COVID-19 mitigation measures in primary schools and association with infection and school staff wellbeing: An observational survey linked with routine data in Wales, UK.

INTRODUCTION: School-based COVID-19 mitigation strategies have greatly impacted the primary school day (children aged 3-11) including: wearing face coverings, two metre distancing, no mixing of children, and no breakfast clubs or extra-curricular activities. This study examines these mitigation measures and association with COVID-19 infection, respiratory infection, and school staff wellbeing between October to December 2020 in Wales, UK.
METHODS: A school staff survey captured self-reported COVID-19 mitigation measures in the school, participant anxiety and depression, and open-text responses regarding experiences of teaching and implementing measures. These survey responses were linked to national-scale COVID-19 test results data to examine association of measures in the school and the likelihood of a positive (staff or pupil) COVID-19 case in the school (clustered by school, adjusted for school size and free school meals using logistic regression). Linkage was conducted through the SAIL (Secure Anonymised Information Linkage) Databank.
RESULTS: Responses were obtained from 353 participants from 59 primary schools within 15 of 22 local authorities. Having more direct non-household contacts was associated with a higher likelihood of COVID-19 positive case in the school (1-5 contacts compared to none, OR 2.89 (1.01, 8.31)) and a trend to more self-reported cold symptoms. Staff face covering was not associated with a lower odds of school COVID-19 cases (mask vs. no covering OR 2.82 (1.11, 7.14)) and was associated with higher self-reported cold symptoms. School staff reported the impacts of wearing face coverings on teaching, including having to stand closer to pupils and raise their voices to be heard. 67.1% were not able to implement two metre social distancing from pupils. We did not find evidence that maintaining a two metre distance was associated with lower rates of COVID-19 in the school.
CONCLUSIONS: Implementing, adhering to and evaluating COVID-19 mitigation guidelines is challenging in primary school settings. Our findings suggest that reducing non-household direct contacts lowers infection rates. There was no evidence that face coverings, two metre social distancing or stopping children mixing was associated with lower odds of COVID-19 or cold infection rates in the school. Primary school staff found teaching challenging during COVID-19 restrictions, especially for younger learners and those with additional learning needs.

Introduction

The COVID-19 global pandemic caused by the transmission of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) resulted in the temporary closure of educational settings worldwide [1]. Implemented worldwide from mid-April 2020, school closures were used as a public health measure to reduce social contacts and the risk of transmission amongst pupils, school staff, families and the wider community. However, recent evidence indicates that children below the age of 14 appear to have lower susceptibility to infection and display fewer clinical symptoms [2-5]. Population-level data suggests that whilst transmission risks within school exists, risks are lower compared to within households [6]. Adults living with young children (0–11 years) during the period after schools reopened encountered no greater risk of COVID-19 infection [7], and school staff were at no greater risk of COVID-19 infection than other working-age adults [8].

Educational settings reopened for face-to-face teaching and learning from September to December 2020. In Wales, one of the four nations of the UK, education is a devolved responsibility of the Welsh Government. Operational guidance to schools in Wales in the 2020 autumn term [9] (1 September to 22 December) included widespread adaptation to social behaviours and a variety of school-based mitigation measures. This included encouraging wearing face coverings, reducing contacts, maintaining social distancing between pupils and staff, segregating classes and guidance on breakfast clubs, extra-curricular activities and outdoor learning [9].

Research examining the implementation of guidelines by schools highlights major challenges, including the ability of school staff to maintain a two metre distance from staff and pupils [10]. School staff highlight the conflict between balancing preventative measures with learning, with measures such as physical distancing policies negatively impacting on teaching quality. A rapid scoping review assessing the impacts of school-based measures concluded that there is an urgent need for research assessing the effectiveness of these measures on directly affected populations (e.g. pupils and school staff) [11], and on the psychosocial well-being and mental health of school populations. This is important as evidence suggests teacher wellbeing is a critical factor in creating stable environments for children to thrive [12] and is positively associated with academic achievement [13].

This study linked routinely collected COVID-19 polymerase chain reaction (PCR) test results data with survey data to examine the association between COVID-19 positive cases within the primary school setting and different school-based mitigation measures aligned to guidance, implemented between October to December 2020. It also examined these measures with school staff’s self-reported (a) cold symptoms in the previous seven days, as a proxy for infection rates; and (b) levels of anxiety and depression. Secondary qualitative data exploring the impacts of wearing face coverings are also presented to complement quantitative findings.

Methods

Study design

This study adopted a mixed methods design. Participants were recruited through the HAPPEN primary school network (Health and Attainment of Pupils in a Primary Education Network) [14] in September 2020. An online survey (open 9 October 2020 to 16 December 2020) with school staff captured self-reported implementation of school-based COVID-19 mitigation measures and individual level outcomes of cold symptoms and anxiety/depressive symptoms. The survey findings were linked with routine data on COVID-19 test results for staff and pupils within the respective school of the staff participant for the school-level outcome. Linkage was performed using the SAIL (Secure Anonymised Information Linkage) Databank [15,16]. Data were linked at the individual level using the School Workforce Annual Census (SWAC) to assign each school staff to their school, and the Pupil Level Annual School Census (PLASC) to identify pupil by school and link COVID-19 test results to the appropriate school [17]. In addition, open-ended survey responses were used to examine views of school staff using a content analysis approach [18,19]. The RECORD checklist [20] for this study is presented in S1 Appendix.

Ethics

Ethical approval was granted by the Swansea University Medical School Research Ethics Committee (2017-0033E). Information sheets and consent forms were distributed via email to participants detailing the aims of the study. To participate in the survey, primary school staff were required to provide written informed consent. All participants were able to withdraw from the research at any point. All participants were assigned a unique ID number, and any personal data such as names were removed. Electronic data (survey responses) were stored in password-protected files that were only accessible to the research team. The routine data used in this study are available in the SAIL Databank [21] and are subject to review by an independent Information Governance Review Panel (IGRP), to ensure proper and appropriate use of SAIL data. Before any data can be accessed, approval must be received from the IGRP. When access has been approved, it is accessed through a privacy-protecting safe haven and remote access system referred to as the SAIL Gateway. SAIL has established an application process to be followed by anyone who would like to access data via SAIL. This study has been approved by the SAIL IGRP (project reference: 0911).

School staff survey and linked data

A convenience sample of primary school staff were recruited by contacting members of the HAPPEN network and directly emailing all primary schools in Wales, UK (n = 1,203) in September 2020. The survey was promoted through existing partnerships with stakeholders including regional education consortia groups. The online survey was open for responses from 9 October 2020 to 16 December 2020 (study period) when schools returned for face-to-face teaching. Inclusion criteria for participation was any primary school staff working within a local authority maintained (publicly funded) primary school. The development of the survey was based on input from the research team specialising in child health and education research (authors EM, MJ, SB), education stakeholders (regional education consortia curriculum staff) and a headteacher and teacher from two primary schools to ensure appropriate wording and usability. The final survey contained 41 questions consisting of demographic, categorical and open-ended questions. The survey included the validated questionnaires Generalized Anxiety Disorder (GAD-7) [22] and Patient Health Questionnaire (PHQ-9) [23] to assess the presence and severity of anxiety and depressive symptoms. The survey was conducted online and could be completed by a member of school staff at a convenient time via an electronic device including mobile phone, tablet, laptop and computer. Responses were downloaded to an Excel spreadsheet. Quantitative data responses were uploaded to the SAIL Databank [15,16] to be linked with COVID-19 school testing data [17], and analysed using Stata (version 16). A full copy of the survey is presented in S2 Appendix, and detail regarding survey item, item response categories and item coding for analyses are presented in S3 Appendix.

