Implementation of COVID‐19 Preventive Measures and Staff Well‐Being in a Sample of English Schools 2020‐2021

ABSTRACT BACKGROUND We examined fidelity and feasibility of implementation of COVID‐19 preventive measures in schools, and explored associations between adherence to these measures and staff well‐being, to inform policy on sustainable implementation and staff wellbeing. METHODS Surveys were conducted across 128 schools in England with 107 headteachers and 2698 staff‐members with reference to autumn term 2020, examining school‐level implementation of preventive measures, adherence, and teacher burnout (response rates for headteacher and staff surveys were 84% and 59%, respectively). RESULTS The median number of measures implemented in primary and secondary schools was 33 (range 23‐41), and 32 (range 22‐40), respectively; most measures presented challenges. No differences were found regarding number of measures implemented by school‐level socio‐economic disadvantage. High adherence was reported for staff wearing face‐coverings, staff regularly washing their hands, (secondary only) desks facing forwards, and (primary only) increased cleaning of surfaces and student hand‐washing. Adherence to most measures was reported as higher in primary than secondary schools. Over half of school leaders and 42% (517/1234) of other teaching staff suffered from high emotional exhaustion. Higher teacher‐reported school‐wide adherence with measures was consistently associated with lower burnout for leaders and other teaching staff. CONCLUSIONS Findings indicate a tremendous effort in implementing preventive measures and an urgent need to support investments in improving teacher wellbeing.

T he COVID-19 pandemic has resulted in over 250 million confirmed COVID-19 cases and over 5 million deaths worldwide as of November 2021. 1 Evidence suggests that children are less vulnerable than adults to morbidity and mortality. [2][3][4][5][6] However, school closures have been widely used as a control measure, resulting in the largest disruption to schooling in recent history. 7 School closures have adversely impacted students' learning, physical health, mental health and wellbeing, and societal functioning. [8][9][10][11][12][13][14][15] Mental health has been a concern over the course of the pandemic, including the mental health of children, adolescents, and teachers. 22,23 Informed by the general literature on teacher mental health and burnout, it is likely that increased workloads and stresses associated with enacting complex preventive measures with limited guidance and support may predispose teachers to burnout. [24][25][26] Burnout is defined as a psychological syndrome involving emotional exhaustion, depersonalisation and reduced personal accomplishment resulting from chronic interpersonal job stressors. 27 Burnout could also be exacerbated by teachers' concerns about workplace safety, where measures are perceived as inadequately implemented or insufficiently adhered to. Furthermore, burnout and low morale may in turn be associated with lower fidelity of implementation of preventive measures given that staff commitment is commonly a key influence on the implementation of school-based health interventions. 28 Bi-directional causal relations between staff burnout and poor implementation of preventive measures may therefore trigger feedback cycles which hamper prevention efforts in schools. However, to the best of our knowledge, research to date has not explored whether there is any association between implementation of COVID-19 preventive measures and staff burnout.
Under guidance from England's Department for Education (DfE), primary and secondary schools reopened to all students in September 2020 with multiple preventive measures, including the requirement that individuals who are unwell stay home, use of facecoverings, hand and respiratory hygiene, enhanced cleaning, and physical distancing. 29 Face-coverings in corridors or communal areas were recommended for secondary but not primary schools. Primary school guidance emphasized keeping students in separated small groups (''bubbles''). Secondary school guidance focused on distancing between individuals within larger bubbles. Schools were further required to engage with test and trace systems, exclude infected students and contain outbreaks with advice from local health protection teams. From January 5, 2021, schools were again closed to all except vulnerable or key-workers' children. 30 On March 7, 2021 schools re-opened to all students with revised measures to implement including widespread asymptomatic testing using lateral flow devices and use of face-coverings in classrooms for secondary-school students. 31 Focusing on experiences in the autumn term of 2020 (September-December 2020), this paper examines the fidelity and feasibility of implementation of COVID-19 preventive measures in schools, and explores associations between adherence to these measures and staff well-being, to inform policy on sustainable implementation of such measures and staff mental well-being.

