Multilevel Determinants of HIV Prevention Among Urban Refugee Youth in Uganda: Baseline Findings from the Tushirikiane-4-Uthabiti Trial

Abstract

While urban refugee youth face HIV vulnerabilities spanning socio-ecological levels, knowledge gaps persist in HIV prevention outcomes. We conducted a baseline analysis of a cohort enrolled in Tushirikiane-4-Uthabiti, an intervention focused on HIV testing practices among urban refugee youth aged 16–24 in Kampala, Uganda (N = 330). Using regression models, we examined the societal, community, and interpersonal factors associated with condom use self-efficacy [CUSE], consistent condom use, HIV self-testing [HIVST] kit access, and recent HIV testing. Most participants were women (53.3%), with a mean age of 21.3 years (SD = 2.9). One-fifth reported consistent condom use (19.1%), over half accessed HIVST kits (56.4%), and recent HIV testing (50.6%). Statistically significant results showed that higher education (β = 0.18, p < 0.001) and financial resilience (β = 0.18, p < 0.001) were positively associated with CUSE, whereas food insecurity (β =  − 0.21, p < 0.05) was negatively associated with CUSE. Being in a casual relationship (adjusted odds ratio [aOR] 3.33, p < 0.05) and CUSE (aOR 1.12, p < 0.010) were associated with increased odds of consistent condom use. Higher education (aOR 2.45, p < 0.001), adolescent sexual and reproductive health stigma (aOR 1.16, p < 0.010), and perceived HIV stigma (aOR 1.05, p < 0.05) were associated with increased odds of recent HIVST kit access. Financial resilience (aOR 1.05, p < 0.010) and young parenthood (aOR = 2.32, p < 0.010) were associated with increased odds of recent last year HIV testing. The findings demonstrate suboptimal HIV prevention outcomes and highlight the need for tailored multilevel interventions to improve the sexual health of urban refugee youth.

Introduction

Urban refugee youth in Eastern and Southern Africa (ESA), where more than half of the world’s HIV-affected youth reside [1] and a growing population of refugees live in cities [2–4], face increasing inequities in HIV prevention [5]. In September 2025, Uganda’s refugee population reached nearly 2 million, with a quarter aged 15–24 years, and about 157,000 settled in Kampala [6], a city where the HIV prevalence rate of 11% is more than double the national average [7]. Urban centers such as Kampala offer refugees improved access to education and healthcare but also expose them to heightened HIV risk, driven by pervasive violence, increased sexually transmitted infections (STI) within informal settlements, escalating substance misuse, widespread sex work, and entrenched stigma [8–11]. Systematic reviews [5, 12, 13] show that urban refugee youth are frequently excluded from sexual and reproductive health programs, leaving major gaps in the response to their specific HIV vulnerabilities and needs. To mitigate susceptibility to HIV acquisition among urban refugee youth, contextualized evidence-based HIV prevention interventions are urgently needed.

Guided by the socio-ecological model, multilevel HIV prevention interventions that concurrently address barriers at the societal (e.g., economic insecurity), community (e.g., stigma), and interpersonal (e.g., early parenthood) levels can improve condom and HIV testing practices among youth in humanitarian settings [12, 14, 15]. At the societal level, these interventions could address widespread economic challenges in informal settlements that may prevent refugees from affording transportation to clinics to access HIV testing, HIV self-testing [HIVST] kits, or condoms [5, 16, 17]. For instance, employment is a critical economic determinant of HIV prevention uptake, and multiple studies [18, 19] have found a positive association between employment status and HIV testing. Related infrastructural issues may further exacerbate HIV-related risks. For example, a study [20] analyzing population-based data from 112,955 survey participants aged 15–59 years from Zambia, Eswatini, Lesotho, Uganda, Tanzania, and Namibia found that food insecurity was associated with both transactional and condomless sex. Similarly, a previous mixed-methods study [21] of 340 urban refugee youth in Kampala found that both food and water insecurity were associated with reduced sexual and reproductive health access (including HIV testing), and that water insecurity was linked to lower condom use self-efficacy. Education and literacy also play crucial roles in HIV prevention, with studies finding that individuals with higher education levels tend to engage more with HIV services [19]. In contrast, individuals with lower education levels exhibited lower odds of consistent condom use [22] and HIV testing compared to their higher-education counterparts [23]. Indeed, access to higher education may increase young people’s confidence and ability to make decisions regarding how and when to use condoms or test for HIV, and if using HIVST kits, their ability to correctly interpret results [24–26]. Among refugee youth specifically, other findings indicate that poor literacy acts as a barrier to engaging with HIV testing and prevention services, further emphasizing the downstream effect of robust educational infrastructure on condom and HIV testing practices [19]. The absence of tailored HIV prevention programs for urban refugee youth compounds the adverse effects of societal barriers, intensifying their vulnerability to HIV in the context of ongoing displacement and precarious living conditions.

Community-level factors, particularly stigma and discrimination, exert a profound influence on condom and HIV testing practices among youth living in humanitarian contexts [10, 11]. For instance, increased adolescent sexual and reproductive health (A-SRH)-related stigma (i.e., devaluing social, cultural, and religious norms and practices toward sexually active youth, youth seeking SRH services, and pregnant and parenting youth) is associated with lower condom use self-efficacy and inconsistent condom use [11], as well as lower HIV testing rates [10]. A recent study [11] focused on urban refugees in Uganda found that intersecting HIV-related and A-SRH stigmas were associated with reduced condom use self-efficacy and actual condom use. Additionally, multiple forms of intersecting stigmas faced by urban refugee youth (i.e., HIV-related stigma, refugee stigma, A-SRH stigma, and sex work stigma) constitute barriers to HIV testing [27]. A systematic review [28] of 10 studies enrolling 3,788 participants found that increased HIV-related stigma decreased the likelihood of HIV testing and care engagement. Therefore, to ensure access and effective utilization of HIV prevention strategies among urban refugee youth, there is a need for evidence to inform the development of community-level stigma-reduction interventions.