The process of data coding involved two researchers. The first researcher downloaded the raw data, cleaned the data, checked for duplicates, generated a unique participant ID number and removed identifiable information. This process protects participants’ anonymity by ensuring that the second researcher conducting the analyses could not identify individuals. This coded dataset was uploaded to the SAIL Databank, a national data infrastructure asset of anonymised data about the population of Wales that enables secure data linkage and analysis for research. To link the data, the demographic data are separated from the survey data and sent to a trusted third party, Digital Health and Care Wales and the survey data goes to SAIL using a secure file upload. A unique Anonymous Linking Field (ALF) is assigned to the person-based record before it is joined to clinical data via a system linking field. This dataset was accessible to authors listed from Population Data Science.

Quantitative analysis

A COVID-19 school incident in Wales, UK, is defined as one or more positive COVID-19 cases in a school [24]. The primary outcome was the probability of at least one positive school-level COVID-19 test (pupils or staff) within the school setting linked to the staff participant during the study period. Secondary binary outcomes investigated at an individual level captured by the online survey included self-reported cold symptoms in the previous seven days as a proxy of infection risk as evidence suggests a crossover of symptoms between COVID-19 and the common cold [25], and moderate/severe anxiety (GAD-7) and moderate/severe depression (PHQ-9). Eligibility criteria within final analyses models were any primary school staff participant with complete linked survey and routine records. Participants contracted to multiple schools were excluded from analyses (n = 3) (see Fig 1).

Fig 1

Cohort flow diagram.

Logistic regression analyses adjusting for confounding variables (school size, proportion of pupils eligible for free school meals as an indicator for deprivation) and clustered by school determined the Odds Ratio (OR) at a school level for at least one positive linked COVID-19 test at the respective school during the study period and for individual-level (school staff) secondary outcomes (self-reported cold symptoms, moderate/severe anxiety, moderate/severe depression).

All exposure measures relating to government guidance were captured through self-report by school staff via the online survey and were analysed in individual models (univariable) and then in a combined model (multivariable). Items with multiple category responses or continuous numerical values were assigned ordinal categories to ease interpretation. For example survey response categories for keep two metres from pupils/staff included i) never, ii) rarely, iii) some of the time, iv) most of the time, v) always, with combined ordinal categories for analyses of i) never/rarely, ii) some of the time, iii) most of the time/always. For these variables, likelihood-ratio tests of variables as whole were performed to assess goodness of fit between models including and excluding variables for the primary outcome. Further detail of exposures for all survey items within analyses including possible response category, grouping and coding can be found in S3 Appendix. This study assumed self-reported mitigation measures to be in effect for the duration of the study period based on operational guidance issued to schools at the time of the study [9].

Qualitative analysis

Secondary qualitative content analysis was conducted to explore the impact of wearing face coverings on teaching, attained from item 29 (see S2 Appendix). Content analysis aims to make contextual inferences of data by condensing text into related concepts to provide knowledge to describe a phenomenon [18]. Conceptual content analysis was chosen to quantify the frequency of reoccurring words/themes and offer a descriptive lens of the quantitative data in terms of the most significant impacts of wearing face coverings for school staff [26,27]. An inductive approach was used as knowledge of this subject is limited due to the new and rapidly evolving nature of the COVID-19 pandemic. The lead researcher (EM) followed the steps of preparation, organising and reporting outlined by Elo and Kyngäs [26].

During the preparation stage, words or sentences were chosen as the unit of analysis to represent related concepts. The lead researcher (EM) who was female and had previous experience in qualitative data analysis read the open-ended responses several times to facilitate immersion in the data [28] and to gain an understanding of ‘what is going on’ [29]. The use of memoing recorded notes of patterns and emerging insights relating to coding ideas. Thoughts relating to decision processes were documented in a reflexive journal [30,31]. In the case of inductive content analysis, an open coding process to organising the data was applied by manually assigning freely generated open codes, consisting of words and sentences representing key conceptual responses. The initial list of words and sentences were grouped under higher order headings [28], with each heading named using content-characteristic words that describe the phenomenon [26]. The categories produced were discussed and reviewed with the research team to develop the final list of category headings characterising any impacts of face coverings on teaching. The researchers did not have any interaction with participants.

Results

Reponses were obtained from 353 participants from 59 primary schools located within 15 local authorities in Wales, UK (Table 1). A cohort flow diagram is presented in Fig 1. 87 (24.7%) participants had a linked COVID-19 positive test, 31 (8.8%) reported cold symptoms, 62 (17.6%) and 67 (19.0%) reported moderate/severe anxiety and depression respectively. Participants were removed from the regression analyses due to missing values for the following outcomes; cold symptoms outcomes (n = 8), anxiety (n = 49) and depression (n = 125) (multivariable models). Missing values of exposure variables ranged from 0 to 19 (see Table 2). Complete case analyses are presented below. Sensitivity analyses where missing responses are coded as 0 are presented in S4 Appendix.

Table 1

Demographics of survey respondents; *obtained from Welsh Government data online [32].

Characteristics	% (n)
Number of participants (school staff)	353
Number of schools	59 (1,203 national total*)
Number of local authorities	15 (22 national total*)
School characteristics
Mean Percentage of Free School Meals	20.6% (national average 19%*)
Free School Meal category
0–10%	28.8% (17)
11–20%	25.4% (15)
21–30%	23.7% (14)
31%+	22.1% (13)
School size (number of pupils)	(national average 223*)
0–100	8.5% (5)
101–200	32.2% (19)
201–300	23.7% (14)
301–400	16.9% (10)
401–500	15.3% (9)
501+	3.4% (2)
Participant characteristics
Job role
Support staff	4.1% (14)
Supply teacher	1.2% (4)
Teaching assistant	35.1% (120)
Teacher	53.2% (182)
Headteacher (teaching)	1.2% (4)
Headteacher (non-teaching)	5.3% (18)
Missing	3.2% (11)
Full time	78.8% (278)
Part time	18.4% (65)
Missing	2.8% (10)
Year group
Foundation phase (ages 3–7) Reception	25.6% (90)
Key Stage 2 (ages 7–11)	30.0% (106)
Combination of years	35.7% (126)
Missing	8.8% (31)
Outcomes
Positive COVID-19 school test	24.7% (87)
Missing	0
Report cold symptoms previous 7 days	8.8% (31)
Missing	2.3% (8)
Report moderate/severe anxiety (GAD-7)	17.6% (62)
Missing	13.9% (49
Report moderate/severe depression (PHQ-9)	19.0% (67)
Missing	35.4% (125)

Table 2

Distribution of individual school staff responses to mitigation survey items and school-level response agreement (see S3 Appendix).

