key: cord-0926625-du64gkxx
authors: Daniels, S.; Wei, H.; Han, Y.; Catt, H.; Denning, D. W.; Hall, I.; Regan, M.; Verma, A.; Whitfield, C. A.; van Tongeren, M.
title: Risk factors associated with respiratory infectious disease-related presenteeism: a rapid review
date: 2021-04-14
journal: nan
DOI: 10.1101/2021.04.12.21255302
sha: c8b0b645a84106534e90eb9dd43faf3be13467cd
doc_id: 926625
cord_uid: du64gkxx

Background: Workplace transmission is a significant contributor to severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) outbreaks. Previous studies have found that infectious illness presenteeism could contribute to outbreaks in occupational settings and identified multiple occupational and organizational risk factors. Amid the COVID-19 pandemic, it is imperative to investigate presenteeism particularly in relation to respiratory infectious disease (RID). Hence, this rapid review aims to determine the prevalence of RID-related presenteeism, including COVID-19, and examines the reported reasons and associated risk factors. Methods: The review followed a Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) search approach and focused on studies published in English and Chinese. Database searches included MEDLINE, EMBASE, Web of Science, China Knowledge Resource Integrated Database (CNKI) and preprint databases MedRxiv and BioRxiv. Results: The search yielded 54 studies, of which four investigated COVID-19-related presenteeism. Prevalence of work presenteeism ranged from 14.1% to 55% for confirmed RID, and 6.6% to 100% for those working with suspected or subclinical RID. The included studies demonstrated that RID-related presenteeism is associated with occupation, sick pay policy, age, gender, health behaviour and perception, vaccination and organisational factors including workplace culture and peer pressure. Conclusions: This review demonstrates that presenteeism or non-adherence to isolation guidance is a real concern and can contribute to workplace transmissions and outbreaks. Policies which would support workers financially and improve productivity, should include a range of effective non-pharmaceutical inventions such as workplace testing, promoting occupational health services, reviewing pay and bonus schemes and clear messaging to encourage workers to stay at home when ill. Future research should focus on the more vulnerable and precarious occupational groups, and their inter-relationships, to develop comprehensive intervention programs to reduce RID-related presenteeism.

to compliance failure with control measures during an influenza pandemic because of job insecurity and financial problems associated with missing work [5, 6] . More recently, low rates of self-isolation behaviours were reported in key worker sectors during the COVID-19 pandemic, likely due to greater financial need, social pressure to attend work or inability to work from home [7] . (PRISMA) guidelines and adjustments made to accommodate qualitative research [8] .

Presenteeism is commonly defined as people who attend work while ill [2] . Though frequently measured as prevalence from an epidemiological perspective, i.e. the percentage of workers who attended work while ill, it can be measured as productivity loss from a health economics perspective, i.e. the number of hours or days worked with compromised productivity due to the . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 14, 2021. ;  https://doi.org/10.1101/2021.04.12.21255302 doi: medRxiv preprint illness, with some converted into economic loss. In this review, focus is on prevalence and five types of presenteeism behaviour:

1. Working with an RID infection (confirmed by clinical diagnosis or laboratory testing).

2. Working with RID symptoms (suspected or subclinical). 3 . Going to work with a history of exposure to RID. 4 . Non-adherence to guidance to stay at home from work with infected, suspected or exposed to RID. 5 . Propensity to attend the workplace with confirmed, suspected or exposure to RID.

Houghton et al [9] defined RIDs as diseases that cause acute respiratory tract infection (RTI) and severe respiratory disease in susceptible people with apparently normal immune systems.

Population: working women and men, any age.

Exposure: confirmed or suspected RID or close contact with confirmed or suspected cases (i.e. family members or shared accommodations).

Comparator: none.

Outcomes:

1. Prevalence of presenteeism in the following sub-populations: a) workers attending work with confirmed, suspected or exposure to RID. b) propensity to attend the workplace with confirmed, suspected, or exposed to RID. c) adherence to guidance (e.g. government or physician) to stay at home from work with confirmed, suspected or exposure to RID.

2. Reported reasons for presenteeism in any of the three sub-populations listed in 1. (a-c).

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted April 14, 2021. ; https://doi.org/10.1101/2021.04.12.21255302 doi: medRxiv preprint

English language searches were conducted by two researchers (SD and HW). Search terms for presenteeism were 'presenteeism', 'going to work while ill/sick', 'suspected, subclinical or mild symptoms', 'non-compliance or violating guideline/guidance/protocol' and 'exposed to or contact with confirmed/diagnosed case'. We also included terms such as 'isolation', 'quarantine' 'social distance' or 'lockdown'. For RID and COVID-19 diseases, we used 'COVID', 'coronavirus', 'nCoV', 'SARS', 'MERS', 'flu/influenza/influenza-like', 'respiratory infectious disease' and 'respiratory tract infection'. Search terms were translated into Chinese by two Chinese speaking researchers (HW and YH) and search strategies were adapted for CNKI.