At the interpersonal level, growing evidence [29, 30] indicates that greater social support is positively associated with increased HIV prevention practices among youth in humanitarian contexts. For example, involving family members and peers, especially those who serve as trusted peer navigators, is linked to a higher incidence of HIV testing [29]. Power dynamics within relationships also influence condom use self-efficacy and HIV testing practices. For instance, partner communication about HIV testing is associated with an increased likelihood of testing [31]. However, there are limited quantitative studies on how relationship dynamics influence HIV prevention outcomes among urban refugee youth. Such relationship dynamics should account for parenthood or caregiving, as the presence of children within a home is also associated with improved HIV testing and prevention practices [19, 32]. For example, one study [33] suggested that childcare responsibilities may motivate parents to seek HIV testing and care, while other research indicates that parenthood can act as a barrier to accessing HIV services, particularly for single parents who may face additional financial or resource insecurities. Scant evidence exists on how young parenthood influences HIV prevention among urban refugee youth, which is a key gap in tailoring interventions for this population.

Given the complexity of the obstacles experienced by urban refugee youth across all levels of the socio-ecological framework in Kampala and beyond, there is a critical need for novel interventions to overcome persistent barriers to HIV prevention and diagnosis. HIVST, recommended by the World Health Organization [34] for adolescents, offers a confidential, accessible, and youth-friendly approach that directly addresses challenges related to stigma, privacy, and service accessibility. This approach enables individuals to privately collect and test their own specimens, such as blood or oral fluid, and to interpret the results themselves. HIVST holds substantial promise for improving HIV status awareness and the uptake of prevention services among urban refugee youth [34]. Encouragingly, a global review [35] found that HIVST implementation not only doubled HIV testing uptake compared to facility-based testing but also increased service providers’ capacity to reach priority populations, including young people aged 16–24 years. In a recent pragmatic controlled trial [26] examining the effectiveness of HIVST with digital support in increasing HIV testing uptake and HIV status knowledge among refugee youth aged 16–24 in Kampala, Uganda. The researchers found that HIVST was highly acceptable among urban refugee youth, with intervention groups exhibiting an uptake rate exceeding 90%. In another study [36] evaluating the effectiveness of combining accessing HIVST kits with 'edutainment' comics among rural settlement-based refugees, findings showed that the intervention increased HIV testing and knowledge of HIV status among refugee youth aged 16–24. Taken together, this emerging body of evidence indicates that HIVST can be effectively implemented in urban and rural refugee contexts. However, there is a need to address multilevel factors that may contribute to the effectiveness and scalability of HIVST interventions among vulnerable youth populations, such as urban refugee youth in low- and middle-income countries, such as Uganda.

To contribute to the evidence-base, this study reports on the baseline HIV prevention outcome data from the Tushirikiane-4-Uthabiti (translation: Supporting Each Other For Resilience) trial, a multilevel intervention, described elsewhere [37]. The intervention was co-designed with and for urban refugee youth in Kampala, Uganda. In this study, we describe the participants' sociodemographic characteristics and assess the societal, community, and interpersonal factors associated with HIV prevention outcomes (i.e., condom use self-efficacy, condom use, HIVST kit access, and HIV testing) among urban refugee youth (aged 16–24 years) enrolled in the study. The findings will enable us to 1) gather insights into HIV prevention strategies and vulnerability in a cohort of urban refugee youth in Kampala, Uganda, and 2) assess the potential of addressing multilevel factors (interpersonal, community, and societal levels) to reduce barriers and increase access to and uptake of HIV prevention services.

Methods

Study Design and Setting

We analyzed baseline cohort data of urban refugee youth aged 16–24 collected between December 2023 and February 2024, from the ongoing three-arm Tushirikiane-4-Uthabiti trial. Further details regarding the interventions and study design are described elsewhere [37]. The study was conducted in five informal settlements in Kampala, grouped into three clusters based on geographic proximity: 1) Kabalagala and Kansanga, 2) Katwe and Nsambya, and 3) Rubaga. The settlements were chosen based on specific criteria: hosting a large population of urban refugees or displaced persons, similarity in socioeconomic status, healthcare access, languages, and living conditions, and a high prevalence of depressive symptoms among urban refugee youth.

Study Population and Eligibility Criteria

We recruited 330 youth (110 participants per geographic cluster) aged 16–24 for the study. The eligibility criteria included: 1) residing in one of the five selected informal settlements in Kampala at the time of the study; 2) identified as a displaced person, refugee, or having a parent(s) who were a refugee or displaced; 3) aged 16–24 years; 4) owned or had daily access to a mobile phone; and 5) spoke French, English, Kirundi, Kinyarwanda, or Swahili.

Participant Recruitment

Peer navigators facilitated participant recruitment through phone calls, in-person meetings, and WhatsApp. The peer navigators (12 in total: six young women and six young men) were experienced health and peer educators from the study communities, identified and recruited by community-based collaborators for their respected status and active involvement in their communities. We employed purposive recruitment methods, such as word-of-mouth and venue-based sampling at community events and refugee agencies to recruit participants. Data collection was conducted in collaboration with refugee (Young Africans for Integral Development [YARID]) and government (Uganda AIDS Control Program, Ministry of Health) agencies. The study received ethical approval from the University of Toronto (#37,496) and Mildmay Uganda Research Ethics Committee (#MUREC-2021–41) and was registered with the Uganda National Council for Science & Technology (#SS1021ES). The clinical trial is registered at ClinicalTrials.gov (NCT06270160).

