Maintaining laboratory quality assurance and safety in a pandemic: Experiences from the KEMRI-Wellcome Trust Research Programme laboratory’s COVID-19 response [version 2; peer review: 2 approved]

Laboratory diagnosis plays a critical role in the containment of a pandemic. Strong laboratory quality management systems (QMS) are essential for laboratory diagnostic services. However, low laboratory capacities in resource-limited countries has made the maintenance of laboratory quality assurance, especially during a pandemic, a daunting task. In this paper, we describe our experience of how we went about providing diagnostic testing services for SARS-CoV-2 through laboratory reorganization, redefining of the laboratory workflow, and training and development of COVID-19 documented procedures, all while maintaining the quality assurance processes during the COVID-19 pandemic at the Kenya Medical Research Institute (KEMRI) Wellcome Trust Research Programme (KWTRP) laboratory. The KWTRP laboratory managed to respond to the COVID-19 outbreak in Kenya by providing diagnostic testing for the coastal region of the country, while maintaining its research standard quality assurance processes. A COVID-19 team comprising of seven sub-teams with assigned specific responsibilities and an organizational chart with established reporting lines were developed. Additionally, a total of four training sessions were conducted for county Rapid Response Teams (RRTs) and laboratory personnel. A total of 11 documented procedures were developed to support the COVID-19 testing processes, with three for the pre-analytical phases, seven for the analytical phase, and one for the post-analytical phase. With the workflow re-organization, the development of appropriate standard operating procedures, and training, research laboratories can effectively respond to pandemic outbreaks while maintaining research standard QMS procedures. Competing competing interests were disclosed.

through laboratory reorganization, redefining of the laboratory

Amendments from Version 1
The revised manuscript incorporates the reviewers comments and recommendations. The warm-base concept has been introduced and training approach explained in the discussion forum. The red-designing approach has been revised to provide a guide that can be used by any laboratory performing testing in a pandemic.

Introduction
Laboratory diagnosis forms an essential component of any disease outbreak (Nkengasong et al., 2018;Perkins et al., 2017;Wilson et al., 2018). The importance of well-equipped and prepared laboratories in a pandemic response cannot be understated (Kelly-Cirino et al., 2019). Studies have shown that weak laboratory diagnostic capacity has contributed to significant delays in the identification of the recent outbreaks of Ebola (Perkins et al., 2017), Yellow Fever (Berkley, 2018, and Zika (Lowe et al., 2018). This, therefore, emphasizes the importance of investing in stronger laboratory capacity that can respond to disease outbreaks. This should include elaborate and strong quality management systems (QMS). Weak laboratory quality systems in low-and middle-income countries (LMICs) coupled with the low laboratory capacity, and the technical challenges of developing suitable diagnostic tests for pandemic prevention and containment, has made the implementation of laboratory quality assurance (QA) processes during a pandemic a daunting task (Nkengasong et al., 2018;Olmsted et al., 2010).
Coronavirus disease 2019 , was declared a pandemic by the World Health Organization (WHO) (Cucinotta & Vanelli, 2020) on 11 th March 2020 following its insidious global threat. It was first reported in Wuhan, China in late December 2019 and later spread to more than 200 countries and territories by December 2020. The disease is caused by a virus from the Coronaviridae family, defined as 'severe acute respiratory syndrome coronavirus 2' (SARS-CoV-2) (Ceraolo & Giorgi, 2020). By the time of writing this paper (6th December 2020), COVID-19 had infected over 67.4 million people and caused over one million deaths worldwide (Dennison Himmelfarb & Baptiste, 2020). When the director-general of WHO advised various stakeholders to 'test, test and test' (Adhanom, 2020), it was a clear indication that laboratory involvement in testing would play a vital role in containing the COVID-19 pandemic. Laboratory testing for SARS-CoV-2, the viral agent causing COVID-19, demands that the laboratory produces accurate, reliable, and timely results. In addition, well-trained personnel and a well-established QMS are required to ensure streamlined processes during testing.
Most of the clinical and diagnostic laboratories in LMICs are underfunded and are often overwhelmed during a pandemic (Nkengasong et al., 2018). These laboratories also have feeble and weak quality management systems which if not strengthened makes laboratory diagnosis during a pandemic a daunting task. Thus, redesigning of the research laboratory's workflow and reorganization to increase testing capacity can help provide a better platform in a pandemic response.
The KEMRI-Wellcome Trust Research Programme (KWTRP) conducts integrated clinical, laboratory, epidemiological, and health systems research. Research at the KWTRP is constantly conducted within the local health system; often with health managers and policy makers. This has led to research findings from the programme constantly feeding into local and international health policies (The RESYST/DIAHLS learning site team, 2020). The programme is organized into four scientific departments, namely biosciences, clinical, epidemiology, and health systems and research ethics departments. These are supported by the laboratories, demographic surveillance, clinical services, and stakeholder engagement platforms. The laboratory platform of the programme consists of four main state-of-the-art laboratories. These are the clinical trials laboratory (CTL), Short-Turn-around-Time laboratory (STAT), microbiology laboratory and immunology basic science research laboratory. The laboratory has full accreditation to Good Clinical Laboratory Practice (GCLP) standards, through an accreditation scheme operated by Qualogy (UK) Ltd since 2007, (Gumba et al., 2019) and it has been maintaining compliance to GCLP quality standards to date.
Since the confirmation of the first COVID-19 case in Kenya on 13 th March, 2020 (MoH-Kenya, 2020) KWTRP was mandated by the Kenya government to support the COVID-19 testing for six coastal counties, namely Kilifi, Kwale, Mombasa, Taita Taveta, Tana River and Lamu. To achieve this critical government mandate, we reorganized our laboratory organization, redesigned the laboratory workflow, developed documented procedures, and conducted training for county Rapid Response Teams (RRTs) and laboratory personnel. Since the KWTRP laboratory is primarily a research laboratory in its design, and the laboratory processes and procedures are organized to facilitate research activities, the laboratory was appropriately reorganized to provide COVID-19 diagnostic testing as requested by the government. We leveraged and utilized the existing quality management systems (QMS) in order to maintain the laboratory quality assurance (QA) during the COVID-19 diagnostic testing to provide reliable, valid, accurate and timely results.
In this paper, we provide a detailed description of our experience in reorganizing a research laboratory to provide diagnostic testing, and highlight the lessons learned. The main objective of this paper is to document how to reorganize a laboratory primarily used for research, in order to provide large scale diagnostic testing during a pandemic outbreak, while also maintaining research standard quality management systems in order to provide reliable, valid, accurate and timely results.

