The Rwanda Clinical Mentorship model has evolved considerably over the years (Figure 1). In the first phase (2007-2011), clinical mentorship was focused on operationalizing task shifting to rapidly scale up ART coverage nationwide. The clinical management of PLHIV was transferred from a few national HIV specialists based in Kigali to decentralized general practitioners (GP) and nurses. Periodic group trainings were held to increase the knowledge and skills of new practitioners, and these trainings were reinforced with on-going clinical mentorship.

During the second phase (2012-2017), a team at the national level provided on-site mentoring and supervision to decentralized sites. This model included UMB’s multi-disciplinary mentoring team composed of specialists in infectious diseases, pediatric HIV, and quality improvement who conducted at least one weeklong mentorship visit per quarter per hospital. The UMB team trained district hospitalbased mentors to ensure future sustainability of the mentorship program. Through the use of mentor observation checklists, logbooks, and review meetings, the mentoring team identified and addressed challenges in management of pediatric HIV, disclosure, adherence, treatment failure, and switching treatment options. Mentors complemented the faceto- face visits with e-consults, follow-up phone calls, and targeted continuous medical education sessions. A total of 3,833 healthcare workers were trained from 230 health facilities.

Despite the expansion of HIV services throughout Rwanda, a centralized clinical mentorship model was unablet to sustain on-going support of providers countrywide. To overcome this situation, during the third phase (2017-2020), the UMB team supported Rwanda’s MOH to develop and implement a district-based, decentralized clinical mentorship. In this phase, UMB, in collaboration with the Rwanda Biomedical Center (RBC) team, trained a doctor and nurse for each district to provide higher-level mentorship in a hub-and-spoke type model. Each doctor/nurse pair mentored providers within a single district. Through this collaboration, 1,941 healthcare workers have been trained across the country to assist in strengthening HIV testing, ART initiation, and retention.

Across the years, the Rwanda MOH and RBC have engaged implementing partners in the development and update of national care and treatment guidelines and training materials. District-based mentors participated in this process either by contributing to consultations or training health providers on new materials across affiliated facilities. This participation across the different levels have contributed to the adoption and acceptance of the clinical mentorship.

In 2017, the clinical mentorship was transferred to the MOH. Mentors were initially hired under the PEPFAR grant funds, and by 2018 the Government of Rwanda assumed leadership in recruiting, hiring, and covering the cost of the clinical mentors effectively transitioning the program. Currently, all mentors are paid under civil service payrolls by the government. The gradual and coordinated transition, combined with leadership from RBC, has contributed to the long-term sustainability of the program.

Following the successful rollout of clinical mentorship, RBC has increasingly recognized the importance of integrating continuous quality improvement (CQI) to routinely evaluate whether services offered by the multi-disciplinary team are following established guidelines to achieve intended goals. The 2019 integrated clinical mentorship guidelines have incorporated CQI into routine mentorship and have established a list of quality indicators to ensure standardized reporting. The clinical mentorship program fortified the scale-up and decentralization of Rwanda’s ART program by taking the trainings to non-expert clinicians (i.e. task shifting). This could not have been done as well without the mentorship.

Challenges encountered during the implementation of clinical mentorship include district hospital-based mentor attrition (e.g. nine mentors of 32 left their positions between 2014 and 2015 for other jobs), maintaining mentor competencies, limited allocated time to provide mentorship, and distance to cover to deliver on-site mentorships. To respond to some of these challenges, under the leadership of the MOH, RBC, in partnership with UMB, is developing an e-learning training program, through Moodle, for ART clinic staff. Across some district hospitals, the monitoring and evaluation team visits are combined with mentorship and support visits to reduce travel costs.

Given improving ART outcomes in Rwanda, how can clinical mentorship be optimized?

First, clinical mentorship in Rwanda focuses on providing optimal care for unstable patients who are at high risk for poor HIV outcomes. For example, adolescents and young men are at higher risk of unsuppressed viral load than any other age group.⁵ Adolescent-specific health provider training and tools may improve the quality of care they receive and, hence, their engagement and retention in care. Other groups at high risk for loss to followup include key populations (e.g., men who have sex with men and female sex workers). Key populations con-sistently report stigma and discrimination as barriers they encounter to accessing HIV care.^7;8 Second, the use of technology such as virtual community of practices, online training, and electronic decision support applications, offers an opportunity to reach more providers and reduce geographical barriers impeding the delivery of high-quality mentorship. Third, a data-driven approach to clinical mentorship can inform the training and mentorship needed by healthcare workers. The use of new tools like an integrated CQI reporting system may enable mentors to visualize trends over time, take timely remedial actions, and track progress.