Study design
This trial protocol describes a community-based, participatory arts-based intervention to be implemented in schools and evaluated using a stepped-wedge cluster randomized design. This protocol is reported in compliance with SPIRIT guidelines (Figure 1; Additional File 1) to create and evaluate an arts-based intervention. A formative phase with interviews will provide data to inform the intervention content and delivery. The intervention will be developed using a series of community-based participatory research (CBPR) meetings with a Community Advisory Committee (CAC) and Elders group. A process evaluation will occur throughout the study to evaluate the development, implementation, and effects of the intervention. The study was designed in partnership with the District Health Officer, the District Education Officer, the AIDS Support Organization (TASO), and the Waroco Kwo Elders Association in Omoro District.
[***Insert Figure 1 here***]
The 32-week stepped-wedge cluster-randomized controlled trial will be implemented in three blocks of four schools each stratified by school type (primary, secondary) such that each block will have two schools of each type per block. The clusters are assembled into three blocks that will crossover from control to intervention condition, with crossover order selected at random. Including the control period, each school will participate across four, 8-week periods with crossovers from control to intervention conditions occurring after each period (20, 21). Therefore, clusters will receive the intervention over 24 weeks, 16 weeks or 8 weeks, depending on their block allocation (Figure 2).
| Figure 2. Stepped-wedge design for an evaluation of an arts-based intervention in Omoro, Uganda |
| Clusters | Blocks | 8-week period | 8-week period | 8-week period | 8-week period |
| 1 (primary) | 1 | Control | Intervention | Intervention | Intervention |
| 2 (secondary) |
| 3 (primary) |
| 4 (secondary) |
| 5 (primary) | 2 | Control | Control | Intervention | Intervention |
| 6 (secondary) |
| 7 (primary) |
| 8 (secondary) |
| 9 (primary) | 3 | Control | Control | Control | Intervention |
| 10 (secondary) |
| 11 (primary) |
| 12 (secondary) |
Ethics
This study has been approved by the Thompson Rivers University Research Ethics Board (102240), the University of Saskatchewan Behavioural Research Ethics Board (Beh-REB1701), the University of Alberta Research Ethics Board (00093974), and the University of Lethbridge Human Participant Research Committee (2020-004). The study has also been approved in Uganda by the Uganda National Council for Science and Technology (HS510ES) and The AIDS Support Organization (TASOREC/011/2020-UG-REC-009).
Conceptual framework
The adapted Health Stigma and Discrimination framework (22) will guide several aspects of this study, including the development and implementation of the arts-based intervention. The framework provides a theory of HIV stigma processes by highlighting various domains of health-related stigmatization across the socio-ecological spectrum. We have adapted the framework to the local context through a consultation process with the District Education Officer and the District Health Officer. Adaptation includes applying the framework’s constructs related to the drivers and facilitators as prevention strategies and manifestations to minimize the harmful effects of stigma during data collection and analysis to further inform and evaluate the arts-based intervention.
Community Advisory Committee
A Community Advisory Committee (CAC) will advise the research team on development, implementation, refinement, and evaluation of the arts-based intervention, and participate in the interpretation and dissemination of findings. The 16 member CAC is comprised of research team members and locals that hold various leadership roles (e.g., Village Health Team, education specialists) community members, and youth in rural villages in Omoro District. Though intermittently disrupted by COVID-19 containment measures, community engagement by the research team began on 05/2019 and is ongoing. The CAC will be engaged for in-person and virtual meetings (or as per COVID-19 protocols) as needed. All meetings will be audio recorded with consent of the participants.
Setting and participants
This study will take place in Omoro District, northern Uganda where the HIV prevalence among the 15-24 age group is 7.2%, greater than the national prevalence of 6.0% (23). Omoro District has a slightly higher pregnancy rate among young people (28% among 10-18 years) compared to the national rate of 25%. HIV prevalence among young pregnant mothers who were tested at their first antenatal visit was 8.9% in the District compared to 7.6% for pregnant mothers nationally (23).