Documenting the process
We employed a laboratory management implementation model consisting of four approaches, namely: laboratory reorganization, redesigning of laboratory workflows to accommodate diagnostic testing, training, and development of documents related to COVID-19 diagnostic testing.
Throughout these processes, detailed notes of planning meetings and implementation processes were maintained so as to facilitate both the implementation and monitoring of the processes. The meeting notes included the number of staff working in each area, documentation of testing processes and the testing activities were reviewed during planning meetings. In addition, all the implementation outputs (testing organizational structures, workflow charts, and Standard Operating Procedures (SOPs) were produced in hard copies. These were then used in subsequent review meetings and in monitoring the implementation progress. Moreover, regular review and reflection meetings were also held to discuss the testing process implementation challenges, and, where necessary, adjustments were made to the team composition and/or roles, the laboratory workflows, and the documented procedures. Finally, a review of all the meeting notes and implementation outputs was conducted, and our findings were grouped around common themes for easy presentation. In the section below, we outline in detail each of the four management approaches undertaken; and provide detailed account of the outcomes, and lessons learned.

Ethics statement
This work was conducted as part of the implementation project to document experiences and lessons learned, so as to continuously enhance quality improvements. Formal ethical approval was thus not needed. Institutional authorization and permission to share this experience was however sought and provided by the director-general, KEMRI.

Approach taken
In this section, we outline in detail each of the four management approaches undertaken, and provide a detailed account of the outcomes, and lessons learned. A warm base concept derived from the manufacturing industry was implemented. This concept translated to the laboratory setting refers to using already existing systems to support an outbreak (Association Public Health Laboratories, 2011). We conclude the section by highlighting some of the challenges encountered during the process.

