key: cord-0800135-msvnd5j5
authors: Wei, H.; Daniels, S. A.; Whitfield, C. A.; Han, Y.; Denning, D. W.; Hall, I.; Regan, M.; Verma, A.; van Tongeren, M. J.
title: Agility and sustainability: A qualitative evaluation of COVID-19 Non-pharmaceutical Interventions (NPIs) in the UK logistics sector
date: 2022-01-29
journal: nan
DOI: 10.1101/2022.01.28.22270013
sha: 132f10ae6db6aa193714eaf874c150aea3719b3f
doc_id: 800135
cord_uid: msvnd5j5

Background: The emergence of SARS-CoV-2 triggered a chain of public health responses that radically changed our way of living and working. Non-healthcare sectors, such as the logistics sector, play a key role in such responses. This research aims to qualitatively evaluate the non-pharmaceutical interventions (NPIs) implemented in the UK logistics sector during the COVID-19 pandemic. Methods: We conducted nine semi-structured interviews in July-August 2020 and May-June 2021. In total 11 interviewees represented six companies occupying a range of positions in the UK logistics sector, including takeaway food delivery, large and small goods delivery and home appliance installation, and logistics technology providers. Inductive thematic analysis was completed using NVivo12 to generate emerging themes and subthemes. Themes/subthemes relevant to interventions were mapped deductively onto an adapted Hierarchy of Control (HoC) framework, focusing on delivery workers. Themes/subthemes relevant to the process of implementation were analyzed to understand the barriers and facilitators of rapid responses. Results: HoC analysis suggests the sector has implemented a wide range of risk mitigation measures, with each company developing their own portfolio of measures. Contact-free delivery was the most commonly implemented measure and perceived effective. In addition, a broad range of measures were implemented, including social distancing, internal contact tracing, communication and collaboration with other key stakeholders of the sector. Process evaluation identified facilitators of rapid responses including capacity to develop interventions internally, localized government support, overwhelming external mandates, effective communication, leadership support and financial support for self-isolation, while barriers included unclear government guidance, shortage of testing capacity and supply, high costs and diversified language and cultural backgrounds. Main sustainability issues included compliance fatigue, and the possible mental health impacts of a prolonged rapid response. Conclusions: This research identified drivers and obstacles of rapid implementation of NPIs in response to a respiratory infection pandemic. Existing implementation process models do not consider speed to respond and the absence or lack of guidance in emergency situations such as the COVID-19. We recommend the development of a rapid response model to inform the design of effective and sustainable infection prevention and control policies and to focus future research priorities.

The study offered important insights into the process of the UK logistics sector's 53 response to the COVID-19 pandemic. 54

•

The Hierarchy of Control (HoC) framework was adapted for the evaluation of a 55 collection of non-pharmaceutical interventions (NPIs) implemented in a non-healthcare 56 sector. 57

• Thematic analysis of qualitative data generated themes that were relevant to the 58 process of rapid implementation of NPIs during a public health emergency. 59

• Barriers, facilitators and sustainability issues of the sector's rapid response to the 60 COVID-19 pandemic have been highlighted to inform future research on implementation 61 strategies. 62

The novel severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) virus shocked the 64 world in the last few days of 2019 and we still very much live in this Coronavirus disease 65 2019 pandemic at the time of writing. In the UK, the logistics sector worked 66 together to keep the workers and customers safe and increased capacity to cope with the 67 sustained high level of demands. The sector employs and contracts a large number of workers 68

to deliver a wide range products and goods to private and commercial addresses; many of 69 them are self-employed. They could face both health and financial risks over a pandemic (1) , 70 and contribute to community transmissions (2-4). An analyses of COVID-19 mortality in 71

