Adolescence and emerging adulthood are critical life stages that shape the trajectory of future health. However, adolescents and young adults are underserved by existing health services, which often fail to overcome the unique barriers to care they face. Achieving universal access to healthcare requires moving beyond service availability to understand and address the myriad interconnected structural, sociocultural, and demographic factors that determine individuals’ level of engagement with health services. Each chapter of this dissertation shares the unifying objective of informing real-world implementation strategies to improve access to essential health services among adolescents and young adults.

In Chapter 1, I present the preliminary impacts of a multifaceted loyalty program intervention on the provision of sexual and reproductive health products and health facility referrals to adolescent girls and young women (AGYW, ages 15-24 years) at drug shops in Tanzania. The intervention was evaluated through a 4-month randomized trial at 20 drug shops in 2019. By the end of the study, the rate of patronage by AGYW customers was four times higher at shops that implemented the intervention than comparison shops. Intervention shops also distributed more HIV self-test kits, contraceptives, and health facility referrals to AGYW than comparison shops over the study period. These results suggest that enhancing private sector drug shops with specialized programs designed around the needs and preferences of AGYW may be an effective strategy to reach them with sexual and reproductive health services.

In Chapter 2, I describe the findings of an integrated COVID-19 symptom and exposure monitoring and testing system piloted in a longitudinal cohort of 2,918 university students and employees in California in 2020. At baseline and endline, participants were tested for active SARS-CoV-2 infection via quantitative polymerase chain reaction (qPCR) test and provided blood samples for antibody testing. Throughout the study, participants who self-reported certain symptoms or exposures in daily surveys were automatically notified to return for additional qPCR testing. Analyses reveal that undergraduate students who were 18-19 years old were more likely than other participants to be diagnosed with SARS-CoV-2 infection via qPCR test, in part due to a superspreader event identified among those living in congregate housing. Test positivity was also higher among participants who recently reported symptoms or household exposures that triggered notifications to test than those without recent symptoms or exposures. Among participants with newly developed antibodies for SARS-CoV-2 at endline, 91% had been diagnosed with incident infection via qPCR test over the study period. These results demonstrate that integrated monitoring systems can successfully identify and link at-risk students to COVID-19 testing when combined with low-barrier entry points to services.

In Chapter 3, I assess HIV knowledge and condom-related beliefs and their associations with HIV testing and condom use in a cross-sectional sample of 6,079 Rwandan secondary school students (ages 12-19 years) surveyed in 2021. Participants’ level of HIV knowledge was high overall, with 74% answering at least 6 of 7 HIV knowledge questions correctly. In contrast, beliefs about condoms were mixed, with participants responding favorably to 4 of 8 condom-related statements on average. Most participants endorsed misconceptions about condoms or did not believe that condom use was appropriate in casual or serious relationships. In multivariable models, having high HIV knowledge was associated with recent HIV testing but not condom use. However, participants with supportive or middling condom beliefs were more likely to report condom use than those with prohibitive beliefs. These results motivate renewed attention to educational gaps about condoms’ value in prevention and school-based interventions to create supportive social norms around condom use among youth.

In summary, this dissertation provides rigorous evidence demonstrating the benefits of tailored, multipronged approaches to reduce barriers to health services experienced by adolescents and young adults. This research also underscores the importance of epidemiologic data analysis as a tool to identify disparities, evaluate impact, and enable evidence-based decision-making to reduce inequities experienced by youth and support health and well-being across the life span.

Financial incentives are increasingly used to improve the health of vulnerable populations in the context of HIV in resource-constrained settings; however, little is known about their effect on gender-based violence (GBV) among intervention beneficiaries. My dissertation explores how cash transfers and lotteries, two distinct types of financial incentive interventions, affect GBV among women living with HIV and female sex workers, two groups who experience heightened risks of violence compared to the general population. I conducted three studies among vulnerable women in Tanzania: 1) a qualitative study among partnered women living with HIV who received cash transfers through phase 1 of a two-phase randomized trial aiming to improve viral suppression among people living with HIV; I assessed perceptions of financial incentives and whether and how they affect relationship dynamics and intimate partner violence (IPV); 2) an intent-to-treat analysis using data from phase 2 of the randomized trial to estimate the impact of cash transfers on IPV among partnered women living with HIV; and 3) an intent-to-treat analysis using data from RESPECT II, a randomized trial assessing the effect of a lottery-based incentive intervention on HIV/HSV2 incidence, to determine the effect of the lottery, if any, on GBV. Chapter 1 revealed that cash transfers of $5 or $10 USD were not substantial enough to shift relationship dynamics and did not increase physical or sexual IPV, however controlling behaviors/emotional violence were exacerbated for some partnered women living with HIV. Chapter 2 demonstrated that cash transfers of $10 USD did not significantly alter risks of IPV among partnered women living with HIV. Chapter 3 found that the lottery-based incentive intervention did not significantly affect the prevalence of GBV, including non-partner violence and IPV, among female sex workers. Together, these findings suggest these types of financial incentive interventions, which are designed as small motivational nudges rather than larger social protection or poverty alleviation schemes, do not significantly change the risk of violence for women living with HIV and female sex workers when provided in the context of HIV care. However, interventions must be designed and implemented with caution and should include continuous consultations with local female beneficiaries to ensure that incentives do not pose unintended harms or adverse effects.