Laboratory reorganization
Laboratories require a well-coordinated and organized structure. This helps in constant communication and facilitate clear and sometimes rapid decision-making processes during a pandemic. To achieve this, and to streamline the testing process, we established COVID-19 diagnostic testing teams with assigned clear roles and responsibilities to each (Table 1). We formed these teams by reviewing our laboratory human resource capacity, and by assessing the projected testing demands from the counties we had been assigned. Through this, we developed and produced a COVID-19 diagnostic testing organizational structure for the laboratory (Figure 1), so as to facilitate effective communication and coordination among the teams.
As shown in Figure 1, each of the seven teams had a team lead who acted as a coordinator and focal person for that particular laboratory function/section and reported to the COVID-19 diagnostic testing project manager. The reorganization of the laboratory to form a COVID-19 testing team comprising seven sub-teams also enhanced communication and the coordination of COVID-19 testing activities, which if not clearly addressed could have been a major detractor to the quality assurance system. The reorganization made it possible for us to support testing from six Kenyan coastal counties.
From our experience, the development and implementation of this organizational structure for COVID-19 diagnostic testing significantly enhanced communication and coordination among different teams and greatly facilitated clear and sometimes rapid decision-making processes. This served to promote efficiency and minimize conflict within the laboratory organization and when liaising with the various county RRTs. The day-today running of the COVID-19 testing and related work was coordinated by the KWTRP Biosciences Head of Department (HoD). As the work volume increased and with a strong foundation in place, daily activities were managed by the laboratory project manager with oversight from the HoD. Each team was tasked with developing documentation for their COVID-19 testing process. Furthermore, the strong laboratory QMS and dedicated QA team, who were well versed in quality assurance activities, allowed for the easy transfer of QMS into the testing process, ensuring that accurate and reliable results could be generated during the pandemic. The QA team ensured that SOPs and documentation of the testing process were performed correctly, and that the laboratory complied to the established documented procedures as well as ascribed standards.

Redesigning of the laboratory workflow
We undertook process mapping ( Figure 2 and Figure 3) to outline how the COVID-19 testing workflow would happen, i.e., from specimen reception to the release of results. This involved redesigning the already established research systems of SOPs, flow of samples, waste management and training to create a diagnostic environment to support the COVID-19 testing. Figure 2 and Figure 3 outline the workflows that were eventually developed following various team meetings attended by the COVID-19 steering committee.
From our experience in implementing these workflows, we found that once they were fully implemented, we had better coordination of the testing processes and better turnaround time for producing and communicating lab results to all relevant authorities. Process mapping helped to determine the available laboratory infrastructure, facilities, and personnel needed to support COVID-19 testing, while ensuring all the quality assurance activities within the testing process were not compromised. The laboratory workflow was redesigned using the Sianipar (2019) model. This involved dividing the pre-analytical phase into two categories namely: the 'conventional' category and the    'pre-pre-analytical phase'. Using this model helped in identifying error prone areas within the pre-analytical phase and putting measures to minimize them. The 'pre-pre-analytical phase' was primarily the responsibility of the COVID-19 county RRTs sample collection centers, while the 'conventional' category involved the process of centrifuging and aliquoting the samples (Figure 2) which was primarily the role of our laboratory (KWTRP). This was done to minimize the pre-analytical phase errors during the COVID-19 pandemic. In addition, an inventory assessment was made of the adequacy of all available equipment that would be required for SARS-CoV-2 testing, this included safety equipment (biological safety cabinets), sample processing equipment (centrifuges), molecular biology analyzers (real time-polymerase chain reaction (RT-PCR), QIAcube HT, and QIAsymphony) and other auxiliary equipment such as pipettes and vortex.
Through mapping of the laboratory workflow, the existing systems in place for non-stock items were adapted by the laboratory and supplies team to support the SARS-CoV-2 testing process. Sample collection reagents such as viral transport media (VTM), and RT-PCR testing kits supplied by the government were managed and made available to the COVID-19 RRTs, and documentation of issue and receipt of these items was well recorded. On receipt of supplies, an inventory was taken, and a report was sent back to the supplier outlining the quantities received. Additionally, new reagent consignments were requested in advance to minimize reagent stockouts in the laboratory. The laboratory supplies team performed inventories to ensure there was no understocking or stockouts and that available supplies were distributed to all RRTs. This was an integral part of maintaining quality assurance during SARS-CoV-2 testing.
During the redesigning of the laboratory workflow, the already established Laboratory Information Management System (LIMS) was used to create a database for COVID-19 testing that linked a unique laboratory identifier number of a specimen to its storage in the laboratory's biobank system (Figure 3). The system that was developed allowed for batch loading of samples to aid in sample reception and printing of barcode labels. Access to LIMS was password restricted. To promote a smooth laboratory and data interface, the existing systems for data entry which REDCap (Version 10.5.1) were modified and customized to support SARS-CoV-2 testing and ensure only reviewed results are authorized and released to the relevant authorities. The data team entered the information from the CIFs into the database and reviewed them to check for any errors in data entries. This was then merged with the final PCR results using the LIMS unique lab identifier to generate individual patient results reports that were shared with the county RRT.