Data Collection Procedures

All participants provided informed consent before completing the tablet-based survey. Data were collected by research assistants (RA), who self-identified as refugees and were trained by the Ministry of Health in pre- and post-test HIV counselling, using the SurveyCTO app (Dobility) in all study languages via mobile phones or tablets. The research coordinator conducted daily quality checks on the survey responses to identify any missing or inconsistent data. When such issues were identified, the RAs re-contacted the participants to resolve them, ensuring high data quality and completeness. Given the longitudinal nature of the trial [37], data collection was not anonymized or blinded to the research team. However, strict confidentiality protocols were followed: only authorized research personnel accessed identifying information, and all datasets were de-identified prior to analysis. Data were securely stored, and participant confidentiality was maintained throughout the study period [37]. Each participant received an honorarium of UGX 25,000 (approximately USD 8) for completing a 35–45-min survey.

Measurement of Variables

Outcome Variables

HIV prevention outcomes measured in this study included condom use self-efficacy, consistent condom use, HIVST kit access, and recent HIV testing. Condom use self-efficacy was assessed using the 8-item Condom Use Self-Efficacy Scale (CUSES), which measures an individual's confidence in their ability to effectively use condoms (Cronbach’s alpha = 0.89) [38, 39]. Participants rated their confidence on a Likert scale ranging from 1 (Strongly Disagree) to 5 (Strongly Agree). After reverse scoring the negatively worded items, the scores were summed to yield a total score ranging from 8 to 40. Higher scores indicate greater confidence in condom use, whereas lower scores suggest lower confidence. Consistent condom use was measured using a single item: “In the past 3 months, how often have you or your sex partner used condoms during intercourse?” The response options were all of the time, some of the time, and none of the time. For analysis purposes, the responses were dichotomized into all of the time vs. some of the time and none of the time. HIVST kits access was measured using a single question, “Have you ever received an HIV self-test?” with a yes/no response. Recent HIV testing was also assessed using a single-item question, “Have you tested for HIV infections the last one year?” with a yes/no response.

Exposure Variables

We assessed factors within the societal context (educational level, employment status, water insecurity, food insecurity, and financial resilience), community context (adolescent sexual and reproductive health (A-SRH) stigma and HIV-related stigma), and interpersonal context (relationship status and parenthood).

Societal level factors included education level (less than secondary school vs. post-secondary education) and employment status (employed vs. unemployed). Water insecurity was measured using the 12-item household water insecurity experiences scale (total score range: 0–36); a score of 12 or more is considered water insecure (Cronbach’s alpha = 0.92 in the current study) [40]. Food insecurity was assessed using a shortened version of the 6-item U.S. Household Food Security Survey Module [41], designed to measure food insecurity within households. Each item was scored as "yes = 1" or "no = 0" with affirmative responses indicating some level of food insecurity. The total score ranged from 0 to 6, with higher scores indicating greater food insecurity. A score of 0–1 indicates high or marginal food security, 2–4 represents low food security, and 5–6 signifies very low food security and financial resilience [42, 43] was measured using a 14-item scale, where participants responded on a 5-point Likert scale (1 = strongly disagree to 5 = strongly agree). Total scores were calculated after reversing negatively phrased items, with higher scores indicating greater financial resilience (Cronbach's alpha = 0.90).

Community level factors included adolescent sexual and reproductive health (A-SRH) stigma measured using Logie et al.’s 7-item ‘Sexual Activity and Pregnancy Stigma’ subscale, previously validated with urban refugee youth in Uganda [10] (Cronbach's alpha = 0.85). HIV-related stigma was assessed using Steward et al.’s 10-item perceived stigma subscale (Cronbach's alpha = 0.90), where higher scores indicated greater levels of HIV-related stigma [44]. Interpersonal level factors included relationship status (categorical: no relationship, dating one partner/married, and casual dating/multiple partners), parenthood, where participants were asked if they had children (yes = 1 or no = 0), and sexual relationship power was measured using the 15-item Relationship Control subscale of the Sexual and Relationship Power Scale (SRPS) (Cronbach's alpha = 0.90) [45]. Participants responded on a 4-point Likert scale (1 = strongly agree, 4 = strongly disagree). Total scores were calculated, with higher scores indicating greater sexual relationship power.

We assessed the influence of several potential covariates identified in the pre-existing literature. We considered the impact of sociodemographic characteristics such as age, gender, place of birth, and duration of residence in Uganda.

Data Analysis

After data collection was completed, the dataset was systematically coded, cleaned, and analyzed using Stata version 18.0 (StataCorp, College Station, TX). Descriptive statistics were used to summarize the characteristics of the youth participants and the study outcomes across the three geographic locations. For continuous data, means and standard deviations (SD) were calculated, while categorical data were summarized using frequencies and percentages, with chi-square (χ2) tests applied to assess differences between the groups. Logistic and linear regression analyses were conducted to examine the associations between independent and outcome variables. Unadjusted odds ratios (ORs) and unstandardized beta coefficients (b) were provided for bivariate analyses, while adjusted odds ratios (aORs) and adjusted standardized beta coefficients (β), along with 95% confidence intervals (CIs), were reported for multivariable analyses. Age, gender, and geographic location were included as confounders in the final multivariable models. All statistical tests were two-tailed, with a significance level of 5%. The goodness of fit for the final model were assessed using the Hosmer–Lemeshow test for the logistic regression model and the adjusted R-squared value for the linear regression model. The model's fit and parsimony were carefully evaluated to maintain an optimal balance between complexity and explanatory power.