Increasing evidence demonstrates that short-term conditional economic incentives may improve adherence to antiretroviral therapy (ART) for people living with HIV, thereby conferring individual health benefits and preventing transmission. However, few previous studies have assessed viral suppression, a biomarker of adherence and the ultimate goal of HIV “treatment as prevention” efforts. Moreover, questions remain around pathways of impact, best practices for real-world implementation, and durability of effects. For example, there is a lack of coherent evidence indicating whether such incentives mitigate food insecurity, an underlying barrier on the hypothesized pathway between economic incentives and ART adherence. Additionally, essential details for scaling up implementation remain largely unexplored, including optimal design and distribution aspects that can affect scalability. Finally, the short follow-up period for most prior studies inhibits understanding of any long-term harms or benefits of time-limited incentives. To address these gaps, this work studied the effects of cash and food incentives in two randomized trials of adults initiating ART in rural Tanzania. It evaluated outcomes including food insecurity, viral suppression, and long-term retention in care, and investigated the roles of incentive size and delivery mode.

Chapter 1 assessed effects of short-term cash and food incentives for HIV care on food security, nutrition, and livelihoods. This analysis used data from a 2013-2016 study of 800 food-insecure ART initiates at three HIV primary care clinics in Shinyanga region, Tanzania. Participants were randomized to receive usual HIV care (control group) or to additionally receive cash or food transfers for up to 6 consecutive months, conditional on timely attendance at scheduled clinic visits. The primary study results demonstrated that both cash and food incentives increased medication possession and retention in care at 6 and 12 months compared to the control group. The analysis herein found that food security, nutritional status, and work status improved over time in all groups, potentially due to the benefits of ART on physical health and the ability to work. Incentives further reduced severe food insecurity at 6 months relative to usual care but had no effects on these other measures at 6 or 12 months. These results suggest that alleviating severe short-term food insecurity via a modest income effect is one plausible pathway for incentives to bolster ART adherence. The lack of impact on other outcomes indicates that small incentives likely operate primarily via a price effect, lowering costs associated with clinic attendance. Moreover, the overall improvements in well-being after starting ART underscore the importance of strategies to support adherence both for ending the HIV epidemic and strengthening livelihoods.

Chapter 2 evaluated impacts of two different sized financial incentives for HIV care on viral suppression. A randomized trial was conducted at four health facilities in Shinyanga region, Tanzania (2018-2019), whereby 530 adult ART initiates were randomized to receive usual HIV care or to additionally receive a financial incentive for monthly clinic attendance in the amount of 10000 (US $4.50) or 22500 (US $10) Tanzanian Shillings for up to 6 months. Mobile health technology (mHealth) was used to implement biometric attendance monitoring linked to automated electronic payments. The study found that both incentive amounts improved viral suppression at 6 months, with a trend toward larger effects with increasing incentive size. As the first randomized trial to evaluate the impact of clinic-based financial incentives for treatment-seeking behavior on HIV viral suppression in a low- or middle- income country, these findings contribute critical evidence for understanding the promise of financial incentives to improve HIV treatment adherence.

Chapter 3 ascertained the long-term effects of time-limited incentives for HIV care. Individuals who had participated in the previous study of cash and food incentives in Tanzania were followed up after 3 years had elapsed since enrollment in the original trial. These former participants were located using gold standard tracing procedures including phone calls, home visits from community health workers, and triangulation with other facilities. Clinic attendance records obtained at follow-up were used to measure retention in care and mortality at 24 and 36 months. Contributing to the scientific gap regarding long-run incentive effects, the results showed no significant differences between study groups in retention or mortality, indicating neither lasting benefit nor eventual harm from short-term incentives.