Survey item	Response	% (n)	% (n) of schools with ≥80% agreement of responses (for school-level outcome)
Keep two metres from pupils	Never/rarely	67.1% (237)	61% (36)
	Sometimes	23.5% (83)
	Most of the time/always	7.9% (28)
	Missing	1.4% (5)
Keep two metres from staff	Never/rarely	9.1% (32)	59% (35)
	Sometimes	22.1% (78)
	Most of the time/always	66.9% (236)
	Missing	2.0% (7)
Wear face covering	No	56.1% (198)	83% (49)
	Mask	31.4% (111)
	Visor	11.3% (40)
	Missing	(<5)
Non-household contacts within one metre	0	24.7% (87)	41% (24)
	1–5	38.8% (137)
	≥ 6	36.5% (129)
	Missing	0
Non-household contacts direct	0	81.9% (289)	73% (43)
	1–5	8.5% (30)
	≥ 6	9.6% (34)
	Missing	0
Classes mixing at play	No	72.8% (257)	88% (52)
	Yes: outdoors in a field or large outdoor space	22.4% (79)
	Yes: in the hall	3.7% (13)
	Missing	(<5)
School offers breakfast club	No	36.3% (128)	95% (56)
	Yes	58.4% (206)
	Missing	5.4% (19)
School offers extra-curricular clubs	No	71.7% (253)	91% (54)
	Yes	26.6% (94)
	Missing	1.7% (6)
Teaching outdoors	Never/hardly ever	25.2% (89)	58.6% (34)
	Some of the time	61.8% (218)
	Most of the time	11.1% (39)
	Missing	2% (7)

Quantitative results

Exposure variables were examined individually (univariable) for association with outcomes and then all variables were entered together (multivariable) in the final combined models for the outcomes of school-level COVID-19 (Table 3), self-reported cold symptoms (Table 4), moderate/severe anxiety (Table 5) and depressive (Table 6) symptoms. Models were adjusted for school size and free school meal proportion, and clustered by school (see S3 Appendix for exposure response coding).

Table 3

Univariable (model 1) and multivariable (model 2) logistic regression models of self-reported school-based mitigation measures (survey) and school-level probability of any positive COVID-19 case in school (SAIL).

At least one positive COVID-19 test at school (pupils and staff) during study period (SAIL) (school-level)
Self reported measures from survey		Univariable (model 1)		Multivariable (model 2) R2 = 0.12
		OR	95% CI	OR	95% CI
Face covering (reference no face covering)	Mask	2.82**	1.11 to 7.31	2.10*	0.87 to 5.05
	Visor	1.65	0.47 to 5.74	1.42	0.40 to 5.2
Keep two metres from pupils (reference never/rarely)	Sometimes	1.01	0.50 to 2.02	0.79	0.36 to 1.75
	Most of the time/always	0.97	0.39 to 2.38	0.89	0.33 to 2.38
Keep two metres from staff (reference never/rarely)	Sometimes	1.58	0.47 to 5.32	1.82	0.63 to 5.26
	Most of the time/always	2.46	0.76 to 7.96	2.85*	0.97 to 8.37
Non-household contacts within one metre (reference 0 contacts)	1–5	0.97	0.57 to 1.67	0.89	0.47 to 1.66
	≥ 6	1.47	0.78 to 2.79	1.17	0.53 to 2.56
Non-household contacts direct (reference 0 contacts)	1–5	2.27*	0.98 to 5.22	2.89**	1.01 to 8.31
	≥ 6	1.58	0.86 to 2.89	1.70*	0.93 to 3.10
Classes mix at play (reference no classes mixing)	Yes	0.89	0.40 to 1.98	1.06	0.53 to 2.13
School offers breakfast club (reference no breakfast club)	Yes	0.58	0.23 to 1.48	0.67	0.28 to 1.64
School offers extra-curricular clubs (reference no extra-curricular clubs)	Yes	1.67	0.73 to 3.86	1.99	0.85 to 4.71
Teach outdoors (reference never/hardly ever)	Sometimes	0.89	0.58 to 1.38	0.88	0.52 to 1.47
	Most of the time	0.65	0.23 to 1.84	0.45	0.11 to 1.81

OR: Odds Ratio; 95% CI: 95% confidence intervals; p<0.05**, p<0.1*; adjusted for school size, proportion of pupils eligible for free school meals, clustered by school. Model 2 likelihood-ratio test keep two metres from pupils (p = 0.1) and staff (p = 0.03). See S3 Appendix for variable codebook.

Table 4

Univariable (model 3) and multivariable (model 4) logistic regression models of self-reported school-based mitigation measures (survey) and individual level (school staff) self-reported cold symptoms (survey).

Reported cold symptoms in previous 7 days (individual level: school staff)
Self reported measures from survey		Univariable (model 3)		Multivariable (model 4) R2 = 0.07
		OR	95% CI	OR	95% CI
Face covering (reference no face covering)	Mask	1.66	0.89 to 3.10	1.98**	1.02 to 3.88
	Visor	2.16	0.76 to 6.17	2.35	0.81 to 6.86
Keep two metres from pupils (reference never/rarely)	Sometimes	0.46	0.16 to 0.31	0.50	0.15 to 1.62
	Most of the time/always	0.79	0.20 to 3.14	0.81	0.22 to 2.96
Keep two metres from staff (reference never/rarely)	Sometimes	0.66	0.16 to 2.76	0.59	0.11 to 3.10
	Most of the time/always	0.57	0.20 to 1.60	0.51	0.14 to 1.81
Non-household contacts within one metre (reference 0 contacts)	1–5	0.92	0.41 to 2.10	0.86	0.35 to 2.09
	≥ 6	0.85	0.30 to 2.46	0.68	0.16 to 2.89
Non-household contacts direct (reference 0 contacts)	1–5	2.53*	0.85 to 7.51	3.09*	0.96 to 9.93
	≥ 6	0.78	0.20 to 2.97	1.14	0.20 to 6.34
Classes mix at play (reference no classes mixing)	Yes	0.49	0.19 to 1.27	0.53	0.22 to 1.28
School offers breakfast club (reference no breakfast club)	Yes	0.98	0.46 to 2.07	1.15	0.51 to 2.58
School offers extra-curricular clubs (reference no extra-curricular clubs)	Yes	1.59	0.82 to 3.10	1.19	0.53 to 2.64
Teach outdoors (reference never/hardly ever)	Sometimes	0.54	0.23 to 1.26	0.60	0.26 to 1.36
	Most of the time	1.17	0.36 to 3.77	0.86	0.26 to 2.90

OR: Odds Ratio; 95% CI: 95% confidence intervals; p<0.05**, p<0.1*; adjusted for school size, proportion of pupils eligible for free school meals, clustered by school. See S3 Appendix for variable codebook.

Table 5

Univariable (model 5) and multivariable (model 6) logistic regression models of self-reported school-based mitigation measures (survey) and individual level (school staff) moderate/severe anxiety symptoms (survey).

