This study was approved by the Ethics Committee of the University of Botswana (Reference: UBR/RES/IRB/BIO/GRAD/244). All data experiments were performed in accordance with relevant guidelines and regulations such as the Declaration of Helsinki. All participants were informed of the objective of the exercise as well as their voluntary participation, and all gave informed consent for this study. Those who consented to participate were immediately sensitized and granted access to the REDCap instance at the UB. No compensation was provided for participating in the study.

NVivo 11 software was used for thematic analysis [21] of data to determine participants’ requirements for an ideal COVID-19 data management system and their postimplementation experiences.

The literature emphasizes the critical need for robust ICT infrastructure to serve as the backbone for successful and sustainable digital health implementation [22-25]. The national eHealth strategy tool kit by the World Health Organization (WHO) and International Telecommunication Union (ITU) identifies essential eHealth infrastructure components as high-speed data connectivity, computing infrastructure, identification and authentication services, directory services, health care provider systems, electronic health record repositories, and health information data sets, all of which underpin a national eHealth environment [26]. The tool kit further urges countries to secure long-term funding for investment in national eHealth infrastructure and services. It is against this backdrop that the WHO recognizes a pressing need for countries to invest in infrastructure to support digital transformation; Internet connectivity; and issues related to legacy infrastructure, technology ownership, privacy, and security while adapting and implementing globally recognized standards and technologies [27].

In this study, infrastructural challenges experienced include weak Internet speed and misalignment of computer network firewalls between the GDN and the UB network. These issues resonate with the identified ICT infrastructural challenges within the Botswana National eHealth strategy [28]. During implementation of the REDCap platform at the NHL, initial Internet download and upload speeds were 12.59 Mbps and 16.84 Mbps, respectively (Table 3). The Federal Communications Commission (FCC) considers minimum Internet speeds of 5 Mbps to 25 Mbps as ideal to support online tasks such as file download and telecommuting [29]. However, despite meeting the FCC requirement, study participants reported the need for the Internet speed to be upgraded to adequately support frequent data transactions and data sharing between REDCap and the DHIS2 platform. This could be due to multiple factors. The following are some example quotes from study participants pertaining to their dissatisfaction with the Internet bandwidth:

As a mitigation for the slow Internet connectivity, relevant government departments were engaged, and measures were put in place to increase the Internet speed. This includes procurement of mobile Internet routers from private internet service providers (ISPs) for use during the study. The engagement of private sector stakeholders to support implementation of eHealth initiatives is also highlighted as essential within the “Strategy and Investment” pillar of the Botswana National eHealth Strategy, which emphasizes “eHealth planning, with involvement of major stakeholders and sectors” [28].

Another challenge was the restriction introduced by computer network firewalls resulting in data flow constraints between the GDN and UB networks. A similar issue was encountered in another study, in which implementation of the REDCap platform incurred network firewall challenges hindering participants from using computers outside the Veterans Health Administration network to complete a survey on REDCap [30]. According to Nagpure et al [31], firewall policy conflicts can be complex to eliminate but could be addressed through practical resolution methods such as the “first-match resolution” involving “identifying which firewall policy rule involved in a conflict situation should take precedence when multiple conflicting rules (with different actions) filter a particular network packet simultaneously.” At the NHL, the authors resolved the network firewall challenges by engaging relevant ICT authorities at both governmental and UB IT departments to facilitate reconfiguration of the firewall settings to allow data flow between the 2 networks. This necessitated effective measures for data security, privacy, and confidentiality throughout the study. To achieve this, the authors ensured that the REDcap and DHIS2 servers communicate through an encrypted Internet connection using a 128-bit Secure Sockets Layer (SSL) certificate. The use of SSL has been previously considered a standard technology for securing electronic commerce and electronic banking transactions over the Internet [32]. However, recent advances in cybersecurity attacks call for using technologies to detect compromised SSL network traffic [33]. This consideration was brought to the attention of both governmental and UB IT departments. Compliance with the Botswana-specific Data Protection Act [34] could improve the necessary safeguards for the right to privacy of individuals and the collection and transfer of their personal data.