Results

Sample Characteristics

A total of 330 participants were enrolled in the Tushirikiane-4-Uthabiti trial, consisting of 53.3% women (n = 176) with a mean age of 21.3 years (SD = 2.9). Most participants were from the Democratic Republic of Congo (n = 257, 77.9%) and had lived in Uganda for six or more years (n = 250, 76.4%). Regarding education, approximately three-fifths (n = 205, 62.1%) had post-secondary education, and roughly two-thirds of the participants were unemployed (n = 227, 69.8%) (Table 1).

Table 1.

Baseline demographic characteristics of urban refugee youth living in the slums of Kampala, Uganda stratified by geographic location (N = 330)

Participants’ Characteristics	Full sample (N = 330)	Kansanga/Kabalagala (n = 110)	Rubaga (n = 110)	Nsambya/Katwe (n = 110)
	N (%)/ Mean (SD)	n (%)/ Mean (SD)	n (%)/ Mean (SD)	n (%)/ Mean (SD)
Socio-demographic variables
Age (missing: n = 2)	21.3 (2.9)	21.1(2.6)	21.6(3.1)	21.1(3.0)
Gender
Women	176(53.3%)	58(52.7%)	64(58.2%)	54(49.1%)
Men	174(46.7%)	52(47.3%)	46 (41.8%)	56(50.9%)
Country of Birth
Democratic Republic of Congo	257 (77.9%)	90 (81.8%)	61 (55.4%)	106 (96.4%)
Burundi	31 (9.4%)	0 (0%)	29 (26.4%)	2 (1.8%)
Others	42 (12.7%)	20 (18.2%)	20 (18.2%)	2 (1.8%)
Length of time in Uganda
< 5 years	78 (23.6%)	28 (25.5%)	24 (21.8%)	26 (23.6%)
6–10 years	148 (44.8%)	43 (39.1%)	46 (41.8%)	59 (53.7%)
> 10 years	104 (31.5%)	39 (35.4%)	40 (36.4%)	25 (22.7%)
Societal factors
Education level
Less than secondary	125 (37.9%)	33(30%)	37(33.6%)	55(50.0%)
Post-secondary	205 (62.1%)	77(70%)	73(66.4%)	55(50.0%)
Employment Status (missing: n = 5)
Full-time employed	37 (11.4%)	8 (7.5%)	17 (15.6%)	12 (11.0%)
Part-time employed	61 (18.8%)	20 (18.7%)	21 (19.3%)	20 (18.4%)
Unemployed	227 (69.8%)	79 (73.8%)	71 (65.1%)	77 (70.6%)
Water insecurity (missing: n = 11)
water secure	139 (43.6%)	35 (33.3%)	66 (61.1%)	38 (35.8%)
water insecure	180 (56.4%)	70 (66.7%)	42 (38.9%)	68 (64.2%)
Food security (missing: n = 3)
Food secure	28 (8.6%)	4 (3.7%)	13 (11.8%)	11(10.1%)
Food insecure	299 (91.4%)	104 (96.3%)	97 (88.2%)	98 (89.9%)
Financial resilience (missing: n = 13)	38.98 (8.28)	40.0 (7.4)	36.1 (10.3)	40.9 (5.5)
Community level factors
A-SRH Stigma (missing: n = 12)	4.93 (2.28)	4 (2.45)	6.39 (1.15)	4.36 (2.24)
HIV stigma attitude (missing: n = 21)	22.7(6.1)	20.7(6.7)	25.3(5.9)	21.8(4.6)
Interpersonal level factors
Relationship Status
No current partner	127 (38.5%)	39 (35.4%)	49 (44.6%)	39 (35.5%)
Dating one partner/married	159 (48.2%)	56 (50.9%)	35 (31.8%)	68 (61.8%)
Casual dating/multiple partners	44 (13.3%)	15 (13.6%)	26(23.6%)	3 (2.7%)
Parenthood
No	270 (81.8%)	98(89.1%)	78 (70.9%)	94 (85.4%)
Yes	60 (18.2%)	12(10.9%)	32 (29.1%)	16 (14.6%)
Sexual relationship power (SRP) (missing: n = 32)	48.22 (7.85)	48.03(8.28)	47.17(8.24)	49.76 (6.58)
Study Outcomes
Condom use self-efficacy (missing: n = 21)	28.34(6.48)	26.4(7.67)	29.32(6.37)	29.36 (4.42)
Consistent condom use
No
Yes	123 (58.0%)	44 (63.8)	40 (56.3)	39 (54.2)
HIV testing (missing: n = 4)
No	161 (49.4%)	49 (45.8%)	41 (37.3%)	71 (65.1%)
Yes	165 (50.6%)	58 (54.21%)	69 (62.7%)	38 (34.9%)
HIV self testing (missing: n = 2)
No	143 (43.6%)	66 (60.0%)	28 (25.4%)	49 (45.4%)
Yes	185 (56.4%)	44 (40.0%)	82 (74.6%)	59 (54.6%)

SD standard deviation, n missing case

Multilevel Factors Associated with Condom Practices

Condom Use Self-efficacy

The mean score of condom use self-efficacy (CUSE) was 28.34 (SD = 6.48) on a scale ranging from 8 to 40, with significant differences observed among geographic locations. Table 2 shows that after adjusting for age, gender, and geographic location, societal factors associated with CUSE: participants with secondary education or higher had significantly higher CUSE than those with less than secondary education (β = 0.18, 95% CI: [0.08, 0.29]; p = 0.001). Conversely, low food security (vs. high/marginal food security) was associated with a lower CUSE (β =  − 0.21, 95% CI: [− 0.30, − 0.11]; p = 0.039). Greater financial resilience was positively associated with higher CUSE scores (β = 0.18, 95% CI: [0.07, 0.17]; p = 0.001).

