Implementation of a REDCap-based Research Data Collection System in Cameroon

Abstract

Implementing electronic data collection for health research can be challenging in resource-limited settings, where electricity, internet access, and study staff with computer training may be limited. Our team has established a successful research data infrastructure using the REDCap software at three HIV clinics and a data coordinating center in Cameroon. We desribe our recommended network architecture and guidance for study data teams working in similar settings.

Introduction

Cameroon is a member country of the Central Africa International epidemiology Databases to Evaluate AIDS (CA-IeDEA), one of seven regional consortia of HIV cohorts encompassing 45 countries globally [1]. Clinics participating in IeDEA contribute longitudinal patient care data for observational HIV research and the development of care guidelines for groups like the World Health Organization. Participating Cameroon HIV clinics used paper charts, so the Cameroon IeDEA research team needed to build a low-cost electronic cohort database for three clinics plus an in-country data center in order to send data to the US-based coordinating center. The solution selected should not interfere with the government’s long-term planned rollout of electronic health records. Furthermore, the Cameroon National Ethics Committee advised that all data should be housed in-country before sharing with outside collaborators.

Methods

We implemented our research data collection system using REDCap [2], a free, secure and flexible web-based clinical research data capture platform. Unlike the single webserver solution typically seen in high-resource settings, we implemented three tiers of REDCap servers (Figure 1). To accommodate electricity and internet outages, we set up a Windows desktop at each clinic as a “REDCap server” on a local area network (LAN) with battery backup. Patient enrollment and routine visit data were transcribed from paper forms into REDCap by data entry personnel working on laptops. This enabled data entry work to continue despite recurrent internet and power interruptions. We configured a REDCap server (with fixed IP address) at the coordinating center in Yaounde to receive data from all sites over Internet. Data officers at each clinic pushed data to the central server weekly ensuring timely data availability at the in-country coordinating center. We implemented REDCap forms with integrated quality checks (e.g. range and valid values, date formats, skip logic) and conducted a monthly data quality control at the central level. At the end of every data quality review, a report was sent to the sites for verification and correction. The data quality control procedure was repeated to ensure all corrections were implemented. Cleaned data were uploaded quarterly to a Regional Data Center. Variables names and codes were the same across REDCap projects on all server tiers, allowing easy transfer and harmonization of data.

Figure 1 –

Diagram of the 3-step data flow from HIV clinics in Cameroon to the regional data coordinating center

Results

This multi-tier REDCap architecture has enabled us to capture high quality, longitudinal data on over 6,000 unique patients despite weeks of Internet interuptions at both the Cameroon data center and participating clinics. Challenges with data collection and entry and REDCap data uploads have been resolved collaboratively with increased site engagement and training[3]. The Cameroon-based server meets requirements for in-country data storage and secure data sharing.

Conclusions

A multi-tier REDCap infrastructure can be a workable data capture solution for research in resource-limited settings.