Together, these findings contribute an improved understanding of the effectiveness and optimal implementation of conditional economic incentives to promote HIV treatment adherence. The short-term effects on food insecurity show that incentives may help stabilize vulnerable households at the time of ART initiation, thereby supporting improved adherence. Next, the viral suppression results definitively show that incentives have impact on a biological measure of ART adherence, and that these effects can be achieved using mobile health technology to deliver incentives. A trend toward increasing effectiveness with larger incentives was established, while even the smallest incentive tested resulted in significant improvement over the standard of care. Lastly, the long-term results found no eventual harm from incentives, while benefits gradually reduced over time. These findings demonstrate that economic incentives are a safe and effective strategy to promote adherence at the time of ART initiation, adding a major contribution to evidence-based approaches to ending the HIV epidemic.

Despite global advances in the treatment and prevention of HIV, adolescent girls and young women (AGYW, aged 15-24) remain vulnerable to HIV acquisition in sub-Saharan Africa (SSA). There, AGYW outnumber their male peers in terms of HIV incidence three to one, and account for nearly two-thirds of all incident HIV infections. Pre-exposure prophylaxis (PrEP), which is 99% effective at preventing HIV, could contribute to reducing the burden of HIV among AGYW if used at scale and effectively. Nevertheless, awareness of PrEP in high HIV burden settings—especially among populations most vulnerable to HIV acquisition such as AGYW—remains unacceptably low and presents a primary barrier for effective PrEP coverage. Compounding this low awareness of PrEP are three interrelated drivers of behaviors that put AGYW at risk of adverse sexual and reproductive health (SRH) outcomes: 1) inequitable gender views, 2) poor sexual and reproductive empowerment, and 3) low reproductive autonomy. These three drivers collectively form a set of documented upstream determinants to poor sexual health outcomes, including reduced engagement in preventive health care, such as condom, contraceptive, and/or PrEP use. This dissertation is thus motivated by the overarching question: “How might AGYW-centered interventions improve the realization of SRH intentions and subsequent SRH outcomes among AGYW in east and southern Africa?”. To address this question, I use data from studies conducted in three countries and designed in collaboration with AGYW and with the purpose of understanding how to reduce inequitable gender views, improve sexual and reproductive empowerment, and increase reproductive autonomy as intermediaries to larger, downstream health behaviors (Chapters 1 and 3). For Chapter 2, I conduct primary data collection for a discrete choice experiment to understand AGYW’s preferences for potential HIV prevention implementation models, specifically for PrEP.

In Chapter 1, I leverage data from a Type 2 Effectiveness-Implementation cluster randomized controlled trial (CRCT) of CyberRwanda, a digital SRH intervention for Rwandan youth, to examine the intervention’s impact on: 1) inequitable equitable views, and 2) reproductive autonomy. Both inequitable gender views and reproductive autonomy are upstream determinants of a host of SRH outcomes, including contraceptive uptake, experiencing intimate partner violence, and unintended pregnancy, highlighting the potential impact that improving these domains may have on long-term AGYW health and wellbeing. To evaluate the intervention’s impact on these two outcomes, sixty secondary schools in eight districts in Rwanda were randomized 1:1:1 across three arms: two CyberRwanda implementation models (self-service, with unstructured tablet access to the digital platform; or facilitated, which included both unstructured and structured tablet access plus CyberRwanda clubs led by peer facilitators), and a comparison arm. Leveraging data from 2,955 AGYW at 12-months post intervention start and compared to AGYW at comparison schools, AGYW from schools assigned to the CyberRwanda facilitated model experienced small improvements in inequitable gender views and reproductive autonomy; no effect was observed among AGYW at self-service schools. These findings suggest that facilitated digital health interventions among school-age AGYW may be a promising strategy for improving inequitable gender views and reproductive autonomy.

In Chapter 2, I report the results of a forced choice discrete choice experiment (DCE) that I designed and conducted with local partners to evaluate preferences for different PrEP modalities and delivery models for AGYW at risk for HIV in Lake Zone, Tanzania. There exist numerous barriers to uptake and retention in PrEP, compounded by low awareness, resulting in preventable HIV cases. With newer PrEP modalities (i.e. injectable, vaginal ring) under regulatory review, we are at an opportune juncture where an understanding of AGYW preferences could ensure that implementation of future PrEP modalities could be grounded in users’ preferences. We enrolled 696 AGYW who met the criteria of being recently (n=138) or currently (n=558) PrEP eligible, and evaluated five attributes: (i) PrEP modality, (ii) screening location, (iii) refill location, (iv) provider identity, and (v) cost per refill. Results show that prior to the educational module explaining PrEP (which preceded the DCE), knowledge of PrEP was low—less than 1 in 5 respondents had ever heard of PrEP. Modality of PrEP, specifically injectable PrEP, was the most preferred and influential component driving AGYW preferences. Respondents reported high levels of interest in taking up PrEP across various implementation scenarios, demonstrating that newer modalities of PrEP, such as long-acting injectables, may help close the gap between unmet need for PrEP and barriers to uptake of current PrEP modalities.