Different search strategies were trialled with consideration for both specificity and sensitivity.

HW and SD carried out preliminary searches on different databases testing a variety of search strategies. These were finalised in discussions with HC and MvT. Our final search strategy used terms and associated words for 'COVID-19' or 'respiratory infectious diseases' and 'presenteeism', joined by the AND function. A copy of our search strategy in MEDLINE is included as Additional file 1.

Data from the final set of studies was extracted by SD and HW using a data extraction table agreed by all reviewers. Data extracted included citation, study design, objectives, sample size, population, and results regarding the prevalence of RID-related presenteeism and reported reasons or statistical risk factors associated with it.

Two authors (SD and HW) performed the risk of bias assessments, which were verified by one author (HW). The Newcastle-Ottawa Scale (NOS) was used for the longitudinal cohort studies . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted April 14, 2021 [10] was used for the cross-sectional studies.

For both NOS and CASP, the sum of scores for all subscale items were used to categorise the study quality as either low (<3), moderate (4-6) and high (>7). The cross-sectional studies were categorised as either low (<4), moderate (5-7) and high (>8).

Study designs and outcome measures of the included literature were heterogeneous.

Consequently, we used narrative synthesis for data analysis rather than meta-analyses. The studies were grouped into the four types of RID-related presenteeism behaviour. The reported reasons and risk factors for RID-related presenteeism were structured into over-arching themes related to work factors (occupation type, work and employment, and organisational culture) and individual factors (sociodemographics, health behaviours or perception and vaccination uptake). Data synthesis and analysis was performed by SD and HW.

Our initial search yielded 794 papers after deduplication, with an additional 13 papers identified through reference list searches. After title and abstract screening, 65 papers were taken to full text review. Of these, 54 papers were selected for inclusion for data extraction.

Three of the included studies were published in Chinese and 51 studies were English. See . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. Full-text articles assessed for eligibility (n = 65)

Full-text articles excluded, with reasons (n = 11)

• Indirect measure of RID-related presenteeism (n = 6) • Qualitative data not associated with RIDrelated presenteeism (n = 1) • Not RID-related presenteeism (n = 4) Included studies (n = 54)

Records identified through database searching (n = 893)

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted April 14, 2021. ; https://doi.org/10.1101/2021.04.12.21255302 doi: medRxiv preprint

Of the 54 included studies, 44 were cross-sectional studies (see Additional file 2). Study locations were North America (n = 32) [5, , Asia (n = 9) [42] [43] [44] [45] [46] [47] [48] [49] [50] , Europe (n = 7) [51] [52] [53] [54] [55] [56] [57] , Australia (n = 4) [58] [59] [60] [61] , worldwide population (n = 1) [62] , and an unspecified location (LinkedIn members) (n = 1) [63] . See Figure 2 for the distribution of included studies by country.

The sample size ranged from 14 to 550,360, ages ranged from 18 to >65 years old, and the percentage of females ranged from 22.5% to 99.2%. Included studies focused on healthcare occupations (n = 30) [13-18, 22, 23, 26, 27, 30, 31, 33-40, 42-44, 47, 49, 54, 57-59, 62] , other specified occupations (n = 10) [12, 20, 21, 24, 25, 28, 32, 45, 46, 55] or general working populations (n = 14) [5, 11, 19, 29, 41, 48, 50-53, 56, 60, 61, 63] . Studies investigated influenza-and/or ILI-related presenteeism (n = 38) [5, 14-26, 28, 29, 31, 33, 34, 37, 39, 40, 42- . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 14, 2021. 46, 48, 49, 52, 55, 57, 58, [60] [61] [62] [63] , general upper respiratory infection (n = 10) [11, 27, 30, 32, 35, 36, 38, 47, 53, 59] , common cold or RID-type symptoms such as cough and sore throat (n = 7) [21, 34, 45, 55, 56, 60, 63] , COVID-19 (n = 4) [12, 41, 50, 51] , breath infections (n = 1) [54] and streptococcal infection ( n = 1) [13] .