Training
A total of four different types of training were conducted for two different staff categories ( Table 2). All the four types of training were entirely focused on the personnel safety and laboratory preparedness during COVID-19 testing throughout the pandemic. The first training was conducted for county RRTs to equip them with knowledge on proper sample collection and triple packaging to safely transport samples to the laboratory. The rest of the training was performed for laboratory personnel and included risk assessments, infection control procedures, and COVID-19 testing SOPs. These training was designed to equip both the RRTs and the laboratory personnel with knowledge of procedures and laboratory safety, including competency testing.
Conducting training that covered the COVID-19 testing process was critical to the successful maintenance of laboratory quality assurance during the COVID-19 testing. Using the existing laboratory system for training and competency assessment, the COVID-19 testing team and the RRTs were trained on sample collection and transport, risk assessment, infection control procedures, and developed COVID-19 testing SOPs. A risk assessment was carried out for both the county RRTs and the laboratory personnel; and involved identifying laboratory equipment to be used and a possible sample path flow to minimize the risk of infections in the laboratory. Four class 2 biological safety cabinets and a fridge were used only for COVID-19 sample processing and temporary specimen storage, respectively, and for the flow of samples as they came from the field/gates into the (v/v) bleach was provided in areas where COVID-19 samples were processed. Furthermore, staff were encouraged to maintain social distancing and to use personal protective equipment (PPE), which included N-95 masks, disposable laboratory coats, and goggles. Laboratory personnel were briefed on the existing structure for reporting accidents and incidents using the established procedure within the QMS. A risk communication protocol was also developed to ensure that information is modulated and checked for consistency before relaying it, to avoid the creation of panic during a pandemic. The COVID-19 teams were also trained on infection control procedures. This involved refresher training on use of PPEs, waste management, and triple packaging of COVID-19 samples. Due to the close proximity and close working relationship of the Kilifi County Hospital (KCH), several documents such as tracking logs and SOPs were shared between the institutions to support the Kilifi RRTs. The goal of the infection control procedure training was to ensure safety of the healthcare worker, the laboratory personnel, and the environment to help prevent infections.

Development of COVID-19 testing standard operating procedures
To ensure the laboratory maintains the quality assurance within the COVID-19 testing process, we developed and documented various new SOPs; and reviewed the existing ones to support the COVID-19 diagnostic testing. The SOPs were categorized according to the three phases of the COVID-19 testing process (Table 3).
Procedure for specimen reception and rejection was developed to ensure samples of high integrity were received. The specimen rejection criteria involved checking and excluding leaking samples, duplicate samples, samples collected using the wrong swabs or swabs with no Viral Transport Media (VTM), samples without case investigation forms (CIFs) or specimen tubes with illegible labels (Figure 3). The county RRTs were informed by the project manager of samples with missing CIFs and allowed 48 hours for communication to be received on the missing CIFs. Due to the limitations in sample collection kits and the need for rapid contact tracing from the positive cases, the rejection criteria were leniently revised to accept samples with legible labels but with no CIFs, hence the 48-hour window for communication. Consequently, the existing health and safety documented procedures for specimen transport, processing, and waste management were reviewed with reference to the United States Centers for Disease Control and Prevention (CDC) and the WHO COVID-19 laboratory safety guidelines (Centers for Disease Control and Prevention, 2020; WHO, 2020a; WHO, 2020b; WHO Laboratory Biosafety Manual, 2020).
Development of the SOPs for the COVID-19 testing activities enhanced training and promoted consistency in performing the COVID-19 testing procedures. Since this is a novel virus and the SOPs (Table 3) had to be developed from scratch, there was need to develop validation protocols and have them approved to avoid the risk of generating RT-PCR false-negative or positive results that would undermine the reliability of test results. The validation protocols were developed, and assay performance was optimized to ensure valid and reliable results were generated (Said et al., 2020). Lot numbers of the testing kits and the enzymes were verified, and parallel testing was performed to ensure the kits and the reagents were in good working condition and consistent in producing reliable results. Procedure-for pooled testing was also developed to enable the laboratory perform testing of approximately 554 samples daily with limited resources (Agoti et al., 2021). The identified positive pools were, on the same day, expanded and ran as singlets to identify the positive samples within the pools. To manage and streamline the pooling of samples during sample preparation, a working tool (Excel spreadsheet) was developed to assist in the pooled testing. This is available in Extended data, (Gumba, 2021). Sample IDs were entered into a specimen list worksheet that automatically generated several worksheets with the same specimen unique identifier. The worksheets included a PCR plate map, a worksheet that uploads the specimen list to the RT-PCR analyzer, and a worksheet for the acquisition of raw results from the analyzer to generate an interpreted output ( Figure 3). The interpreted results worksheet contained formulas that incorporated the validity criteria (cycle threshold cut-off values) of all negative and positive controls, yielding 'VALID' in green or 'INVALID' in red. The results were reviewed by highly trained and competent personnel who continually monitored results against all controls and carefully defined thresholds for defining a positive or negative test result. This automated approach minimized errors and was an efficient way of handing the results over to the data team, and for handing notes to testing team the next day. Moreover, it provided a standardized tool for the pooling system and results interpretation, and at the same time it took care of the full documentation of all the specimens tested that day. The system also recorded the performance of the QCs, as is required for compliance with GCLP.