METHODS
The COVID-19 Schools Infection Survey (SIS) was an observational cohort study conducted in state primary and secondary schools across 15 selected local authorities (LAs) in England during the school year 2020/2021. Multi-stage stratified sampling was used to select schools, oversampling LAs where SARS-CoV-2 infection prevalence was higher (top 20% of LAs ranked by population rate of confirmed infection) in September 2020, and oversampling secondary schools within selected LAs. The study aimed to assess the role of schools in SARS-CoV-2 transmission and school implementation of preventive measures. SIS involved: repeat SARS-CoV-2 antigen and antibody testing; web-based questionnaire surveys of staff, parents, and older students; and semi-structured telephone interviews with a sub-set of schools. Six rounds of data collection were planned over the 2020/2021 school year, with rounds 1 and 2 taking place in November and December 2020, respectively. Further details on the overall study are reported elsewhere. 32 We considered data from the autumn 2020 school term (rounds 1 and 2) for this analysis where 128 schools provided headteacher or staff questionnaire data.

Participants and Procedures
As schools enrolled in the study, headteachers or their nominated representatives (usually other school senior leaders) were sent a school-level headteacher questionnaire and each school was asked to complete one. All staff in participating schools were eligible to enroll in the SIS study. All enrolled staff were initially sent a demographic questionnaire followed by a more extensive questionnaire in mid-December 2020. Surveys were administered online. Multiple reminder emails were sent and follow-up telephone calls were made to participants. A telephone helpline was made available to participants to answer queries.
In this paper, we report on school implementation of preventive measures in the autumn term (September-December) of 2020, when schools were open to all students. This cross-sectional analysis draws on data from headteacher and staff questionnaires with reference to autumn term 2020, issued to participants in round 1 and 2, completed November 2020-March 7, 2021. Questionnaires remained open past the end of the autumn term to allow time for more participants to complete them.

Measures
Fidelity is defined as the degree to which a program is implemented as intended 33 and in this context refers to the extent to which the suite of preventive measures recommended by the government were implemented in schools. Adherence refers to the extent to which students or teachers implement measures consistent with advice from the school. Feasibility considers the level of challenge associated with implementation and adherence. These measures of implementation were assessed using indicators developed specifically for this project, informed by: previous research conducted in summer 2020; 19 public documents providing government guidelines on school preventive measures; and discussion with experts.
To understand fidelity and feasibility of implementation, the headteacher questionnaire included a list of potential preventive measures, 46 of which were applicable to primary schools and 47 to secondary schools, and participants were asked whether each of these was implemented at their school and how challenging implementation was. For example, headteachers were asked to indicate whether the measure ''All desks face forward'' was implemented and if so, whether implementation involved major challenges, some challenges or was easy to implement. The headteacher questionnaire also included questions on bubble sizes, procedures for internal risk assessments and review, and school and headteacher characteristics.
The staff questionnaire included questions on adherence with measures and burnout. Teaching staff were asked about their perception of school-wide adherence with a subset of 14 measures relating to hygiene and sanitation, social distancing, and facecovering use among staff and students. They were asked how well, in their experience, each preventive measure was being followed (always, sometimes, rarely, never, not applicable, do not know). A ''perceived adherence'' score was calculated for each respondent, summing their responses across 10 of these measures classed as ''strongly recommended'' or ''recommended.'' For each measure, ''always'' was scored as 2, ''sometimes'' as 1, and ''never,'' ''rarely,'' ''not applicable,'' or ''do not know'' as 0. The potential range of scores was thus 0-20.
Teacher burnout was assessed using the 22-item Maslach Burnout Inventory (MBI) Educators Survey. MBI is a validated measure designed to assess 3 dimensions of burnout using subscales focused on emotional exhaustion (9 items), depersonalisation (5 items), and personal accomplishment (8 items). 27,34 Emotional exhaustion is often considered a key aspect of burnout, while depersonalisation refers to negative, cynical attitudes and feelings toward students and colleagues. 27 Reduced personal accomplishment refers to a tendency to evaluate oneself negatively and feelings of incompetence and poor achievement in one's work. For each statement of job-related feelings, participants were asked how often they felt that way, with responses scaled from ''never'' (0) to ''every day'' (6). Subscale scores were calculated by summing scores across the relevant items, with higher scores indicating greater burnout for emotional exhaustion and depersonalisation, and lower scores indicating greater burnout in relation to personal accomplishment. In addition to the total scores, a binary measure of high burnout was also created for each subscale, indicated by validated scores of 27 or over for emotional exhaustion, 14 or over for depersonalisation, and 30 or below for personal accomplishment. 35