Moderate/severe anxiety (GAD-7) (individual level: school staff)
Self reported measures from survey		Univariable (model 5)		Multivariable (model 6) R2 = 0.07
		OR	95% CI	OR	95% CI
Face covering (reference no face covering)	Mask	1.35	0.78 to 2.33	1.10	0.51 to 2.39
	Visor	2.41*	0.87 to 6.72	2.58	0.82 to 8.08
Keep two metres from pupils (reference never/rarely)	Sometimes	0.64	0.31 to 1.30	0.62	0.29 to 1.35
	Most of the time/always	2.12	0.67 to 6.68	2.31	0.72 to 7.35
Keep two metres from staff (reference never/rarely)	Sometimes	0.50	0.14 to 1.76	0.53	0.14 to 2.06
	Most of the time/always	0.63	0.21 to 1.91	0.77	0.21 to 2.76
Non-household contacts within one metre (reference 0 contacts)	1–5	0.90	0.42 to 1.89	0.85	0.39 to 1.87
	≥ 6	1.31	0.59 to 2.88	1.41	0.64 to 3.08
Non-household contacts direct (reference 0 contacts)	1–5	0.58	0.18 to 1.92	0.62	0.18 to 2.13
	≥ 6	1.59	0.47 to 5.34	2.03	0.55 to 7.52
Classes mix at play (reference no classes mixing)	Yes	0.99	0.49 to 1.99	0.93	0.43 to 2.02
School offers breakfast club (reference no breakfast club)	Yes	0.70	0.38 to 1.27	0.77	0.38 to 1.55
School offers extra-curricular clubs (reference no extra-curricular clubs)	Yes	1.22	0.50 to 2.94	1.25	0.44 to 3.56
Teach outdoors (reference never/hardly ever)	Sometimes	0.65	0.34 to 1.22	0.62	0.31 to 1.25
	Most of the time	0.70	0.26 to 1.87	0.70	0.25 to 1.94

OR: Odds Ratio; 95% CI: 95% confidence intervals; p<0.05**, p<0.1*; adjusted for school size, proportion of pupils eligible for free school meals, clustered by school. See S3 Appendix for variable codebook.

Table 6

Univariable (model 7) and multivariable (model 8) logistic regression models of self-reported school-based mitigation measures (survey) and individual level (school staff) moderate/severe depressive symptoms (survey).

Moderate/severe depression (PHQ-9) (individual level: school staff)
Self reported measures from survey		Univariable (model 7)		Multivariable (model 8) R2 = 0.07
		OR	95% CI	OR	95% CI
Face covering (reference no face covering)	Mask	1.78	0.93 to 3.42	1.70	0.83 to 3.48
	Visor	3.38**	1.31 to 8.77	4.81**	1.52 to 15.22
Keep two metres from pupils (reference never/rarely)	Sometimes	1.03	0.50 to 2.15	0.97	0.40 to 2.36
	Most of the time/always	1.18	0.50 to 2.78	1.95	0.61 to 6.21
Keep two metres from staff (reference never/rarely)	Sometimes	1.26	0.29 to 5.36	0.68	0.13 to 3.48
	Most of the time/always	1.05	0.28 to 3.97	0.73	0.16 to 3.26
Non-household contacts within one metre (reference 0 contacts)	1–5	1.44	0.73 to 2.84	1.88	0.74 to 4.75
	≥ 6	1.65	0.76 to 3.59	2.70**	1.11 to 6.56
Non-household contacts direct (reference 0 contacts)	1–5	1.12	0.45 to 2.77	0.90	0.27 to 3.00
	≥ 6	1.28	0.45 to 3.68	1.17	0.35 to 3.98
Classes mix at play (reference no classes mixing)	Yes	0.82	0.41 to 1.64	0.82	0.30 to 2.22
School offers breakfast club (reference no breakfast club)	Yes	0.73	0.40 to 1.34	0.89	0.32 to 2.44
School offers extra-curricular clubs (reference no extra-curricular clubs)	Yes	1.03	0.35 to 3.05	0.87	0.24 to 3.21
Teach outdoors (reference never/hardly ever)	Sometimes	0.86	0.40 to 1.84	0.75	0.30 to 1.91
	Most of the time	1.84	0.56 to 6.06	1.59	0.39 to 6.50

OR: Odds Ratio; 95% CI: 95% confidence intervals; p<0.05**, p<0.1*; adjusted for school size, proportion of pupils eligible for free school meals, clustered by school. See S3 Appendix for variable codebook.

Number of non-household contacts (1-metre, direct)

In the multivariable models, compared to reporting 0 contacts, reporting more non-household direct contacts was associated with higher odds of COVID-19 at the school level (1–5 contacts OR = 2.89, Table 3, model 1), and a trend to higher general infection (Table 4, model 3). Reporting 6 or more contacts within 1-metre was associated with higher depression (OR = 2.70, Table 6, model 8).

Face covering

In the univariable model there was evidence that reporting to wear a face covering was associated with an increased odds of a school-level COVID-19 case; OR = 2.82. Compared to reporting no face coverings, masks were associated with increased odds of reporting cold symptoms (multivariable model: OR = 1.98), Table 4, model 4). Reporting wearing a visor was associated with higher odds of depression (multivariable model: OR = 4.81, Table 6, model 8).

Two metre distance from pupils or staff

In the univariable models there were no statistically significant results to support a reduced odds for any of the outcomes when using two metre distancing. In the multivariable models we found a trend to an increased odds of a COVID-19 positive test for the grouped exposure of staff maintaining a two metre distance from other staff most of the time/always compared to never/rarely.

Classes mixing, breakfast club, extra-curricular clubs and teaching outdoors

There was no significant difference in terms of infection (COVID-19 and cold) or anxiety/depression for staff in schools that allowed classes to mix, offered breakfast or extra-curricular clubs or taught outdoors most of the time.

Qualitative results

There were 129 responses from primary school staff relating to impacts of wearing face coverings. The final categories conceptualising the impacts of wearing face coverings and frequency counts were; (i) difficulty being heard/understood–having to talk louder (n = 71); (ii) difficulty understanding body language/facial expressions (n = 25); (iii) physical impacts of wearing a face covering including impacts on health and vision (n = 22); (iv) social/emotional impacts affecting relationships with pupils (n = 12); (v) challenges for pupils with additional learning needs and English as an additional language (n = 9); and (vi) impact on teaching phonics (n = 6). In some instances, quotes were coded within multiple categories due to the open-ended nature of the survey question allowing long text responses. A summary of each category is discussed below and additional key quotes are presented in S5 Appendix.

Difficulty being heard/understood—Having to talk louder

The most frequent impact of wearing face coverings was the challenge of being heard or understood by pupils. This required staff to have to stand closer to pupils and to raise their voice to be heard. School staff reported that they found it difficult to hear others wearing a mask, and this was a particular issue for staff with hearing problems.

“Pupils can’t always hear me so I have to lift the visor…when two meters away and talk louder when I am closer to support pupils” (teaching assistant)

Difficulty understanding body language/facial expressions

School staff noted a challenge for pupils in understanding the body language or interpreting facial expressions of adults. This impacted staff in this study to communicate with children and was particularly challenging for younger pupils.

“I find it extremely difficult to wear a mask/visor whilst teaching. They are young children and need to see facial expressions. It also affects my hearing and their ability to hear me clearly” (teacher)

Physical impacts of wearing a face covering including impacts on health and vision

School staff reported physical impacts and negative complaints including feelings of discomfort. Other common negative effects included their vision, headaches and sore throat. Underlying medical conditions including asthma contributed to challenges experienced by staff with perceived restrictions to breathing.

“Visors are really difficult, they make me feel enclosed and stressed. The children cannot hear me and the vision is not brilliant either” (teacher)

Social/emotional impacts affecting relationships with pupils

Those that wore a face covering and particularly mask use commented on the emotional impact of children not being able to interpret emotions. Staff perceived that this had an impact on their relationship with pupils.

“Yes, the children would not be able to see my expression, if they are upset they wouldn’t be able to see my reaction or compassion” (teaching assistant)

Challenges for pupils with additional learning needs and English as an additional language

Additional challenges were presented with supporting children with additional learning needs (ALN) or English as an additional language (EAL), with mask use impacting communication and language development.

