chlamydia and gonorrhea testing. Diagnoses associated with physician visits were classified according to International Classification of Diseases chapters. Conditional logistic regression models were used to compare cases/controls, with adjusted odds ratios (AORs) and their 95% confidence intervals (95% CI) reported. RESULTS: A total of 193 cases and 965 controls were included. Physician visits and hospitalizations were higher for cases, compared to controls. In the 2 years prior to case date, cases were more likely to be diagnosed with "blood disorders" (AOR: 4.2, 95% CI: 2.0-9.0), be treated for mood disorders (AOR: 2.4, 95% CI: 1.6-3.4), and to have 1+ visits to a hospital (AOR: 2.2, 95% CI: 1.4-3.6). CONCLUSION: Opportunities exist for prevention, screening, and earlier diagnosis. There is a need for better integration of healthcare services with public health.

Elucidating Drivers for Variations in the Explosive Human Immunodeficiency Virus Epidemic Among People Who Inject Drugs in Pakistan.

BACKGROUND: Pakistan's explosive human immunodeficiency virus (HIV) epidemic among people who inject drugs (PWID) varies widely across cities. We evaluated possible drivers for these variations. METHODS: Multivariable regression analyses were undertaken using data from 5 national surveys among PWID (n = 18 467; 2005-2017) to determine risk factors associated with variations in city-level HIV prevalence. A dynamic HIV model was used to estimate the population-attributable fraction (PAF; proportion of HIV infections prevented over 10 years when that risk factor is removed) of these risk factors to HIV transmission and impact on HIV incidence of reducing their prevalence. RESULTS: Regression analyses suggested that city-level HIV prevalence is strongly associated with the prevalence of using professional injectors at last injection, heroin use in last month, and injecting ≥4 times per day. Through calibrating a model to these associations, we estimate that the 10-year PAFs of using professional injectors, heroin use, and frequent injecting are 45.3% (95% uncertainty interval [UI], 4.3%-79.7%), 45.9% (95% UI, 8.1%-78.4%), and 22.2% (95% UI, 2.0%-58.4%), respectively. Reducing to lowest city-level prevalences of using professional injectors (2.8%; median 89.9% reduction), heroin use (0.9%; median 91.2% reduction), and frequent injecting (0.1%; median 91.8% reduction) in 2020 reduces overall HIV incidence by 52.7% (95% UI, 6.1%-82.0%), 53.0% (95% UI, 11.3%-80.2%), and 28.1% (95% UI, 2.7%-66.6%), respectively, over 10 years. CONCLUSIONS: Interventions should focus on these risk factors to control Pakistan's explosive HIV epidemic among PWID, including a concomitant expansion of high-coverage needle/syringe provision, opioid substitution therapy, and antiretroviral therapy.

Comparing the Ecological Niches of Chlamydial and Gonococcal Infections in Winnipeg, Canada: 2007-2016.

BACKGROUND: Previous studies have shown substantial differences in geographic clustering of sexually transmitted infections (STI), such as chlamydia (CT) and gonorrhea (NG), conditional on epidemic phase. Chlamydia and NG have recently shown resurgent epidemiology in the northern hemisphere. This study describes the recent epidemiology of CT and NG in Winnipeg, Canada, combining traditional surveillance tools with place-based analyses, and comparing the ecological niches of CT and NG, in the context of their evolving epidemiology. METHODS: Data were collected as part of routine public health surveillance between 2007 and 2016. Secular trends for CT and NG, and CT/NG coinfection were examined. Gini coefficients and population attributable fractions explored the distribution, and concentration of infections over time and space. RESULTS: Rates of CT increased from 394.9/100,000 population to 476.2/100,000 population from 2007 to 2016. Gonorrhea rates increased from 78.0/100,000 population to 143.5/100,000 population during the same period. Each pathogen had its own ecological niche: CT was widespread geographically and socio-demographically, while NG was clustered in Winnipeg's inner-core. CT/NG co-infections had the narrowest space and age distribution. NG was shown to be undergoing a growth phase, with clear signs of geographic dispersion. The expansion of NG resembled the geographic distribution of CT. CONCLUSIONS: We demonstrated that NG was experiencing a growth phase, confirming theoretical predictions of geographic dispersion during a growth phase. During this phase, NG occupied similar geographic spaces as CT. Knowledge of different ecological niches could lead to better targeting of resources for subpopulations vulnerable to STIs.