England showed that, similar to other essential workers, van drivers had an increased risk of 72 death from COVID-19, compared to non-essential workers (5). It is therefore important to 73 introduce risk mitigation measures (RMMs) within this sector. Non-pharmaceutical 74 interventions (NPIs) are often significant investments that require well-coordinated actions by 75 multiple stakeholders across organizations and society (6, 7). To cope with imminent threats, 76

such as a novel disease pandemic, interventions must be deployed rapidly to ensure 77 behavioral and mindset changes occurring within a short time frame. In the case of 19, mathematical models suggested that restrictive measures to reduce social mixing could 79 reduce virus transmission and must take effect in a matter of days in order to save lives (8-80 10) . While research about the health systems' response to public health emergencies has 81 provided good quality evidence (11, 12) , similar evidence on the contribution of control 82 measures in non-healthcare sectors, such as the logistics sector, to control work-related 83 transmission is so far lacking (13) (14) (15) . Hence, it is imperative to learn more about what 84

RMMs were implemented by the UK logistics companies, the barriers and facilitators of 85 implementation and whether the control measures are sustainable in the long-term. The aim 86 of this study was to answer these questions through interviews that explored the company 87

representatives' opinions and experiences. 88

We have generally followed the Consolidated Criteria for Reporting Qualitative Studies 90

(COREQ) to report the methods and findings (16). A checklist can be found in Additional file 91

1. 92

We recruited companies through a variety of approaches, such as direct contact, approaching 94 trade and industry associations, via personal and professional networks and a social media 95 campaign on LinkedIn. All recruitment activities were carried out using phones, emails or 96 online facilities. We completed nine semi-structured interviews with six companies between 97

July and August of 2020 (Round 1) and May and June of 2021 (Round 2), with three 98 companies interviewed twice. Each of the interviews lasted between 60 and 90 minutes. 99

There were in total 11 participants as four companies had two or three representatives. 100

All participants received a study scope and Participant Information Sheet and gave verbal 101 consent before the interviews began. We used the Zoom teleconferencing facility to audio 102 record the interviews. Three trained postdoctoral researchers (HW, SD, CW) carried out all 103 the interviews, with attendance by other members of the study team. Interview schedules 104

were developed in advance, with open ended questions which included inquiries on the type 105 of RMMs implemented, facilitators and barriers of implementation, recommendations for 106 possible future pandemics and potential health impacts of coping with a long pandemic. The 107 interview schedules for both round 1 and 2 are available in Additional file 2. A summary 108 report was emailed to each participating company for comments and corrections. One 109 company returned written comments and another discussed feedback with us over Zoom. 110

HW, SD and CW edited and anonymized the auto-transcripts generated by Zoom. One 112 company supplied a detailed list of events from February 2020 to July 2020, which was also 113 analyzed. Thematic analysis was carried out using NVivo12 software following the latent 114 approach (17, 18). HW and SD studied the transcripts and events list and identified emerging 115 themes and sub-themes, which were developed into new codes to be fitted within the 116 codebook. . Thereafter the themes and subthemes relevant to RMMs for delivery workers 117

were deductively matched, if appropriately, with the levels of the Hierarchy of Control (HoC) 118 (19, 20) . HoC ranks preventative measures according to their expected level of protectiveness 119 against one particular hazard, moving from the most protective measures that eliminate the 120 hazard completely from the work environment, down to personal protective equipment 121 (PPE), the last layer of protection for workers (see Figure 1 ). The mapping exercise was 122 reviewed and discussed extensively within the team and with experts from the Health and 123

Safety Executive (HSE) and Public Health England (PHE, now known as UK Health Security 124 Agency). Codebooks were developed separately for the two rounds of interviews to allow for 125 changes that occurred over the course of the pandemic. Based on the agreed codebooks, HW 126 and SD then coded the transcripts independently, with coding results merged to assess inter-127 coder reliability. The percentage of agreement between the two coders was very high (>90%) 128 and the average Kappa coefficient was 0.61 for the first round and 0.51 for the second round 129 (0.41-0.75 is considered fair to good (21)). Individual codes that showed higher discrepancy 130

were discussed and consensus was reached. 131 132 Figure 1 HoC: COVID-19 -Delivery workers (adapted from HSE websites i ) 133

The HoC analysis focused on the delivery workers who would collect deliveries from a 134 workplace (i.e. warehouses or depots) and deliver them to customer premises, using a certain 135 type of vehicle. For large items, they might also enter customer premises in order to drop the 136 deliveries to a designated room (Room of Choice) or to complete the installation. 137

