Feedback on the Afya-Yetu experience in Goma, Democratic Republic of Congo (DRC)

The onset of the COVID-19 pandemic triggered widespread movement and contact restrictions by most governments to curb the virus's spread. Our teams, like many healthcare actors in LMICs (Low and Middle Income Countries), feared that these measures would negatively impact the care and follow-up of patients living with HIV (PLHIV), particularly the most vulnerable, such as patients suspected of treatment failure, those coinfected with HIV and tuberculosis, pregnant and breastfeeding women, and patients with comorbidities (hypertension, diabetes), etc. The challenge was to improve the phone-based manual follow-up of a large cohort with reduced staffing.

In this context, and to address this critical issue, the MSF Foundation proposed the use of an automated SMS messaging system to facilitate individualized, automated, and prioritized follow-up of PLHIV remotely, as a supplement or replacement for medical consultations that had become more complex during the COVID-19 pandemic. Based on this proposal, a digital tool, "SMS AFYA YETU," was designed and developed.

The goal of this tool was to enable healthcare staff to better monitor patients with chronic illnesses, particularly those with medical and/or social vulnerabilities, and to quickly identify and report issues that could lead to difficulties adhering to their treatment or deterioration of their health.

The tool allowed:

• The regular sending of simple and automated questions via SMS through the digital platform to patients, who could respond with a multiple-choice answer by pressing a number on their phone's keypad.
• Assigning medical follow-up tasks to MSF medical team members, with alert levels based on symptoms and issues identified.
• Creating a 24/7 updated database accessible to the MSF medical team to track field progress and provide additional support if needed.

Very quickly, the project's pilot phase began and demonstrated the tool's effectiveness.

The Afya Yetu project targeted PLHIV enrolled in the program set up by MSF and the Ministry of Health in Goma, with risk factors that made them more vulnerable to symptoms associated with COVID-19, notably:

• HIV-related risk factors: Tuberculosis coinfection, first-line treatment failure, recent hospitalization.
• COVID-19-related risk factors: hypertension, diabetes, advanced-stage HIV.

The total number of patients targeted by the Afya Yetu project was around 1,000 PLHIV (young adolescents living with HIV, unstable patients followed as outpatients and/or post-hospitalization, HIV-Tuberculosis coinfected patients, and pregnant and breastfeeding women). The pilot phase was designed to test the tool with a small population requiring close monitoring to see the concrete results on the ground before expanding the application to other populations if necessary.

Over four months in 2021, the Afya Yetu platform was tested on 30 patients. These patients volunteered to participate in this evaluation and were provided with a phone and an operational Airtel or Orange SIM card supplied by the project.

A large-scale deployment that could not be realized.

The results of this pilot phase were very satisfactory. When we, along with our partners, were ready to move into the larger-scale deployment phase, the teams faced several technical, administrative, and contextual challenges. Several factors explain the eventual failure of the deployment.

Contextual difficulties:

• Our technical partners (Medic) realized late in the process that to facilitate large-scale deployment and ensure the service was free for patients, it would be necessary to operate through a toll-free number. MSF successfully negotiated its acquisition (granted in March 2021), but this process took several months of negotiation, follow-ups, and monitoring.
• It then took several months to find a local IT company capable of providing the API needed to translate the medical algorithm into SMS language and capable of sending bulk SMS.
• Next, it was necessary to negotiate prepaid credits with each telecom operator. These negotiations also took several months, with no clear explanation of the delay. Our last hypothesis was that our request was seen as too small compared to those of larger humanitarian actors.

Contextual challenges:

Unfortunately, when these technical issues were finally resolved after a long delay, an operational decision by MSF halted the deployment: the closure of the HIV project in Goma in 2023. Without a strong operational base, the MSF Foundation naturally decided to stop the deployment process.

This project, which was subject to a capitalization report, greatly advanced the teams' understanding of the project cycle and the stakeholders involved in a multi-partner project like this.

Achievements and satisfaction:

• Rapid action capacity and transversal mobilization: The MSF Foundation was able to quickly mobilize a known technical partner (Google.org), MSF's technical referents (HIV, Tuberculosis), the Goma field operations, and an Africa-based technical partner (Medic Mobile). The evaluation of the needs, issues, and risks, as well as the modeling and design of the solution to counter the indirect harmful effects of the COVID-19 pandemic on the health of an extremely vulnerable population, was completed in just a few months.
• In a context of impossible travel, the tool was developed entirely remotely, thanks to the enthusiasm of a multidisciplinary team and the strong involvement of our field teams in Goma.
• The solution provided corresponded well to what the field teams wanted: maintaining close monitoring of the entire cohort (3,000 patients) during the pandemic. It met the specifications and allowed nurses, without additional workload, to monitor all patients while automating the prioritization of the most vulnerable individuals needing calls and/or visits, all with the same number of nurses (automation of SMS vs. time spent on the phone).
• The solution was designed and discussed with the patients from our test cohort (30 patients). It was innovative and unprecedented in the context of essential therapeutic monitoring for such chronic conditions. This pilot phase proved that it was possible to improve the quality of monitoring and the prioritization of needs in a precarious and volatile context. It was based on our patient-centered approach, and the success of this phase was primarily due to their involvement in the design stage of the application and the consideration of their feedback to improve the tool.

What we learned from this failure:

The teams learned a great deal about the importance of clearly identifying technical and administrative responsibilities between ourselves and our partners in projects with a strong mobile component. The technical constraints must be thoroughly understood in such precarious contexts: what equipment the target population needs, network quality in the target area, the cost of purchasing a phone fleet, data storage capacity, and understanding the relationships between various government institutions, and commercial interests.

Moreover, the interoperability of the developed platforms was key and required in-depth analysis: in this project’s case, the platform managing SMS (RapidPro) vs. the tool developed by Medic (CHT) vs. the prototype developed by Google.org to create the follow-up algorithm.

These elements are now integrated into all future MSF Foundation project explorations.

Lastly, this project has emphasized the importance of patient involvement in the design of our projects. While our tools must meet the needs of medical teams, they are also, when necessary, tools for our patients. It is crucial to address the social and economic environment to ensure adherence to the system.

The capitalization of our learnings from this pilot phase now enables us to provide optimal support to MSF actors who wish to try a new approach to monitoring a large cohort and/or chronic disease follow-up.