“Yes, it’s affecting my teaching. I work with pupils who are learning English as an additional language and they ideally need to be able to see my facial expressions and lip movements in order to help them understand and develop the language themselves” (teacher)

Impact on teaching phonics

School staff specifically made references to teaching phonics, including the challenges of teaching reading, writing and language skills. Some felt that face masks restricted modelling of words and demonstrating pronunciation.

“Pupils in my class have low language development. They need to see my mouth to support the modelling of words and phonics. Greater effort in delivering modelled speech can become tiring very quickly” (teacher)

Discussion

This study aims to examine the association of different school-based mitigation measures reported by primary school staff between October to December 2020 on the likelihood of any school-level COVID-19 infection (pupils and staff) at the linked school during this period. This study also examined the association of these measures with individual-level self-reported infection (cold symptoms), anxiety and depression of school staff. Findings suggest that reporting more direct non-household contacts was associated with higher odds of COVID-19 at the school level, and a trend towards self-reported infection. Reporting six or more non-house contacts within 1-metre was also associated with higher depression in school staff. We found no evidence that reporting wearing face coverings or maintaining a two metre distance from pupils or other staff during the study period was associated with lower odds of COVID-19 in the linked school setting.

Whilst this observational study offers a real-world evaluation of the school setting, findings highlight the challenge for staff in implementing and adhering to school guidelines. This study assumes that reported measures were in place for the duration of the study period in line with operational guidance issued to schools at that time. However, changes in day-to-day school practice brings methodological challenges of evaluating compliance with and effectiveness of national-level guidance. Our findings of within-school agreement suggests some measures are implemented at a school-level (face coverings, mixing classes at play, breakfast and extra-curricular clubs). In comparison, agreement of other measures (number of contacts, maintaining two metre distance from pupils and staff and teaching outdoors) suggest individual-level influences of adherence to measures, reflecting the challenge of implementing generic guidance in a dynamic school environment.

The finding that reduced contacts may be protective at the school-level is important within the contexts of different settings where the implementation and adherence to different blanket mitigation measures varies. Specifically, this study finds an association between the number of direct physical contacts and increased likelihood of COVID-19 school infections. It is well established that contact patterns of close proximity, prolonged contact and contact frequency are strongly associated with increased risk of transmission [33]. Our finding is consistent with the evidence base regarding contact patterns where reducing number of contacts is associated with a reduction in the basic reproduction number (R0) [34]. A crossover between COVID-19 and common cold symptoms has been established [25], and the current study also found an association between direct physical contacts and self-reported cold symptoms. As this study suggests variation of school-based mitigation measures between and within-schools, encouraging individual behaviours of school staff such as reducing direct contacts may be of benefit in reducing transmission of COVID-19 or general infection in the school setting.

Relating to proximity, qualitative findings from this study suggest challenges for staff wearing face coverings including pupils having difficulty hearing and understanding, and this required them to talk louder or move physically closer to pupils to be heard. Research demonstrates that people speak louder when wearing masks [35]. Staff also noted that pupils were unable to interpret facial expressions or emotions, impacting their relationship with pupils and children’s perception of compassionate emotions conveyed by staff. Challenges were cited for ALN or EAL pupils particularly regarding speech and language development. As facial expressions and gestures are largely responsible for verbal, non-verbal and emotional face-to-face communication, face masks may hinder interpersonal communication with pupils [36].

Type of face mask was not captured in this study (e.g. medical/non-medical grade). Guidance to primary schools during the study period (autumn term 2020) did not enforce medical-grade face coverings [9,37]. The type of face covering worn by staff in this study may include cloth masks which have been found to increase respiratory infection risk due to moisture retention, reuse and poor filtration [38]. This may explain individual-level findings that staff wearing face masks had higher odds of reporting cold symptoms in the previous seven days. In the context of SARS-CoV-2 transmission, the main purpose of face coverings is to prevent onward transmission to others as opposed to protecting the individual wearing the face covering [39]. A systematic review and meta-analysis showed a reduction in COVID-19 incidence with mask wearing, though type of fask mask, compliance and frequency of use were not captured [40]. It is important to note the many confounding variables of face covering usage that were not measured in this study. This includes background prevalence in the area which may influence wearing face coverings. Evidence suggests that mandating face covering use alone may not increase usage and thus, individual behaviours and other influences are likely to play a role in face covering behaviour [41]. In addition, this observational study assumes reported mitigation measures were in effect for the duration of the study period. It is possible that reverse causality occurred, that is a school staff practitioner may have chosen to wear a face covering following the onset of common cold symptoms.

The use of visors was associated with higher anxiety/depression for staff in this study. Impacts on teacher wellbeing have been highlighted in previous research by HAPPEN during school closures and the phased reopening of schools in the summer term of 2020, with primary school staff advocating for their wellbeing to be prioritised [42]. This is important as teacher wellbeing is associated with academic achievement [13]. School staff in the current study also commented on the physical impacts of wearing face coverings, including negatively affecting their vision, causing headaches and breathing difficulties. Qualitative research exploring face covering behaviour has highlighted the wide range of motivations, including individual and community protection, and barriers such as physical challenges and discomfort [43]. It is possible that the physical discomforts expressed by staff in this study influence face covering behaviour.

We found no evidence in this study that maintaining a two metre distance from pupils reduces the odds of a COVID-19 school-level incident. However, few staff were able to achieve this. Research examining the implementation of preventive school-based measures in primary schools in England highlights the challenge of maintaining physical distancing from pupils and the negative impact of distancing measures on teaching including teaching letter formation [10]. This finding is mirrored in the current study, with specific references to the challenges of teaching phonics and those discussed previously. The potential consequences of failing to address these pedagogical impacts include pupils falling further behind in their learning [44].

This study did not find evidence of higher odds of COVID-19 school incidents where children from different classes mix, including breakfast club, extra-curricular clubs and mixing different classes at playtime. School provision during the COVID-19 pandemic encompasses balancing transmission risks against the benefits for children’s social and emotional development, wider skill development, educational attainment and reducing inequalities. The COVID-19 pandemic has exacerbated pre-existing inequalities including food insecurity, child poverty and child hunger [45,46] which negatively impact educational attainment [47]. Provision such as breakfast clubs that address socio-economic inequalities are of great public health, education and economic importance and this was reflected in guidance at the time of the study encouraging breakfast clubs [9].

The World Health Organization (WHO), UNICEF and UNESCO recently updated advice to policymakers and educators, issuing a set of risk-based considerations regarding school provision since reopening during the COVID-19 pandemic [48]. Whilst the principles aim to prevent and minimise transmission risks within the school setting, the WHO advocate that at the forefront, educational settings should prioritise “the continuity of education for children for their overall well-being, health and safety”, the “social learning and development of children” and to consider implications of decisions on school staff. Findings from this study highlight the challenges of evaluating the implementation of guidance and the variation in implementation at an individual and school-level. Governments continually review available evidence to inform risk-based approaches to education delivery that safeguard children’s learning, health and wellbeing and support school staff. This must consider the risk of transmission in addition to the impacts on pupils, teachers and senior school leaders. Finally, both the Welsh and UK governments have recently announced plans to reverse some of these guidelines for schools in the upcoming 2021/22 academic year starting in September 2021. This includes the removal of isolation policies for children in close contact with confirmed cases, removing the use of school ‘bubbles’ to segregate year groups, and face coverings will no longer be recommended.