Results 138

Participants represented one takeaway food delivery platform, four logistics companies that 140 delivered large and small items and one technology provider for food and grocery chain 141 stores i.e. supermarkets and restaurant chains. Most of the representatives that we recruited 142 were directly involved in the day-to-day running of the logistics business. However, for 143 grocery store deliveries, we only managed to recruit a technology developer that served the 144 food and grocery chains. All the delivery companies were large employers (500+) except the 145 technology developer. The roles of the participants covered a range of functions in the 146 companies, including health and safety, operation, operational support, communication, 147 marketing and external affairs. Delivery of large items was normally fulfilled by two-person 148 teams, while parcel and takeaway food deliveries were fulfilled by lone drivers or bicycle 149 riders. Of the five delivery companies, delivery workers were engaged as self-employed in 150 four, with one large items delivery company employing drivers directly. 151

. CC-BY-NC-ND 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted January 29, 2022. ; https://doi.org/10.1101 https://doi.org/10. /2022 HoC analysis focuses on the interventions. A wide range of RMMs were designed and 153

implemented by the interviewed companies. Through the pandemic, they continued to do so 154 to tackle newer challenges, such as the emergence of new variants, risks of increased 155 transmission during the winter season, and adapting to new government measures, such as 156 mass testing and vaccination. HoC analysis excluded the technology provider as they were 157 not directly involved in delivery work. We summarized the RMMs that were discussed in the 158 interviews in Table 1 . Food 1 refers to the takeaway platform, Parcel 1 and 2 refer to the two 159 parcel delivery companies, and Large 1 and 2 refer to the two large items delivery companies. 160

Food 1 engages couriers using an app and does not operate any physical sites, while the other 161 four companies do, of which, Large 1 and 2 also provide company vehicles. 162 is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint refers to fixing each two-person delivery team permanently. Before the pandemic these pairs 214 would change every day or in some cases multiple times per day. Drivers and warehouse staff 215 would be grouped by location to establish working group bubbles, with no rotation between 216 sites. The key was to keep the same teams together as much as possible to reduce the number 217 of contacts, and to make workplace contact tracing more effective. When a case was 218 confirmed, the workers who had been in close contact with the infected individual would be 219 notified immediately to go into self-isolation. The other AC measures reported included 220 staggered working where breaks and beginning of shifts were staggered at intervals, i.e. 15 221 minutes to minimize contact. All of the interviewed companies demonstrated a strong 222 capacity in workplace infection rate monitoring, especially in the second round of interviews. 223

Four of them stated infection rates in the workforce merely reflected community infection 224 rates, indicating limited workplace transmission. One reported they had outbreaks within 225 workplaces when the Alpha variant emerged in winter 2020. They then immediately 226 deployed third-party testing facilities to test the entire workforce at those sites. 227

For personal protection and personal hygiene, participants reported they provided drivers 228 with face coverings, gloves and hand sanitizers. 229

In this section, we investigate the process of implementation and identify barriers, facilitators 231 and issues that might affect sustainability. The process had prominent features, such as the 232 speed to action, the overwhelming external pressure, improvised interventions, ad hoc 233 approach, a fast-evolving situation and steep learning curves for all stakeholders. Based on 234 the emerging themes of our thematic analysis, we summarized 15 key characteristics of rapid 235 responses that can be categorized into five domains, with relevant barriers and facilitators 236 identified in Table 2 . 237 is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted January 29, 2022. ; https://doi.org/10.1101/2022.01.28.22270013 doi: medRxiv preprint (Table 2 is presented at the end of the manuscript) 240 241

The themes that emerged in relation to interventions characteristics included source of 243

interventions, strength & quality of evidence and costs. 244

The companies developed the interventions drawing from both external and internal sources. 245

External sources were mostly government guidance such as social distancing, face covering 246 and hand washing, which were relatively standardized. For companies that operate in 247 multiple countries, signals from other countries also provided sources of intervention. For 248 example, Parcel 1 mentioned they had secured a supply of facemasks (described as three-249 layer paper masks) for their UK workers, as colleagues from across the world recommended 250 this as a preventative measure at the early stages of the pandemic. Internally developed 251 measures generally followed the principal of minimizing contact but with customized 252 characteristics. Contact-free delivery is an example of an internally developed intervention 253

with slightly different features designed by each company. Both Parcel 1 and 2 used 254 photographs to replace customer signatures, while Large 1 required no signature and Large 2 255 asked the driver's "mate" (the other personnel in a two-person delivery team) to sign as a 256

proof of delivery. Food 1 required no signature and strongly advised online payment. When 257 cash payment was necessary, they then asked the money to be put into an envelope. 258