Challenges
We experienced a major challenge of long lead times between ordering and receiving of the RNA extraction kits. The unavailability of qPCR testing kits was another challenge which was due to the high demand of these kits during the pandemic.
Moreover, there was limited information on validation and verification studies since most methods were still under development. Additionally, long lead times of global laboratory supplies were experienced, thus we also obtained resources independently and maximized on the available resources (Agoti et al., 2021;Said et al., 2020). Similarly, it was a big challenge of ensuring that individuals were confident in the health and safety measures that were put in place to minimize the risk of infection in the workplace.

Discussion
The reorganization of the laboratory to form the COVID-19 team with seven sub-teams was our proactive role to meet and sustain the overwhelming and urgent need for testing during the pandemic (Lippi & Mattiuzzi, 2019;Lippi et al., 2020). This allowed us to cope with the intense pressure brought about by testing hundreds of specimens a day, coupled with the large amounts of paperwork and documentation. The formation of the COVID-19 team with assigned responsibilities and the redistribution of the workload across the SARS-CoV-2 testing process enhanced the maintenance of quality assurance and enabled a high standard of testing. Through this process, personnel were trained and empowered to confidently perform their tasks within the established COVID-19 organizational and management structure. More importantly, involvement of laboratory personnel in all aspects of the SARS-CoV-2 testing process from the onset promoted responsibility, accountability and ownership of processes and decisions made within each team.
Redesigning the laboratory workflow was a critical factor that enhanced maintenance of our laboratory QA processes during the pandemic. This was done by employing the 'warm base' concept which is a concept derived from the manufacturing industry and means being ready for production when a service or a product is needed (Association Public Health Laboratories, 2011). Translated to the laboratory context, it means using the already existing systems to support an outbreak. We did not set up systems from scratch for the novel SARS-CoV-2 pandemic, but we used the already established systems employed in our research and clinical trials laboratory in the laboratory where COVID-19 testing was being performed. The successful maintenance of QA and safety process during the COVID-19 testing was due to mapping of the COVID-19 testing process, conducting training, performing competency and risk assessments. Additionally, the laboratory strengths in health and safety, laboratory supplies and equipment management equally made this possible., Moreover, using the already established equipment management system, the documentation (validation records, installation records, service records and preventive maintenance records) for the identified equipment were rechecked to ensure the equipment were in good working conditions and maintenance activities were performed routinely during the SARS-CoV-2 testing process.
The laboratory also performed interlaboratory comparison for other testing laboratories within the coastal region. In this process, the testing laboratories were sending five already analyzed samples to our laboratory so that we could test and check if the results were comparable. We also enrolled in the molecular SARS-COV-2 external quality assurance (EQA) scheme provided by the Royal College of Pathology Australia (RCPA) and also participated in the proficiency testing organized by the World Health Organization and provided by the National Influenza Laboratory. All these were to ensure that the results being generated by the laboratory are valid, accurate and reliable.

Conclusions
Ensuring quality assurance in the laboratory during a pandemic and ensuring staff work safely is not a trivial process. However, capitalizing on the existing laboratory systems greatly supported the rapid set up of the COVID-19 testing, yielding reliable, accurate and timely results in a safe work environment. The involvement of key laboratory personnel, the development of documented standard operating procedures and the formation of a COVID-19 team that is organized and properly coordinated, led to the successful maintenance of laboratory quality assurance during these challenging times. The implemented steps, discussions and suggestions highlighted in this paper can be transferable to any laboratory planning to take up testing during a pandemic.

Underlying data
No underlying data are associated with this article.