Leaving no one behind? An equity analysis of the HIV care cascade among a cohort of people living with HIV in Manitoba, Canada.

BACKGROUND: Manitoba is a central Canadian province with annual rates of new HIV infections consistently higher than the Canadian average. National surveillance statistics and data from the provincial HIV care program suggest that epidemiological heterogeneity exists across Manitoba. New HIV cases are disproportionately reported among females, Indigenous-identifying individuals, and those with a history of injection drug use. Given the heterogeneity in acquisition, it is of interest to understand whether this translates into inequalities in HIV care across Manitoba. METHODS: A sample of 703 participants from a clinical cohort of people living with HIV in Manitoba, with data current to the end of 2017, was used to conduct cross-sectional, disaggregated analyses of the HIV care cascade to identify heterogeneity in service coverage and clinical outcomes among different groups receiving HIV care in Manitoba. Equiplots are used to identify and visualize inequalities across the cascade. Exploratory multivariable logistic regression models quantify associations between equity variables (age, sex, geography, ethnicity, immigration status, exposure category) and progression along the cascade. Adjusted odds ratios (AOR) and 95% confidence intervals (95%CI) are reported. RESULTS: Equity analyses highlight inequalities in engagement in and coverage of HIV-related health services among cohort participants. Equiplots illustrate that the proportion of participants in each cascade step is greater for those who are older, white, non-immigrants, and report no history of injection drug use. Compared to those living in Winnipeg, participants in eastern Manitoba have greater odds of achieving virologic suppression (AOR[95%CI] = 3.8[1.3-11.2]). The odds of Indigenous participants being virologically suppressed is half that of white participants (AOR[95%CI] = 0.5[0.3-0.7]), whereas African/Caribbean/Black participants are significantly less likely than white participants to be in care and retained in care (AOR[95%CI] = 0.3[0.2-0.7] and 0.4[0.2-0.9], respectively). CONCLUSIONS: Inequalities exist across the cascade for different groups of Manitobans living with HIV; equiplots are an innovative method for visualizing these inequalities. Alongside future research aiming to understand why inequalities exist across the cascade in Manitoba, our equity analyses can generate hypotheses and provide evidence to inform patient-centred care plans that meet the needs of diverse client subgroups and advocate for policy changes that facilitate more equitable HIV care across the province.

The HIV care cascade in Manitoba, Canada: Methods, measures, and estimates to meet local needs.

BACKGROUND AND OBJECTIVE: We describe the development of the first HIV care cascade for Manitoba, Canada, detailing steps taken to establish indicator definitions for each cascade step, and derive a full complement of local estimates. METHODS: Manitoba is a Canadian Prairie Province with disproportionately high annual HIV incidence. In 2013, a clinical cohort of people living with HIV was established within the primary HIV care program in Manitoba. Using cohort data from 2017, we describe the creation of a set of indicator definitions and calculate estimates for each cascade step to create the first Manitoban cascade model. RESULTS: Of the 703 cohort participants categorized as alive and diagnosed, 638 (90.8%) were in care, 606 (86.2%) retained in care, 573 (81.5%) on treatment, and 523 (74.4%) virologically suppressed. The greatest point of leakage occurred between the first and second steps; 9.3% of those alive and diagnosed in 2017 were not in care in the same calendar year. CONCLUSION: This is the first comprehensive examination of HIV clinical epidemiology in Manitoba using a cascade framework, with the potential inform programming to improve service coverage within Manitoba and significantly contribute to evidence informing provincial policies to support these efforts.

HIV-1 phylodynamic analysis among people who inject drugs in Pakistan correlates with trends in illicit opioid trade.

