Health workers play a critical role in providing gender-sensitive HIV care. However, without adequate training, they may inadvertently reinforce negative gender norms that hinder engagement in care. To improve ART services in sub-Saharan Africa, effective implementation strategies, including trainings, are essential.

This study evaluated the feasibility and acceptability of a novel 4-session, gender-sensitivity training for HIV health and lay workers in rural and peri-urban Uganda. We conducted a pilot quasi-experimental controlled trial in 2022–23, utilizing data from three sources: process data from 144 providers and 238 clients, qualitative post-intervention focus groups and interviews with 53 training attendees, and stakeholder interviews with 12 community leaders and facilitators. We employed thematic analysis guided by the Consolidated Framework for Implementation Research (CFIR).

The training was highly rated by attendees via brief post-session surveys, with 99% of sessions deemed satisfactory. Qualitative feedback indicated that the sessions were perceived as appropriate for addressing providers’ gaps in delivering gender-sensitive care and as beneficial for participants. High feasibility was evidenced by 100% of intervention steps delivered (fidelity checklists), 86% attendance (attendance logs), and 98.6% and 94.8% recruitment of eligible providers and clients (study records), respectively. Through qualitative exit focus groups, factors that positively affected acceptability and feasibility included tactics to gain support from supervisors, the use of experienced facilitators and active learning approaches, and the consideration of the health system context in delivery (e.g., cadre mix, location). Suggested improvements encompassed addressing interpersonal and structural barriers to client-centered care, incorporating community sensitization, streamlining content, and enhancing mentorship follow-ups. Stakeholders recognized the training’s alignment with national health priorities and emphasized the need for clarity regarding its integration with existing training curricula and ongoing stakeholder engagement.

This study underscores the acceptability and feasibility of a gender-sensitivity training program for HIV providers; further evaluation of the program is warranted. Identified factors influencing implementation can inform enhancements to gender-focused training initiatives to improve the quality of HIV services and enhance client engagement and health outcomes.

Clinicaltrials.gov, NCT05178979, registered: November 15, 2021.

Significant progress has been made in increasing access to HIV testing and antiretroviral treatment (ART) globally and in sub-Saharan Africa, with the goal of getting 95% of all people on ART virally suppressed to achieve population-level epidemic control, following UNAIDS’ 95–95-95 targets [1]. Uganda, with 5.1% of the adult population living with HIV, has reached its 95–95-95 goals through concerted efforts to scale up and make HIV services accessible [2]. Despite progress, UNAIDS points to the need to close gaps that exist in HIV outcomes between groups, including disparities by gender [3]. In sub-Saharan Africa, where HIV is transmitted primarily through heterosexual sex [4], a large literature highlights gender norms as a driver of HIV risk and disengagement from HIV services. Gender norms are culturally grounded expectations for men and women’s roles, relationships, and behavior [5]. In Uganda and similar settings, masculine norms (e.g., toughness, independence) can be counter to men’s health seeking behavior, gender inequitable relationship dynamics and gender-based violence increase women’s exposure to HIV and limit their receipt of HIV prevention and treatment, and the fear that HIV status disclosure will compromise masculine and feminine role fulfillment discourages both men and women’s engagement in the HIV care continuum [6,7,8,9]. As such, it is critical that HIV services recognize and respond to the strong influence that gender norms have on men and women’s HIV outcomes.

Health workers play a critical role in the provision of ART and can help men and women overcome known gender-influenced barriers to HIV care engagement, such as HIV stigma, HIV status disclosure issues, and gender-based violence [10,11,12]. At the same time, research from sub-Saharan Africa highlights how providers can also exacerbate the negative effects gender norms can have on client outcomes when their treatment of people living with HIV is not gender-sensitive, or responsive to men and women’s individual needs and preferences. For example, adherence to traditional masculine norms is associated with greater experience of HIV stigma and worse engagement in HIV care among men in sub-Saharan Africa [6, 7]. Providers can increase men’s feelings of shame or embarrassment associated with their diagnosis by not respecting their desire for privacy and reprimanding them for poor treatment adherence [13]. For women, particularly young unmarried women or women engaged in sex work or transactional sex, providers can similarly be a source of HIV stigma, holding prejudice grounded in social expectations related to women’s sexuality [14].

In addition to stigma, many other aspects of the client-provider relationship that are vital to clients’ satisfaction with care and engagement outcomes are also shaped by gendered constructs, such as communication, power dynamics, empathy, and trust [15]. For example, men and women’s communication styles are shaped by dominant norms about men and women’s acceptable behaviors, which can vary depending on the gender of the provider/client in any given relationship. Studies show men’s socialization to be direct and women’s to be passive can be experienced differently by providers and result in the perception that men are rude and impatient, and women are compliant but may be deceitful [15, 16]. Social norms may also dictate conversations about sexual health as taboo based on gender—limiting women’s ability to speak openly on such topics in some settings and limiting men/women dyads from communicating on these topics with each other in other settings [15, 17].

To mitigate these issues, ART service providers need to be aware of the role that gender norms can play in their clients’ care engagement and in their own bias towards and treatment of clients, while being motivated and skilled to provide gender-sensitive care. Like other cultural competencies, gender-sensitivity can be strengthened through provider capacity building [18, 19] and may be improved by strengthening HIV care providers’ client-centered care and communication skills [20, 21]. The Institute of Medicine names patient-centered care one of its six domains of healthcare quality, defined as “care that is respectful of and responsive to individual patient preferences, needs, and values and ensuring that patient values guide all clinical decisions” [22]. Its principles lend themselves to improving cultural-sensitivity, as they can help providers understand and empathize with clients’ socio-cultural barriers to care engagement (i.e., in this case, gender norm-related barriers). Further, client-centered care can enhance skills to communicate with men and women and reduce gendered power dynamics affecting provider–client relationships. The broader literature assessing sensitivity interventions related to gender, race/ethnicity, and sexual orientation, as well as client-centered care interventions, supports their potential to improve providers’ knowledge and attitudes, with subsequent effects on improved client outcomes [23, 24]. However, both areas are understudied in sub-Saharan Africa. In Uganda, sensitivity trainings on key populations, including sex workers and gender and sexual minorities, have reported improvements in providers’ attitudes and empathy toward these groups [25, 26]. A preliminary trial on client-centered care also yielded promising results in Kenya, including reduced viral suppression in clients in the clinics that received the training [27]. However, the authors underscore significant structural barriers to client-centered care delivery in resource-limited settings, and the dearth of research in this area [27]. To our knowledge, no studies in Uganda or the region have specifically aimed to increase client-centered skills through a gender norms lens.