Table 2.

Multilevel factors associated with condom use self efficacy among urban refugee youth living the slums of Kampala, Uganda (N = 330)

Variables	Condom self-efficacy
	Unadjusted b coefficient (95% CI)	P-value	Adjusted β coefficient (95% CI)	P-value
Sociodemographic factors
Age	0.19 (0.13, 0.51)	0.001	0.16 (0.05, 0.27)	0.004
Gender (ref Women)
Cisgender Men	0.22 (0.16, 0.27)	< 0.001	0.20 (0.09, 0.30)	< 0.001
Place of Birth (ref Democratic Republic of Congo)
Burundi	0.23 (0.20, 0.26)	< 0.001	0.22 (0.12, 0.33)	< 0.001
Others	0.14 (0.10, 0.18)	0.013	0.15 (0.04,0.26)	0.006
Length of time in Uganda (Ref. < 5 years)
6–10 Yrs	0.01 (− 0.07, 0.08)	0.938
> 10 Yrs	− 0.07 (− 0.14, − 0.01)	0.325
Interpersonal factors
Relationship Status (ref No current partner)
Dating one partner/married	0.10 (0.60, 0.16)	0.116	0.04 (− 0.08, 0.15)	0.561
Casual dating/multiple partners	0.19 (0.15, 0.23)	0.002	0.18 (0.06, 0.291)	0.003
Parenthood (ref No)
Yes	0.04 (− 0.01, 0.08)	0.519
Sexual Relationship Power	0.04 (− 0.86, 0.95)	0.464
Community factors
SRH Stigma	− 0.01 (− 0.27, 0.25)	0.877
HIV Stigma	0.07 (− 0.64, 0.77)	0.254
Societal factors
Education (ref. Less than secondary)
Post-secondary	0.16 (0.11, 0.22)	0.004	0.18 (0.08, 0.29)	0.001
Employment (ref. Full-time employed)
Part-time employed	0.12 (0.05, 0.18)	0.162
Unemployed	− 0.13 (− 0.20, − 0.05)	0.123
Water insecurity (ref. water secure)
water insecure	− 0.09 (− 0.15, − 0.04)	0.104
Food insecurity (ref High or marginal food security)
Low food security	− 0.21 (− 0.31, − 0.11)	0.048	− 0.21 (− 0.30, − 0.11)	0.039
Very low food security	− 0.15 (− 0.25, − 0.05)	0.148	− 0.11 (− 0.21, − 0.01)	0.289
Financial resilience	0.19 (0.74, 1.12)	0.001	0.18 (0.70, 0.17)	0.001

β-beta coefficient, OR-odds ratio; CI, confidence interval; ref, reference group(s); Adjusted standardized beta coefficient- adjusted for geographic location, age and gender

Bold are socio-ecological factors that are significant

Consistent Condom Use

Overall, only one-fifth of the study participants (n = 39, 19.1%) reported consistent condom use, with significant differences observed between the geographic locations. As illustrated in Table 3, in multivariable analyses, interpersonal factors, such as youth engaged in casual dating or having multiple partners, had higher odds of consistent condom use (aOR = 3.33, 95% CI: [1.12, 9.90]; p = 0.030). Higher CUSE scores were also significantly associated with consistent condom use (aOR = 1.12, 95% CI: [1.02, 1.22]; p = 0.011) (Table 3).

Table 3.

Multilevel factors associated with consistent condom use among refugee and displaced youth participants in Kampala, Uganda (N = 212)

Variables	Consistent condom use
	Unadjusted Odd ratio (95% CI)	P-value	Adjusted Odd ratio (95% CI)	P-value
Sociodemographic factors
Age	0.90 (0.81,1.00)	0.061	0.89 (0.79, 0.99)	0.040
Gender (ref Women)
Cisgender Men	2.15 (1.23, 3.74)	0.007	2.35 (1.32, 4.16)	0.004
Place of Birth (ref Democratic Republic of Congo)
Burundi	1.81 (0.61,5.38)	0.284
Others	0.99 (0.45,2.16)	0.975
Length of time in Uganda (ref. < 5 years)
6–10 Yrs	0.88 (0.44, 1.76)	0.716
> 10 Yrs	0.58 (0.27,1.22)	0.148
Interpersonal factors
Relationship Status (ref no current partner)
Dating one partner/married	1.23 (0.64,2.33)	0.533	1.45 (0.74, 2.86)	0.280
Casual dating/multiple partners	2.32 (0.94, 5.75)	0.069	2.50 (0.97, 6.47)	0.059
Parenthood (ref No)
Yes	0.39(0.21, 0.72)	0.003	0.55 (0.28, 1.09)	0.088
Sexual relationship power (SRP)	1.02 (0.99, 1.06)	0.213
Condom use self efficacy	1.08 (1.03, 1.14)	0.003	1.08 (1.02, 1.14)	0.005
Community factors
A-SRH Stigma	1.05 (0.93, 1.18)	0.427
HIV stigma	1.01 (0.96, 1.05)	0.832
Societal factors
Education (ref Less than secondary)
Post-secondary	0.69 (0.39,1.23)	0.212
Employment (ref. full-time employed)
Part-time employed	0.71(0.27, 1.91)	0.503
Unemployed	0.88(0.37, 2.07)	0.765
Financial resilience	0.99 (0.96, 1.02)	0.657