The overall quality of the cross-sectional studies ranged from poor to moderate (see Additional file 3). Six studies did not describe the assessment of exposure to RID or used a survey with hypothetical questions such as "would you attend work whilst ill?" [5, 24, 35, 46, 60, 61] .

Fourteen were not representative of the target population due to recruitment from selected groups or organisations [18, 24, 26, 27, 30, 31, 36, 37, 39, 45, 49, 55, 62, 63] and only six had justified sample sizes [5, 19, 23, 43, 60, 61] . In addition, only nine studies compared respondents with non-respondents and weighted the data to the population distribution to avoid response bias [5, 15, 17, 26, 43, 55, 56, 60, 61] .

The quality of the seven cohort studies ranged from moderate to high. Six studies were truly or somewhat representative of the target population [25, 32, 38, 51, 57, 58] with one study having invited participation from a selected group [28] . Only three cohort studies used formal records to ascertain the exposure to an RID. Imai et al [58] and Kuster et al [57] acquired laboratoryconfirmed influenza data, and Jain et al [51] obtained data from the London Coronavirus

Response Centre. No studies demonstrated that RID-presenteeism was not present at the start of the study. Four studies followed participants over an influenza season to account for the disease of interest [25, 28, 32, 57] . Of the five prospective cohort studies, only one study had a low follow-up rate (23.8%) [38] .

Of the three qualitative studies, two were assessed to be moderate [12, 33] and one was of high quality [59] . The methodology and research design of the two studies, assessed to be of . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 14, 2021. ; https://doi.org/10.1101/2021.04.12.21255302 doi: medRxiv preprint moderate quality, was unclear and neither study included a statement related to ethical approval.

None of the qualitative studies indicated whether the researcher critically examined their own role, leaving potential bias and influence during analysis and selection of data for presentation.

Only Mitchell and Coatsworth [59] gave an in-depth description of the data analysis process.

Varying measurement methods and recall periods were used to quantify presenteeism. Fortysix studies reported frequency of RID-related presenteeism which included prevalence and mean days/hours worked while ill, and five studies measured presenteeism as workplace productivity levels (see Additional file 4).

Six studies reported that prevalence of presenteeism ranged from 14.1% to 55% for respondents with RID confirmed by laboratory test or clinical diagnosis [13, 14, [51] [52] [53] 58] . Prevalence ranged from 6.6% to 100% for symptoms of RID (suspected or subclinical) in 27 studies [15-18, 20-27, 29, 30, 37-44, 47, 49, 50, 57, 59] . For history of exposure, the prevalence was 77% for trainee physicians [31] . The propensity to attend work while ill ranged from 14% to 100% [34-36, 46, 48, 56, 62] . While 50.7% to 96.6% reported that they would adhere to guidance to stay at home from work with RID [5, 60, 61] . Within this same sub-group, a high proportion (94-96.1%) hypothesised they would stay at home with RID for at least 7 days during a confirmed pandemic.

We grouped presenteeism reasons and analyses into themes by work and individual factors (see Additional files 5 and 6).

Occupation type . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) Five studies reported a sense of duty or professional obligation as reasons for RID-related presenteeism, particularly in healthcare workers and school employers [16, 17, 20, 37, 59] .

The percentage of participants who chose it as one of the main reasons for presenteeism ranged from 28% to 56% within these five studies.

Eight studies measured the association between occupations and RID-related presenteeism [22, 27, 29, 38, 40, 50, 51, 58] . Healthcare workers were more likely to report previously working with ILI than other workers including those in education in Kansas, U.S. [29] .

Results from a study in Japan, by Machida et al [50] , showed that company employees are more likely to return to work within 7 days after symptom onset, compared to the selfemployed, part-time workers and government workers. In contrast, Jain et al [51] found no differences in workplace attendance after COVID-19 symptom onset between occupational sectors in the UK, including office, retail and construction.

Comparing sub-groups of healthcare workers, Canadian physicians were significantly more likely to work with RTI than medical students and residents, and considered the risk of transmitting infection to others to be the lowest [27] . Nursing and health practitioners had longer sick leave than medical doctors in Australia [58] . However, Bhadelia et al [22] reported that U.S. physicians and nurses were equally likely to work while symptomatic.

Eight studies reported "lack of cover" as one of the main reasons for presenteeism [17, 24, 26, 39, 42, 47, 49, 59] and one reported an association in healthcare workers [40] . The percentage of participants who chose this as one of the main reasons ranged from 24% to 96%. Four studies reported concerns about "pay loss" as a main reason [16, 33, 38, 49] and one study reported that attendance bonuses incentivised employees to work while ill [12] . Three studies cited . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 14, 2021. ; https://doi.org/10.1101/2021.04.12.21255302 doi: medRxiv preprint workload and fear of falling behind at work felt as main reasons, categorised as 'job demand' [24, 38, 39] .