As such, we developed and pilot-tested a gender-sensitivity training for HIV care providers and clinic staff in central Uganda. The development of the gender-sensitivity training was originally based on an existing set of evidence-based strategies grounded in social cognitive psychology research to reduce bias in clinical interactions —adapted to be specific to gender [28]. For the present study, these strategies were adapted to focus on gender in the local context of Uganda, with the goal of increasing HIV providers’ knowledge, motivation and empathy, and skills to deliver Uganda Ministry of Health (MOH) ART program guidelines to cisgender men and women in a gender-sensitive, client-centered way [29]. In this paper, we present the findings of the primary aim of this trial to examine the intervention’s feasibility and acceptability and factors affecting implementation and the potential future adoption of the program, guided by the Consolidated Framework for Implementation Science Research (CFIR).

Implementation frameworks such as CFIR can widen the scope of factors studied that drive gender disparities by focusing on multilevel factors beyond the individual affecting the implementation and uptake of an evidence-based intervention [30, 31]. They can also guide the development of implementation strategies, or methods to enhance the implementation of an evidence-based intervention, to improve client outcomes [31, 32]. Considering ART an evidence-based intervention, we used CFIR to guide this study with the aim of understanding the role of gender norms in ART’s delivery and uptake, and subsequently create an intervention strategy (i.e., training intervention) to improve HIV care providers’ gender-sensitivity. CFIR outlines 5 domains (intervention characteristics, outer setting, inner setting, characteristics of individuals, process) and 39 sub-domains that can affect intervention implementation. In the current paper, CFIR is used to guide the evaluation of the pilot trial, specifically through the selection of relevant implementation outcomes and factors explored that may have affected implementation.

To pilot test the gender-sensitive implementation strategy, we used a mixed methods embedded experimental design, inclusive of a quantitative pilot quasi-experimental controlled trial with qualitative methods embedded at the end of the trial [33]. Following recommendations for Stage 1b pilot studies (pilot testing of a new behavioral intervention), the primary aims of our trial were to examine feasibility, acceptability, and factors affecting implementation/future adoption of the gender-sensitivity training [34]. We will report the results of the secondary aim (i.e., to explore the preliminary effects of the implementation strategy on relevant HIV care provider and client outcomes) in a separate manuscript (under review). However, some of the themes reported in the outcome paper are included here to provide context needed to fully interpret our results.

The quantitative component was a quasi-experimental controlled trial, implemented from 2022–2023, which compared two matched clusters (health facilities) randomly allotted to receive the gender-sensitivity training intervention or no intervention (standard-of-care). Process data were also collected through this trial to gather metrics for feasibility/acceptability, such as participant retention and participant satisfaction with each intervention session (collected after each session through brief surveys, “How satisfied or dissatisfied were you with today’s session overall?” (4-point Likert scale, very dissatisfied to very satisfied). The qualitative component within the larger experiment included focus group discussions with intervention participants after the 12-month quantitative assessment, as well as individual interviews with intervention facilitators and HIV and community stakeholders at the end of the study. The purpose of the qualitative component was to interpret and expand on the quantitatively measured indicators of acceptability and feasibility, and identify factors affecting implementation success [33].

The University of Texas at San Antonio and Makerere University School of Public Health’s institutional review boards approved the study. The Ugandan National Council for Science and Technology also granted study approval. The trial’s protocol is published [35], and it was registered with clinicaltrials.gov (NCT05178979) on November 15, 2021. All participants provided written, informed consent before participation.

The study was carried out in six public health facilities serving rural and peri-urban populations in Luwero, Mityana, and Wakiso Districts. These sites were supported by Mildmay Uganda, an implementing partner to the Ugandan Ministry of Health (MOH) in the delivery of free HIV treatment and services. Mildmay Uganda was a partner to the study and helped with the selection of health facilities to ensure balance based on multiple factors (e.g., clinic size, populations served, services offered). The six clinics included two clusters (three clinics close in proximity were considered one cluster)—and the two clusters were randomly allocated to either the intervention arm or the control arm via coin toss. As a pilot trial, only two clusters could be included, but cluster randomization was necessary to avoid contamination in a clinic-wide intervention – it also allowed us to test the feasibility and acceptability of procedures for a future, multi-site trial. We ensured clinics in the intervention arm were at least one hour away from clinics in the control arm to reduce the risk of contamination.

The study enrolled two purposively sampled cohorts of participants, including HIV care providers and clients enrolled in HIV care. The provider cohort included HIV care providers and staff (age 18 and above) with regular interaction with clients living with HIV (e.g., clinical officers, nurses, nurse midwives, counselors, and lay health workers, such as peer educators). Providers in the intervention arm received the training intervention, while those in the control did not. The client cohort was made up of people living with HIV enrolled in care at the participating clinics. They were 18 years of age or older, or emancipated minors (married/living with children and catering for oneself), fluent in Luganda, and either new on ART (pre-ART or initiated within one year) or struggling with ART. Struggling with ART was defined as those virally unsuppressed based on most recent viral load results (within prior 6 months) obtained through clinic records, or self-reported ART non-adherence in the prior 2 weeks. We included only those with a detectable viral load or new on/struggling with ART to increase the study’s power to see change in HIV engagement-related outcomes with our pilot sample size (relevant to our effectiveness aims, to be reported in a separate manuscript).

