Many countries have repurposed pre-existing public venues or infrastructure as community care facilities (CCFs) to provide treatment for those infected with coronavirus disease 2019 (COVID-19), reserving hospital care for patients with serious sequelae. ^{1- 3} An inherent challenge in using non-healthcare public venues, is to ensure adaptations to layout, infrastructure, workflows, and work practices are designed to support infection prevention and control (IPC) for staff and patient safety. Reports of frontline healthcare professionals who contracted COVID-19 and lost their lives, shock and sadden. ^{4- 8} Although no official worldwide tally is available, conservative estimates are that between 7,000 to >20,000 healthcare professionals have died in the line of duty, as of October 2020. ^{9, 10}

In Singapore, two waves of imported and community cases of COVID-19 in March/April 2020 were dwarfed by a sharp rise in the number of infected migrant workers housed in dormitories. ¹¹ The COVID-19 multi-ministry taskforce commissioned CCFs to be rapidly set-up as a hospital-sparing strategy in April 2020 amidst rising cases, to provide care for patients with mild symptoms. ¹² Given fatalities in healthcare professionals reported around the world due to COVID-19, ensuring staff safety in CCFs was a priority. We were tasked to set-up a CCF within ten days, repurposing halls in an exhibition centre (hereafter CCF@EXPO) to isolate and treat infected patients in the community. ¹³ However, due to the speed required in setting up CCFs, a prospective theory-informed approach to implementation is uncommon. At the time of writing, we found no published studies using implementation science to plan, set-up, and operate a large-scale CCF in this pandemic.

We achieved our goal of zero transmission of COVID-19 to healthcare and non-healthcare staff working at CCF@EXPO, while providing high quality patient care. This paper is a retrospective analysis using implementation science to elicit critical factors in the achievement of the stated goal, and to inform similar endeavors in the future. ¹⁴ Contextual barriers and implementation strategies used to address these are surfaced. Findings are reported in line with the Standards for Reporting Implementation Studies (StaRI) checklist. ¹⁵

The assessments were conducted by three of the authors (AT, NT, SLC). ³¹ Baseline scores for innovation characteristics indicated that ‘physical environment’, ‘work practices’, ‘tools and technology’, and ‘staff training’ needed to be addressed from an IPC perspective ( Figure 1a and extended data file 4 ³¹ ). The main pre-commencement/baseline challenges were ( Table 1): inadequate facilities for PPE donning/doffing and inadequate space for staff to rest/have meal breaks (physical environment); hospital IPC workflows and practices required modifications to suit CCF@EXPO (work practices); missed opportunities to reduce dwell time in patient areas (suboptimal use of tools and technology); and training especially for non-healthcare staff in PPE use, ²¹ safe workflows, and safe behaviours (staff training). At baseline, only basic IPC audits were in place. Figure 1. Table 1.

i-PARIHS constructs	Main contextual barriers for infection prevention and control a	Key ERIC Implementation strategies used b [ERIC list number]
Characteristics of the Innovation
Physical environment	Inadequate area for PPE donning/doffing (pre-commencement). Inadequate space for staff rest and meal breaks to allow segregation between teams and safe distancing within teams (pre-commencement).	Assess for readiness and identify barriers and facilitators [4], Tailor strategies [63], Visit other sites [72], Change physical infrastructure and equipment [11], Facilitation [33]. 30
Work practices	Hospital IPC work practices not applicable/required modifications to be suitable for this site. Not all work practices were clearly documented. Safe workflows required. Safe behaviours required.	Conduct local needs assessment [18], Conduct cyclical small tests of change [14], Audit and provide feedback [5], 30 Promote adaptability [51], Identify and prepare champions [35], Visit other sites [72], Facilitation [33]. 30
Tools and technology	Suboptimal use of tools and technology to enhance staff safety by reducing dwell time of staff in ‘hot zones’ (patient areas), improve workflow, and increase efficiency in provision of care.	Conduct local needs assessment [18], Conduct cyclical small tests of change [14], Identify and prepare champions [35], Tailor strategies [63], Promote adaptability [51], Audit and provide feedback [5], 30 Provide local technical assistance [54], 30 Centralised technical assistance [8], Facilitation [33]. 30
Staff training	Non-healthcare staff required training in correct use of PPE, safe workflows, and safe behaviours. Healthcare staff required similar training/refresher training to meet site requirements. Staff rotations and impact on training. Staff acceptability of myriad new systems, processes, protocols, practices, and policies. Only a fire contingency plan was in place. Other contingency plans pertinent to this site are required.	Conduct local needs assessment [18], Audit and provide feedback [5], 30 Conduct cyclical small tests of change [14], Conduct ongoing training [19], Develop educational materials [29], Distribute educational materials [31], Facilitation [33], 30 Identify and prepare champions [35].
Audits	Only basic audits for IPC were in place. Other audits pertinent to this site are required.	Develop and implement tools for quality monitoring [26], 30 Develop and organise quality monitoring systems [27], 30 Audit and provide feedback [5], 30 Facilitation [33], 30 Identify and prepare champions [35].
Recipients
Motivation to change, individual level	No key barriers.	To maintain and enhance: Facilitation [33]. 30
Ability to change, individual level	No key barriers.	To maintain and enhance: Facilitation [33]. 30
Motivation to change, team level	No key barriers.	To maintain and enhance: Identify and prepare champions [35].
Ability to change, team level	Challenges in building a resilient, well-functioning team. Staff deployed were from different institutions, had not worked as a team previously, diverse range of disciplines and specialties. Boundaries experienced include professional, syntactic, semantic, and pragmatic boundaries (pre-operational). Significantly reduced staff to patient ratio compared to pre-COVID-19 (pre-operational). Involvement of multiple external vendors for bed and facility management, cleaning services, security services, and food catering.	Facilitation [33], 30 Identify and prepare champions [35], Recruit, designate, and train for leadership [57], Inform local opinion leaders [38], Build a coalition [6], Conduct local consensus discussions [17], Capture and share local knowledge [7], Create a learning collaborative [20], Create online learning communities [--], 30 Model and simulate change [45], Obtain formal commitments [47], Promote network weaving [52], Provide ongoing consultation [55].
Context
Inner context, local	No key barriers.	To maintain and enhance: Facilitation [33], 30 Identify and prepare champions [35], Capture and share local knowledge [7], Organise clinician implementation team meetings [48], Promote network weaving [52].
Inner context, organisational	No key barriers.	To maintain and enhance: Facilitation [33], 30 Build a coalition [6], Identify and prepare champions [35], Obtain formal commitments [47], Use advisory boards and workgroups [64], Involve executive boards [40], Develop academic partnerships [24].
Outer context	No key barriers.	To maintain and enhance: Facilitation [33], 30 Involve executive boards [40].