Barrier 1 Barriers to rapid development of interventions here appeared to be the lack of and 259 changing government guidance. 260

The resourcefulness and capacity to design and develop interventions internally 261 appeared to be a facilitator. 262

The companies reported how they actively collected data to monitor the effectiveness of 263 communication and infection rates. They mentioned customer and staff surveys, monitoring 264 message click rate and dwelling time, and monitoring infection and self-isolation rates. 265

Participants appeared to be more confident about the quality and validity of the evidence in 266 round 2. During round 1, they generally reported a very low number of confirmed cases, 267

while during round 2, participants provided more details about how they collected and 268 analyzed data systematically. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted January 29, 2022. ; https://doi.org/10.1101/2022.01.28.22270013 doi: medRxiv preprint the self-employed and additional bonuses, were direct costs. There was also other investment 283 such as communication systems, posters and markings, sanitary stations, physical barriers and 284 alteration of workplaces. One participant mentioned that costs related to were in the range of hundreds of thousands of pounds every month at that time. However, 286 participants were also in universal agreement that the companies had been performing well 287 since the first lockdown as demand increased and sustained at high level, which might have 288 helped to absorb the costs. 289

Three themes emerged in relation to external environment, including prioritization of COVID 291 safety, unprecedented collaboration within the industry and overwhelming external mandates 292

to enact rapid responses. 293

The UK government imposed lockdown measures in March and November 2020 and January 294 2021 to stop non-essential contact and travel. Nevertheless, delivery of food and other 295 essential supplies was recognized as essential work by the government. Hence worker and 296 customer safety must be prioritized and the companies modified work procedures to reduce 297 work contact, including suspension of services, such as installation or Room of Choice, and 298

stopped procedures, such as signing on documents or equipment. 299

The level of collaboration within the industry was unprecedentedly high as reported by the 300 participants. It included working with the sector including competitors, the government and 301

international collaboration within the organizations. 302

Facilitator 2 Localized government support was a facilitator of the rapid response. 303

Participants described working with the local police, Department for Environment, Food and 304

Rural Affairs (DEFRA), HSE, PHE, National Health Service (NHS) and local authorities. 305

When there was a high level of uncertainty, the companies appreciated the support from local 306 authorities and local branches of HSE, PHE and unions. They would send their internal 307 guidance and risk assessment to these bodies and obtain their opinions. The support was 308 personalised to the companies, which then provided the companies with confidence to 309 implement these measures. 310

Networking in this sector was strengthened especially at the beginning of the pandemic. 311

Participants spoke highly about the industry forum organised by DEFRA that occurred 312 weekly and then bi-weekly. It was unprecedented as all the main competitors of the industry 313

joined. Participants reported that they shared best practices with an open mind and worked 314 together to contribute to the development of government guidance. Email groups were set up 315

to facilitate exchange of ideas and questions. 316

Facilitator 3 In addition to the networked collaborative activities, the numerous government 317 recommendations, guidelines and updates, and that COVID-19 dominated the media and the 318

Internet for a substantial period of time, all created an overwhelming incentive for the 319 companies to respond rapidly. 320

Four themes emerged in this domain, including effective communications, safety culture, 322

facilitating implementation climate and leadership commitment for implementation. 323 is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted January 29, 2022. ; https://doi.org/10. 1101 /2022 Facilitator 4 Effective communications were emphasized by many participants as an 324

important facilitator of rapid responses. They reported that communications were highly 325 valued by the staff because the situation had been a fast-evolving one. Uncertainties and lack 326 of specific guidance meant that workers needed that information to guide their everyday 327 work. Every time new threats, guidance or public measures came out, such as regional 328 lockdowns, new variants, mass testing or the vaccination program, it was important the 329 company communicated guidance and recommendations that were specific to their work 330