Four studies tested associations between "paid sick leave" and RID-related presenteeism.

Ahmed et al [11] reported that U.S. workers who had access to paid leave were significantly less likely to work during the first 3 days of illness with an acute respiratory infection (ARI).

Piper et al [19] reported that U.S. employees with paid sick days had a higher probability of staying home for their own and child's ILI or influenza. Jiang et al [38] reported that Canadian healthcare workers without paid sick leave were significantly more likely to choose "can't afford to stay home" while symptomatic with an ARI. Johnson et al [40] identified that access to paid sick leave increases adherence to absenteeism for ILI in U.S. healthcare workers.

Four studies tested flexible work or leave policy as a risk factor for RID-related presenteeism [11, 25, 50, 61] . Inflexible work conditions, such as work that does not accommodate home working, appears to be a driver for presenting at work with RID symptoms. Machida et al [50] demonstrated that "unable to work from home" was a significant factor for going to work within 7 days of symptom onset during the COVID-19 outbreak in Japan. Rousculp et al [25] also reported that an employee without a "work from home" policy is significantly more likely to attend work when ILI symptoms are most severe. Furthermore, the intention to comply with home quarantine following exposure to pandemic influenza was much lower for the employed who are unable to work from home, compared to people not in paid employment [61] .

Organisational culture Ten studies linked presenteeism and organisational culture, but only one was outside the healthcare sector [24] . Five studies reported "avoid burdening colleagues" [15, 33, 34, 37, 59] with 57% to 100% of participants choosing this as a main reason for presenteeism. Five studies investigated "peer pressure" as a main reason and concerns were expressed as "avoid or afraid . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 14, 2021. ; https://doi.org/10.1101/2021.04.12.21255302 doi: medRxiv preprint of appearing weak or lazy", or "feeling pressure or judgement from colleagues or supervisors" [24, 26, 34, 40, 59] . Rebmann et al [39] reported that "perceived pressure from colleagues or supervisor" was a significant predictor of presenteeism behaviour among school nurses. Four studies cited reasons such as "had a perception that they were encouraged to work while ill", "seeing other colleagues working when similarly unwell", or "avoid negative repercussions" [33, 34, 38, 59] .

Five studies found significant associations between gender and RID-related presenteeism. Two studies showed females were more likely to report adherence to public health guidance to stay at home following exposure to influenza pandemic, compared to males [60, 61] . Additionally, women are more likely to stay at home for children's ILI and influenza [19] . Conversely, UK studies showed that males were 66% less likely to attend the workplace with COVID-19 symptoms [51] and ILI-related presenteeism rates were higher in female healthcare workers in the U.S. [18] .

Findings from three studies showed that younger age is linked with higher level of presenteeism and lower rate of anticipated compliance. Mossad et al [18] found that presenteeism was significantly higher among those aged below 40 years. Brown et al [60] reported that those over 55 years were significantly more likely to report anticipated compliance for both the common cold and seasonal influenza. Ablah et al [29] reported that those 30 years old or younger were 2.8 times more likely to report previously working while ill. In comparison, Jain et al [51] reported no association between age and workplace attendance after COVID-19 symptom onset.

Health behaviour or perception . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 14, 2021. ; https://doi.org/10.1101/2021.04.12.21255302 doi: medRxiv preprint "Not sick enough or considered a minor disease" is the most reported reason for presenteeism, cited by ten studies with this listed as the main reason in 38% and 84% settings [16, 17, 24, 26, 38, 39, 42, 47, 49, 59] , and tested as a significant association with non-adherence to absenteeism in clinical healthcare workers [40] . In addition, Rebmann et al [39] demonstrated that school nurses were significantly more likely to work with symptoms of ILI, if it was considered a mild illness by the participant. Similarly, Jiang et al [38] reported that Canadian healthcare workers were significantly less likely to work as symptom severity increased.

Qualitative research on Philadelphian physicians and clinicians highlighted beliefs that it is "unreasonable" to expect staff will take sick leave until symptoms of RTI resolve as resolution can take several days [36] .

Three studies tested the association between vaccinated and non-vaccinated groups with inconsistent results. One study showed higher prevalence of RID-related presenteeism among the influenza vaccinated healthcare workers [17] , one reported no significant difference among nursing staff [44] , and one found vaccination is associated with statistically significant reductions in days of working while ill among 50-64 years old university employees [28] .