Data Analysis
Data were stored on secure software and exported into the Office for National Statistics (ONS) Data Access Platform. De-identified datasets were accessed by accredited members of the study team via ONS secure servers. Descriptive tables were generated using Stata MP 16.1 (StataCorp LLC) within ONS's Secure Research Service platform.
We present the proportion of schools implementing each COVID-19 preventive measure, as reported in the headteacher questionnaire ( Figure 1). Measures were grouped within 8 domains aligning with government guidance. 29 We also calculate the median number of measures implemented by schools (overall and by level of DfE recommendation), alongside interquartile range (IQR) and range. We further explored whether the number of measures implemented differed according to school-level socio-economic disadvantage assessed via student eligibility for free school meals (FSM): up to 20% eligible = low; 21-30% eligible = medium; >30% eligible = high.
Staff-reported school-wide adherence to each preventive measure was assessed among teaching staff working in schools in which headteachers reported that the measure in question was implemented. We present the overall frequency with which staff report that each measure was ''always'' or ''sometimes'' followed and the school-level intra-cluster correlation coefficient (ICC) for the ''always'' response.
The analysis of teacher burnout was performed separately for those with leadership positions (ie, headteachers, and senior and middle leaders) and other teaching staff. We present mean scores and standard deviations for each subscale of the MBI, and the percentage of teachers with ''high'' burnout for each subscale. School-level ICCs for the mean scores were estimated using an intercept-only multi-level linear regression model.
To explore the association between teacher burnout and teacher-reported school-wide adherence to COVID-19 preventive measures, we considered each MBI subscale separately. We conducted multi-level mixed-effects linear regression with random intercept for school and the continuous score for each subscale as the dependent variable. In addition to teacher-reported school-wide adherence to COVID-19 prevention measures (calculated among all teaching staff irrespective of whether headteacher reported measure as implemented), independent variables considered in the unadjusted analysis were school type (primary/secondary), percentage of students eligible for FSM (low/medium/high) and government inspection rating (inadequate or requires improvement/good/outstanding). Those variables found to be associated with any of the MBI subscales in these models (p < .1) were then included, alongside teacher's age and gender, in adjusted models for all subscales.

RESULTS
In total, 128 schools (45 primary and 83 secondary) participated and provided questionnaire data in rounds 1 and 2 of the SIS study. Of these, 107 (84%) schools (40 primary and 67 secondary schools) completed the headteacher questionnaire by March 7, 2021. Among 4566 teaching staff across all 128 schools enrolled in rounds 1 and 2 of the study, 59% (640 primary and 2058 secondary teaching staff) completed the questionnaire by March 7, 2021 and were included in the analysis (Table 1).

School and Participant Characteristics
Secondary schools with a median of 936 students were larger than primary schools that had a median of 297 students (Table 1). Twenty-nine percent (13/45) of primary and 23% (19/83) of secondary schools in this analysis were in the high FSM band (>30% eligibility), while over half were in the low FSM band. Thirtyseven (82%) primary and 48 (58%) secondary schools were rated by government inspectors as ''good'' or ''outstanding.'' Most headteacher questionnaires were completed by headteachers themselves (primary: 83%, 33/40; secondary: 55%, 37/67). The median number of teaching staff completing the questionnaire per school was 13 (IQR: 10-17) for primary and 24 (IQR: 15-33) for secondary schools. Most participants in both the headteacher and staff questionnaires were women.