In Omoro District, there are seven secondary and 65 primary government-funded schools that serve six sub-counties (Koro, Lalogi, Odek, Bobi, Ongako, Lakwana, Omoro). Each grade level enrolls approximately 50 students per school, with approximately 12,000 students enrolled across these grades. In this study, we will approach students in six grade levels within the 10-19+ age group for recruitment. Primary grades 5-7 include approximately n=150 students per school; secondary grades 1-3 include n=150 students per school. Therefore, n=1,800 is estimated to be the total number of students attending 12 schools within the grade levels that are eligible.
Randomization
In consultation with a representative of the District Education Officer, we identified a sampling frame of all government-funded schools situated within Omoro District (65 primary and seven secondary schools). One secondary school (Awere Secondary School) and nine primary schools (Awere, Dino, Odek, Agweno, Jing Komi, Acet, Awali, Aromo Wanglobo, Binya) were removed if travel time from the research office in Omoro District to the school was estimated to be one hour or longer. From an eligible sampling frame of 56 primary and six secondary schools, one Co-Principal Investigator (Co-PI) independently used a random number generator in Stata 11.0 to randomly select (without replacement) and allocate six primary and six secondary schools to three crossover blocks. Other investigators, members of the local research team, and Head Teachers did not have advance knowledge of the allocation prior to enrollment. After allocation to blocks, Head Teachers were invited to enroll their schools in the study. Three primary schools (Ajuri, Laminonami, and Kweyo ) were resampled due to a land issue that closed the school, a bad road that blocked transit to the school, and an administrative redistricting that moved the school to a neighboring District (Gulu District). All secondary schools agreed to participate. Participating schools have remained blinded to their block allocation. The research team will coordinate with Head Teachers to facilitate participation of all classrooms in each eligible grade level.
Recruitment of students
Initial contact with students will occur after schools are randomized and enrolled in the study. The research team will read the information letter in the local language in the classroom before inviting students to participate. A copy of the information letter will be sent home with students for parents or guardians to review with an option to opt-out. We will return the next week to obtain a list of students who are interested in participating. Verbal consent (age 10 and over with parental opt-out process) will be sought and documented in a tracking form. Students who decline to participate will join another classroom for the duration of the study.
We will recruit three classrooms per primary school and three per secondary school (one classroom per grade-level, and one participating teacher per classroom. Elders will also take part in the arts-based intervention and will be assigned one per each classroom. Each Elder will be responsible to participate in the arts-based intervention in three different classrooms per week. An Elder, as defined by the CAC members, is a person who is knowledgeable in traditional and customary issues, a custodian of cultural knowledge, and presides over may societally issues on behalf of their community. Elders will be recruited through consultation with each Headteacher.
Arts-based HIV stigma reduction intervention
Formative research phase
We will conduct six pilot formative semi-structured interviews in Luo with two students and teachers (one each from primary and one each from secondary school) and two Elders to inform the arts-based intervention recruited through the schools’ Headteacher. We will then conduct 50 one-on-one formative semi-structured interviews, lasting 30-60 minutes, with purposively selected eligible students, Elders and teachers at participating schools.
The initial interview guide will be created by the research team and revised with feedback from pilot interviews and the CAC to ensure cultural and contextual relevance. The guide will include questions for students, teachers and Elders, such as ‘How are people who are known to be HIV-positive treated in your community, family, school?’ and ‘How does your village treat HIV-positive persons?’. Data from the full set of formative qualitative interviews will be used to inform modifications to quantitative survey instruments.
In coordination with Head Teachers, 24 students aged 10-19+ years balanced by gender and attending a participating primary or secondary school in grades primary 5-7 and secondary 1-3, will be invited to participate in one-on-one formative interviews. Additionally, 16 teachers have been recruited by the Head Teacher. Elders in the community have been identified based on competency in local language fluency, and demonstrated skills and practices in intergenerational knowledge transfer. Approximately five elder women and five elder men were recruited through the Waroco Kwo Elders Association. The research team will attend an Elders’ meeting to facilitate discussions about the study and provide copies of the information letter. Interested Elders will contact the study’s research officer. Verbal informed consent will be obtained through information letters translated and read to participants in the local language (Luo). Verbal consent from parent/guardian will be obtained for students. Even after consenting, participants will have the choice to opt-out at anytime.