Some study participants lacked the requisite skills to effectively use the REDCap platform, which was another challenge encountered during the study. The lack of competent and experienced REDCap users in some low and middle-income countries (LMICs) has been linked to poor “REDCap penetration” in those countries [35]. For example, in Botswana, only 3 institutions have REDCap instances [36], and unless affiliated with any of these institutions, most health care workers are unfamiliar with the platform. Low ICT literacy is common among health care workers in most LMICs [25] and could affect the ability of some participants to competently use the REDCap platform. Recognizing this challenge, the Botswana National eHealth strategy included “Workforce Development” as one of its pillars dedicated to eHealth capacity building among health care workers [28].

A reported benefit of having a health workforce trained in ICT prior to engaging with a health information system is minimizing avoidable errors, which, in this case, may be influenced by a lack of familiarity with the REDCap platform or poor ICT competency. To ensure the effective use of REDCap at the NHL, the authors continuously trained participants at all stages in the study. The ever-evolving nature of the REDCap platform [16,36] also necessitates that those working with the system must continuously familiarize themselves with its core functionalities and data workflows through support from the REDCap Consortium. Almost all training needs for this study were identified during weekly virtual progress report meetings with all study participants. Training content was also informed by the identified system requirements for the REDCap platform (Table 2). According to Nsaghurwe et al [37], “functional requirements describe what the system should do—such as its ability to exchange client-level data in a single repository, search for records with data quality issues; while non-functional requirements describe how the system should perform,” including system performance, reliability, and user-friendliness.

In essence, a central aspect to the successful implementation of any health information management system is the need to train users on how to appropriately utilize the system to capture and manage data [25]. Recommended approaches from the existing literature include ensuring continuous engagement and availing training materials for trainees' access and reference at a later stage. In this study, the authors made all training material and manuals available via the REDCap “File Repository” feature for the participants to access at their convenience posttraining.

Informed by the level of participants’ familiarity with the REDCap platform, other measures were put in place such as denying participants “data deletion” privileges on the REDCap system. Instead, participants could add, edit, and view data they required to complete their respective tasks. This minimized accidental data deletion and resonated with the famous “principle of least privilege” by Saltzer and Schroeder [38], which states that “every program and every user of the system should operate using the least set of privileges necessary to complete the job.”

An important lesson learned was that, although REDCap is often regarded as a complete solution for data management [16], in some instances, support from other software applications could augment its limitations. In this study, the REDCap platform was not able to generate barcodes for the NHL personnel to scan to retrieve previously captured information. This limitation was resolved by developing a web-based application supported by the Angular framework [39] for quick production of scannable barcodes required during the Results and Verification Lab stage at the NHL. Some documented benefits of the Angular framework include support for lightweight web applications; faster software development capabilities; and easily readable, testable, and interoperable software solutions [39]. For this study, the authors leveraged the ability to easily link the Angular web application with the REDCap platform. Moreover, only one username and password combination was used to access the data capture forms between the Angular-based web application and REDCap.

Another constraint encountered was the inability of the REDCap platform to automatically store COVID-19 specimen barcodes on the data collection form. Further, the REDCap system requires that the barcode scanner accept a “tab key” to move to the next field, but the scanner configuration used at the NHL accepted a “carriage return key” instead to move to the next field. This meant that each time a barcode was scanned by the study participants, REDCap would not capture it into the appropriate field unless the user pressed the “tab key” on the keyboard to trigger a move to the next field. This became tedious to participants considering the high number of barcodes requiring scanning at the NHL during the COVID-19 pandemic. To resolve this, the authors reconfigured the barcode scanners to align with REDCap requirements, that is, use of the “tab key” to capture barcodes versus the initial configuration of the “carriage return key.” Moreover, the MOH procured dedicated barcode scanners for use with the REDCap platform as the previous ones accompanied a different system at the NHL. Overall, REDCap ensured auto-saving of the barcodes as they are being scanned, minimizing data loss in case of network or power failure.