context. 331

Barrier 4 A number of participants reported that language and the complexity of the 332 guidance could be a challenge as English is the second language for many workers within this 333

sector. To tackle this issue, they simplified the language and added infographics to illustrate 334 the meaning. A couple of participants mentioned the cultural background of the workforce 335 could be a barrier to enforce social distancing as certain cultures tend to socialise more and 336

workers of that background were likely to share transport to work or accommodations. This 337 however could also be a facilitator in some circumstances as Large 1 explained how they 338 paired "sons and dads, mums and dads" or drivers from the same households or family 339 bubbles for two-person deliveries. 340 COVID-19 safety was discussed by the participants as a belief rather than something they 341 reluctantly comply to. One participant articulated it particularly well. 342

"We have a culture in our leadership of putting safety first… we track our Facilitator 5 Three of the companies mentioned they provided financial support such as sick 353 leave pay to support the self-employed drivers to take COVID-related self-isolation. It can 354 facilitate adherence among delivery workers as many of them were self-employed and did not 355 enjoy statutory sick pay. They also mentioned that they promoted intangible incentives such 356 as customers' appreciation messages and exemplar stories to be put on their websites and 357 communication channels. 358

Facilitator 6 Key stakeholders' commitment for implementation appeared high. Leadership 359 engagement was evident in all the interviews. Two participants particularly emphasized the 360 influence from the leadership team that keeping workers safe from COVID-19 infection was 361 the right thing to do and would reward the business in the long-term. This is then linked to 362 resources dedicated for implementation. It appeared that the companies allocated adequate 363 resources timely to support the interventions. 364 . CC-BY-NC-ND 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted January 29, 2022. ; https://doi.org/10.1101/2022.01.28.22270013 doi: medRxiv preprint

The implementation process can be characterized as an unplanned rapid response, full 366 engagement, strong execution and continuous reflecting & evaluating. 367 "Rapid response" was a prominent feature emphasized by all of the participants. From early 368

March 2020, the volume of home deliveries "went through the roof". is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint 

A rapid response mode may be effective in the short-term but can run into problems if it 419 lasted longer than expected and hence introduce questions about sustainability. 420

As the pandemic continued into 2021, some workers developed compliance fatigue and this 421 became a barrier to effective implementation. In round 2 interviews, we asked the 422 participants whether they observed any relaxed attitudes towards the COVID measures. 423

Participants agreed that to some extent attitudes had relaxed and described how they took 424 actions to mitigate this. They highlighted the need to maintain effective communications by 425 providing a "permanent alert" or "constant reminder" to workers. Two participants mentioned 426 they added extra monitoring, that is, sending out staff to walk around the workplaces and 427

giving colleagues a reminder whenever they observed behaviours not meeting the standard. 428

In round 2, all participants stated that high volume of home deliveries continued even when 429 lockdown was lifted. They told us that the industry was used to working on full speed during 430 the Christmas peak that was normally from late October to the end of December. Participants also mentioned that the companies were surveying workers regarding to their 442 mental wellbeing and trying to offer some support. 443

We have provided a schematic diagram to illustrate the important findings in Figure 2 . It 444 highlights the key characteristics of rapid responses, grouped into five domains (interventions 445 . CC-BY-NC-ND 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted January 29, 2022. The process of implementation had prominent features, such as the speed to action, the 456 overwhelming external pressure, improvised interventions, and steep learning curves for all 457 stakeholders. We scoped the literature to identify an appropriate theoretical model to inform 458 the analysis. Multiple existing frameworks offered some useful insights, including the RE-459 AIM (Reach, Efficacy, Adoption, Implementation and Maintenance) (22, 23), CFIR 460

(Consolidated Framework for Implementation Research) (24, 25), PRECEDE-PROCEED 461 (26, 27) and other process evaluation models that generally included components such as 462 recruitment, dose delivered, dose received, fidelity, satisfaction, maintenance and context 463 (28). However, they generally assumed a systematically developed intervention program 464 . CC-BY-NC-ND 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted January 29, 2022. ; https://doi.org/10. 1101 /2022 implemented with some extent of control, and none of them fully captured the characteristics 465 of this sector's response to COVID-19. It suggests the urgency of developing a rapid 466 response model that can first, analyze a collection of NPIs implemented in occupational 467 settings. When responding to a pandemic, NPIs are likely to be implemented simultaneously 468 with many other measures and a single measure would not be sufficient (20) . Second, the 469 model should take into account the barriers and facilitators of rapid responses to a public 470 health emergency (29). 471