Strengths and limitations

All primary schools in Wales (n = 1,203) were contacted however the findings in this study are a convenience sample, only representing those that participated and may not be representative of non-participating schools. A range of school-based measures have been implemented and the findings in this study may not encapsulate all approaches. School-based mitigation measures included in analyses were obtained from a self-report survey and may result in recall bias. This is an observational study and so cannot show cause and effect. As with all observational studies, unmeasured confounders and reverse causality may influence findings, e.g., face covering usage may increase due to a previous COVID-19 case in the school, higher community prevalence and individual behaviours. Thus, face covering use and future COVID-19 cases may be linked by an unmeasured confounder. This study assumed that reported measures were in effect for the duration of the period of study based on national-level guidance issued to schools by the Welsh Government at the start of the autumn term 2020. It is possible that within-schools’ day to day practice varied. Despite this, the sample consists of a range of primary school staff including headteachers, teachers and support staff working in schools in 15 of 22 local authorities in Wales, of varying school size and ranges of pupils eligible for free school meals. This study was able to examine all COVID-19 PCR test results in Wales and link these to the relevant school setting and so gives an objective assessment of the association of self-reported adherence to mitigation measures and COVID-19 test positive cases.

Conclusions

Implementation of COVID-19 mitigation measures was variable and challenging in primary schools in Wales. This study did find evidence that reducing the number of direct non-household contacts is associated with lower risk of COVID-19 in the school and general infection for the individual. This study did not find evidence that face coverings, two metre social distancing, stopping children mixing or removing breakfast clubs are associated with fewer COVID-19 cases in the school or with lower general infection rates and did find evidence that these measures can affect teaching quality.

RECORD statement.

(DOCX)

Click here for additional data file.

Full survey copy.

(DOCX)

Click here for additional data file.

Staff survey variable codebook.

(DOCX)

Click here for additional data file.

Sensitivity analyses.

(DOCX)

Click here for additional data file.

Additional qualitative quotes.

(DOCX)

Click here for additional data file.

Transfer Alert

This paper was transferred from another journal. As a result, its full editorial history (including decision letters, peer reviews and author responses) may not be present.

1 Nov 2021

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

Amitava Mukherjee, ME, Ph.D.

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. In your Data Availability statement, you have not specified where the minimal data set underlying the results described in your manuscript can be found. PLOS defines a study's minimal data set as the underlying data used to reach the conclusions drawn in the manuscript and any additional data required to replicate the reported study findings in their entirety. All PLOS journals require that the minimal data set be made fully available. For more information about our data policy, please see http://journals.plos.org/plosone/s/data-availability.

Upon re-submitting your revised manuscript, please upload your study’s minimal underlying data set as either Supporting Information files or to a stable, public repository and include the relevant URLs, DOIs, or accession numbers within your revised cover letter. For a list of acceptable repositories, please see http://journals.plos.org/plosone/s/data-availability#loc-recommended-repositories. Any potentially identifying patient information must be fully anonymized.

Important: If there are ethical or legal restrictions to sharing your data publicly, please explain these restrictions in detail. Please see our guidelines for more information on what we consider unacceptable restrictions to publicly sharing data: http://journals.plos.org/plosone/s/data-availability#loc-unacceptable-data-access-restrictions. Note that it is not acceptable for the authors to be the sole named individuals responsible for ensuring data access.

We will update your Data Availability statement to reflect the information you provide in your cover letter.

3. Please note that in order to use the direct billing option the corresponding author must be affiliated with the chosen institute. Please either amend your manuscript to change the affiliation or corresponding author, or email us at plosone@plos.org with a request to remove this option.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Partly

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: I Don't Know

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: Thank you for the opportunity to review this paper. COVID-19 mitigation measures in primary schools and associations with infection is an interesting and important topic. The low risk of severe disease and death in children has led to debates on whether school closures and mitigation measures affecting learning are justified.

Despite the density of data and the important topic, my overall impression is that the paper would benefit from being more focused. There are four separate outcomes for associations with staff self-reported mitigation measures; (i) school-level positive COVID-19 cases (primary outcome), (ii) self-reported cold symptoms, (iii) moderate/severe anxiety and (iv) depressive symptoms, and adding qualitative data to this.

Comments:

• The univariable and multivariable OR were presented separately rather than in the same table for the same outcome. I had to scroll back and forth several times to follow the flow. As a reader I would prefer that univariable and multivariable results are presented in the same table with one table per outcome.

• Numbers already presented in the tables were repeated numerically in the text, adding to information overload.

• The questions was phrased with multiple alternatives (see below) whereas the analyses are done separately as dichotomous exposures. It is not clear whether the research question is relative or related to the exposure as whole. I would have expected that Keeping 2 metres from STAFF was treated as one variable (exposure as whole), and that this variable should be tested for significance before interpreting the significance of the subvariables (rarely, some of the time, most of the time, always) I think authors need to clarify this.

Keep 2 metres from STAFF 34. How often are you able to keep 2 metres away from other members of staff during the teaching day? i) Never

ii) Rarely

iii) Some of the time

iv) Most of the time

v) Always

• It is not clear whether the authors corrected for multiple comparisons, which would be relevant if the outcome is treated as categorical variables. The outcome is not frequent and most associations are not significant.

• The authors refer to two different levels of significance <0.5 and <0.1, which is fine. However, these levels are only presented with bold or italic OR and CI in the tables. It would be easier for the reader if the p-value information was presented numerically, or by one and two asterix depending on the level of significance.

• Adding qualitative data to the manuscript is interesting and important. However, the analysis is drowning in all the other information and would deserve better.

• Self-reported cold symptoms in the previous seven days is included as a separate outcome (proxy for infection risk). However, it is not clear how this should be interpreted and what it adds to the research question.

• In the discussion they find that staff wearing face-masks had higher odds of reporting cold symptoms in the previous seven days. It is not clear whether they were more likely to use masks due to cold-symptoms, rather than vice-versa.

• The authors suggests that asymptomatic transmission from children could be an explanation. However, asymptomatic transmission from children in school settings is not very common.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

4 Jan 2022

Reviewer #1: Thank you for the opportunity to review this paper. COVID-19 mitigation measures in primary schools and associations with infection is an interesting and important topic. The low risk of severe disease and death in children has led to debates on whether school closures and mitigation measures affecting learning are justified.

Despite the density of data and the important topic, my overall impression is that the paper would benefit from being more focused. There are four separate outcomes for associations with staff self-reported mitigation measures; (i) school-level positive COVID-19 cases (primary outcome), (ii) self-reported cold symptoms, (iii) moderate/severe anxiety and (iv) depressive symptoms, and adding qualitative data to this.

We would like to thank you for taking the time to read our manuscript and to provide your comments. The manuscript covers two themes: infection (COVID & cold like symptoms), and wellbeing (anxiety and depression) with complementary qualitative data to give context to quantitative findings. The impacts of the COVID-19 pandemic are very much on risk of infection and the impact of mitigation measures on mental health. We feel these aspects need to be addressed together as one impacts on each other. Therefore, whilst we appreciate and agree that this may feel like a data dense manuscript, we believe the risk of infection has to be discussed alongside the impact to mental health. We also believe that the qualitative context is very important in understanding the full picture, and adds a significant contribution to the literature.

Please find section, page and line numbers presenting revisions below, with manuscript text in italic.

Comments:

• The univariable and multivariable OR were presented separately rather than in the same table for the same outcome. I had to scroll back and forth several times to follow the flow. As a reader I would prefer that univariable and multivariable results are presented in the same table with one table per outcome.

