Remote Interviewer Training for COVID-19 Data Collection: Challenges and Lessons Learned From 3 Countries in Sub-Saharan Africa

"Remote interviewer training for large-scale surveys can be an effective replacement to in-person learning in low-resource contexts when data are urgently needed and in-person learning is impossible."

COVID-19 research is particularly important for countries in sub-Saharan Africa, where disease surveillance systems can be limited. In much of sub-Saharan Africa, in-person interviewer training and face-to-face data collection remain the norm for population-based surveys, largely due to inconsistent internet connectivity and barriers to technological literacy. However, during the COVID-19 pandemic, in-person activities may be risky and/or prohibited by government restrictions. This article presents Performance Monitoring for Action (PMA)'s development of a remote training system for COVID-19 surveys in the Democratic Republic of the Congo (DRC), Kenya, and Nigeria, including challenges faced and lessons learned.

The article begins by explaining PMA's process of collecting rapid-turnaround data on family planning and other reproductive health indicators for policymakers across Africa and Asia since 2013. In short, with the support of the Johns Hopkins Bloomberg School of Public Health (JHSPH) and Jhpiego, universities and research institutes in each country ("country teams") recruit and train female community members to conduct mobile-phone-based population and facility interviews surveys. In late 2019 and early 2020, teams in 4 of the 8 PMA countries collected baseline survey data in-person, obtaining consent and phone numbers from women willing to participate in follow-up surveys. These teams also had previous experience with conducting remote surveys over the phone as part of PMA Agile, a separate project within the PMA platform.

As the pandemic hit, in March 2020, PMA developed a COVID-19 survey and began preparing for remote training. The survey consisted of a short phone-based questionnaire administered to women of reproductive age (15-49 years) who consented to follow-up at baseline. The survey consisted of 6 sections: COVID-19 awareness and information sources; perception of personal infection risk; knowledge of COVID-19 symptoms, transmission, and prevention; social consequences of COVID-19-related restrictions; and the impact of COVID-19 on accessing health services and on reproductive health outcomes. Data are representative of Kinshasa province in the DRC, nationally representative in Kenya, and representative of Kano and Lagos States in Nigeria.

The article describes the remote training process. For instance, recognising that interviewers had limited familiarity with online learning, the training system mimicked the in-person experience through video lectures, reinforcement through small-group activities, evaluation through electronic assessments, and active monitoring via one-on-one phone calls between facilitators and interviewers. Interviewers accessed all training materials from PMA smartphones also used for data collection. Each team adapted the training system to fit within their own context, as described in the article.

In implementing the remote training system, PMA encountered 3 central challenges, which they attempted to resolve as follows:

1. Rapidly preparing remote training facilitators - PMA worked to reduce interviewers' learning burden by, for example, relying on platforms with which they were already familiar - i.e., WhatsApp, Google Drive, Open Data Kit (ODK), and YouTube.
2. Ensuring interviewers are able to fully participate in the training - Anticipating internet connectivity concerns, country teams distributed print manuals, adapted training to include recorded (not live) sessions, and designed training lectures to be 20 minutes or less. In consideration of women's competing household priorities, interviewers were permitted to go through a day's training at their own pace yet were required to finish all sessions for the day by a designated time.
3. Ensuring interviewers are actually learning - PMA created quizzes in Google Forms and ODK, and country teams used small group-discussions via WhatsApp and recorded, live debriefs via video call to gauge interviewers' comprehension.

Two overarching lessons learned from the experience include:

• Acknowledge the tension between fostering local ownership over training design and meeting timelines for data collection - PMA suggests that training system designers engage with country teams earlier in the planning - e.g., by organising a facilitator orientation session to introduce the system and get feedback before delving into the details of content creation. Another path forward is to promote pursuit of a "minimal viable product", a concept that originated in business, which holds that imperfect, unpolished materials can be created in less time while still being effective for learning.
• Acknowledge the limitations of remote trainings - PMA concedes that in-person learning has significant paedagogical advantages, including easier observation of learner engagement, greater flexibility to address learners' needs, more opportunities for organic discussion and practical exercises, and greater socialised learning. That said, when remote learning is the only option, certain factors can enable successful learning. PMA's structure and history as a project was key, in that considerable time and resources had already been invested into training interviewers and fostering an environment of collaboration across the project. Such factors "undoubtedly contributed to interviewer confidence in attempting a new way of learning."

In conclusion, although PMA does "not advocate for systematic replacement of in-person trainings with remote learning, demonstrating that a remote approach is possible in these settings is an important step toward ensuring the availability of high-quality data during the COVID-19 pandemic."