The search and screening process yielded 54 studies. Of these, 94% (n = 51) were observational studies and 56% (n = 30) were from the U.S. Furthermore, 91% (n = 49) of the studies were of low to moderate quality, with a predisposition to response bias and poor sampling strategies.

This limits the robustness of the observations, as bias may have occurred to an extent that the results do not adequately represent the target population. In addition, geographical applicability may also be affected by the predominant U.S. bias, as countries have different working practices, health & safety legislation and culture. Despite the limitations, this review identified . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 14, 2021. ; https://doi.org/10.1101/2021.04.12.21255302 doi: medRxiv preprint clear behavioural trends that may impact the effectiveness of workplace COVID-19 control and prevention policies and offers potential solutions for consideration.

In the UK, as in other countries, one of the response measures to control the spread of COVID-19 within the workforce was a requirement to stay at home and self-isolate where an employee or their family has any COVID-19 symptoms [51] . Reported prevalence of presenteeism for those with confirmed RID ranged from 14.1% to 55%, while for suspected RID or potential exposure to RID this ranged from 6.6% to 100%. Similar results were reported in a review of presenteeism relating to all infectious illness, including gastroenteritis [2] .

These findings contrast with other included studies where a high percentage (>94%) of respondents reported intent to adhere with guidance to stay at home, for at least 7 days, during an influenza pandemic [5, 60, 61] . This may indicate that, while workers may acknowledge a workplace or national policy to stay at home with confirmed or suspected RID, their actual behaviour will be adversely affected by other factors, and so additional occupational health policies may be required to enable policies within the workplace. The COVID-19 Rapid Survey of Adherence to Interventions and Responses [CORSAIR] study demonstrated that adherence rates to test, trace and isolate in the UK was low (18.2%), but intention to carry out these behaviours was much higher (around 70%), and this was associated with a number of factors including gender, age, lower socioeconomic grade and working in a key sector [7] .

A clear isolation from work policy should be supplemented with a range of IPC measures to reduce workplace transmission and could be accompanied with a policy of RID testing of employees in the workplace. As an example, laboratory confirmation of influenza in medical doctors led to an increase in associated sick leave from 16.6% (baseline) to 76.6% [58] .

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 14, 2021. ; https://doi.org/10.1101/2021.04.12.21255302 doi: medRxiv preprint Moreover, 96% of healthcare workers agreed that it is important for ill employees with confirmed influenza infection (positive test result) to stay at home [40] .

Our review was unable to distinguish similar symptoms shared between common colds, such as a sore throat, headache, cough and muscle ache, from influenza [64] or other significant RIDs. However, while colds symptoms can include elevated temperature, they rarely include sudden high temperature of   C or above (fever), so this criterion can be used to reduce influenza transmission. This can be undertaken through the form of temperature checks either at home (self-assessment), randomised workplace checks or as a condition of entry. Workplace based symptom testing and self-isolation polices can form part of an effective methodology to decrease the impact of RID in high-risk occupational settings.

Machida et al [50] determined that most company workers in Japan did not practice strict selfisolation during the COVID-19 pandemic, compared to part-time workers, self-employed and government workers. This may indicate that, in this study company workers felt more obligated to go into work with milder symptoms of RID than other workers, possibly due to organisational or peer pressures. This feeling of obligation was also observed in other studies. The most common reason given by respondents for RID-presenteeism was a feeling of sense of duty or professional obligation, particularly healthcare workers [16, 17, 20, 37, 59] . While other studies [15, 24, 26, 33, 34, 37, 39, 59] listed organisational culture-based reasons such as "avoid burdening colleagues", "peer pressure" and a "perception that the organisation encouraged working while ill".

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 14, 2021. ; https://doi.org/10.1101/2021.04.12.21255302 doi: medRxiv preprint Ahmed et al [11] found that employees who were actively discouraged from attending work if they had influenza-like symptoms were significantly less likely to attend their usual workplace compared to those who were not. Clear messaging throughout all levels of the organisation is therefore vital for an effective transmission prevention policy.

Role of occupational health services "Not sick enough" or "considered a minor disease" was reported as a justification for RIDpresenteeism in, mostly, healthcare workers [16, 17, 24, 26, 38, 39, 42, 47, 49, 59] . This indicates that organisations should consider staff education and training regarding the consequences of presenteeism, even when symptoms are minor or employees are asymptomatic, to limit the risk to other employees, visitors, clients or patients when they attend work with RID. Mitchell and Coatsworth [59] noted that there is a role for occupational health services in educating staff regarding the risk of coming to work with an RID. This may be especially relevant where only minor RID symptoms are identified, and to separate common colds from more severe RID.