Clinic staff referred potential participants in both cohorts to the study staff. Research assistants informed individuals of the study, screened for eligibility, and for those eligible and interested, obtained written informed consent. Participants in both cohorts completed an interviewer-administered questionnaire at baseline and two follow-ups after intervention implementation (6-months and 12-months post-baseline). Participants in both cohorts received compensation for their time (baseline & 6-months: 10,000 Ugandan Shillings (UGX) [~ 3 USD], 12 months: 18,000 UGX [~ 5 USD]).

The sample size was based on guidelines for Stage 1b studies (pilot testing of new behavioral interventions), which suggest 15–30 participants per cell [34]. For the provider cohort, we aimed to recruit the entire population of health workers, lay health workers, and staff with regular interaction with clients living with HIV, which fit within these guidelines. For the client cohort, we explored the study’s power to detect change in clients’ main outcomes of ART adherence (assuming power = 0.80, α = 0.05) using G*Power. We determined a sample of 100 clients per intervention arm (estimating 17% attrition from n = 120) would detect a moderate effect size (d = 0.4) in ART adherence [36].

Control

Health workers and staff in the control clinics received no intervention; thus, the intervention was compared to the standard-of-care implementation of MOH ART guidelines [29]. All clinics in intervention and control arms were government health facilities and supported by the same implementing partner, Mildmay Uganda, with the goal of equalizing other external trainings related to client-centered care or related topics. For equity, we provided the gender-sensitivity training to the control clinics after the end of the study.

Intervention

We adapted the aforementioned existing strategies by Dovidio and colleagues [28]. The strategies are based on evidence that increasing providers’ motivation [37,38,39,40], awareness [41,42,43], skills [44,45,46,47], empathy [48,49,50,51,52,53], and emotional regulation [54,55,56,57] can prevent implicit racial and gender biases from affecting clinical judgment and behavior. The adaption of these strategies, along with the development of additional intervention content, was guided by the literature from sub-Saharan Africa, as well as a local needs assessment. As a precursor to the trial, we collected qualitative data from ART clients and community members, as well as HIV care providers and stakeholders. We also elicited multiple rounds of direct feedback on the approach and materials from HIV stakeholders involved in HIV prevention and care, including Mildmay Uganda, MOH officials, and local District Health Teams.

The intervention package, detailed in Table 1, comprised of four training sessions led by two experienced national trainers. The first three sessions featured structured education, group discussions, interactive role plays, and case studies. The fourth session served as a practicum to review and apply the material from the initial sessions. The training was delivered in both Luganda and English, based on the participants' needs, and included a workbook for notetaking and reference. Following standard practices locally, attendees were provided 30,000 Ugandan Shillings (~ 8 USD) as transportation refund, along with light morning refreshments and lunch for the full-day sessions 1 and 2, and light refreshments and 15,000 Ugandan Shillings (~ 4 USD) for the half-day session 3. The appropriate rates were determined in consultation with the study implementing partner, Mildmay Uganda.

Sessions 1 and 2 were held at an off-site location over two consecutive days. Session 1 covered an introduction to gender concepts, such as the distinction between sex and gender, gender norms and roles, and their relation to HIV acquisition, barriers to HIV care, and gender disparities in HIV. The session included contextually relevant examples and introduced participants to gender-transformative health programs [58, 59]. An interactive workshop used role plays to illustrate how HIV might challenge traditional gender norms, creating stigma, referred to as the "HIV threat framework," which was based on prior research by Naugle et al. [60] and other relevant literature [6, 7]. Providers were trained to use gender-sensitive, gender-specific, and gender-transformative strategies to mitigate HIV stigma, including referring clients to appropriate support services, counseling them on how antiretroviral therapy (ART) can positively impact their gender role fulfillment, and counseling them away from gender norms that conflict with HIV care engagement.

Session 2 focused on locally relevant bias and stereotypes related to gender and HIV. Using an adaptation of Bower et al.’s "three-perspective framework,"facilitators guided providers to assess client needs from the client’s perspective, their own perspective (including self-reflection on biases), and the societal perspective, incorporating gender norms [61]. The session included exercises to help providers recognize how personal biases and emotions might affect their care and to enhance empathy for client challenges. Evidence on how external stressors can impact clinical interactions was presented, along with stress management techniques to mitigate bias.

The remainder of Session 2 explored how gender intersects with power dynamics in provider–client relationships and influences communication styles. The session introduced client-centered care as a strategy for delivering gender-sensitive care, addressing power imbalances, and improving communication. Role plays focused on gendered barriers to HIV care, allowing participants to practice client-centered techniques for building rapport, communication, and collaborative decision-making, with some content adapted from Hofert et al. [62].

Session 3 was a 2-hour on-site training at clinics, concentrating on gender-based violence. This session built upon existing MOH training focused on responding to gender-based violence HIV in clinical care [63]. We emphasized gender-based violence’s connection to HIV and the broader gender norms covered in earlier sessions, and promoted responding to gender-based violence as central to providing gender-sensitive, client-centered care.

The final session was a review of Sessions 1–3 with an emphasis on applying the content in practice. A facilitator observed and evaluated providers' interactions with clients using a checklist adapted from a client-centered communication assessment [62]. Following observation, individual feedback was given, and a group discussion was held to review core content, address practical challenges, and discuss solutions based on providers’ experiences.

Our main, quantitatively measured, implementation outcomes measure the acceptability and feasibility of the intervention and relevant trial procedures. These quantitative data were collected from process data during the intervention’s pilot trial. Each outcome is described in Table 2, along with the data collection procedures used to obtain them.