Pakistan is considered by the World Health Organization to currently have a "concentrated" HIV-1 epidemic due to a rapid rise in infections among people who inject drugs (PWID). Prevalence among the country's nearly 105,000 PWID is estimated to be 37.8% but has been shown to be higher in several large urban centers. A lack of public health resources, the common use of professional injectors and unsafe injection practices are believed to have fueled the outbreak. Here we evaluate the molecular characteristics of HIV-1 sequences (n = 290) from PWID in several Pakistani cities to examine transmission dynamics and the association between rates of HIV-1 transmission with regards to regional trends in opioid trafficking. Tip-to-tip (patristic) distance based phylogenetic cluster inferences and BEAST2 Bayesian phylodynamic analyses of time-stamped data were performed on HIV-1 pol sequences generated from dried blood spots collected from 1,453 PWID as part of a cross-sectional survey conducted in Pakistan during 2014/2015. Overall, subtype A1 strains were dominant (75.2%) followed by CRF02_AG (14.1%), recombinants/unassigned (7.2%), CRF35_AD (2.1%), G (1.0%) and C (0.3%). Nearly three quarters of the PWID HIV-1 sequences belonged to one of five distinct phylogenetic clusters. Just below half (44.4%) of individuals in the largest cluster (n = 118) did seek help injecting from professional injectors which was previously identified as a strong correlate of HIV-1 infection. Spikes in estimated HIV-1 effective population sizes coincided with increases in opium poppy cultivation in Afghanistan, Pakistan's western neighbor. Structured coalescent analysis was undertaken in order to investigate the spatial relationship of HIV-1 transmission among the various cities under study. In general terms, our analysis placed the city of Larkana at the center of the PWID HIV-1 epidemic in Pakistan which is consistent with previous epidemiological data.

Filling the gaps in the Peruvian care continuum for HIV-infected pregnant mothers: a case-control study in metropolitan Lima-Callao, Peru.

Mother-to-child transmission of HIV (MTCT) accounts for a significant proportion of new HIV infections in Peru. The purpose of this case-control study was to examine maternal and infant factors associated with MTCT in Peru from 2015 to 2016. For each biologically confirmed case infant, we randomly selected four birth year- and birth hospital-matched controls from five hospitals in Lima-Callao. Maternal and infant information were gathered from medical records. Simple conditional logistic regression was utilized to examine possible maternal and infant characteristics associated with MTCT. The rate of MTCT was 6.9% in 2015 and 2.7% in 2016. A total of 63 matched controls were identified for 18 cases. Protective factors included higher number of prenatal visits (odds ratio [OR]: 0.72; 95% confidence interval [CI]: 0.55-0.94, p = 0.012) and having more children (OR: 0.10, 95% CI: 0.01-0.79, p = 0.029). Risk factors included later maternal diagnosis (OR: 1.19; 95% CI: 1.06-1.34; p = 0.001) and greater viral load at the time of maternal diagnosis (OR: 1.05; 95% CI: 1.01-1.10; p = 0.022). Our study highlights the importance of targeting early and continued prenatal care as specific areas to target to prevent gaps in the HIV treatment cascade for pregnant HIV-infected women. These strategies can ensure early screening and initiation of antiretroviral therapy to reduce MTCT rates.

Potential contributions of an on-site nurse mentoring program on neonatal mortality reductions in rural Karnataka state, South India: evidence from repeat community cross-sectional surveys.

BACKGROUND: We assessed the effects of a nurse mentoring program on neonatal mortality in eight districts in India. METHODS: From 2012 to 2015, nurse mentors supported improvements in critical MNCH-related practices among health providers at primary health centres (PHCs) in northern Karnataka, South India. Baseline (n = 5240) and endline (n = 5154) surveys of randomly selected ever-married women were conducted. Neonatal mortality rates (NMR) among the last live-born children in the three years prior to each survey delivered in NM and non-NM-supported facilities were calculated and compared using survival analysis and cumulative hazard function. Mortality rates on days 1, 2-7 and 8-28 post-partum were compared. Cox survival regression analysis measured the adjusted effect on neonatal mortality of delivering in a nurse mentor supported facility. RESULTS: Overall, neonatal mortality rate in the three years preceding the baseline and endline surveys was 30.5 (95% CI 24.3-38.4) and 21.6 (95% CI 16.3-28.7) respectively. There was a substantial decline in neonatal mortality between the survey rounds among children delivered in PHCs supported by NM: 29.4 (95% CI 18.1-47.5) vs. 9.3 (95% CI 3.9-22.3) (p = 0.09). No significant declines in neonatal mortality rate were observed among children delivered in other facilities or at home. In regression analysis, among children born in nurse mentor supported facilities, the estimated hazard ratio at endline was significantly lower compared with baseline (HR: 0.23, 95% CI: 0.06-0.82, p = 0.02). CONCLUSION: The nurse mentoring program was associated with a substantial reduction in neonatal mortality. Further research is warranted to delineate whether this may be an effective strategy for reducing NMR in resource-poor settings.

Geographical Variation and Factors Associated With Inflammatory Bowel Disease in a Central Canadian Province.