School-Level Implementation of Preventive Measures
According to the headteacher questionnaire, most primary schools (75%, 30/40) kept students in bubbles of normal class sizes, while bubbles in most secondary schools (72%, 48/67) consisted of yeargroups ( Table 2). In about half of both primary and secondary schools, students were seated less than 0.5 m apart in the classroom. Only 1 primary and 3 secondary schools reported that students were seated at least 2 m apart.
The frequencies with which primary and secondary schools reported implementing specific COVID-19 preventive measures (headteacher questionnaire) are  9 Toilets restricted to parƟcular bubbles 9 9 Staggered break Ɵmes X X Stop team sports 0 0 S t o p i n -p e r s o n s t a ff m e e Ɵ n g s 9 9 Staggered start/end Ɵmes 0 0 One-way systems in school corridors X X Not providing aŌer-school acƟviƟes 9 9 Staff can work from home if able to do job X X Teaching assistant works with single class throughout week 9 Students in same bubble on school transport (secondary only) N/A X X Teacher works with single class throughout week 9 Same We present the proportion of schools implementing each COVID-19 preventive measure, as reported in the headteacher questionnaire. Measures were grouped within 8 domains aligning with government guidance: minimizing contact with symptomatic individuals; hand and respiratory hygiene; enhanced cleaning/minimizing potential for fomite transmission; minimizing mixing between bubbles; maintaining distance between individuals; face-coverings; ventilation; and other. The research team further categorized measures, separately for primary and secondary schools, as ''strongly recommended,'' ''recommended,'' ''not mentioned,'' or ''not recommended'' based on their reading of government guidance.  presented in Figure 1. Measures to minimize contact with symptomatic individuals and measures relating to hand/respiratory hygiene were implemented by almost all schools. Enhanced cleaning was also adopted by almost all, though primary schools were more likely to implement additional measures to limit fomite transmission. The most commonly adopted measures to minimize contacts between individuals included keeping students in bubbles and stopping large gatherings. Overall, implementation of measures to prevent mixing between bubbles was slightly lower in secondary than primary schools. However, measures to maintain distance between individuals were more widely implemented in secondary than primary schools (eg, desks faced forwards, teachers stayed at the front of the class, and staff maintained at least 2 m distance from students). Primary schools were unlikely to implement distancing measures (these measures were not emphasized in the guidance for primary schools). Under a quarter of secondary schools reported implementing distancing between students within the same bubble.
All secondary-school headteachers reported that staff and students wore face-coverings in corridors and other communal areas. Forty-seven percent of primary and 33% of secondary schools reported that teachers were required to use face-coverings in classrooms. Policy requiring use of face-coverings by students in the classroom was much less frequent, reported by under one-fifth of secondary schools and 1 primary school.
All primary and most secondary-schools opened windows to ventilate classrooms. Mechanical ventilation systems were present in very few primary and 36% of secondary schools.
Most headteachers reported conducting regular internal reviews of COVID-related policies and measures within their schools: Of the 40 primary and 67 secondary schools, 38% of primary and 30% of secondary schools did these reviews once to twice per month, while 33% of primary and 31% of secondary schools did them once or twice per week.

Challenges Associated With Implementing Preventive Measures
Headteachers reported that most measures presented challenges to schools (Supporting Information). Regular handwashing for staff and installing hand sanitisers at school entrances were considered easy to implement. Other measures considered easy to implement by a majority of headteachers were: (in primary) keeping students in the same classroom all day, stopping large gatherings and team sports, and not providing after-school activities; (in secondary) staff wearing face-coverings in corridors and staff rooms. Secondary schools tended to report more challenges than primary schools implementing measures to minimize mixing between bubbles and potential for contamination. For example, keeping students in the same classroom all day was considered majorly challenging in 41% (12/29) of secondary, and easy to implement in 70% (26/37) of primary schools, reporting implementing this measure.

School-Wide Adherence With Preventive Measures
Levels of staff-reported school-wide adherence with prevention measures varied by measure and between primary and secondary schools (Table 3). High levels Table 3. Staff-reported school-wide adherence to each preventive measure was assessed among teaching staff working in schools in which headteachers reported that the measure in question was implemented. We present the overall frequency with which staff report that each measure was ''always'' or ''sometimes'' followed and the school-level ICC for the ''always'' response.