We also use qualitative data collected as part of the trial to explain and expand on the acceptability and feasibility measures detailed on Table 3 by exploring factors that positively and negatively affected implementation and that may affect the future adoption/sustainment of the intervention by the health system. At the end of the intervention trial (after the 12-month quantitative assessments), all providers in the intervention arm were invited to participate in an exit focus group discussion about their experience with the training. For those who could not attend, they were invited to an individual interview. Of the 61 total providers, 53 (87%) participated (46 in focus groups, 7 in interviews). An experienced qualitative interviewer, independent of the study team and trained in by our team in the interview guide, moderated the focus groups in Luganda, and later translated and transcribed their audio-recordings. A research assistant took detailed notes used to aid transcription. The focus groups were approximately 90-min. Participants received 22,000 UGX (~ 6 USD) for their participation.

We also conducted 12 semi-structured qualitative interviews, including 1 with each intervention facilitator (2 total), as well as interviews with local HIV stakeholders, including MOH leaders, district health leaders, Mildmay Uganda representatives involved in HIV/gender programming and health worker capacity building, and representatives of other community-based organizations working in HIV and gender programming. Participants were recruited by the Project Coordinator for participation, and our researcher partners at Mildmay Uganda helped refer some participants to the study team. Participants provided their written informed consent before participation. Interviews were approximately 30-min and participants received 25,000 UGX (~ 7 USD) for their participation. The audio recordings of focus group and individual interviews were transcribed after completion, and interviews that were conducted in Luganda were translated to English.

The interview guides for training participants, facilitators, and HIV stakeholders were tailored to the different groups, but their overall aim was to elicit experience or anticipated challenges and successes in intervention implementation, uptake, and future adoption and scale up by the health system. The development of interview guides was informed by CFIR, with questions developed to elicit factors negatively or positively affecting implementation, uptake, and future adoption based on CFIR constructs across relevant CFIR levels. Table 3 provides an overview of CFIR constructs that questions were developed to elicit.

Quantitative data collected from process data required simple calculations (e.g., means) carried out in Microsoft Excel. If deemed feasible based on data distribution, we planned to make simple comparisons in outcomes by gender but recognized that this may not be feasible given the pilot sample size and generally more women providers/staff than men at health centers in Uganda.

We used thematic analysis to analyze qualitative data [64]. The coding guide was developed a priori based on CFIR (Table 3)—which was revised through iterative review of the transcripts. Three research assistants, trained by the PI (KMS), coded the data (AA, RLL, KW) first deductively, following CFIR constructs, and then inductively, revising the coding guide based on the data. All transcripts were double coded, with one coder first coding, a second coder reviewing and confirming agreement or disagreement of codes, and a third coder (KMS) reviewing the final coded data and facilitating discussions to come to final consensus. The team met weekly to discuss codes, revisions to the coding guide, and potential themes. We kept decision trails to ensure the accuracy and consistency of codes in the early stages. We coded transcripts first, and moved the final codes into a Microsoft Excel spreadsheet, which we used to review data by codes and inform a summary of themes and representative quotations. We also explored whether there were differences in key findings by gender and cadre. We elicited feedback from community stakeholders on the final themes presented here during community and MOH dissemination events.

Following the mixed methods design, we compared the quantitative and qualitative results by implementation outcome at the point of the presentation and interpretation of results. We present them alongside each other in a table, organized by the overarching outcome (e.g., acceptability of the intervention) and in the results narrative. We used the qualitative data to explain, validate/contest, and expand on the quantitative results (e.g., reasons for intervention acceptability or lack thereof).

We display the Consolidated Standards of Reporting Trials (CONSORT) study flow chart in Fig. 1. We present the quantitative and qualitative findings in support of our implementation outcomes alongside each other in Table 4. These findings are described in detail next, by implementation outcome.

CONSORT Diagram

Full size image

Satisfaction with the intervention and perceived benefits to adoption of the training

In post-intervention session surveys, when asked about their level of satisfaction with the session, 99.0% of attendees reported being either satisfied or highly satisfied. The qualitative data also supported intervention acceptability among providers; in exit focus groups and interviews, providers praised the intervention because they perceived it to be relevant and beneficial to their work. Many providers shared that the acquisition of knowledge and skills on providing gender-sensitive care to clients living with HIV translated into improvements in their interactions with clients, with subsequent benefits on client outcomes.

“The content was relevant in that some of us didn’t know the gender issues that might affect people in the communities because we wouldn’t probe about their situations at home or other issues that might result in the person not coming back at the scheduled appointment” (Clinician, man, interview).

However, while providers and stakeholders agreed that a gender-sensitive training focused on HIV providers and staff was important, they expressed that the intervention would be more effective if it worked with the community on gender-related issues in tandem. As one nurse (woman) suggested in a focus group: “We have to engage community leaders and other stakeholders whom we can also train on those issues because they lack knowledge.” Participants recognized harmful gender norms as pervasive and felt that the intervention’s reach would be improved if gender-sensitization efforts included clients, community members, and community leaders, especially for addressing norms underlying GBV.

Appropriateness of program content for addressing clients’ gendered needs

Relatedly, the qualitative data supported the appropriateness, or fit, of the intervention content in addressing issues contributing to HIV gender disparities locally. Providers’ characterizations of their relationships and treatment of clients before the training corroborated the need to increase providers’ motivation and skills to provide gender-sensitive, client-centered care. Participants openly shared their rushed and harsh treatment of clients and lack of empathy for their barriers to HIV care engagement (in general and those related to gender). For example, one lay health worker shared about their experience following-up lost clients:

Like us [peers], they expect us to look for those people [clients lost from care]. If a client is lost, it is your responsibility to look for him wherever he is. So, when he comes – remember you have the responsibility to look for him. Sometimes you call and he cuts you off, you call and he laughs at you, like that. So, when he comes [to the clinic], you feel anger and you pour it out by being harsh [laughter], and yet, the person you were looking for has come (Peer, Man, FGD).

Providers’ narratives also highlighted a multitude of biases they held towards clients—related to gender, as well as prejudice towards stigmatized groups that sometimes intersected with gender. Most relevant to gender, providers discussed how the training helped them to recognize bias and increase empathy towards and communicate better with: men as clients (perceived as difficult, impatient, rude, particularly by young women providers), young women (with judgment towards their manner of dress and sexual relationships), and key populations (i.e., sexual and gender minorities, sex workers).