BACKGROUND: We investigated temporal trends, geographical variation, and geographical risk factors for incidence of inflammatory bowel disease (IBD). METHODS: We used the University of Manitoba IBD Epidemiology Database to identify incident IBD cases diagnosed between 1990 and 2012, which were then geocoded to 296 small geographic areas (SGAs). Sociodemographic characteristics of the SGAs (proportions of immigrants, visible minorities, Indigenous people, and average household income) were obtained from the 2006 Canadian Census. The geographical variation of IBD incidence was modeled using a Bayesian spatial Poisson model. Time trends of IBD incidence were plotted using Joinpoint regression. RESULTS: The incidence of IBD decreased over the study years from 23.6 (per 100,000 population) in 1990 to 16.3 (per 100,000 population) in 2012. For both Crohn's disease (CD) and ulcerative colitis (UC), the highest incidence was in Winnipeg and the southern and central regions of Manitoba, whereas most of northern Manitoba had lower incidence. There was no effect of sociodemographic characteristics of SGAs, other than the proportion of Indigenous people, which was associated with lower IBD incidence. CONCLUSIONS: Although the incidence of IBD in Manitoba is decreasing over time, we have identified geographic areas with persistently higher IBD incidence that warrant further study for etiologic clues.

HIV prevention programme cascades: insights from HIV programme monitoring for female sex workers in Kenya.

INTRODUCTION: HIV prevention cascades have emerged as a programme management and monitoring tool that outlines the sequential steps of an HIV prevention programme. We describe the application of an HIV combination prevention programme cascade framework to monitor and improve HIV prevention interventions for female sex workers (FSWs) in Kenya. METHODS: Two data sources were analysed: (1) annual programme outcome surveys conducted using a polling booth survey methodology in 2017 among 4393 FSWs, and (2) routine programme monitoring data collected by (a) 92 implementing partners between July 2017 and June 2018, and (b) Learning Site in Mombasa (2014 to 2015) and Nairobi (2013). We present national, sub-national and implementing partner level cascades. RESULTS: At the national level, the population size estimates for FSW were 133,675 while the programme coverage targets were 174,073. Programme targets as denominator, during the period 2017 to 2018, 156,220 (90%) FSWs received peer education and contact, 148,713 (85%) received condoms and 83,053 (48%) received condoms as per their estimated need. At the outcome level, 92% of FSWs used condoms at the last sex with their client but 73% reported consistent condom use. Although 96% of FSWs had ever tested for HIV, 85% had tested in the last three months. Seventy-nine per cent of the HIV-positive FSWs were enrolled in HIV care, 73% were currently enrolled on antiretroviral therapy (ART) and 52% had attended an ART clinic in the last month. In the last six months, 48% of the FSWs had experienced police violence but 24% received violence support. National and sub-national level cascades showed proportions of FSWs lost at each step of programme implementation and variability in programme achievement. Hotspot and sub-population level cascades, presented as examples, demonstrate development and use of these cascades at the implementation level. CONCLUSIONS: HIV prevention programme cascades, drawing on multiple data sources to provide an understanding of gaps in programme outputs and outcomes, can provide powerful information for monitoring and improving HIV prevention programmes for FSWs at all levels of implementation and decision-making. Complexity of prevention programmes and the paucity of consistent data can pose a challenge to development of these cascades.

Laboratory Detection of First and Repeat Chlamydia Cases Influenced by Testing Patterns: A Population-Based Study.

OBJECTIVES: The purpose of this study was to describe and explore potential driving factors of trends in reported chlamydia infections over time in Manitoba, Canada. METHODS: Surveillance and laboratory testing data from Manitoba Health, Seniors and Active Living were analysed using SAS v9.4. Kaplan-Meier plots of time from the first to second chlamydia infection were constructed, and Cox proportional hazards regression was used to estimate the risk of second repeat chlamydia infections in males and females. RESULTS: Overall, the number of reported infections found mirrored the number of tests conducted. From 2008 to 2014, the number of first infections found among females decreased as the number of first tests conducted among females also decreased. Between 2008 and 2012, the number of repeat tests among females increased and was accompanied by an increase in the number of repeat positive results from 2009 to 2013. From 2008 to 2016, the number of repeat tests and repeat positive results increased steadily among males. CONCLUSIONS: Chlamydia infection rates consistently included a subset composed of repeat infections. The number of cases identified appears to mirror testing volumes, drawing into question incidence calculations that do not include testing volumes. SUMMARY BOX: 1) What is the current understanding of this subject? Chlamydia incidence is high in Manitoba, particularly among young women and in northern Manitoba.2) What does this report add to the literature? This report suggests that incidence calculated using case-based surveillance data alone does not provide an accurate estimate of chlamydia incidence in Manitoba and is heavily influenced by testing patterns.3) What are the implications for public health practice? In general, improving testing rates in clinical practices as well as through the provision of rapid services in non-clinical venues could result in higher screening and treatment rates. In turn, this could lead to a better understanding of true disease occurrence.