In Chapter 3, I leverage data from the ‘SKILLZ’ study, a Type 1 hybrid effectiveness-implementation CRCT evaluating the impact of SKILLZ, a peer-led, sports-based program intended to empower adolescent girls implemented by Grassroot Soccer (GRS) in Zambia. Sports-based programming is associated with delayed sexual debut and higher empowerment among AGYW, yet these relationships have not been rigorously evaluated via a CRCT. The study enrolled 2,153 AGYW across 46 secondary schools in and surrounding Lusaka, Zambia; my dissertation evaluates the impact of the intervention on sexual and reproductive empowerment, measured via the Sexual and Reproductive Empowerment (SRE) for Adolescents and Young People scale. Results show small to moderate impacts of the intervention on SRE, with these impacts magnified for those who were sexually active and those who attended more than eight of twelve sessions.

The central hypothesis linking the three chapters is that interventions developed by and for youth that reduce barriers to engaging in sexual health care will result in reduced inequitable gender views and improved sexual and reproductive empowerment and reproductive autonomy, all of which are key determinants of subsequent uptake and engagement with SRH and HIV prevention. The development of each of the interventions evaluated herein is consistent with several internationally recognized frameworks guiding AGYW-focused SRH research and intervention design, such as those developed by the Human-centered Design (HCD) Exchange and the World Health Organization. Further, the focus on upstream determinants of SRH health, such as gender views and reproductive autonomy and empowerment, aligns with World Health Organization priorities to ensure universal access to SRH to improve quality of life and reduce health inequalities. Taken together, the knowledge generated from this research may lead to a better understanding of how to address upstream determinants of poor HIV and SRH care among AGYW, as well as provide timely information to influence policy on the roll-out of novel PrEP products. Further, they may provide insight on effective programs and policies related to HIV and SRH care among AGYW among other southern- and eastern- African countries.

Background

Methods/design

Discussion/implications

Background

Methods

Findings

Interpretation

Funding

Background

Cash and in-kind incentives can improve health outcomes in various settings; however, there is concern that incentives may ‘crowd out’ intrinsic motivation to engage in beneficial behaviors. We examined this hypothesis in a randomized trial of food and cash incentives for people living with HIV infection in Tanzania.

Methods

We analyzed data from 469 individuals randomized to one of three study arms: standard of care, short-term cash transfers, or short-term food assistance. Eligible participants were: 1) ≥18 years old; 2) HIV-infected; 3) food insecure; and 4) initiated antiretroviral therapy (ART) ≤90 days before the study. Food or cash transfers, valued at ~$11 per month and conditional on attending clinic visits, were provided for ≤6 months. Intrinsic motivation was measured at baseline, 6, and 12 months using the autonomous motivation section of the Treatment Self-Regulation Questionnaire (TSRQ). We compared the change in TSRQ score from baseline to 6 and 12 months and the change within study arms.

Results

The mean intrinsic motivation score was 2.79 at baseline (range: 1–3), 2.91 at 6 months (range: 1–3), and 2.95 at 12 months (range: 2–3), which was 6 months after the incentives had ended. Among all patients, the intrinsic motivation score increased by 0.13 points at 6 months (95% CI (0.09, 0.17), Cohen’s d = 0.29) and 0.19 points at 12 months (95% CI (0.14, 0.24), Cohen’s d = 0.49). Intrinsic motivation also increased within each study group at 6 months: 0.15 points in the food arm (95% CI (0.09, 0.21), Cohen’s d = 0.37), 0.11 points in the cash arm (95% CI (0.05, 0.18), Cohen’s d = 0.25), and 0.08 points in the comparison arm (95% CI (-0.03, 0.19), Cohen’s d = 0.21); findings were similar at 12 months. Increases in motivation were statistically similar between arms at 6 and 12 months.

Conclusion

Intrinsic motivation for ART adherence increased significantly both overall and within the food and cash incentive arms, even after the incentive period was over. Increases in motivation did not differ by study group. These results suggest that incentive interventions for treatment adherence should not be withheld due to concerns of crowding out intrinsic motivation.

Objective

Design

Methods

Results

Conclusion