Finally, the qualitative data highlighted the need to strengthen the implementation of existing GBV protocols. The MOH training for responding to GBV is meant to be mandatory for all health workers, but many said they had not had it. MOH and district leadership attributed this to high staff turnover and change within clinics. Providers admitted that they were not regularly following procedures for GBV screening and referrals set by the MOH, confirming the need for content to strengthen its implementation: “In my routine, I had never screened on GBV and I would just ignore it. But after the refresher, I started screening for GBV with my HIV clients. Yes!” (Nurse, woman, FGD).

Perceived barriers to translating training into practice

The training aimed to strengthen the provision of gender-sensitive care by reinforcing the principles and techniques of client-centered care, applied to clients’ gender-specific preferences and needs. While there was consensus among the providers and HIV stakeholders interviewed on the importance of client-centered care (described as a national and district priority in HIV programming), structural and interpersonal barriers to implementation were noted. Providers discussed challenges in giving clients the amount of time and privacy needed to be client-centered given heavy client load and limitations to the clinic environment; these limitations were well-known and reiterated by the facility and district leaders interviewed. The following quotation from one provider highlights some of these barriers and conveys frustration towards the program for promoting practices that might not be feasible without broader structural change to support client-centered care:

They [the facilitators] made me happy when they came to see the environment in which we work. But, although they came, [they] never changed [the training approach based on] the environment we are working in, although they saw the weaknesses and the conditions we work in. Something concerning counseling in the environment in which we work, there is very little privacy. You cannot give that service effectively. So, that did not change. It became a barrier towards delivering the service” (Peer, woman, FGD).

Another barrier to client-centered care shared among a subset of providers was the desire to maintain boundaries with clients; these providers said that showing too much care or attention to clients can send the wrong message and make clients “fall in the love” with them.

“They told us to give the clients time, but this man will come and the linkage [facilitator] is giving him time – he will not go back home because he thinks this woman likes him. Then he calls on phone and he puts it in his own thinking that you ‘love’ him (chorus laughter). You see, the way we care for them…he can become a nuisance” (Peer, woman, FGD).

Providers’ interviews suggested broad acceptance of the gender-focused content, without expressing overt resistance or disagreement on the importance of providing gender-sensitive care or the relevance of gender to their work with clients living with HIV. However, a subset of providers did share that being bias-free and always being able to enact client-centered, gender-sensitive behavior with clients was not easy. One provider, for example, conveyed a sense of injustice that they should have to let go of anger towards clients they perceived as difficult or in the wrong:

As a health worker, that transitioning from one [emotional] state to another is not easy and it [the training] found us when we are already old (laughter)! So, to change, you change with pain for the benefit of the client. But as an individual, you feel it is still unfair, like you have not been convinced and you have not got justice” (Linkage Facilitator, woman, FGD).

Acceptability and appropriateness for health setting of intervention delivery approach

Providers shared that the general approach taken to delivering the content and modes of delivery (a mix of directive education paired with practicum-based learning) met their learning styles and needs. They expressed high satisfaction with the integration of active learning approaches throughout the sessions, and the use of experienced and engaging facilitators. Providers explicitly named the use of role plays and vignettes, participatory discussions, and the application of the material to real-world examples relevant to their work as facilitating their understanding of the material. One provider explained how the role plays specifically helped build empathy for clients:

“What I was grateful for is that they put us in [client’s] ‘shoes’ by making you act as clients and then you experience the way they [providers] can treat them. We do these things [ourselves] but we ignore it…You experience the way someone can handle you and you work on your attitude. Those tactics helped us to see that if we have problems we are experiencing and if someone treats us well, we can be set free and we can explain what we need” (Peer, man, FGD).

There was consensus among providers that the mix of cadres in groups added value to the intervention by enriching discussions and facilitating co-learning from each other’s perspectives, as one nurse (woman, focus group) explained, “There were those who had experience and they knew things on the subject and yet some of them were new. So, whenever they would share their experiences, we would also understand better." Participants also recognize the cadre mix as necessary for feasibility, ensuring that the intervention did not interrupt service provision, as one Linkage Facilitator (woman, focus group) explained: “That process…it did not affect any health care service provision because the facility was not left empty…We went in groups, and this supported not leaving the facility without any health worker.”

Providers and facilitators expressed the need for more coaching and to practice applying core content to provider-client interactions. The gap between sessions 3 and 4 was considered too long; by the time of the final refresher, the facilitators perceived a loss of knowledge among providers on some of the core content. “We broke off and then we came back and had a refresher. And some of the participants had really forgotten. That is why I am saying it requires continuous engagement (Intervention Facilitator, interview). Facilitators and providers suggested integrating more practicum-based approaches or mentorship models, a perspective that was corroborated by HIV stakeholders interviewed as best practice for capacity building. Participants agreed that the practicum-based approach taken in session 4 helped them refresh on program concepts and apply them to their work, with the mentorship of the facilitator on-site.

Despite session 4’s overall positive feedback, the use of direct observation elicited mixed feelings. Facilitators first observed client-provider interactions, after which they provided individual feedback on their gender-sensitivity and client-centered communication, followed by a group discussion. Many providers said that the observation provided coaching that helped hold themselves accountable and point out things they did not know needed improvement, or to reinforce positive changes, as one provider said, “This feedback increased my confidence on how to handle such clients” (Peer, woman, FGD). However, the facilitators felt some providers, clinicians specifically, changed their behavior during observations, minimizing the accuracy and usefulness of the approach. “Some of the clinicians, because you are there assessing them, they changed depending. But for the other cadres of staff, the counselors, the expert clients, I noted a [real] change” (Intervention Facilitator, interview). A few providers, particularly lay counselors, said being observed and assessed made them nervous and some worried that the observations may have made clients uncomfortable.