Thank you for raising this point. We agree that presenting the univariable and multivariable analyses in the same table ensures clarity for the reader. We have addressed this as suggested, presenting each outcome within separate tables with univariable and multivariable analyses. These tables and models have been renumbered as below:

Quantitative results

Page 15-21

Table 3: School-level probability of any positive COVID-19 case in school (model 1: univariable, model 2: multivariable)

Table 4: Individual level (school staff) self-reported cold symptoms (model 3: univariable, model 4: multivariable)

Table 5: Individual level (school staff) moderate/severe anxiety symptoms (model 5: univariable, model 6 multivariable)

Table 6: Individual level (school staff) moderate/severe depressive symptoms (model 7: univariable, model 8 multivariable)

• Numbers already presented in the tables were repeated numerically in the text, adding to information overload.

We appreciate you drawing attention to the presentation of our results and we agree that repeating the results numerically in the text may overload the reader. To address your comment, we have removed the majority of numerical results. We feel that in instances of statistically significant results at the 5% level, stating the Odds Ratio is useful for the reader. However, we have removed the 95% CIs to ensure results are concise and readable. This style of presentation of results can also be found in other Covid published research in PLOS ONE.

Quantitative results

Page 14, line 321-325

In the multivariable models, compared to reporting 0 contacts, reporting more non-household direct contacts was associated with higher odds of COVID-19 at the school level (1-5 contacts OR = 2.89, Table 3, model 1), and a trend to higher general infection (Table 4, model 3). Reporting 6 or more contacts within 1-metre was associated with higher depression (OR = 2.70, Table 6, model 8).

• The questions was phrased with multiple alternatives (see below) whereas the analyses are done separately as dichotomous exposures. It is not clear whether the research question is relative or related to the exposure as whole. I would have expected that Keeping 2 metres from STAFF was treated as one variable (exposure as whole), and that this variable should be tested for significance before interpreting the significance of the subvariables (rarely, some of the time, most of the time, always) I think authors need to clarify this.

Keep 2 metres from STAFF 34. How often are you able to keep 2 metres away from other members of staff during the teaching day? i) Never, ii) Rarely, iii) Some of the time, iv) Most of the time, v) Always

• It is not clear whether the authors corrected for multiple comparisons, which would be relevant if the outcome is treated as categorical variables. The outcome is not frequent and most associations are not significant.

Thank you for raising this point regarding some survey items and associated response categories and exposure coding. We assigned ordinal categories for the purpose of analyses to some items, including the item you have highlighted (staff social distancing). We treated these variables as categorical in the analyses, with the reference category indicated in the results tables. An example is demonstrated below and presented in S3 Appendix:

Exposures Survey item Survey responses categories Coding for analyses

Keep two metres from STAFF 34. How often are you able to keep 2 metres away from other members of staff during the teaching day? i) Never

ii) Rarely

iii) Some of the time

iv) Most of the time

v) Always Ordinal:

- Never/rarely (i, ii)

- Some of the time (iii)

- Most of the time/always (iv, v)

Example taken from S3 Appendix

Analyses with ordinal responses used the grouped never/rarely category as the reference group.

We did not conduct sub-variable analyses but included the distributions of response categories in Table 2. To improve the clarity for readers and address the points you have raised, we have updated Table 2 to reflect the exposure categories used within analyses. For example:

Survey item Response % (n) % (n) of schools with �  80% agreement of responses (for school-level outcome)

Keep two metres from pupils Never/rarely 67.1% (237) 61% (36)

Sometimes 23.5% (83)

Most of the time/always 7.9% (28)

Missing 1.4% (5)

Example taken from table 2 (page 12-13)

We have also updated the layout and presentation of tables 3 to 6 to aid the interpretation of tables. For example:

Self reported measures from survey Univariable (model 1) Multivariable (model 2)

OR 95% CI OR 95% CI

Keep two metres from staff

(reference never/rarely) Sometimes 1.58 0.47 to 5.32 1.82 0.63 to 5.26

Most of the time/always 2.46 0.76 to 7.96 2.85* 0.97 to 8.37

Example taken from table 3 (page 15-16)

We have presented a full breakdown of exposures included within analyses, survey item, response and coding in S3 Appendix. However, we agree that further clarity is required within the manuscript. To address this within the manuscript text, we have included additional explanations as below:

School staff survey and linked data:

Page 7, line 160-162

A full copy of the survey is presented in S2 Appendix, and detail regarding survey item, item response categories and item coding for analyses are presented in S3 Appendix.

Quantitative analysis

Page 8-9, line 193-206

All exposure measures relating to government guidance were captured through self-report by school staff via the online survey and were analysed in individual models (univariable) and then in a combined model (multivariable). For the purpose of analyses, items with multiple category responses or continuous numerical values were assigned ordinal categories. For example survey response categories for keep two metres from pupils/staff included i) never, ii) rarely, iii) some of the time, iv) most of the time, v) always, with combined ordinal categories for analyses of i) never/rarely, ii) some of the time, iii) most of the time/always. Further detail of exposures for all survey items within analyses including possible response category, grouping and coding can be found in S3 Appendix. This study assumed self-reported mitigation measures to be in effect for the duration of the study period based on operational guidance issued to schools at the time of the study [9].

We have also amended the column title in S3 Appendix to Coding for analyses.

Finally, we have addressed this within the results write up:

Quantitative results

Page 14, line 336-339

In the multivariable models we found a trend to an increased odds of a COVID-19 positive test for the grouped exposure of staff maintaining a 2-metre distance from other staff most of the time/always compared to never/rarely

• The authors refer to two different levels of significance <0.5 and <0.1, which is fine. However, these levels are only presented with bold or italic OR and CI in the tables. It would be easier for the reader if the p-value information was presented numerically, or by one and two asterix depending on the level of significance.

We appreciate that bold or italic is unclear for the reader. To address this, we have amended tables 3-6 (pages 15-21) presenting univariable and multivariable analyses with one (p<0.1) and two (p<0.05) asterisks.

For example:

At least one school positive COVID-19 test (pupils and staff) during study period (SAIL) (school-level)

Self reported measures from survey Univariable (model 1) Multivariable (model 2)

OR 95% CI OR 95% CI

Face covering

(reference no face covering)

Mask 2.82** 1.11 to 7.31 2.10* 0.87 to 5.05

Visor 1.65 0.47 to 5.74 1.42 0.40 to 5.2

Taken from table 2

• Adding qualitative data to the manuscript is interesting and important. However, the analysis is drowning in all the other information and would deserve better.

We agree with the importance of providing complementary qualitative data within our manuscript, and believe it offers a rich perspective of school staff that have been required to adapt their teaching practice and adhere to a range of school-based mitigation measures. Qualitative research can offer a meaningful contribution to shaping policy and practice. In the case of this study, we explored staff perspectives of the impacts of face coverings to further explain the quantitative findings.

As this is a secondary outcome of the manuscript we were conscious to ensure that the qualitative results were discussed succinctly and clearly. We also felt it was important to include one example verbatim quote to represent each theme, with additional verbatim quotes available for the reader in S5 Appendix. Whilst it would be possible to dedicate an entire qualitative paper to these findings, we feel they complement the quantitative data and offer potential mechanisms to explain these findings, as outlined in the discussion.

Furthermore, the immediacy of the COVID-19 pandemic requires rapid research to provide evidence informing emerging policy and practice, which we believe we should include all findings in this manuscript to ensure research findings are delivered in a timely way so that they can be used to inform practice in schools.