^a

Main contextual barriers at baseline, unless otherwise stated.

Several key implementation strategies were used to address baseline challenges for innovation characteristics ( Table 1), driven by Deming’s Plan-Do-Study-Act (PDSA) methodology. ²² For example, Plan—assessing local needs for safer work practices (PPE donning/doffing, Figure 2); Do—developing enhanced practices (conduct training in the use and sequence of PPE donning/doffing, provide gowning stations, posters as visual reminders); Study—audit and feedback (PPE donning/doffing audit, ‘PPE buddy’ (staff were assigned a staff ‘buddy’ for verification of correct donning/doffing of PPE), mirrors); and Act to refine work practices further (trial of Smart mirror). The practice of using posters as a checklist was normalized. A no-blame culture of continual learning and looking out for one another was modelled and encouraged by facilitators and leaders. By the 100 ^th day, PPE compliance had improved from 97% to 100% and from 93% to 98%, respectively, for healthcare and non-healthcare staff. These outcomes were more favourable when compared to another Singaporean study in a hospital setting. ²³ Assessments showed positive shifts in all five elements of innovation characteristics by the 100 ^th day. Figure 2.

For recipients, baseline scores although favourable ( Figure 1b and extended data file 4 ³¹ ), could be strengthened given the challenge of caring for >3000 COVID-19 patients with lean human resources. ¹³ The ability to change at the individual level remained stable between the two timepoints—individual staff understood the modifications in workflows and practices for delivery of patient care, and each had the ability, resources (e.g. PPE supplies), and support to enact these workflows and practices. However, positive shifts in scores from baseline to 100 ^th day were observed for motivation to change at individual level, motivation to change at team level, and ability to change at team level.

Key implementation strategies ( Table 1) used to strengthen individual and team efficacy include: facilitators being skillful in identifying champions and domain leads, team building, and creating virtuous cycles centred on cultures of collaborative learning, multi-disciplinary teamwork, as well as transparency in communications and decision-making. Key strategies were enacted mainly through rapid PDSA cycles serving as vehicle and enabler of staff initiatives to improve processes, to identify and raise relevant issues impacting staff safety and patient care, as well as to develop, test, refine and own the solutions. Experiencing themselves as empowered agents of change individually and collectively was critical for rapid ground-up innovation in a pandemic. In the end, increased efficacy and agency of individuals and teams allowed out-of-box thinking beyond comfort zones. New tools and technologies, accompanied by new procedures/protocols refined via PDSA cycles, were adopted as routine work practices. For example, an autonomous robot ( TEMI ^® , New York, USA) was used to provide medication counselling and tele-triage to reduce staff dwell time in patient areas, and fluorescent marker (Glo-Germ) was used for objective assessment of environmental cleaning during audits ( Figure 3). Figure 3.