Presenteeism due to a "lack of cover" was reported in eight studies, predominantly in healthcare workers [17, 24, 26, 39, 42, 47, 49, 59] . This can be prevented by having contingency plans in place. Miwa et al [42] found that improvements in workplace logistics during times of high RID prevalence, such as providing additional human resources and back-up systems, enabled healthcare workers to feel they could take sick leave when necessary.

Four studies reported "pay loss" as a reason for presenteeism [16, 33, 38, 49] . Babcock et al [33] suggested that, where organisations combine paid vacation and sick days, a worker might decide to go to work with RID rather than claim it as a sick day and, thus, lose a vacation day.

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 14, 2021. ; https://doi.org/10.1101/2021.04. 12.21255302 doi: medRxiv preprint Their recommendation is that this policy is avoided as it created "a perverse incentive for presenteeism".

Attendance bonuses may also incentivise employees to work while ill [12] , with the potential for bonuses to be missed regardless of cause [33] . This may, however, have direct and indirect effects on organisations due to health-related productivity loss which can lead to economic costs. For example, Letvak et al [65] found that nurse presenteeism in U.S. hospitals raised health care costs, with estimated costs of about $2 billion dollars annually from increased patient falls, medication errors and lower quality-of-care scores. Hence, organisations should consider unintended consequences of RID and other forms of presenteeism, such as economic cost and employee health and wellbeing, before implementing attendance bonus schemes, respectively, due to presenteeism.

Paid sick leave was significantly negatively associated with lower RID-presenteeism in three studies [11, 19, 38] . While annual leave and paid sick leave is statutory in many European countries, as more workers become self-employed contractors to online platforms such as Amazon and Uber [66] , access to employee benefits may decline. Most self-employed workers do not receive paid sick leave and have less collective bargaining power against platform companies [67, 68] . Although companies such as Just Eat and Deliveroo offered some financial support, i.e., for a maximum of 14 days with certain eligibility criteria, for their takeaway couriers who were required to self-isolate over the COVID-19 pandemic [69, 70] .

Gender was significantly associated with RID-presenteeism. Women had a higher prevalence of RID presenteeism [18, 51] but conversely demonstrated a higher probability of staying home for a child's RID [19] . Studies show that economic slowdowns disproportionately affect women. According to analysis by McKinsey and Company, women represented 39% of the . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 14, 2021. ; https://doi.org/10.1101/2021.04.12.21255302 doi: medRxiv preprint global workforce but accounted for 54% of job losses in 2020 [71] . Mothers, particularly lone mothers, were more likely to work for sectors that had been shut down by the UK government during the COVID-19 pandemic [72] . This may account for the higher incidence of COVID-19 presenteeism in women in the UK, identified by Jain et al [51] . Conversely, women not working in lockdown sectors are twice as likely as men to be key workers, and over four times as likely to work for the health and social care sector [72] . Mothers of younger children in the U.S. reduced their work hours four to five times more than fathers during the COVID-19 pandemic [73] . Therefore, the ability for women to do paid work may be adversely impacted by the pressures from childcare even if their jobs remain active. Organisations should engage with employees when reviewing contingency plans and policies to ensure they do not generate unequal impacts.

This review identified inconsistent results for the association between receiving influenza vaccination and presenteeism. Askarian et al [44] found no significant difference in continuing to work with ILI symptoms between vaccinated and unvaccinated nurses. Chiu et al [17] reported that the uptake of influenza vaccine was associated with ILI-related presenteeism and that vaccination may reduce symptomatic healthcare workers' perceived risk of having influenza and onwards transmission. Conversely, Nichol et al [28] reported that influenza vaccination uptake is associated with a reduction in days of working with ILI.

Although influenza vaccinations interrupt viral transmission and reduce illness, there is an annual variation in effectiveness. For example, studies have found vaccine effectiveness to be reduced in those who received repeated prior influenza vaccinations [74, 75] . There is the potential for an adverse relationship to develop between vaccination uptake and presenteeism, . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 14, 2021. ; https://doi.org/10.1101/2021.04.12.21255302 doi: medRxiv preprint with vaccine recipients willing to work while ill, because they have been vaccinated, yet unaware of the limitations on the effectiveness.

It is recommended that where an organisation has a vaccination policy consideration, or takes advantage of a national programme, employees should still be advised on the steps to take if COVID-19 symptoms are reported with regards to attending the workplace.