• Journal of School Health • 2022
of adherence were reported for measures such as staff wearing face-coverings in corridors and communal areas, staff regularly washing their hands, secondary schools having desks facing forwards, and increased cleaning of surfaces and student handwashing in primary schools. Many measures were frequently reported as ''sometimes'' followed. These included staff wearing face-coverings in classrooms and staffrooms, staff distancing from students and other adults, student catching coughs and sneezes in a tissue, and students not sharing equipment in class or carrying equipment between school and home.
Secondary-school staff reported higher adherence than primary schools for some measures, such as staff distancing from students, desks facing forwards and students not sharing equipment. However, for most measures, adherence was reported as worse in secondary than primary schools. More staff in secondary schools also answered ''do not know'' regarding adherence with measures relating to staff and student behaviors. Across primary and secondary schools, ICCs indicated that staff perception of adherence to measures related to school environments or mandated observable behaviors (such as face-covering use and increased cleaning of surfaces) tended to exhibit more within-school clustering than measures related to individual behavior (such as maintaining distancing, students catching coughs/sneezes and students carrying equipment between school and home).

Teacher Burnout and Well-Being
Over half those in leadership positions and 42% (517/1234) of other teaching staff suffered high burnout regarding emotional exhaustion (Table 4a). Six percent (40/629) of leaders and 15% (n = 180/1226) of other teaching staff reported high burnout regarding personal accomplishment. Few (3%) reported high burnout regarding depersonalisation. Mean scores indicate higher burnout among leaders than other teachers regarding emotional exhaustion and depersonalisation, and lower burnout among leaders regarding personal accomplishment. Lower teacher-reported school-wide adherence with measures was consistently associated across all 3 domains with higher burnout among leaders and other teaching staff (Table 4b). School-type was also associated with teacher burnout, with secondaryschool leaders and other teachers more likely than those in primary to experience greater emotional exhaustion, greater depersonalisation and lesser personal accomplishment. Teachers in more deprived (high-FSM) schools were less likely to experience emotional exhaustion and depersonalisation than teachers at schools with low FSM.

DISCUSSION
Despite concerns about the feasibility of implementing preventive measures in schools 36,37 and notwithstanding challenges reported, most schools in our study implemented many preventive measures. Fidelity of implementation varied across schools, likely due to challenges interpreting guidance received, physical constraints of school environments and resources. 19,22,38 Though not all secondary schools in our study implemented all strongly recommended measures, many schools did more than advised in national guidance at the time. For example, as reported in other studies, 39 some primary schools implemented use of face-coverings.
Secondary-school staff were more likely to report poorer adherence to many measures related to staff and student behavior. This is likely indicative of the challenges of implementing measures in large institutions. A study in Germany postulated more flexible adaptation in primary schools and a more dismissive attitude among older students toward regulations on social behavior as possible reasons for better implementation of preventive measures in primary than secondary schools. 39 Furthermore, some guidance for secondary schools may have been unrealistic, making poor adherence likely, as implementation theory would predict. 40 For example, less than a third of secondary schools reported implementing distancing for students within their bubbles, despite this being strongly recommended in the guidance. On the other hand, the older age of students in secondary schools may facilitate adherence to other distancing measures such as staff distancing from students and desks facing forwards.
We also found that secondary-school teachers were more likely to experience burnout and teachers working in schools with more economically disadvantaged students were less likely to suffer from emotional exhaustion and depersonalisation. Other research looking into why teachers stay in challenging schools found that wanting to make a difference, a strong sense of social responsibility and rewarding relationships with students, even where challenging, were key factors, 41,42 and it is possible that fulfillment of these ideals may act as protective factors against burnout among teachers in these schools. However, this analysis is focused on the association between reported adherence and burnout, and is not a comprehensive assessment of factors influencing teacher burnout; these findings, therefore, must be interpreted accordingly.
The high levels of burnout among teachers, and especially school leaders, in this study are consistent with anecdotal reports of the challenges teachers have faced and the negative impacts on their well-being, 22 although we have no comparison to levels of burnout Table 4. 19 (3) 40 (6) 517 (42) 36 (3) 180 ( before the pandemic. We found a consistent association between burnout and lower teacher-perceived school-wide adherence with COVID-19 measures. Though the cross-sectional nature of this analysis precludes inference about causality, it is plausible both that school environments in which preventive measures are poorly adhered to may contribute to teacher burnout and that teacher burnout could undermine implementation of COVID-19 measures. Finally, as adherence with preventive measures was a subjective measure based on teachers' perceptions, it is possible that teachers with higher burnout were more likely to perceive poor adherence with preventive measures.