Baseline scores for context were favourable ( Figure 1c and extended data file 4 ³¹ ), which signaled good alignment between culture, leadership, systems, policies, and strategic priorities at local, organisational, and national levels in Singapore. To maintain alignment, facilitators used implementation strategies such as identifying champions at various levels, engaging with executive/advisory boards (for clinical governance, resource allocation, accountability reporting), and developing partnerships (for CCF@EXPO operations or research). Favourable scores for context were maintained at the 100 ^th day ( Figure 1c and extended data file 4 ³¹ ).

Facilitation featured strongly as an implementation strategy that ensured success. The findings of the assessment indicate that these are seasoned/expert facilitators (extended data file 5 ³¹ ). Facilitators used various communication techniques and channels (formal and informal means) to engage, involve, inform, update, encourage, acknowledge, gather support, and celebrate wins/milestones with staff and senior stakeholders (extended data files 3 and 6 ³¹ ). A ‘flat organisational structure’ was intentionally adopted to enable open communication and ease of access to facilitators. This non-hierarchical structure was important and necessary for encouraging a constant flow of ground-up innovations and initiatives, especially as few in the team were experts in managing COVID-19 patients.

Importantly, facilitators were given authority, autonomy, support, and resources to implement the CCF@EXPO IPC innovation. Facilitators displayed desirable attributes (16/18, 89%), performed essential tasks (11/13, 85%), and demonstrated a diverse range of required skill sets (14/15, 93%). Facilitators were skilled in addressing task, team, and individual needs ( Table 2), and were able to iteratively move through the four phases of facilitation articulated in the i-PARIHS framework: (a) clarify and engage (identify the problem, get the right people together); (b) assess and measure (appraise the evidence, consider recipients and context); (c) action and implementation (use rapid PDSA cycles); and (d) review and share (audit/re-audit, timely feedback, sustain, spread). ¹⁹ Apart from working with staff on the ground, facilitators actively engaged SingHealth organisational stakeholders and decision-makers at the National/Ministry level. Facilitators focused on goals, thus acted as translators and bridges across professional, semantic, syntactic, and pragmatic boundaries. Table 2.

CCF@EXPO: Community care facility repurposed from the Singapore exhibition centre; EXPO: Refers to the Singapore exhibition centre; PPE: Personal protective equipment.

The SingHealth CCF@EXPO required rapid planning and set-up to deliver healthcare to >3000 COVID-19 patients. Clinician facilitators worked closely with experts in IPC to design a tightly woven set of strategies to overcome contextual challenges to ensure staff safety while delivering high quality patient care. To actualize the goal of zero transmission of COVID-19 to staff, the following principles were used as a basis for the planning, set-up, and operations of CCF@EXPO: (a) reduce opportunities for transmission by minimizing staff dwell time in patient areas; (b) reduce risks of transmission by ensuring that physical infrastructure, environmental cleaning, workflows and practices support IPC; (c) enhance staff protection by ensuring staff competency and adherence to correct PPE use; and (d) constant monitoring via audit and feedback mechanisms to quickly identify emerging issues that threaten staff or patient safety. We reflected these principles in the ‘characteristics of the innovation’ construct within the i-PARIHS framework.

The goal of zero transmission of COVID-19 to staff at a large CCF was achieved via skilful facilitation amidst significant contextual challenges. The use of rapid PDSA cycles anchored by principles to ensure staff safety was instrumental for implementation success. The i-PARIHS framework is relevant and useful for deconstruction of rapid implementation in a pandemic. It can inform preparedness planning for future scenarios of volatility, uncertainty, complexity, and ambiguity.

The data referenced by this article are under copyright with the following copyright statement: Copyright: � 2021 Chow W et al.

Data associated with the article are available under the terms of the Creative Commons Attribution Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original data is properly cited.

EL conceptualized the paper and methods, provided critical input into the use of implementation science, and developed the assessment rubric. WEC piloted the assessment rubric and developed the 14PS Matrix which provides high-level documentation and rationale for strategies used. NT and HKT obtained research ethics approval. LCL and MLL were instrumental in the development and implementation of IPC practices, including audits. Data was collected, collated, and provided by WEC, AT, LCL, HKT, and NT. WEC, AT, SLC, HKT, and NT provided critical review of the assessment rubric; EL refined the rubric. Assessment was conducted by AT, SLC, and NT; results were collated by SLC. EL and WEC had full access to the data in this study; and verified, analyzed, and interpreted the data. Radar diagrams were prepared by SLC. Photos were provided by LCL and WEC. EL prepared the figures, tables, and boxes. EL and WEC drafted and refined the manuscript. All authors reviewed the manuscript, provided critical input, and approved the final manuscript for submission.

No competing interests were disclosed.

The authors declared that no grants were involved in supporting this work. The CCF@EXPO was funded by the Singapore Ministry of Health (sponsor). The sponsor was not involved in the study design, data collection, analysis and interpretation, reporting of the study, or in the decision to submit this paper for publication.