Phase 1 is a repeated panel survey of prison residents and staff in a sample of 28 prisons across England. The survey will be repeated to enable the calculation of incidence rates of SARS-CoV-2 infection in the study prisons as well as the proportion of positive tests of SARS-CoV-2 infection among residents and staff within the selected prisons.

Phase 2 comprises an enhanced outbreak investigation in 3 prisons in which a SARS-CoV-2 outbreak is declared. An outbreak is defined as two or more cases of SARS-CoV-2 that are epidemiologically linked [16].

This study will be conducted in prisons across England. The sample of 28 prisons will be selected to be as representative of the closed prison estate as possible with regard to important features. These features are prison function and security category, geographical area, whether or not a COVID-19 outbreak was previously reported in that prison, staff numbers, resident population, and the proportion of the population classified as “older residents.” Operational factors, including capacity to undertake whole prison testing, will also be given consideration. In phase 1, all prisons will be selected. In phase 2, 3 prisons from the sample of 28 will be included when an outbreak is declared.

When rounds 1 and 2 of antigen testing have finished, if an outbreak of COVID-19 is declared by the local health protection team in any prison in England, further antigen testing of all prisoners and staff will take place as soon as is feasible after the outbreak has been declared. Testing will be repeated 7 days later and again at 28 days after the first test. This periodicity will give detailed information on outbreak dynamics. Staff will also be tested for antibodies to SARS-CoV-2 12 weeks after the start of the outbreak.

In both phases, the governor of each study prison will appoint a member of custodial staff to be the single point of contact (SPOC) to support the research project within the prison. The SPOC will be there to support staff and residents, and answer any questions or concerns they may have about the process.

For the primary objective of estimating the proportion of positive tests of SARS-CoV-2 infection within the selected prisons, the precision (margin of error) that various sample sizes (assuming 80% response) provide around various estimates of infection rates is illustrated in Figure 1.

The sample sizes (N) represent a range of individual prison population sizes from the smallest participating prison (approximately 100 prisoners) to the largest (approximately 1500 prisoners). For a given proportion of positive tests and sample size (N), the expected margin of error corresponds to the expected half-width of the 95% CI associated with the point estimate of the proportion of positive tests obtained using an exact binomial test. As illustrated, the precision of the SARS-CoV-2 positivity estimates in participating prisons is expected to be smaller for larger prisons. The 2 rounds allow for the secondary objective, the incidence of SARS-CoV-2, to be estimated; however, there will be limited power to detect small incidence rates.

The team comprises individuals with expertise in epidemiology and statistics, infectious diseases, public health, the prison system, and the conduct of research in prisons.

For all testing rounds in each participating prison (whether phase 1 or 2), all residents and staff present on the first day of data collection in that prison (the census date) are eligible to be included in the study (Textbox 1).

All data collection will be done within the prison. Private contractors will undertake the testing. In each prison, there will be a team of 6 to 10 testing staff fully trained in testing procedures by their employer and trained in the study procedures and ethical practice by the University of Southampton. Each tester will be escorted around the prison by a member of the custodial staff, who will open the cell door to each participant. At the first contact, the tester will take written consent for the whole study from the participant and collect the completed paper questionnaire. At subsequent testing rounds, verbal consent will be obtained. In phase 1, participants will complete a self-swab of their nose and throat as instructed and observed at 2 m by the tester. In phase 2, participants will self-swab twice. They will provide a nasal swab for testing using a lateral flow device and a nasopharyngeal swab for polymerase chain reaction testing. A member of the health care team will swab any residents with a physical disability that makes it difficult for them to self-swab. For later rounds of testing in an outbreak situation, the testing process will be supervised as previously, by externally contracted staff, or by the prison health care team, dependent on local circumstances.

A private provider, commissioned by HMPPS to provide occupational health services to prison staff, will take a swab of their nose and throat or, alternatively, staff will swab their own nose and throat if the occupational health provider is unable to do this. For later rounds of testing in an outbreak situation, staff will self-administer the test as part of HMPPS’s ongoing weekly testing program. They will also self-administer a second nasal swab that will be used by the testing team to test using a lateral flow device. They will receive an information sheet and a consent form to enable their data to be shared with the study team.