The Healthcare Information and Management Systems Society (HIMSS) defines interoperability as “the ability of different information systems, devices, and applications (‘systems’) to access, exchange, integrate and cooperatively use data in a coordinated manner, within and across organizational, regional and national boundaries, to provide timely and seamless portability of information and optimize the health of individuals and populations globally” [40]. Some documented benefits of interoperable health care information systems include improved patient management, quality of care, and decision-making, as well as reduced health care costs [41]. However, several challenges have been noted that hinder the interoperability of health care systems, affecting their successful and sustainable implementation, especially in resource-limited countries [10].

Despite the documented interoperability challenges, REDCap has been successfully linked to other systems previously. For example, a study in the United Kingdom automated the process by which COVID-19 clinical trial registration records were exported from the WHO International Clinical Trials Registry Platform into external software [42]. A linking script subsequently pulled relevant data, aligned it with the data dictionary in REDCap, and directly imported it, removing any duplicates in the process. Other researchers utilized REDCap as a data harmonizer, managing and combining cancer research data from multiple registries and allowing for the reconciliation of disparate data sets as well as their conversion [43,44]. In another previous study, REDCap served as the platform housing all global data collected with modified survey questions to capture nuances and allowed for individualization to study public health interventions for COVID-19 [45]. Consequently, these applications reinforced REDCap’s utility for cross-continental collaborations, extending beyond any singular institution or setting.

In this study, although linking the REDCap platform with the DHIS2 platform was achieved, the failure of the REDCap system to automatically pull data from the DHIS2 platform was noteworthy. This limitation was a result of constraints within the Dynamic Data Pull (DDP) feature in REDCap. The REDCap DDP is a special feature for importing data into REDCap from an external source system [46]. The DDP feature provides an adjudication process whereby REDCap users can approve all incoming data from the source system before the data are officially saved in their REDCap project. Because the REDCap DDP assumes that all incoming data from the source system may not be trusted as valid or that only a subset of the data coming from the source system needs to be imported, it utilizes an adjudication web page inside REDCap’s interface to allow manual review of the data obtained from the source system before confirming that it be imported into the REDCap repository [46]. It is precisely the adjudication process that hindered the automatic data import from the DHIS2 system into REDCap. To address this challenge, a custom API was developed by the authors to support accessing health records from the DHIS2 platform, staging them for quality assurance as a batch and automatically loading them into the REDCap system. Further, this approach helped minimize network traffic whenever data synchronization between REDCap and the DHIS2 systems occurred.

The REDCap system was implemented at the NHL in response to the emergency COVID-19 pandemic. Consequently, a systematic approach to its implementation could have been compromised, and essential considerations could have been overlooked. Most notably, multiple training and feedback sessions had to be conducted virtually using the Zoom platform due to restrictions on in-person gatherings. Hence, the authors were unable to frequently visit the NHL in person to provide the necessary technical support and training. Participation in the study was limited and often disrupted, as some NHL personnel were reassigned to efforts such as vaccine distribution or other scenarios not applicable for REDCap use. Last, increased workloads for participants due to COVID-19 surges in Botswana likely contributed to the lack of thorough quality checks when interacting with the system.

Implementation of the REDCap platform to support COVID-19 data management at the NHL in Botswana was successful, albeit with challenges. It is worth noting that, like any software, REDCap as a system possesses limitations that were addressed with other applications to meet requirements for this study. Most challenges encountered were exacerbated by the lack of pandemic response preparedness, as was the case in Botswana and around the world. As such, most of these challenges will not be unique to this specific case of REDCap implementation but will continue to affect the sustainable implementation of eHealth innovations until conscious efforts using key digital health strategies are made that create an enabling environment to support implementation of digital health innovations. Another lesson learned is the essential need for collaborations with key stakeholders to minimize technological barriers while implementing eHealth solutions. Despite the challenges encountered, the REDCap and DHIS2 platforms served as readily available and customizable platforms to address COVID-19 data management at the NHL. To this end, effective planning is essential, including the engagement and training of key personnel to optimize the use of eHealth systems beyond the COVID-19 pandemic.