In addition to the well-known COVID-19 NPIs, such as face coverings, hand washing and 472 social distancing (14), our HoC analysis identified measures that were important to the 473 delivery work setting, including contact-free delivery, fixed pairing, effective 474 communications/IIT and sectoral collaboration. Contact-free delivery and fixed pairing (for 475 two-person deliveries) were new measures improvised by this sector during this COVID-19 476 pandemic and became established practices as the participants told us. Working 477 collaboratively with key stakeholders of the sector, including the competitors and local and 478 state authorities was considered an important measure and a facilitator in outer setting (25). 479

We identified important barriers and facilitators to rapid responses. Financial support for self-480

isolation was considered a facilitator for delivery workers especially the self-employed, as a 481

previous study found sick leave pay was associated with adherence to infection, prevention 482 and control measures among healthcare workers (30). In addition, COVID-19 infection rates 483 among delivery and warehousing workers from the developed and developing countries 484 varied significantly. For example, in Canada, it was as low as 0% (31), whilst in Ecuador it 485 was 15.2% (32). Although the sample of the two studies may not be directly comparable, it is 486 possible that financial conditions served a social determinant of COVID-19 related health 487 outcome (33). The sector's capacity to design and develop interventions internally was also a 488 key facilitator. As SARS-CoV-2 was a novel virus and the pandemic was fast-evolving, a 489 response protocol or prevention guidance for the logistics sector was not available in the UK 490

initially. Hence, internal knowledge and assessment was an important source of intervention 491 development. Companies also used their judgement to decide not to adopt certain measures, 492

such as workplace testing. This echoes the concern that people without COVID-19 self-493

isolating due to false-positive lateral flow test results could be a cost to the individual, their 494 household, and their workplace (34). In addition, localized government support, effective 495 communication and leadership support were considered facilitators. This is in line with 496 findings from existing studies that evaluated the implementation of interventions programs 497 (35, 36) . Overwhelming external mandates was probably a prominent facilitator associated 498

with the situation of a pandemic as few other health interventions received media attention 499

like those for COVID-19. 500

Major barriers included unclear and changing government guidance, lack of testing capacity, 501

shortage of facemasks, and diversified language and cultural backgrounds. Barriers 502 associated with government guidance, testing capacity and supply of PPE mainly affected the 503 rapid response at the early stages (37, 38). Language and cultural barriers were also identified 504 by multiple intervention studies previously (36, 39). Carefully designed trainings were 505 recommended, which were consistent with the measure took by the companies we 506

interviewed. We identified compliance fatigue in the second interview round. Such 507 behavioral changes reflected a response to adjustments in individuals' risk assessment (40, 508 41), especially when the government announced their Roadmap to lift restrictions. Our 509 participants suggested adding more behavior monitoring measures and reminders to maintain 510 the level of alert. Participants mentioned the high costs associated with these NPIs but also 511 . CC-BY-NC-ND 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted January 29, 2022. ; https://doi.org/10.1101 https://doi.org/10. /2022 believed such costs were compensated by increased volume. Going forward, a more 512 systematic approach should evaluate such costs from health economics perspective. 513

The prolonged WFH measure and sustained high workload both add to work stress (42). It 514 highlighted a key sustainability issue associated with the current approach to dealing with the 515 pandemic. The concern is consistent with findings from studies that examined healthcare staff 516 burnout during . It is not sustainable, and a more systematic approach and 517 coherent sectoral strategy is urgently needed. 518

This paper is based on views expressed by those in managerial roles rather than the delivery 519 drivers. We recognize that their views could differ significantly from the frontline workers' 520

perspective. For example, surveys among app-based drivers reported concerns of infection 521 risks from interactions with the public and insufficient workplace protections such as access 522

to personal protective equipment (PPE) (31, 46) . Delivery workers in the French gig 523 economy also expressed concerns of financial precarity and lack of union support (1) . 524