Limitations
Of 289 schools initially approached for the SIS study, 128 schools participated and provided questionnaire data in rounds 1 and 2. Although response rates to the headteacher questionnaire were high, with completion by 107 of the 128 participating schools, it is possible schools that did not sign-up for the study or where headteachers did not respond differed in important ways from those in which headteacher questionnaires were completed. They may, for example, have had less rigorous protocols around COVID-19 prevention than those included in this analysis. Response rates were also lower than expected for the staff questionnaire, again leading to a potentially non-representative sample of teaching staff and biased estimates for adherence to measures and teacher burnout. Participants were allowed to complete the questionnaires after the end of the autumn term to allow sufficient time to complete the questionnaire as they were made available to participants only toward the end of the autumn term. We do not believe asking participants to reflect on the previous term is likely to have been too problematic given that schools were closed to most students during the following term. Oversampling of LAs where SARS-CoV-2 infection prevalence was high in Sep 2020 was done to facilitate transmission studies; this may limit the ability to generalize findings for England. However, our sample of schools from 15 different LAs and varying Covid-19 prevalence over the year makes it unlikely that implementation practices in our schools differed widely from others in England. Finally, the cross-sectional nature of this analysis does not allow us to fully explore whether poor adherence with preventive measures resulted in greater teacher burnout or vice versa.

IMPLICATIONS FOR SCHOOL HEALTH POLICY AND PRACTICE
The large number of new policies and preventive measures that schools had to innovate and implement, the challenge they have presented, and the task of regularly reviewing policies and ensuring adherence to them have increased the workload of school leaders and teachers. Findings suggest the need for clear technical guidance based on consultation with schools. 19,22 However, the finding that a similar number of preventive measures were implemented at school-level irrespective of socio-economic advantage of schools, reassuringly suggests an equitable provision of guidance.
Our findings support the case for investing in teacher mental health and well-being. This investment is essential in itself to decrease teacher burnout and is likely also key to ensuring sustainability of effective implementation of preventive measures. Furthermore, our findings may also suggest that supporting better implementation of and adherence to preventive measures may decrease teacher burnout. The wider teacher burnout literature, which suggests that classroom environments and discipline, lack of clear rules, and ineffective implementation and enforcement of rules may impact burnout, supports these findings. 43,44 Our findings also suggest the importance of ensuring school working environments feel safe for staff. Other research recommends supporting teachers to feel autonomous, competent and connected with colleagues, students and their families for their well-being 37 and investment in training programs for stress reduction. [45][46][47] Measures that may be worth considering for ongoing implementation, if needed, as schools transition to normal functioning are those that were relatively less challenging to implement and secured higher adherence such as, face-coverings for staff, frequent handwashing, enhanced cleaning, and (secondary schools) forward-facing desks.
Further research is required to explore fidelity and feasibility of measures introduced after autumn term 2020 such as regular asymptomatic testing, ongoing maintenance of those COVID-19 preventive measures that continue to be required of schools, as well as to assess the mental health of staff and students as schools gradually return to normal functioning.

Human Subjects Approval Statement
This study received approval from research ethics committees at the UK Health Security Agency (ref: NR0237) and the London School of Hygiene and Tropical Medicine (ref: 22657). Participants provided consent on the online portal before beginning the survey questionnaire.
in this work does not imply the endorsement of the ONS in relation to the interpretation or analysis of the statistical data. This work uses research datasets which may not exactly reproduce National Statistics aggregates.