Workshops
Following the needs assessment, one 3-day intervention training workshop will be delivered in collaboration with Elders, CAC members, teachers and Head Teachers from participating schools. Workshops will serve to establish relationships between teachers and Elders, review the findings of the formative phase, provide research ethics training, review the HIV-related stigma curriculum, and deliver several of the HIV stigma educational activities in order to practice how to facilitate class discussions. The workshops will build capacity for teachers and Elders to facilitate the intervention sessions with participating students in their classrooms.
Intervention components
The intervention will be delivered two hours per week over 24 weeks and will consist of three interrelated activities: (1) transformative educational activities; (2) arts-based activities; and (3) participatory theater. These activities will address the drivers of HIV-related stigma (i.e., fear and misconceptions about HIV, lack of awareness, social judgment, shaming, prejudice, negative attitudes & reluctance to seek treatment/help), which manifest as types of stigma practices (i.e., enacted, internalized, anticipated, courtesy, perceived) (22).
Students will be eligible to participate in the arts-based intervention if they are 10-18 years, live in Omoro District in Northern Uganda, attend a government-funded primary and/or secondary school in Omoro District during the study period, and are enrolled in primary grades 5-7 or secondary grades 1-3.
Transformative educational activities
These are based on Dewey’s (24) educational approach by grounding educational activities in the daily experiences of young people where stigmatizing experiences and practices become manifest.
The first hour of the weekly session transformative educational activities will engage learners in problem-based discussions initiated with pictures from an HIV stigma toolkit that was created, implemented and evaluated in sub-Saharan Africa with youth and adults (25). The toolkit addresses drivers, facilitators and manifestations of stigma as a means of articulating a “problem.” Learning through doing/practice (i.e., participatory learning) will be emphasized through role play, discussion and issue analysis.
Arts-based activities
These activities are anchored in aesthetics (i.e., theory of artful expression (26)) and influenced by Eisner’s view that arts-informed research entertains, educates, and acknowledges individuals as knowledge-makers (27). A key advantage of using art is to access “ways to perceive and interpret the world…that would otherwise go unknown” (27).
During the second hour of the weekly session, a focus on addressing the “problem” through finding “solutions” and encouraging action will be accomplished by learning traditional Ugandan songs, dance, told and acted stories, proverbs and sayings that incorporate empowering messages and moral teachings connected with respect for the relationship between oneself and the community, peers and authority, and responsibility for others. In grounding the intervention in local culture, this knowledge will reinforce the transformational learning needed to address the root causes of HIV stigma (e.g., stigma toward sex, adolescent sexual and reproductive health stigma), and directly address each type of stigma (enacted, internalized, anticipated, courtesy, and perceived). Connecting problem-based identification and discussion in the first hour with problem-solving skills in the second hour is an approach that draws from the work of Paulo Freire (28). Students will also be invited to write stories, music lyrics, or draw pictures of experiences with HIV-related stigma and will be invited to share their creations during the intervention period. The audio recordings of these weekly sessions will provide data for the process evaluation. Filming of Elders will occur during the arts-based intervention as their stories are considered knowledge preservation activities to promote intergenerational inclusion.
Participatory theater
After five weeks of the arts-based HIV-related stigma intervention, a participatory theatre intervention (PTI) will occur over a 3-week period. Developed in collaboration with students, the PTI will summarize learnings and cultural traditions from previous sessions. The first week will be dedicated to the development of the forum theater event. The remaining two weeks will be dedicated to performances in front of audiences including students, parents, and community members. Five topical scenes will be developed per school in collaboration with theatre specialists, researchers, students, Elders and teachers. Scenes will be developed through performance techniques (28) that will include a summary of qualitative data findings (formative interviews). Each scene will incorporate examples of stigma types (i.e., enacted, internalized, anticipated, courtesy, and perceived) with an HIV-positive and HIV- affected participant (student actor) who acts out each stigma domain (e.g., a participant (student actor) living with HIV who has experiencing internalized stigma; a participant living with HIV anticipating stigma; a participant not living with HIV enacting stigma; a participant experiencing courtesy stigma because of a mother living with HIV; and a participant who perceives stigma in their community). Additionally, as part of the PTI, situations will be enacted as a problem with no solution, inviting an audience member to come in and enact a more positive solution.