Recent research on RID-related workplace presenteeism, including COVID-19, has provided further understanding of the associated risk factors. While there is a strong intention among workers to adhere to non-pharmaceutical interventions, such as requirements to stay at home with an RID, studies still showed high levels of presenteeism. Factors linked to presenteeism include organisational culture, such as organisational or peer pressure, ineffective resource planning preventing workers from taking time off, inadequate sick leave / sick pay, attendancebased bonus schemes, lack of occupational health services and gender inequality. The interrelationship of the factors associated with presenteeism means effective non-pharmaceutical interventions require a comprehensive review of related supporting workforce and organisational policies.

There was insufficient research on the potential impact of onsite RID testing as a preventative factor for presenteeism, while impact of vaccine was inconclusive. With a bias towards health and social care workers, future research should focus on the role of respiratory infection testing and vaccination as intervention strategies for vulnerable and precarious occupational groups, including self-employed and gig workers.

Additional file 1: Search strategy used in MEDLINE.

Additional file 2: Study characteristics.

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted April 14, 2021 

Not applicable . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted April 14, 2021. ; https://doi.org/10.1101/2021.04.12.21255302 doi: medRxiv preprint

This review was supported by the UK Research and Innovation (UKRI) and National Institute for Health Research (NIHR), Grant Ref: MC_PC_19083. The funder did not contribute to the study design, analysis or writing of the report.

Ethical approval was not required because no sensitive or confidential information was collected from participants.

Authors declare no conflicts of interest.

SD and HW made substantial contributions to the design of the study and carried out the full search. SD, HW and YH screened the literature for inclusion and exclusion, with support from HC and MvT. SD and HW performed the analysis and interpretation of the data. SD and HW drafted the first version. Substantial revisions were provided by HC, DD, IH, MR, AV, CW and MvT. All authors read and approved the final version of the manuscript.

Modes of transmission of virus causing COVID-19: implications for IPC precaution recommendations: scientific brief

A systematic review of infectious illness Presenteeism: prevalence, reasons and risk factors

Distribution of Influenza-Like Illness (ILI) by Occupation in Washington State

Increased risk of SARS-CoV-2 infection in staff working across different care homes enchanced COVID-19 outbreak investigations in London care homes

Public response to community mitigation measures for pandemic influenza

Employment and compliance with pandemic influenza mitigation recommendations. Emerging Infectious Diseases

Adherence to the test, trace and isolate system: results from a time series of 21 nationally representative surveys in the UK (the COVID-19 Rapid Survey of Adherence to Interventions and Responses [CORSAIR] study)

Developing PRISMA-RR, a reporting guideline for rapid reviews of primary studies (Protocol)

Barriers and facilitators to healthcare workers' adherence with infection prevention and control (IPC) guidelines for respiratory infectious diseases: a rapid qualitative evidence synthesis

Are healthcare workers' intentions to vaccinate related to their knowledge, beliefs and attitudes? a systematic review

Paid Leave and Access to Telework as Work Attendance Determinants during Acute Respiratory Illness

COVID-19 Among Workers in Meat and Poultry Processing Facilities -19 States

A Cluster of Group A Streptococcal Infections in a Skilled Nursing Facility-the Potential Role of Healthcare Worker Presenteeism

Investigation of an outbreak of 2009 pandemic influenza A virus (H1N1) infections among healthcare personnel in a Chicago hospital

Understanding drivers of influenza-like illness presenteeism within training programs: A survey of trainees and their program directors

Nosocomial outbreak of influenza A H3N2 in an inpatient oncology unit related to health care workers presenting to work while ill

Working with influenza-like illness: Presenteeism among US health care personnel during the 2014-2015 influenza season

Working Despite Having Influenza-Like Illness: Results of An Anonymous Survey of Healthcare Providers Who Care for Transplant Recipients

Paid sick days and stay-at-home behavior for influenza

Influenza-like illness and presenteeism among school employees

Managerial practices regarding workers working while ill

Impact of the 2009 Influenza A (H1N1) Pandemic on Healthcare Workers at a Tertiary Care Center

Respiratory virus shedding in a cohort of onduty healthcare workers undergoing prospective surveillance

Health care versus nonhealth care businesses' experiences during the 2009 H1N1 pandemic: financial impact, vaccination policies, and control measures implemented

Attending work while sick: Implication of flexible sick leave policies

Assessing physicians' in training attitudes and behaviors during the 2009 H1N1 influenza season: a cross-sectional survey of medical students and residents in an urban academic setting. Influenza and Other Respiratory Viruses

Do you come to work with a respiratory tract infection?