Introduction

Page 5, line 102-104

Secondary qualitative data exploring the impacts of wearing face coverings are also presented to complement quantitative findings.

• Self-reported cold symptoms in the previous seven days is included as a separate outcome (proxy for infection risk). However, it is not clear how this should be interpreted and what it adds to the research question.

Thank you for raising this point. We agree that additional clarity is required to explain the inclusion of this outcome and have addressed this within the manuscript as listed below. We have used self-reported cold symptoms in the previous seven days as a proxy for infection (either COVID-19 or general infection). Evidence from the UK ZOE COVID study shows the crossover of symptoms between the common cold and COVID-19, particularly following two vaccine doses. It is also possible that Covid cases in the school (either pupil or staff) were not detected, including asymptomatic transmission or not being tested. Therefore, reporting cold symptoms indicates that transmission of either COVID-19 or other general infections is occurring. In the case of the school setting, this could suggest that some school-based mitigation measures may not be effective in preventing transmission, or are not being adhered to. In addition, if staff reported general cold or other viral infection symptoms, it is possible that this was transmitted within the school environment and would suggest they could also be exposed to COVID-19 or be transmitting an infection onwards to others.

We have addressed this as outlined below, including the addition of a reference to support the crossover between COVID and cold symptoms.

Quantitative analysis

Page 8, line 180-183

Secondary binary outcomes investigated at an individual level captured by the online survey included self-reported cold symptoms in the previous seven days as a proxy of infection risk as evidence suggests a crossover of symptoms between COVID-19 and the common cold [25], and moderate/severe anxiety (GAD-7) and moderate/severe depression (PHQ-9).

Discussion

Page 26, line 523-525

A crossover between COVID-19 and common cold symptoms has been established [25], and the current study also found an association between direct physical contacts and self-reported cold symptoms. As this study suggests variation of school-based mitigation measures between and within-schools, encouraging individual behaviours of school staff such as reducing direct contacts may be of benefit in reducing transmission of COVID-19 or general infection in the school setting.

• In the discussion they find that staff wearing face-masks had higher odds of reporting cold symptoms in the previous seven days. It is not clear whether they were more likely to use masks due to cold-symptoms, rather than vice-versa.

Thank you for raising this important point. Our study assumes that reported measures including the use of face coverings were in effect for the duration of the study period. We agree that reverse causality may be occurring in some instances. Whilst we acknowledged this within the strengths and limitations section (below), we have addressed this by also outlining this within the discussion:

Discussion

Page 27, line 551-555

In addition, this observational study assumes reported mitigation measures were in effect for the duration of the study period. It is possible that reverse causality occurred, that is a school staff practitioner may have chosen to wear a face covering following the onset of common cold symptoms.

Strengths and limitations

Page 30, line 615-618

As with all observational studies, unmeasured confounders and reverse causality may influence findings, e.g., face covering usage may increase due to a previous COVID-19 case in the school, higher community prevalence and individual behaviours. Thus, face covering use and future COVID-19 cases may be linked by an unmeasured confounder.

• The authors suggests that asymptomatic transmission from children could be an explanation. However, asymptomatic transmission from children in school settings is not very common.

Thank you for drawing our attention to this. We have removed the sentence that suggests asymptomatic transmission could be occurring as it is not possible to ascertain this from the current study.

Submitted filename: Response to reviewers.docx

Click here for additional data file.

26 Jan 2022

28 Jan 2022

Reviewer #1:

The questions was phrased with multiple alternatives (see below) whereas the analyses are done separately as dichotomous exposures. It is not clear whether the research question is relative or related to the exposure as whole. I would have expected that Keeping 2 metres from STAFF was treated as one variable (exposure as whole), and that this variable should be tested for significance before interpreting the significance of the subvariables (rarely, some of the time, most of the time, always) I think authors need to clarify this.

This was not related to ordinal data, but to the validity of the analysis. Before doing pairwise comparisons of separate values within a variable (i.e. testing “sometimes” versus “never/rarely”) the variable “keeping 2 meters distance” should be tested for significance. The p-value for the whole variable can be calculated in two ways:

1. A likelihood ratio test comparing a model containing the variable vs a model not containing the variable

2. A wald test that tests the joint significance of beta1 = beta2 = 0 (where beta1 is the coefficient for sometimes and beta2 is the coefficient for most of the time/always)

I believe pairwise comparisons (i.e. testing “sometimes” versus “never/rarely”) should be limited to significant variables to make sense, and the the p-value for the whole variable should be presented in the table alongside the pairwise comparisons.

I have no further comments

On my behalf of the co-authors, we would like to thank you for taking the time to read our revised manuscript. We are glad that you are satisfied that we have addressed the useful comments provided and that this has improved the flow and readability of the paper.

Thank you for drawing further attention and providing additional clarity regarding your previous comment about the multivariable analysis. Further to the changes in revision one, we have addressed this point as you have suggested by performing a likelihood ratio test for the variable keep two metres as a whole variable i.e. survey response categories. We have assessed the goodness of fit of two alternative models including and excluding this whole variable. The output from this test shows a significantly improved difference with the inclusion of the keep two metres from staff variable (p=0.03), a variable that we found to be associated with increased likelihood of COVID-19 positive test in the school. For keep two metres from pupils, the output presents a likelihood ratio of p=0.1 (borderline significant at the 10% level), though we did not report on this variable as we found no association with outcomes in multivariable models. To ease interpretation of Odds Ratios, we created ordinal categories relating to level of exposure.

We have addressed this within the manuscript, including stating within the methods, results and in the Table 3 footnote. Please find below further information of these revisions:

Quantitative analysis, page 8-9, line 191-198.

Items with multiple category responses or continuous numerical values were assigned ordinal categories to ease interpretation. For example survey response categories for keep two metres from pupils/staff included i) never, ii) rarely, iii) some of the time, iv) most of the time, v) always, with combined ordinal categories for analyses of i) never/rarely, ii) some of the time, iii) most of the time/always. For these variables, likelihood-ratio tests of variables as whole were performed to assess goodness of fit between models including and excluding variables for the primary outcome.

Table 3, page 16, line 282-287.

Univariable (model 1) and multivariable (model 2) logistic regression models of self-reported school-based mitigation measures (survey) and school-level probability of any positive COVID-19 case in school (SAIL). OR: Odds Ratio; 95% CI: 95% confidence intervals; p<0.05**, p<0.1*; adjusted for school size, proportion of pupils eligible for free school meals, clustered by school. Model 2 likelihood-ratio test keep two metres from pupils (p=0.1) and staff (p=0.03). See S3 Appendix for variable codebook.

Finally, we have replaced 2-metre to two metre throughout the manuscript, and have updated table layouts.

Submitted filename: Response to reviewers.docx

Click here for additional data file.

2 Feb 2022

COVID-19 mitigation measures in primary schools and association with infection and school staff wellbeing: an observational survey linked with routine data in Wales, UK

PONE-D-21-27732R2

Dear Dr. Marchant,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Amitava Mukherjee, ME, Ph.D.

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

18 Feb 2022

PONE-D-21-27732R2

COVID-19 mitigation measures in primary schools and association with infection and school staff wellbeing: an observational survey linked with routine data in Wales, UK

Dear Dr. Marchant:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Professor Dr. Amitava Mukherjee

Academic Editor

PLOS ONE