Another potential limitation of this study is the small sample size and the size of the 525 participated companies. The sector was extremely busy throughout the pandemic and our 526 invitations were declined by the majority of companies we approached. We were not able to 527 directly assess the effectiveness of the interventions, but the perceived effectiveness of the 528 participants. 529

This qualitative study provides a rich source of contextualized data to evaluate rapid 531 implementation of COVID-19 NPIs in the UK logistics sector. We assessed the interventions 532

against an occupational health and safety standard and identified barriers, facilitators and 533 sustainability issues in the process of a rapid response. In conclusion, the UK's logistics 534 sector rose to the challenge and rapidly developed and implemented a wide range of RMMs 535 in a fast-evolving pandemic. They closely followed national and local guidelines available to 536 them at the time and developed RMMs resourcefully when guidelines were lacking. 537

Elimination of the risk was not practical for the delivery workers and most control measures 538

were considered administrative controls. Contact-free delivery was commonly implemented 539 and considered effective. Participants were confident that the RMMs played an important role 540 in reducing workplace transmission risk for delivery workers. Further research is now needed 541 to design and evaluate models and tools to apply sustainable respiratory infection prevention 542

and control measures across work settings, as well as taking into account the more vulnerable 543 work and social groups. 544

Ethics statement 545

The project was reviewed and approved by the University Research Ethics Committee at 546

University of Manchester, Ref: 2020-9787-15953. Consent to participation was verbally 547 obtained before the commencement of the interviews. 548

Data availability statement 549

The datasets generated for this study are available on request to the corresponding author. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted January 29, 2022. ; https://doi.org/10.1101/2022.01.28.22270013 doi: medRxiv preprint Disciplinary action + + 5. Personal protection: Protect workers with certain equipment, depending on expert risk assessment d Face coverings + + + + + Gloves + + + "+" indicates the measure was reported as implemented. This table is not a complete list of RMMs implemented by the companies. When some of the measures were not ticked by certain companies, it meant that this measure was neither applicable to the company's situation nor discussed during the interviews. a. Time period was estimated by the interviewers during analysis. b. RoC: room of choice c. Lateral Flow Device d. Neither face coverings nor normal gloves were considered PPE. They were issued to prevent transmission rather than protecting workers from getting infected. 

Gig Workers During the Covid-19 Crisis in 572

Financial Precarity and Mental Well-Being

Logistics Workers Are a Key Factor for Sars-Cov-2 Spread in Brazilian Small Towns

Riders on the Storm: Amplified Platform Precarity and the Impact of 579

Covid-19 on Online Food-Delivery Drivers in China

Work-Related Covid-19 582 Transmission in Six Asian Countries/Areas: A Follow-up Study

Occupation and Covid-19 Mortality in England: A National Linked Data Study of 14.3 586 Million Adults

Equity) Research Programme

Together Stronger: Boundary Work within an 594

Australian Systems-Based Prevention Initiative

Control Strategies to Reduce Social Mixing on Outcomes of the Covid-19 Epidemic in 598

China: A Modelling Study. The Lancet Public Health

Marc Baguelin, et 601 al. Impact of Non-Pharmaceutical Interventions (Npis) to Reduce Covid-19 Mortality

Estimating the Effects of Non-Pharmaceutical Interventions on Covid-19 in Europe

Multidisciplinary Response to an Ebola Epidemic

Taking Forward a 610 'One Health

Zoonotic Pathogens with Epidemic Potential

Narrative Review of Models and Success Factors for 614 Scaling up Public Health Interventions

Public Health Measures in Reducing the Incidence of Covid-19, Sars-Cov-2 Transmission, 618 and Covid-19 Mortality: Systematic Review and Meta-Analysis

Using Theory of Change to Design and 621

Evaluate Public Health Interventions: A Systematic Review

Consolidated Criteria for Reporting Qualitative 624 Research (Coreq): A 32-Item Checklist for Interviews and Focus Groups. International 625 Journal for Quality in Health Care