Changes in HIV prevention programme outcomes among key populations in Kenya: Data from periodic surveys.

The Kenya National AIDS and STI Control Programme (NASCOP) conducted annual polling booth surveys (PBS) in 2014 and 2015 to measure outcomes from the national HIV prevention programme for key populations (KPs), comprising behavioural, biomedical and structural interventions. KPs included female sex workers (FSWs), men who have sex with men (MSM) and people who inject drugs (PWID). We compared survey results from the first and second rounds. Comparing the second to the first round, significantly more FSWs (93% vs. 88%, p<0.001) and MSM (77% vs. 58%, p<0.001) reported condom use at last sex with a paying client, and at last anal sex among MSM (80% vs. 77%, p<0.05) and PWID (48% vs. 27%, p<0.01). However, condom use with regular partners remained low, at less than 53% for FSWs and 69% for MSM. Among PWID, there was a significant increase in use of new needles and syringes at last injection (93% vs. 88%, p<0.001), and a significant decrease in reported non-availability of clean needles (23% vs. 36%, p<0.001). The number of overdoses in the past six months reduced significantly but remained high (40% vs. 51%, p<0.001). FSWs and MSM reported significantly higher HIV testing, and in all KP groups, over 93% reported ever having been tested for HIV. Among the respondents self-reporting to have tested HIV positive (24% of FSW, 22% of MSM and 19% of PWID), 80% of FSWs, 70% of MSM, and 73% of PWID reported currently taking antiretroviral therapy (ART). While the experience of forced intercourse by partners declined among FSWs (18% vs. 22%, p<0.01) and MSM (13% vs. 17%, p<0.01), more FSWs reported violence by law enforcement personnel (49% vs. 44%, p<0.001). These findings provide valuable information on the programme's progress, and a signpost for the integrated behavioural, biomedical and structural interventions to achieve their HIV prevention targets.

Vulnerabilities at First Sex and Their Association With Lifetime Gender-Based Violence and HIV Prevalence Among Adolescent Girls and Young Women Engaged in Sex Work, Transactional Sex, and Casual Sex in Kenya.

BACKGROUND: Adolescent girls and young women (AGYW) experience high rates of HIV early in their sexual life course. We estimated the prevalence of HIV-associated vulnerabilities at first sex, and their association with lifetime gender-based violence (GBV) and HIV. METHODS: We conducted a cross-sectional biobehavioral survey among AGYW (14-24 years) in Mombasa, Kenya in 2015. We compared the prevalence of first sex vulnerabilities across AGYW who self-identified as engaging in sex work (N = 408), transactional sex (N = 177), or casual sex (N = 714) and used logistic regression to identify age-adjusted associations between first sex vulnerabilities and outcomes (GBV after first sex; HIV). RESULTS: The median age at first sex was 16 years (interquartile range 14-18). A total of 43.6% received gifts or money at first sex; 41.2% and 11.2% experienced a coerced and forced first sex, respectively. First sex vulnerabilities were generally more common among AGYW in sex work. GBV (prevalence 23.8%) and HIV (prevalence 5.6%) were associated with first sex before age 15 [GBV adjusted odds ratio (AOR) 1.4, 95% confidence interval (CI): 1.0 to 1.9; HIV AOR 1.9, 95% CI: 1.1 to 1.3]; before or within 1 year of menarche (GBV AOR 1.3, 95% CI: 1.0 to 1.7; HIV AOR 2.1, 95% CI: 1.3 to 3.6); and receipt of money (GBV AOR 1.9, 95% CI: 1.4 to 2.5; HIV AOR 2.0, 95% CI: 1.2 to 3.4). CONCLUSIONS: HIV-associated vulnerabilities begin at first sex and potentially mediate an AGYW's trajectory of risk. HIV prevention programs should include structural interventions that reach AGYW early, and screening for a history of first sex vulnerabilities could help identify AGYW at risk of ongoing GBV and HIV.