The feasibility of implementation was supported by measures of intervention fidelity. Based on our assessment of randomly selected session transcripts, 100.0% of all intervention steps were delivered. Of those delivered, 96.9% of delivered intervention steps that we expected would elicit active participation from attendees did. Using a more nuanced measure of fidelity (rating delivery as fully delivered as expected, delivered but with room for improvement, and not or inadequately delivered), 78% of intervention steps were delivered with the highest fidelity (i.e., fully delivered as expected).

In addition to fidelity, a secondary quantitative metric to support the feasibility of the intervention is session attendance. As depicted in the CONSORT diagram (Fig. 1), across all sessions, there was 86% attendance, with the highest attendance in the first sessions (90% attended both sessions 1 and 2), and lowest in the last refresher session (79%). The qualitative data, outlined in Table 4 and described next, sheds light on these findings.

Intervention complexity

The intervention content itself was perceived by facilitators as moderately easy to deliver, with concepts they understood and could communicate well (confirmed by the fidelity assessment). The two facilitators were highly skilled trainers, both with expertise in HIV and gender, and with considerable experience in health worker capacity building. Thus, their experience should be considered as an important contributor to the high fidelity of content delivery. However, facilitators noted several areas of the original intervention manuals that they perceived to be too complex, cumbersome, and inclusive of terminology that was technical or too “academic.” The facilitators commended the study for eliciting their feedback on the materials after the initial pilot and making modifications to these areas to improve content delivery and attendees’ comprehension. “We were engaged in the review process. We cut out what we thought was duplication. We [revised] what would be too much for the participants to understand. The beauty is that we were engaged in the review process” (Intervention Facilitator, interview).

Facilitators also noted that there was a difference in the ease of delivering content to clinical health workers vs. lay health workers. It took longer to explain some concepts to lay health workers, which facilitators attributed to their relatively lower literacy and education levels. There was a need for all content to be in Luganda among lay health workers (whereas health workers often preferred to go between English and Luganda). While the overall perception was that lay health workers understood the content—facilitators explained that the mix of cadres with different learning needs required careful facilitation and extra time to ensure comprehension:

“Comprehension meant it required us to really take some time for these people as compared to the other [lay vs. non-lay health workers]. We sought, because it was more of an interaction, we also sought help from the [clinical] health workers. ‘So, have you understood this? Can you explain to your friends what this means? You understand!’ So, in terms of comprehension… it might have taken some time, but, in terms of use at the end of the day, they were all the same” (Intervention Facilitator, interview).

Time limitations

Based on the assessment of transcripts, when intervention steps were rated as not delivered or not fully delivered, it most often was in the context of limited time. Raters were able to elicit from the context of the transcript that, in these cases, facilitators were often rushing to keep time.

Session mobilization

In the qualitative interviews, the overall levels of attendance by providers across sessions was attributed to two things that enhanced feasibility. First, the facility and ART clinic supervisors were supportive of provider attendance, encouraging their participation, and worked with the study team to find times that would work around the clinic’s schedule. Second, the study team provided adequate notice for providers on session dates, as well as reminders of the upcoming sessions.

Intervention location

The program itself was delivered across four sessions. The first two were off-site trainings (organized at a near-by location outside of the health facility) in two consecutive days. The third was on-site at the facility during a day/time usually dedicated towards continuing education at the clinics, and the fourth session was also on-site in the ART clinics. The location of the training, namely being off-site vs. on-site at the clinics, offered different pros and cons to feasibility. Off-site sessions were met with high acceptability from providers and facilitators, because there was less distraction than at the clinics and more space, increasing the feasibility and fidelity of delivery. As one provider said: Your boss releases you, knowing you are off-site, and he will not call you on phone trying to ask for something. But now that I am here [at the clinic], the doctor was calling me because he knows I am here at the site. (Laughter) (Counselor, man, FGD). On the other hand, the on-site trainings offered easier logistics and cost-effectiveness, not requiring transportation per diem or space rental. For these reasons, the on-site location was considered more feasible and sustainable among facility and district leadership interviewed.

The ability to recruit the intended study population is an indicator of the feasibility of a future efficacy trial and potentially of the level of acceptability or interest in the intervention. As displayed in CONSORT Fig. 1 and summarized in Table 1, among the provider cohort, 98.6% of eligible providers could be recruited. Only 2 providers declined participation in the intervention arm, and no providers declined in the control arm. In the client cohort, 94.8% of eligible clients at the HIV clinics were recruited, with near equivalence in the two arms (intervention: 95.2%, control: 94.4%). In qualitative interviews, the high acceptance of participation among providers was attributed to interest in the intervention, as well as a health system culture supportive of the continuing education of health workers. In addition, the study followed local customs for gaining entry to the clinics, which aided in garnering clinic-level support. This included gaining study approval through individual meetings, first from district leadership, followed by facility leadership, and then mobilizing providers through group presentations at the clinics.

The pilot trial demonstrated the feasibility of outcome measurement for a future trial for the main pilot outcomes. For providers, 95.8% contributed data for the primary outcome (competence for gender-sensitive care) at 12-months, and for clients, 88.6% contributed data for the primary outcome of ART adherence (Fig. 1). However, in a future, fully powered trial, viral load would be the ideal outcome for clients. We collected viral load as an exploratory outcome from clients’ clinic records, but there was significant variability in the completeness and timing of the data available, preventing the use of this outcome. While the study team worked directly with staff to increase the completeness of reporting, the custom was to not retest viral loads for clients that were known to be missing from care or non-adherent, which would lead to biased data. As such, for a future efficacy trial, relying on clinic records would not be feasible, requiring direct viral load collection from the study.