OR-odds ratio; CI, confidence interval; ref, reference group(s); Adjusted OR- adjusted for geographic location, age and gender

Bold are socio-ecological factors that are significant

Multilevel Factors Associated with HIV Testing Practices

Access to HIV Self-test Kits

Access to HIV self-testing kits among participants was 56.4% (n = 185), with significant variability across the study locations (p < 0.001). As shown in Table 4, bivariate analyses revealed that societal (i.e., education) and community (i.e., stigma) factors were positively associated with HIVST kits access. In multivariable analyses, we found that among societal factors, having post-secondary education (vs. less than secondary education) was significantly associated with increased odds of accessing HIVST kits (aOR = 2.45, 95% CI: [1.44, 4.17], p = 0.001). Community level factors of stigma-related factors also played a significant role: participants experiencing higher A-SRH stigma (aOR = 1.16, 95% CI: [1.03, 1.29], p = 0.011) and higher perceived HIV stigma (aOR = 1.05, 95% CI: [1.01, 1.10], p = 0.025) were also significantly associated with increased odds of accessing HIVST kits.

Table 4.

Multilevel factors associated with lifetime HIV self-testing kits access among displaced youth living the slums of Kampala, Uganda (N = 330)

Variables	Lifetime HIV self test kits access
	Unadjusted Odd ratio (95% CI)	P-value	Adjusted Odd ratio (95% CI)	P-value
Sociodemographic factors
Age	1.38 (1.26, 1.52)	< 0.001	1.40 (1.27, 1.54)	< 0.001
Gender (ref Women)
Cisgender Men	1.40 (0.90, 2.17)	0.135	1.18 (0.72, 1.92)	0.506
Place of birth (ref Democratic Republic of Congo)
Burundi	6.80 (2.31, 20.00)	< 0.001	6.09 (1.90, 19.46)	0.002
Others	2.84 (1.37, 5.90)	0.005	2.98 (1.33, 6.67)	0.008
Length of time in Uganda (ref. < 5 years)
6–10 Yrs	2.84 (1.61, 5.04)	< 0.001	2.77 (1.46, 5.23)	0.002
> 10 Yrs	3.04 (1.65, 5.61)	< 0.001	2.98 (1.49, 5.94)	0.002
Interpersonal factors
Relationship Status (ref no current partner)
Dating one partner/married	1.48 (0.92, 2.37)	0.105	0.86 (0.50, 1.48)	0.596
Casual dating/multiple partners	2.21 (1.07, 4.56)	0.032	1.56 (0.70, 3.49)	0.278
Parenthood (ref No)
Yes	2.04 (1.12, 3.72)	0.020	1.25 (0.62, 2.51)	0.534
Sexual relationship power (SRP)	0.99 (0.96, 1.02)	0.729
Ever had sex (ref No)
Yes	2.04 (1.29, 3.20)	0.002	0.77 (0.44, 1.36)	0.367
Community factors
SRH Stigma	1.10 (0.99, 1.21)	0.059	1.16 (1.03, 1.29)	0.011
HIV stigma	1.06 (1.02, 1.10)	0.003	1.05 (1.01, 1.10)	0.025
Societal factors
Education (ref less than secondary)
Post-secondary	2.52 (1.60, 3.99)	< 0.001	2.45 (1.44, 4.17)	0.001
Employment (ref. full-time employed)
Part-time employed	0.80 (0.32, 1.98)	0.628	0.68 (0.25, 1.84)	0.448
Unemployed	0.38 (0.17, 0.82)	0.013	0.62 (0.26, 1.46)	0.270
Water insecurity (ref. water secure)
Water insecure	0.64 (0.41,1.01)	0.056	0.65 (0.39, 1.08)	0.098
Food insecurity frequency High or marginal food security
Low food security	1.40 (0.58, 3.34)	0.452
Very low food security	0.46 (0.20, 1.04)	0.062
Financial resilience	1.01 (0.98, 1.03)	0.606

OR-odds ratio; CI, confidence interval; Ref, reference group(s); Adjusted OR- adjusted for geographic location, age and gender

Bold are socio-ecological factors that are significant

Recent HIV Testing

Only half of the youth (n = 165, 50.6%) reported having received an HIV test in the past year, with significant differences noted across geographic locations. Factors associated with higher odds of recent HIV testing in the bivariate analyses included age, place of birth, having children, sexual history, water insecurity, education level, and employment status (Table 5). In multivariable analyses, interpersonal factors of young parenthood had over twice the odds of reporting a recent HIV test compared to non-parenthood (aOR = 2.32; 95% CI: [1.17, 4.62]; p = 0.016). The societal factor of higher financial resilience was significantly associated with increased odds of recent HIV testing (aOR = 1.05; 95% CI: [1.02, 1.08]; p = 0.003).

Table 5.

Multilevel factors associated with recent HIV testing among displaced youth living in the slums of Kampala, Uganda (n = 330)

Variables	Recent HIV Test
	Unadjusted Odd ratio (95% CI)	P-value	Adjusted Odd ratio (95% CI)	P-value
Sociodemographic factors
Age	1.30 (1.20, 1.42)	< 0.001	1.31 (1.20, 1.43)	< 0.001
Gender (ref Women)	1.00 (0.65, 1.55)	0.988	0.85 (0.52, 1.37)	0.498
Cisgender Men
Place of Birth (ref Democratic Republic of Congo)
Burundi	2.56 (1.16, 5.65)	0.020	1.93 (0.83, 4.51)	0.126
Others	3.21 (1.54, 6.70)	0.002	2.47 (1.11, 5.49)	0.026
Length of time in Uganda (ref. < 5 years)
6–10 Yrs	1.09 (0.63, 1.90)	0.751	1.04 (0.57, 1.89)	0.901
> 10 Yrs	1.44 (0.80, 2.61)	0.227	1.17 (0.61, 2.23)	0.638
Interpersonal factors
Relationship Status (ref no current partner)
Dating one partner/married	1.38 (0.86, 2.21)	0.181	0.95 (0.56, 1.61)	0.850
Casual dating/multiple partners	1.77 (0.88,3.56)	0.106	1.07 (0.50, 2.29)	0.861
Parenthood (ref no)
Yes	3.30 (1.77, 6.13)	< 0.001	2.32 (1.17, 4.62)	0.016
Sexual relationship power (SRP)	1.00 (0.97, 1.03)	0.915	1.00 (0.97,1.03)	0.946
Ever had sex (ref no)
Yes	2.80 (1.76, 4.46)	< 0.001	1.65 (0.95, 2.86)	0.073
Community factors
A-SRH Stigma	1.04 (0.95, 1.15)	0.382	1.09 (0.98, 1.22)	0.094
HIV stigma	0.99 (0.96, 1.04)	0.956	0.99 (0.95, 1.03)	0.678
Societal factors
Education (ref less than secondary)
Post-secondary	2.06 (1.31,3.26)	0.002	1.47 (0.89,2.43)	0.129
Employment (ref full-time employed)
Part-time employed	0.96 (0.40, 2.30)	0.927	0.84 (0.33, 2.13)	0.715
Unemployed	0.37 (0.18, 0.78)	0.009	0.53 (0.24, 1.16)	0.112
Water insecurity (ref water secure)
Water insecure	0.63 (0.40, 0.98)	0.040	0.63 (0.39, 1.03)	0.066
Food insecurity (ref high or marginal food security)
Low food security	0.63 (0.27,1.45)	0.278	0.55 (0.22, 1.38)	0.206
Very low food security	0.83 (0.37, 1.85)	0.646	0.73 (0.30, 1.79)	0.496
Financial resilience	1.04 (1.01, 1.07)	0.009	1.05 (1.02, 1.08)	0.003

OR-odds ratio; CI, confidence interval; ref, reference group(s); Adjusted OR- adjusted for geographic location, age and gender

Bold are socio-ecological factors that are significant

Discussion

Guided by the socio-ecological model, this study provides a baseline analysis of HIV prevention practices and multilevel determinants among a cohort of urban refugee youth participating in the Tushirikiane-4-Uthabiti trial in Kampala, Uganda. Our findings show that both past-year HIVST kit access, HIV testing and consistent condom use remain markedly below national [19] and international targets [46], despite participants reporting moderate condom use self-efficacy. These prevention gaps may reflect COVID-19 related disruptions to health services, as well as ongoing barriers such as limited educational opportunities, food insecurity, and intersectional stigma [21, 47–49]. Societal factors, including post-secondary education and financial resilience, were associated with increased condom use self-efficacy and HIVST access, whereas food insecurity hindered these protective behaviors. At the community level, greater A-SRH and higher perceived HIV-related stigma were associated with increased HIVST access. Consistent condom use was influenced by interpersonal dynamics, particularly relationship context and negotiation skills. Taken together, these findings highlight the pressing need for multilevel, contextually tailored interventions that address intersecting vulnerabilities to improve HIV prevention outcomes among urban refugee youth.

Corroborating prior studies, we found that post-secondary education was associated with higher condom use self-efficacy [50, 51] and greater odds of accessing HIVST kits [52], highlighting a key opportunity for societal interventions that increase access to education for urban refugee youth. Lower education levels may be linked to poor literacy rates among refugee youth [19, 53], which could act as a barrier to negotiating condom use or effectively accessing and using HIVST kits. Indeed, multiple studies [50, 51] of ESA adolescents and young adults have identified an association between higher education levels and increased condom use self-efficacy. Further corroborating prior studies [24, 54], our analysis found that refugee youth with higher education levels had greater confidence in their ability to access HIVST kits. Evidence suggests that higher education is associated with increased awareness and access to information about HIV and HIVST, and potentially increased willingness to access and use HIVST kits [55]. Thus, access to higher education may bolster refugee youth’s health literacy skills, enabling them to understand and properly interpret HIVST instructions and results [56]. Prior studies [27, 57] have also shown that low literacy levels represent a key barrier to access and usage of HIVST kits, as individuals with low literacy levels may be susceptible to misinformation or misinterpretation of results. Therefore, improving literacy and, by extension, health literacy among urban refugee youth represents an opportunity to expand sexual health knowledge and skills, which, in turn, will boost their ability to access, negotiate for, and consistently and effectively use condoms and HIVST kits. Such an initiative would not only align with the goals of non-HIV-related initiatives but also leverage existing infrastructure (e.g., schools and other literacy programs) to promote HIV preventive practices among a vulnerable population. Therefore, to maximize HIVST access and uptake among urban refugee youth, new programs should use diverse distribution and knowledge access points, such as pharmacies, shopping centers, clinics/hospitals, schools, bars/clubs, sporting venues, and sexual partners [58, 59]. These programs should employ vernacular languages [26, 37, 57] and graphic medicine (comics) [36, 60, 61] to increase the reach and usage of HIV prevention information by low-literacy displaced youth.

We also observed an association between high financial resilience and increased condom use self-efficacy and likelihood of HIV testing among our urban refugee youth sample, underscoring the critical role of economic stability in shaping sexual health practices and access to healthcare services in humanitarian settings. Economic stability increases refugee youth’s ability to exercise greater control over their sexual decision-making.For instance, a related finding of this study, namely that food insecurity was inversely related to condom use self-efficacy, corroborates prior findings [62], illustrating how, in the context of economic instability, youth may have less ability to negotiate condom use with partners upon whom they may depend economically. Financial resilience may also mitigate barriers (e.g., transportation costs, lost wages) to accessing health services while reducing urban refugee youth’s engagement in high-risk practices (e.g., transactional sex), which are linked to lower rates of HIV testing in this population. Consequently, these results highlight the importance of integrating economic strengthening components into HIV prevention and testing interventions targeting urban refugee youth living in informal settlements.

We found that high community-level A-SRH and perceived HIV-related stigma were associated with a greater likelihood of accessing HIVST kits. This finding reinforces the notion that young people who are more stigmatized may prefer the confidentiality of HIVST, as it can reach priority populations by providing increased privacy, confidentiality, and autonomy [35]. Prior refugee youth research in Uganda [27, 63] have documented that intersecting stigma increases interest in HIVST over clinic-based testing. Furthermore, studies [26, 36, 64] have shown that HIVST is highly acceptable, especially among populations experiencing greater vulnerability due to structural and social factors (e.g., urban refugee youth). In these populations, the use HIVST can increase testing rates by empowering individuals to test themselves privately, and (depending on testing results) can increase linkage to HIV treatment or prevention services. However, some studies [27, 29, 59] have highlighted the need to diversify the distribution of HIVST kits beyond public facilities due to fears of stigma and discrimination in public sites. Thus, to fully realize the stigma-reducing potential of HIVST, distribution strategies leveraging pharmacies, supermarkets, online stores, and youth-friendly testing centers could minimize exposure to stigma and enhance both the acceptability and uptake of HIVST among populations most affected by HIV-related stigma. These cross-sectional findings, though promising, require further longitudinal exploration to better determine whether stigma predicts increased use of HIVST and whether stigma changes following HIVST access and utilization.

Finally, our findings indicate that at the interpersonal level, individuals in committed relationships and those with children were more likely to report consistent condom use and HIV testing than those without children [19, 32]. These associations may be attributed to improved communication, trust, or shared responsibility between partners, which facilitate mutual decision-making concerning sexual health and service utilization [19]. When interpreted through a rights-based framework, healthy sexual practices and access to HIV prevention are regarded as fundamental rights that must be ensured for all young individuals, irrespective of their relationship or parental status. Thus, opportunities for positive interpersonal dynamics are influenced by broader social and structural determinants, including stigma, gender norms, and access to supportive services. These findings underscore the significance of interventions that enhance partner communication and negotiation skills while addressing the intersecting barriers that limit youth agency in diverse relational contexts. Further, we found that condom use self-efficacy is a critical factor in promoting consistent condom use. Strategies to increase condom use self-efficacy should span socio-ecological levels, integrating intrapersonal (e.g., knowledge), interpersonal (e.g., relationship dynamics), and structural (e.g., condom access) factors [39, 65, 66]. These strategies can empower urban refugee youth with the knowledge, relationships, and infrastructural support they need to effectively negotiate safer sex practices [31]. Future HIV prevention interventions must account for the unique social contexts and relationships of urban refugee youth to improve their overall sexual health outcomes.

Limitations

This study has several strengths, including the large recruitment and enrollment of a cohort of urban refugee youth with diverse HIV prevention needs. Despite this, our findings should be interpreted in light of their limitations. One potential limitation of this baseline analysis is the cross-sectional approach. Although it provided prevalence data, namely condom (i.e., condom self-efficacy and consistent condom use) and HIV (i.e., HIVST access and HIV testing) practices, the observed associations must be interpreted with caution, as they do not imply causal relationships. Second, the absence of data on other key populations, such as sexually diverse individuals, and people who use substances, limits our ability to assess how intersecting identities and forms of stigma that may shape HIV testing practices in these groups. Future research should consider these intersectional dynamics among urban refugee youth in low- and medium-income countries. Third, the outcome measures were self-reported, which may have been affected by social desirability bias. However, self-reported measures have been shown to be reliable for reporting HIVST and HIV testing [67]. To improve the quality of our condom practice assessment, we used validated measures [19, 37, 39]. Furthermore, to limit the impact of social desirability bias on our findings, the survey was contextualized to the setting and administered in a local language selected by the participants. In addition, participants were provided with survey tablets to complete HIV vulnerability-related questions privately, further minimizing concerns regarding the impact of social desirability bias on our study findings.

Conclusion

The uptake of HIV preventive measures among urban refugee youth living in the slums of Kampala, Uganda during the baseline period was suboptimal. Our findings underscore the need for targeted multilevel HIV prevention strategies for displaced populations, particularly urban refugee youth, to address their unique vulnerabilities and ensure progress towards the UNAIDS goal of 95% of people globally knowing their HIV status by 2030 [46]. Our findings regarding the association between socio-ecological factors and reduced uptake of HIV preventive measures underline the need for multilevel, multicomponent interventions, such as initiatives combining economic strengthening interventions that address structural factors of education, food security, and financial resilience with stigma-reduction strategies that address community factors of HIV-related stigma and A-SRH. At the interpersonal level, there is a need to develop young refugees’ condom negotiation strategies [31, 65, 68] to increase consistent condom use, thereby reducing HIV acquisition and transmission. Future research studies from the Tushirikiane-4-Uthabiti project will provide longitudinal evidence from a cohort of urban refugee youth living in urban slums in Africa's largest refugee-hosting country, Uganda.

Funding

The study was funded by the Canadian Institutes of Health Research (CIHR: WI3- 179958). Funders played no role in study design.

Declarations

Conflict of interest

The authors have not disclosed any competing interests.