The contributions and future direction of Program Science in HIV/STI prevention.

BACKGROUND: Program Science is an iterative, multi-phase research and program framework where programs drive the scientific inquiry, and both program and science are aligned towards a collective goal of improving population health. DISCUSSION: To achieve this, Program Science involves the systematic application of theoretical and empirical knowledge to optimize the scale, quality and impact of public health programs. Program Science tools and approaches developed for strategic planning, program implementation, and program management and evaluation have been incorporated into HIV and sexually transmitted infection prevention programs in Kenya, Nigeria, India, and the United States. CONCLUSION: In this paper, we highlight key scientific contributions that emerged from the growing application of Program Science in the field of HIV and STI prevention, and conclude by proposing future directions for Program Science.

Heterogeneity in geographical trends of HIV epidemics among key populations in Pakistan: a mathematical modeling study of survey data.

BACKGROUND: Assessing patterns and trends in new infections is key to better understanding of HIV epidemics, and is best done through monitoring changes in incidence over time. In this study, we examined disparities in geographical trends of HIV epidemics among people who inject drugs (PWIDs), female sex workers (FSWs) and hijra/transgender/male sex workers (H/MSWs), in Pakistan. METHODS: The UNAIDS Estimation and Projection Package (EPP) mathematical model was used to explore geographical trends in HIV epidemics. Four rounds of mapping and surveillance data collected among key populations (KPs) across 20 cities in Pakistan between 2005-2011 was used for modeling. Empirical estimates of HIV prevalence of each KP in each city were used to fit the model to estimate prevalence and incidence over time. RESULTS: HIV incidence among PWIDs in Pakistan reached its peak in 2011, estimated at 45.3 per 1000 person-years. Incidence was projected to continue to rise from 18.9 in 2015 to 24.3 in 2020 among H/MSWs and from 3.2 in 2015 to 6.3 in 2020 among FSWs. The number of people living with HIV in Pakistan was estimated to steadily increase through at least 2020. HIV incidence peak among PWIDs ranged from 16.2 in 1997 in Quetta to 71.0 in 2010 in Faisalabad (per 1000 person-years). Incidence among H/MSWs may continue to rise through 2020 in all the cities, except in Larkana where it peaked in the early 2000s. In 2015, model estimated incidence among FSWs was 8.1 in Karachi, 6.6 in Larkana, 2.0 in Sukkur and 1.2 in Lahore (per 1000 person-years). CONCLUSIONS: There exists significant geographical heterogeneity in patterns and trends of HIV sub-epidemics in Pakistan. Focused interventions and service delivery approaches, different by KP and city, are recommended.

Key Programme Science lessons from an HIV prevention 'Learning Site' for sex workers in Mombasa, Kenya.

OBJECTIVES: In 2013, Kenya's National AIDS and STI Control Programme established a Learning Site (LS) in Mombasa County to support and strengthen capacity for HIV prevention programming within organisations working with sex workers. A defining feature of LS was the use of a Programme Science approach throughout its development and implementation. We provide an overview of the key components of LS, present findings from 23 months of programme monitoring data, and highlight key Programme Science lessons from its implementation and monitoring. METHODS: Routine monitoring data collected from September 2013 through July 2015 are presented. Individual-level service utilisation data were collected monthly and indicators of interest were analysed over time to illustrate trends in enrolment, programme coverage and service utilisation among sex workers in Mombasa County. RESULTS: Over the monitoring period, outreach programme enrolment occurred rapidly; condom distribution targets were met consistently; rates of STI screening remained high and diagnoses declined; and reporting of and response to violent incidents increased. At the same time, enrolment in LS clinics was relatively low among female sex workers, and HIV testing at LS was low among both female and male sex workers. CONCLUSION: Lessons learnt from operationalising the Programme Science framework through the Mombasa LS can inform the development and implementation of similar LS in different geographical and epidemiological contexts. Importantly, meaningful involvement of sex workers in the design, implementation and monitoring processes ensures that overall programme performance is optimised in the context of local, 'on-the-ground' realities. Additionally, learnings from LS highlight the importance of introducing enhanced monitoring and evaluations systems into complex programmes to better understand and explain programme dynamics over time.