Perceived relative priority

In exploring what factors stakeholders and providers perceived might positively or negatively affect the future adoption of this training by the health system, we aimed to gather perceptions on the relative priority of HIV gender programming generally, and of this specific training, relative to other MOH and district priorities. MOH and district leaders’ interviews described national priorities in achieving UNAIDS’ 95–95-95 objectives as driving their program agenda, and considered gender-focused programming, as well as programming to improve client-centered care, as key to achieving those goals. As such, they said a gender-sensitive training such as this one aligns with their overall priorities and was viewed as an area for improvement, as one District Health Officer said: “Gender; we know what gender is. I know they [providers] would be on the forefront of the whole issue. But, we need to get them to understand the role of gender mainstreaming in all services basically.” Similarly, individuals interviewed from HIV-focused community-based organizations described a strong emphasis in their programming on issues related to gender. However, when providers were asked for their views on whether the health system considers a gender-sensitive training like this a priority, there were mixed responses, and their own personal priorities before the training were more narrowly focused on the provision of ART without the consideration of gender.

Stakeholder engagement

When asked about what factors might aid or prevent the future adoption of the program by the health system, key informant interviews overwhelmingly pointed to the importance of stakeholder engagement. They largely perceived the intervention development process, inclusive of formative research and efforts to garner stakeholder input, as appropriate (CFIR construct: intervention source)—despite it having been scaled back to virtual engagement due to it coinciding with the early phase of the COVID-19 pandemic. Participants emphasized the importance of engaging MOH, district, and other HIV experts in continuing to refine of the intervention. They equally emphasized the importance of stakeholder engagement in the dissemination of the findings, and the use of data to inform evidence-based decision-making. MOH partnership and buy-in were considered vital to ensure the training could be integrated with existing MOH training curriculum, and as the next quotation highlights.

“I think the way forward is integration. You give us the evidence. We compare with what is in our guidelines, then we see where the gaps are and [what] can be addressed. Or you do a policy synthesis, and you look at what is the gap from the policy, the guidelines and the practice. Then you give us what solutions and we pick from there” (MOH Official, interview).

Perceptions of intervention value

The stakeholders interviewed emphasized the importance of ensuring that leaders understand the additional value that the intervention has relative to existing MOH curriculum or programming (i.e., that it does not duplicate, but complements and expands on existing programming). While participants agreed that there is a lot of programming aiming to address gender issues in the broader community, they considered this intervention’s focus on training health workers in gender-sensitivity new and to be filling a gap in programming.

In key informant interviews with people who did not attend the intervention, interviewees sometimes struggled to fully understand the core focus and aims of the training. This seemed largely attributed to a conflation of the terms “gender” and "GBV". As much as the interviewer explained that the intervention approached gender more broadly, inclusive of not only GBV but other ways gender intersects to affect HIV care and outcomes, many stakeholders defaulted back to discussing the intervention as narrowly focused on GBV. This highlighted the need find ways to communicate the program’s aims in an easy-to-understand way, while being careful that it wasn’t perceived as a GBV-only focused intervention.

Our study reports on the implementation of a gender-sensitivity training focused on building HIV care provider and staff’s capacity to deliver gender-sensitive, client-centered HIV care. Using mixed methods data gathered from a pilot quasi-experimental controlled trial, this study finds that the intervention is acceptable and feasible, while identifying factors affecting implementation and future adoption, with implications for future refinements to intervention delivery. The study represents an important attempt to develop a training program that will improve the provision of gender-sensitive care, so that HIV care providers and staff can better address gender norm-related barriers to HIV care engagement for clients in sub-Saharan Africa and recognize and change how gender norms shape the quality of counseling they provide to clients themselves. Overall, the findings support the acceptability and feasibility of the intervention. The findings reported meet thresholds across implementation outcomes that were previously set by the study team to determine if this pilot supports progression to a future efficacy trial [35]. Our findings highlight areas for refinement and improvement before this next step, to further improve implementation.

Findings show the acceptability of the training intervention through reports of high satisfaction among training attendees, corroborated, and expanded on, by qualitative data. Participants viewed the training as relevant to their practice and felt that participating improved their relationships with clients in ways that translated to improved client outcomes. Further, results showed that the intervention was perceived to address important gaps in providers’ competence to provide gender-sensitive care. Providers, their supervisors, and leaders in the health system discussed a lack of current client-centered care provision, considerable biases held towards clients, and difficulty addressing specific gendered barriers to care for women and men. These findings align with those of the broader pilot trial’s secondary aim of testing the intervention’s preliminary effects on provider and client outcomes, reported separately (under review).

These data also highlighted perceived gaps in the training – providers and health stakeholders alike noted interpersonal (desire to keep boundaries with clients) and structural (privacy issues, time) barriers to translating gender-sensitive, client-centered care into practice, which others have underscored as a pervasive challenge in resource-limited settings [65]. More research is needed on feasible ways to implement and sustain client-centered care interventions in sub-Saharan Africa, which are limited, with special attention paid to human resource, time, and space constraints [65]. Participants and stakeholders also expressed the need for gender-sensitivity training that works with clients directly, alongside the broader community, to change community norms that underlie clients’ gendered barriers to care. Our focus on providers responds to calls to focus on the health system’s responsiveness to gender [66,67,68], which indeed should be considered as one part of multilevel gender transformative HIV programming aimed at gender norms change among individuals and through community mobilization [69,70,71,72]. If found efficacious, this training could be integrated into a broader package that explicitly works with clients directly and the surrounding communities to reshape harmful gender norms that contribute to HIV infection and deter use of HIV services.

Providers’ satisfaction with the intervention also related to elements of the intervention’s design, including the use of active learning approaches, experienced facilitators, and integration of an applied refresher training and use of coaching that took place in the clinic setting. We received feedback to add even more opportunities for direct coaching, given consensus that this support improved the translation of content into practice. This finding can guide our future adaption of this program to enhance its impact, and aligns with a systematic review of strategies for health worker trainings in low-and-middle income countries that reported programs that incorporate training into clinical practice and at the providers’ work-site had greater mean effectiveness than those that did not [73]. Relatedly, the use of direct observation of provider–client interactions was mostly well-received by participants and facilitators, who noted its positive impact on strengthening providers’ implementation of the training content in their practice. However, these benefits were countered by some concerns about observation not yielding authentic interactions due to social desirability, a phenomenon known as observational bias or the “Hawthorne effect” [74, 75]. Some lay health workers also said they felt pressure and discomfort from being observed, a group that may be less experienced with this type of mentorship. These findings can guide our use of this tactic in the future, which may need refinement in its implementation, using strategies to reduce the intrusiveness of the observer [76].