A semi-structured focused discussion will be held immediately after the PTI to explore participant and audience experiences. These discussions will be recorded and transcribed, providing qualitative data for the process evaluation.
Process evaluation
Drawing on Linnan and Steckler’s framework (29), a process evaluation will be conducted throughout the intervention period. The process evaluation will be used to explain results and potential impacts on participants by assessing implementation per-protocol across study sites. This project will address four components of process evaluation: fidelity, sustainability, reach, and context (29). Fidelity to protocol will be achieved through standardized training and measurement of skill acquisition among Elders, teachers, and students throughout the intervention. Standardization of delivery will occur by routine monitoring and review of the manualized intervention, observation of the classroom, and brief discussions with participants to assess facilitators and barriers to implementation. A protocol and manual outlining the process and content of HIV stigma curriculum including the traditional knowledge delivery for HIV stigma reduction will be developed to support fidelity to intervention protocols (29).
Students, teachers, and Elder participants will be invited to participate in summative interviews for the process evaluation. Five to ten students will be purposively selected from each of the three groups. We will also randomly select 10 CAC members to interview at the end of the study for process evaluation data.
Outcomes
Outcomes will include stigma-related and clinical measures. The primary outcome is enacted anticipated, and internalized stigma as measured using the Adolescents Living with HIV – Stigma Scale (30). Secondary stigma-related outcomes will include courtesy and perceived stigma measured using Stewart et al.’s 10-item subscale (31) and the Brief Stigma by Association Scale (32). Secondary clinical outcomes will include HIV testing frequently among sexually-active participants; and linkage to care, time to initiation of antiretroviral therapy (ART), ART adherence, and viral suppression among HIV-positive participants.
Power calculation
Using the power and detectable-difference calculations for stepped-wedge cluster-randomized controlled trials (91) (Stata 14.2), we calculated 98% power to detect a 20% reduction (i.e., an absolute risk difference of 12%) in the proportion of young people reporting a stigma endpoint (i.e., endorsing elements in the primary stigma measure) given an estimated baseline proportion of 60% having experienced HIV-related stigma, a 5% two-tailed Type I error rate, and a conservative intra-cluster correlation coefficient (ICC) of 0.01 given a lack of estimates in this setting. Prior work in sub-Saharan Africa and rural Uganda found that 50% of participants experienced stigma in some form (33). Given our team’s prior experience working among young people in areas recovering from conflict, we have assumed a higher baseline proportion of 60% to have experienced HIV-related stigma. We targeted a total cluster size of 12, with four schools (two primary and two secondary) grouped to crossover from control to intervention at three specified crossover points. We estimate that there are 150 students enrolled per school across eligible grade levels. Assuming a 50% participation rate after attrition, we estimate that the survey evaluation will reach 75 students per school for a total sample of n=900 who will complete the evaluation survey at the end of the final data collection timepoint (34).
Data collection and management
Survey design and measurements
A survey questionnaire will assess primary and secondary outcomes, along with selected covariates. The questionnaire will be developed in English, translated into Luo and back translated to ensure the Luo questionnaire is semantically valid. The questionnaire will be refined with input from the CAC and pretested with 6-10 students to evaluate ease of comprehension, suitability of questions, and completion issues. Internal consistency of relevance scales (e.g., attitudes, stigma) will be assessed using Cronbach’s alpha.
Study outcomes will be assessed by an independent research team members at baseline, following each crossover period, and at 5-months after completion of the final intervention block. The questionnaire will take up to 60 minutes to complete, and will include questions about socio-demographics, reproductive and sexual health behaviour, HIV testing, and HIV care. Stigma will be measured using three validated scales:
-
Enacted, anticipated and internalized stigma among HIV-positive students will be measured using the Adolescents Living with HIV – Stigma Scale (30).
-
Courtesy stigma will be measured using the Brief Stigma by Association Scale in HIV-affected students (31).