Burden of influenza-like illness and effectiveness of influenza vaccination among working adults aged 50-64 years

Influenza vaccine coverage and presenteeism in Sedgwick County

Working with symptoms of a respiratory infection: Staff who care for high-risk individuals

Evaluation of 2009 pandemic influenza A (H1N1) exposures and illness among physicians in training

Effect of influenza-like illness and other wintertime respiratory illnesses on worker productivity: The child and household influenza-illness and employee function (CHIEF) study. Vaccine

Reported variability in healthcare facility policies regarding healthcare personnel working while experiencing influenza-like illnesses: An emerging infections network survey

Attitudes about sickness presenteeism in medical training: Is there a hidden curriculum? Antimicrobial Resistance and Infection Control

Preventing Respiratory Viral Transmission in Long-Term Care: Knowledge, Attitudes, and Practices of Healthcare Personnel

Reasons Why Physicians and Advanced Practice Clinicians Work While Sick: A Mixed-Methods Analysis

Why physicians work when sick

Which healthcare workers work with acute respiratory illness? Evidence from Canadian acute-care hospitals during 4 influenza seasons

Presenteeism Attitudes and Behavior Among Missouri Kindergarten to Twelfth Grade (K-12) School Nurses. The Journal of school nursing : the official publication of the National Association of School Nurses

Barriers and facilitators to influenza-like illness absenteeism among healthcare workers in a tertiary-care healthcare system

Disproportionate Incidence of COVID-19 Infection, Hospitalizations, and Deaths Among Persons Identifying as Hispanic or Latino -Denver

Healthcare workers' presenteeism and chemoprophylaxis against nosocomial influenza in patients hospitalized during the 2018-2019 season

Influenza Vaccination Coverage among Registered Nurses in China during 2017-2018: An Internet Panel Survey. Vaccines (Basel) 2019

Facilitators for influenza vaccination uptake in nurses at the Shiraz University of Medical Sciences

Impact on Employee Productivity From Presenteeism and Absenteeism: Evidence From a Multinational Firm in Sri Lanka

Risk of zoonotic transmission of swine influenza at the human-pig interface in Guangdong Province, China

Analysis of factors and measures for hospital respiratory tract infection of medical staff in department of respiratory medicine)

东莞市中小学生、工人和居民甲型 H1N1 流感防治知信行现况调查 (Cross-sectional Study on Knowledge,Attitude and Behavior of Influenza A(H1N1) Among

Students,Workers and Residents in Dongguan). 实 用 预 防 医 学 (Practical Preventive Medicine

The actual implementation status of self-isolation among Japanese workers during the COVID-19 outbreak. Tropical medicine and health

Attendance at London workplaces after symptom onset: a retrospective cohort study of staff members with confirmed COVID-19

Refusal to Take Sick Leave after Being Diagnosed with a Communicable Disease as an Estimate of the Phenomenon of Presenteeism in Poland

Refusal to take a sick leave as an estimate of the phenomenon of presenteeism in Poland

Sick at work: presenteeism among nurses in a Portuguese public hospital

The cumulative impact and associated costs of multiple health conditions on employee productivity

Hygiene perception changes during the influenza A H1N1 pandemic in Germany: incorporating the results of two cross-sectional telephone surveys

Absenteeism and presenteeism in healthcare workers due to respiratory illness

Presenteeism among health care workers with laboratory-confirmed influenza infection: A retrospective cohort study in Queensland

Sick leave accessibility in junior doctors at an Australian health service. Infection, Disease & Health

Self-reported anticipated compliance with physician advice to stay home during pandemic (H1N1) 2009: results from the

Knowledge about pandemic influenza and compliance with containment measures among Australians

Influenza-like" symptoms and work-related behavior among healthcare workers and other professionals: Results of a global survey

A preliminary investigation of health and work-environment factors on presenteeism in the workplace

Nurses' presenteeism and its effects on self-reported quality of care and costs

The Gig Economy and Contingent Work: An Occupational Health Assessment

Evaluating Employment Quality as a Determinant of Health in a Changing Labor Market

Precarious employment and mental health: a systematic review and metaanalysis of longitudinal studies

McKinsey and Company: COVID-19 and Gender Equality: Countering the Regressive Effects

COVID-19 and Inequalities. Fiscal studies 2020

COVID-19 and the Gender Gap in Work Hours. Gender, Work & Organization

A Perfect Storm: Impact of Genomic Variation and Serial Vaccination on Low Influenza Vaccine Effectiveness During the

The impact of repeated vaccination on influenza vaccine effectiveness: a systematic review and meta-analysis

Not applicable