The training’s acceptability was also supported by a high level of attendance, which is also indicative of the feasibility of implementation. Factors that positively affected participant attendance include strategies well-supported by implementation science research [30, 77], such as concerted efforts to mobilize the clinics in advance and work with clinic supervisors to gain support for the study, as well as schedule the sessions and hold them in places that did not disrupt service provision. These same factors were also crucial to the success of eligible participant recruitment, which was high. Other evidence that the intervention can be delivered was obtained from our assessment of intervention fidelity, which was high overall, but highlighted intervention content to streamline to better deliver the sessions as intended within time constraints and to reduce intervention complexity. While the mix of cadres (lay health and clinical health professionals in the same sessions) reportedly facilitated co-learning and added to the feasibility of delivery within interrupting service provision, the levels of different baseline education and language abilities sometimes interrupted the flow of session delivery, requiring facilitators to take special care to ensure lay health workers’ comprehension. Based on the findings, we feel the approach to cadre mix was ultimately the best choice when considering the balance of both logistics and positive effects on the learning environment. However, as we refine the intervention, we will consider ways to collectively meet the needs of diverse groups, such as break-out sessions by cadre.

This study generated data on factors expected to affect the future adoption and scale up of this training by the health system, if focus efficacious in a future, fully powered trial. Collecting these data during the early phases of intervention development can enhance the translation of research into practice by allowing early refinements of the intervention that consider the context of real-world implementation [78]. We found overall support of gender-focused HIV programming among providers, as well as implementing partners. Representatives from CBOs supporting the provision of ART or HIV programming locally expressed strong recognition of the intersections of gender and HIV. At the district-level, leaders also considered gender-focused programming to be high priority, if it was clear how the programming complemented and enhanced their ultimate goals to achieve 95–95-95. However, there was a consistent assumption among stakeholders that our program was specifically focused on GBV-reduction. Some stakeholders interviewed had a difficult time conceptualizing what a “gender” program might focus on outside of GBV (given heavy emphasis on GBV in HIV programming locally), which highlighted future opportunities to communicate the value of the intervention to leaders in charge of adoption by emphasizing the broader scope of gender and its relevance to HIV outcomes. The data also echoed what is already known in the literature about implementation and adoption [79] – continued stakeholder engagement throughout the research process, as well as the comprehensive dissemination of the research findings to MOH and other stakeholders, will be key in garnering support for the program’s adoption in the future.

A strength of this study is the use of mixed methods that allowed us to explore not just quantitative metrics of acceptability and feasibility related to the intervention and pilot trial procedures, but qualitative data to understand and expand on these findings. CFIR offered an appropriate framework to guide our exploration of factors affecting implementation. These findings will guide our adaption of this intervention for a trial to fully test this intervention’s effects, while continuing to gather data on implementation. Our findings may also provide valuable insight for other researchers or programmers intending to develop or implement gender-sensitivity programs with a focus on provider capacity building in sub-Saharan Africa.

Study limitations include the small sample size and limited number of clusters/health centers in a pilot quasi-experimental controlled trial, which limits the generalizability of our findings to other settings and populations and introduces the risk that the two clusters were imbalanced on important factors. To minimize the time burden of data collection on our participants, we used a single item to measure intervention satisfaction. Future research should consider the use of validated, multidimensional quantitative measures for intervention acceptability. Social desirability bias could have affected participant’s reporting of intervention acceptability. Although participants were open about intervention areas to improve, and we used a qualitative interviewer independent of the study team to interview participants, it is possible that participants still felt the need to provide positive feedback. Additionally, we provided monetary reimbursement and food for attendance. While these are standard practices locally, even outside of research studies, they may have influenced the high levels of recruitment, session attendance, and satisfaction observed. Finally, no major differences were found in our analysis of the qualitative data by provider gender. However, we were limited by the pilot sample size and greater number of women participants to fully explore sub-analyses by gender. These same factors, paired with low variation in quantitative outcomes, also limited the ability to conduct meaningful sub-analyses by gender with the quantitative data. A future, fully powered trial should explore differences in the acceptability and perceived feasibility of the intervention by gender, as well as cadre and other relevant demographics, which may vary.

Gender-sensitive HIV care is critically important to helping men and women overcome barriers to HIV care engagement, and to ensuring providers do not exacerbate these barriers for clients, especially in settings like Uganda where gender norms are intrinsically tied to HIV outcomes. This study found preliminary support for the acceptability and feasibility of an implementation strategy to improve the delivery of ART services in the form of a gender-sensitivity training for HIV care providers and clinic staff, while identifying areas for refinement to improve its implementation and potential efficacy. The next steps of this research will be to progress this trial to efficacy testing in a fully powered trial, which could have significant implications for improving the quality of HIV care delivery, and ultimately ART clients’ satisfaction with and continuous engagement in HIV care.

Download references

KMS was the study PI and led study conceptualization, study design, data acquisition, analysis, and interpretation and manuscript writing. TSK was the primary K01 Mentored Research Scientist Career Development Award mentor and RWK the site PI and in-country K01 mentor. JFD, SLD, BST, and SHV were additional K01 mentors. All K01 mentors made sustainable contributions to study conceptualization, design, and data acquisition plans and analysis plans. AA, SCM, and BM were collaborators that made contributions to study design, data acquisition, and interpretation. AA, RLL, and KW coded the qualitative data and made contributions to data analysis and interpretation. All authors reviewed and gave feedback on the manuscript and consented to authorship.

Correspondence to K M. Sileo.

The University of Texas at San Antonio and Makerere University School of Public Health’s institutional review boards oversaw the study. The Ugandan National Council for Science and Technology also granted study approval. All participants provided written informed consent. The study adhered to the Declaration of Helsinki.

N/A.

The authors declare no competing interests.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions