Phone-based data collection in a refugee community under COVID-19 lockdown.

Researchers conducting community-based studies face unprecedented challenges in data collection during the coronavirus disease 2019 (COVID-19) pandemic, particularly in low-resource settings under lockdown. War Child Holland, an international non-profit organisation, is doing a randomised controlled trial of its group parenting intervention, the Caregiver Support Intervention, with 480 war-affected refugee caregivers in Lebanon. 3 weeks before the second wave of endline assessment, lockdown began in Lebanon, abruptly ending all field activities. We write to share how we successfully did our endline assessment entirely by telephone, with all research team members working from home. The system described here was developed in 5 days and implemented in 7 days.

Our study is based in Tripoli, Lebanon, where 74% of Syrian refugees live in poverty. Our field research team comprises a seven-person coordination team and ten research assistants located across Lebanon. Our task was to set up a new technological infrastructure for remote data collection, and to convince 240 refugee caregivers to participate in a 30–40 min survey via WhatsApp (an internet-based phone call and instant-messaging application available to smart-phone users) or through regular calls in a context where large families share small spaces, phone calls are expensive, internet is scarce, and electricity cuts are frequent.

Participant scheduling was done 5 days before data collection and was updated continuously. Assignments were constrained by when the research assistant and caregiver could speak privately, and whether caregivers agreed to an interview with a research assistant of the opposite sex. Spouses could not be interviewed simultaneously, as most refugee couples share one phone. Phone credit was transferred to families to compensate for call costs. Research assistants were masked to the participants' group allocation (Caregiver Support Intervention or control).

Pre-formatted spreadsheets (Google Sheets) replaced tablet surveys. Coordinators remotely monitored research assistant data entry live, in read-only mode to prevent accidental data overwrite. Team training was done remotely using Zoom videoconferencing and involved live role-plays, allowing research assistants and coordinators to practice data entry and supervision on Google Sheets, as well as protocols for potential problems, including unreachable or late caregivers, poor phone connections, and lack of privacy. Throughout data collection, the team coordinated through a hierarchy of WhatsApp groups (appendix).

We successfully interviewed 94% of refugee caregivers using this remote methodology. Our recommendations include scheduling sufficient airtime between interviews given connectivity challenges, prioritising team wellbeing given the stress of lockdown for both research team and participants, pacing team energy by scheduling decreasing numbers of interviews each day, and setting a tone of enthusiastic teamwork and agility in response to the unexpected.