-
Perceived stigma will be measured using Stewart’s 10-item subscale on felt-normative stigma towards PLHIV in their community (32).
As intervention benefits may attenuate over time (35), we will examine sustainability of change in attitudes and stigma-associated changes five months after the intervention ends. Sexual risk, HIV testing frequency, and HIV status will also be assessed by self-report.
Administrative health data
In partnership with TASO and community health clinics, health records from community health clinics will be used to assess clinical outcomes including HIV status, date of confirmed diagnosis, time of ART initiation, viral suppression, and concurrent medical conditions. Data will be anonymized by sharing the Study ID file (linked to identifiers) with TASO, then having them return data linked with the study ID stripped of all identifiers. These data will be linked by study ID # within our de-identified dataset and remain confidential and anonymous. All data will be securely stored on a password-protected cloud-based platform, with access limited to approved members of research team.
Data analysis
Analysis of formative Interviews
Interviews will be audio-recorded in Luo (local language), transcribed and translated in English. Transcription and translation will be conducted by local translators. Back-translation from English to Luo will be used to verify the accuracy of the original translation for 5-8 transcripts, enhancing trustworthiness of findings (36, 37). Final English transcripts and field notes will be imported into NVivo™ for analysis.
The Framework Method (38) will be used to analyze qualitative data. Researchers will begin by familiarizing themselves with the data by reading transcripts. The analytical process will begin with deductive coding guided by Stangl et al’s (22) Health Stigma and Discrimination framework, followed by inductive coding for data that does not fit within the framework. Two researchers will independently code a subset of interviews, with a consensus code negotiated through peer-debriefing and coding comparisons in NVivo™. Agreed codes will be categorized to form an analytical framework which can be applied by researchers to code all subsequent interviews. Once all interviews are coded, data will be summarized using a matrix to facilitate interpretation and theming of the data (38). Memos will be kept throughout data analysis to record methodological decisions and analytical insights.
Descriptive and analytical strategies will be used to enhance rigor throughout qualitative data collection and analysis. The research team will maintain an audit trail of analytical steps and decisions throughout the research process, and peer-debriefing between researchers and members of the research team will occur to enhance trustworthiness and credibility of findings. Member checking of findings (39) will occur within a focus group held with CAC members, which includes teachers and youth, to enable researchers to assess interpretive and evaluative validity in study findings (40). CAC members will assist in the interpretation of findings in relation to participants’ narratives, and in validation of study conclusions.
Analysis of survey data
Descriptive statistics will be used to describe schools relative to the required sample size at each time point (planned sample size and achieved sample size), socio-demographic characteristics, primary/secondary outcomes, and to assess balance between randomized clusters. Main analyses will use Linear Mixed Effect Models (LMEM) or Generalized Linear Mixed Effect Models (GLMEM) to assess the intervention effect. LMEM will be used for continuous outcomes that approach normality and GLMEM for binary and other types of outcomes. Normality will be assessed using the model residuals and outliers by Cook's Distance, Leverage and Residuals. Time in 8 weeks from beginning of the trial, time in 8 weeks centered at the beginning of the intervention, time in 8 weeks from the beginning of school semesters, and blocks will be specified as fixed effects. The multilevel structure will be defined by assessments within participants as level 1, participants as level 2, and schools as level 3. Schools and participants within schools will be specified as random effects. Main analyses will follow an intention-to-treat approach, with clusters analyzed as per original treatment assignment. Sensitivity analyses among participants who drop-out and complete will be performed to assess bias arising from attrition. A significance level of 0.05 will be used in all statistical tests. Analyses will be conducted in SAS v9.4.
Data monitoring and adverse events
Participants who report that they are sexually active and have engaged in condomless sex will be supported to attend the nearest health center for an HIV test or will be provided TASO’s contact information for an HIV test. Positive tests will be referred for counselling and follow-up per national protocols. If sexual assault is reported, initial support will be offered through a trained counselor affiliated with the study. If further assessment or support is needed or the participant does not want to see the counselor, a list of referral agencies will be provided through our research team, who will support the participant to connect with the appropriate agency such as health, legal, livelihood and security service providers.