The nose and throat swabs will be couriered to an accredited laboratory where swabs will be tested for the presence of SARS-CoV-2 using reverse transcription polymerase chain reaction (RT-PCR). Before residual material is discarded, if the virus is detected, the material will be couriered securely to a COVID-19 Genomics UK Consortium study laboratory for genomic sequencing of the virus in accordance with the COVID-19 Genomics UK Consortium study protocol [16]. The nasal swabs will be tested on site using Innova lateral flow devices; these provide a result within 30 minutes.

Staff who participate in round 1 or 2 of antigen testing will be invited to participate in antibody testing. Following the second round of testing, the SPOC will issue participating staff with a home testing kit for antibody testing.

Participating residents will complete a brief paper questionnaire that asks about medical risk factors for severe COVID-19, current COVID-19 symptoms, and symptoms in the past 14 days in line with other COVID studies in England. In the second round, the short version of the Warwick Edinburgh Mental Well-Being Scale (WEMWBS), a well-validated measure of mental well-being [17], will be included in the questionnaire. Further information about residents’ age, ethnicity, location, and movements within the prison will be obtained by the SPOC (or appointed administrator) from the secure p-NOMIS. In an outbreak situation, questionnaires will not be completed, but data will be obtained on residents’ ethnicity, age, and sex.

Each time they are tested, each participating staff member will complete a short online questionnaire providing details of medical risk factors for severe COVID-19, current COVID-19 symptoms, and symptoms in the past 14 days. In the second round, the WEMWBS will be included in the staff questionnaire too. In an outbreak situation, questionnaires will not be completed, but data will be obtained on the staff members’ ethnicity, age, and sex.

The SPOC will gather information about the prison for the study team, including geographical layout, age, ventilation systems, sanitation, numbers of residents and staff, and what infection control measures have been implemented. The study team will obtain data on the prevalence of SARS-CoV-2 in the community from the ongoing Office of National Statistics Community Survey [18].

Data will be stored and handled at the University of Southampton and PHE. A pseudonymized database will be created using unique participant IDs.

Summary statistics (N, proportion, mean, SD, and IQR) will be obtained to summarize the prison and participant characteristics. Tables will display the number of participants in each of the 2 surveys, the number of new participants (eg, new residents), and the number of individuals withdrawn/lost to follow-up from the previous survey (eg, withdrawn consent or left the prison).

For the primary objective, simple proportions of individuals testing positive for SARS-CoV-2 following RT-PCR of samples (based on nose and throat swabs) will be calculated for each prison population of staff and prisoners separately. The proportion of positive individuals with symptomatic infection, defined as “yes” to any of the listed symptoms in the questionnaire, will also be calculated for each prison.

Stratified proportions will be estimated to understand the variation in the prevalence of infection among different prison characteristics and different populations (eg, the strata of staff/resident population; prison type; participant age, gender, and ethnicity; current location in prison [prison population only]; and staff role [staff population only]).

Incidence rates within each prison will be calculated. Variation in incidence rates among different prison characteristics and different populations will be explored by estimating stratified incidence rates within the strata of staff/prisoner population; prison type; participant age, gender, and ethnicity; current location in prison (prisoner population only); and staff role (staff population only).

To examine the individual, institutional, system, and community factors associated with SARS-CoV-2 infection within prisons, generalized linear regression models will be fitted to adjust prevalence and incidence rates for relevant hypothesized risk factors. The list of individual-level risk factors includes participant age, gender, ethnicity, and existing comorbidities; and staff role and responsibilities (for staff analyses only). Institutional factors include in-cell sanitation (for resident analyses only), establishment age, ventilation type, use of compartmentalization measures, and regional location. All point estimates will be provided with 95% CIs.

The proportion of staff who test antibody positive at week 6 (overall and according to whether they are prisoner-facing or not) will be calculated for each prison.

The WEMWBS will be scored according to the user guide [19]. The mean scores from the short WEMWBS will be presented for residents and staff, and their associated demographic factors.

Imputation methods may be employed depending on the level of completeness for any specific analysis.

This is a time-pressured study that has high-level UK Government support. All research within prisons has currently been stopped because of the COVID-19 pandemic, and researchers are not allowed into prisons. This study has been allowed to go ahead using professional testing staff who will be wearing appropriate personal protective equipment (PPE). The University of Southampton team will be training (remotely) these staff members to inform them about the study, with a particular emphasis on understanding the principles of ethical research conduct and the specificities of this in a prison setting. The Southampton research team will act as an ongoing resource for advice about the conduct of the study for all testing staff.