Using Thematic Analysis in Psychology

Qualitative Research from Start to Finish

Hierarchy of Controls': Providing a Framework for Addressing 630 Workplace Hazards

The Hierarchy of Preventive 632 Measures to Protect Workers against the Covid-19 Pandemic: A Review

Statistical Methods for Rates and Proportions: john 635 wiley & sons

Planning and Evaluation Framework: Adapting to New Science and Practice with a 20-Year 638

The Re-Aim Framework: A Systematic Review of 640 Use over Time

Planning for Implementation Success Using Re-Aim and Cfir Frameworks: A Qualitative 644 Study

Fostering 646 Implementation of Health Services Research Findings into Practice: A Consolidated 647 Framework for Advancing Implementation Science

Application of Precede-Proceed Planning 650

Model in Transforming the Clinical Decision Making Behavior of Physical Therapists in 651

What Is a Planning Model? An Introduction to Precede-Proceed

Van Mechelen 656 W. What Is Actually Measured in Process Evaluations for Worksite Health Promotion 657 Programs: A Systematic Review

Rapid Versus Traditional Qualitative Analysis Using the Consolidated Framework 661 for Implementation Research (Cfir)

Barriers 664 and Facilitators to Healthcare Workers' Adherence with Infection Prevention and Control 665 (Ipc) Guidelines for Respiratory Infectious Diseases: A Rapid Qualitative Evidence 666

Covid-19 Experiences, Ppe, and Health Concerns in 669

Bicycle Delivery Workers: Cross-Sectional Pilot Survey

High Prevalence of Sars-Cov-2 Infection among Food Delivery Riders. A 673 Case Study from Quito

The Covid-19 Pandemic and Health 676

Sars-Cov-2 Antigen 679 Testing: Weighing the False Positives against the Costs of Failing to Control Transmission

The Lancet Respiratory Medicine (2021)

Enablers and 682 Barriers to the Implementation of Primary Health Care Interventions for Indigenous People 683 with Chronic Diseases: A Systematic Review

Barriers and 686 Facilitators of Implementation of a Community Cardiovascular Disease Prevention 687 Programme in Mukono and Buikwe Districts in Uganda Using the Consolidated Framework 688 for Implementation Research

Emergency Preparedness, Resilience and 691 Response to a Biological Outbreak

Personal Protective Equipment Supply Chain: Lessons Learned from Recent Public Health 695 Emergency Responses

Evaluating and Optimizing the Consolidated Framework for Implementation Research (Cfir) 698 for Use in Low-and Middle-Income Countries: A Systematic Review

A Worldwide 701 Assessment of Changes in Adherence to Covid-19 Protective Behaviours and Hypothesized 702

Pandemic Fatigue

Meta-Analysis of the Relationship between Risk Perception and Health Behavior: The 706 Example of Vaccination

Working from Home 709 and Job Loss Due to the Covid-19 Pandemic Are Associated with Greater Time in Sedentary 710 Behaviors

Factors Associated 712 with Work-Related Burnout in Nhs Staff During Covid-19: A Cross-Sectional Mixed 713

Covid-19) Pandemic 716 in Japan

Healthcare Workers in the Covid 19 Era: A Review of the Existing Literature

The authors declare that the research was conducted in the absence of any commercial or 552 financial relationships that could be construed as a potential conflict of interest. 553

Author Contributions 554 Interview schedules and codebooks were developed by HW, SD, CW and MvT with 555 extensive inputs from the rest of the team and collaborators at HSE and PHE. HW, SD and 556 CW carried out the interviews and completed the analysis. HW and SD drafted the paper. 557MvT, YH, DD, IH, MR and AV commented and edited each version of the manuscript. All 558 the authors reviewed and approved the final version. MvT is the Principal Investigator of the project. 563

The authors would like to thank the project's advisory group that consist of Catherine 565Noakes, Chris Armitage, Sheena Johnson, Jeanette Edwards, Barbara Hockey, Nina Day, 566Nick Gent and Thomas House, for their advice that helped refine the aims and objectives of 567 this article. We would also like to thank Helen Beers and Peter Baldwin from HSE for their 568 advice on the business engagement, thematic analysis and feedback to companies. 569 570 12 References 571