The main study-specific ethical consideration is the inclusion of prisoners, a group regarded as vulnerable in international guidance on the ethical conduct of health research. In addition, some of these prisoners are 16 and 17 years old. There are a number of key considerations as presented below.

The study timeline is as follows: antigen testing in round 1 from July 20 to August 22, 2020; antigen testing in round 2 from September 1 to October 2, 2020; antigen testing in round 3 from February 11 to March 31, 2021; antibody testing from October 5, 2020, to March 31, 2021; analysis of data from August 1, 2020, to December 31, 2021; and dissemination of findings from April 1, 2021, to April 1, 2022.

There are approximately 12 million people imprisoned worldwide [20], and to the best of our knowledge, this is the first national study to examine the incidence and prevalence of SARS-CoV-2 in prisons in the world. COVID-19 in prisons is an important public health issue that has implications for not only those who live and work in prisons, but also the wider society. Prisons are high-risk settings for the transmission of infectious diseases, and this study will provide data important for practice, public health, and policy making across prisons in an entire country during the COVID-19 pandemic. There needs to be an in-depth understanding of the transmission of SARS-CoV-2 within prisons, from the community to prisons and from prisons to the community, in order to inform effective measures to tackle the virus within this unique setting.

With an increasing focus on the mental health implications of the pandemic, the findings from this study will quantify mental well-being in a sample of participants (both residents and staff) using a tool that has been widely validated in the community and in imprisoned people. This will provide important information to enable infection control measures to be appropriately targeted, balancing interventions to prevent infection with the needs of residents and staff to maintain physical and mental health. This unique combination of complimentary data on the epidemiology of SARS-CoV-2 within prisons and on mental well-being will be invaluable in helping adopt a more holistic approach to the implementation of infection control measures in prisons, ensuring that any measures adopted can, as far as is practicable, take into account the impact on the mental well-being of imprisoned people. This is particularly important because of the long-standing nature of this pandemic and can inform health security challenges in the future.

The findings from the study will be made widely available through open-access publications to enable the benefits of the knowledge gained to extend beyond national boundaries. Furthermore, the expertise gained by the research team from the conduct of this unique research in prisons will be shared with other countries using platforms such as the Worldwide Prison Health Research and Engagement Network.

The study does however present huge logistical challenges. Testing up to 10,000 prison staff and 20,000 prison residents at 2 different time points is a considerable undertaking. Researchers are currently not allowed into prisons, and therefore, testing teams, trained in antigen testing for SARS-CoV-2 and the use of PPE, must be deployed. They will need security clearance to enable them to go into prisons. They must be trained remotely in ethical research conduct to enable them to receive informed voluntary consent from participants. This is particularly important as prison residents are classified as “vulnerable” participants according to international ethical guidelines [21] and many challenges in the ethical conduct of prison research have been identified prior to the pandemic, which adds complexity to the processes [22].

The use of information technology in the study is limited because of the prison environment. Questionnaires for residents will be on paper; the use of electronic devices by resident participants is not permitted. This creates delays in receiving data in a form that can be readily analyzed and the potential for data losses through lost questionnaires, transcription errors, etc. Staff will complete an online questionnaire in the prison at the time of testing, but ensuring all staff have access to a computer presents logistical challenges in an environment where the use of information technology, even by staff, is greatly restricted.

The study was developed at great speed and in a fraction of the time that a study of this nature would ordinarily require. This resulted in a level of patient and public involvement (PPI) that was significantly lower than usual for the research team. Furthermore, during the protocol development phase, the research team was not directly working with prison staff or residents. HMPPS liaised directly with the key staff union, the Prison Officers Association, and the governor of each prison was delegated the task of working with resident groups. PPI representation on the independent study steering group was, however, highly engaged and a source of valuable insights.

The COVID-19 in Prison Study presents an important opportunity to gather data in a unique high-risk setting from the whole population (residents and staff alike). It will provide important epidemiological data to develop our understanding of the transmission dynamics of SARS-CoV-2 in prisons and provide preliminary data on the mental well-being of staff and residents. The study will thus provide broad and more holistic data to inform public health policy and practice in prisons during the COVID-19 pandemic.