Approved Projects

Data from the Office for National Statistics (ONS) estimate that as of 2nd January 2023, 2 million people in the UK were living with Long COVID. Within this cohort, 1.5 million people or 77% of those with Long COVID stated their symptoms adversely affected their day-to-day activities, and 380,000 or 19% reporting their ability to undertake day-to-day activities had been ‘limited a lot’. Given the symptoms are wide-ranging and diverse, the impact on health-related quality of life (HRQoL) can be significant, however the impact on quality-adjusted life-years (QALYs) has not yet been quantified.

This study aims to assess the impact of Long COVID on QALYs, using EuroQoL EQ-5D questionnaire results from a smartphone app which is linked to patients registered with TPP general practices to provide GP and health records. Collection of this questionnaire data across 30-day intervals will provide us with estimates of QALYs lost due to Long COVID. Results will evidence whether there are differences in quality-of-life impacts from Long COVID across demographic groups of age, ethnicities, geographic region, socioeconomic status and underlying health conditions.

Converting these results into QALYs, standardised measurements used in the National Health Service (NHS), can assess the impact of Long COVID to help evaluate how much the effects of Long COVID cost the health service to prioritise future interventions.

Hospital admissions due to COVID-19 disease are recorded in multiple sources. This includes the hospital records themselves and specific audits of admissions in participating hospitals that collect more detailed information.

Having multiple ways of capturing the same information provides an opportunity to determine how accurate and reliable each source is, by comparing them with each other. For example, one source may have better capture of clinical characteristics like diabetes or heart disease, while another may offer more complete geographical coverage of relevant COVID-19 admissions.

In this study we plan to explore the relative strengths and weaknesses of each data source so that future researchers are better informed about how best to use them for research.

Blood clots in the vessels of the legs and lungs cause over 12000 deaths per year in the UK. A proportion of these are thought to be preventable. The number of deaths increased significantly following the emergence of COVID-19 with this infection known to increase the risk of the blood clots. This study would help us test the hypothesis that COVID-19 drove an increase in blood-clot related deaths and allow us to understand the characteristics of the individuals that died in more detail. This may help us think about how we might prevent future deaths from blood clots and plan the allocation of national resources appropriately.

For at least a decade, the World Health Organization and the NHS have been encouraging use of digital technology to improve health and care services. This has expanded during the COVID-19 pandemic, where lockdowns and social restrictions have been accompanied by rapid roll-out of digital technologies to maintain access to health and care services. Government policy outlines a ‘digital first’ vision that prioritises the development of digital access to health and care services, e.g., by the NHS App as the ‘front door’ to the NHS. However, there concerns that older people are being disadvantaged and excluded with the increased use of digital technologies. The NIHR Older People and Frailty Policy Research Unit has conducted two evidence reviews exploring the impact of digital technologies on older people’s access to health and care. The first, conducted before the pandemic, found a lack of evidence that digital technologies could support older people’s access to health and care. The second has looked at what research has been conducted on the digitalisation of health and care services since the start of the pandemic. Most research so far has been small-scale and has focused on telephone access. There is an urgent need for detailed analysis of the impacts of pandemic-driven digitalisation of health and care services on older people. In particular, there is a need for a focus on potential health inequalities among different groups of older people (e.g., socioeconomic and ethnic groups). In this study, we propose to analyse electronic medical records in England to explore the impact on older people of digitalised GP access during the pandemic. We will analyse data from adult patients aged at least 18 who were registered with GPs between January 2019 to February 2022, with a specific interest in comparing consultation data for those aged under 65 and 65+.

People with inflammatory rheumatological diseases, such as rheumatoid arthritis, are at increased risk of blood clots, heart attacks and strokes when compared to the general population. Inflammation contributes to the risk of these complications, but other factors can also play a role, including other medical conditions and infections. In particular, we know that infection with COVID-19 increases the risk of blood clots, heart attacks and strokes. What is not yet understood is whether these complications occur more or less frequently in people with inflammatory rheumatological diseases who are prescribed specific medications.

Our goal is to use routinely collected health data to describe rates of the following complications after COVID-19 infection in people with inflammatory rheumatological diseases: 1) blood clots; 2) heart attacks, strokes and other cardiovascular complications; and 3) overall mortality.

We will compare the risks of these complications in individuals who are prescribed different groups of medications for their underlying condition (TNF inhibitors; JAK inhibitors; other biologic medications; non-biologic disease-modifying medications). We will account for other factors that might influence the chance of these complications occurring.

This project will provide clinicians with an evidence base on which to support clinical decisions. It will help determine which individuals are at greatest risk of complications, and who may benefit from closer monitoring.

A considerable proportion of the UK population suffers from long-term health conditions. During the COVID epidemic, a broad range of such conditions (comorbidities) contributed to very severe outcomes (hospitalisation and death) in people infected with the corona virus.

The 2020 shielding policy aimed to protect people at greater risk by asking them to reduce sharply their contact with other people and thus reduce their risk of becoming infected. This public health policy has not been evaluated and this study aims to do so using OpenSAFELY records on patient GP and hospital information on COVID and comorbidities.

The study will implement a novel combination of OpenSAFELY records with a dynamic model of corona virus transmission in the population. This approach accounts for the fast changes in the rate of infection that occurred as the epidemic spread; it also accounts for differing rates among locations, among age groups, and between people with or without comorbidities.

Comparing the actual epidemic with a modelling scenario where the shielding policy would not have taken place, we hope to estimate how many severe outcomes may have been averted by the shielding policy during 2020-21, and to so by age group in the population and NHS region in England.

There are an estimated 2.1 million people in the UK with long COVID as of January 2023 according to data from the Office for National Statistics (ONS). In the absence of effective treatments for Long COVID, it is important to understand the possible effect of vaccination on reducing the burden of Long COVID.

Previous research using OpenSAFELY has found fewer records of Long COVID in primary care data when compared to the ONS estimates of the amount of Long COVID in the UK. Our study aims to provide a description of people with and without a record of Long COVID between 2020 and 2022, broken down by vaccination status of individuals. For those with a Long COVID diagnosis before vaccinations were available, we will also describe the proportion vaccinated in this group.

This descriptive study will provide a detailed description of those with and without a record of Long COVID in electronic health records to guide and motivate future observational research of the condition.

In the UK there was a delay in the uptake of COVID-19 vaccinations for pregnant mothers due to safety concerns of the vaccine and impacts on the unborn child. It is now known that in the height of the pandemic 10% of pregnant women who caught COVID-19 were admitted to hospital for critical care during their pregnancy and 18% of mums with COVID-19 during their pregnancy delivered early (2.5 times that of the background rate).

A lot of non urgent healthcare services were reduced during the pandemic. Little is known about the impact of the pandemic on the care of mothers that did not have COVID-19 infections and how the pandemic impacted their continuity of care.

To date, most studies have looked at summary statistics of secondary care provision (what happened in hospitals). The main objective for work is to investigate the impact of COVID-19 on the treatment pathway for mothers during their pregnancy episode, the impact on pregnancy outcomes and postnatal follow-up. This will look at how mothers interacted with healthcare services and the consequences on maternal health postnatally.

There is widespread concern about winter pressures on NHS services, particularly in the context of continued COVID recovery. People rely on primary care services for initial contact, diagnosis, treatment, and coordination of care. Where primary care is under pressure, it is likely that there will be a negative impact on patient care and the good functioning of the whole health service; for example, if someone cannot get an appointment with their GP, they might be more likely to present to A&E. If we can identify early warning signs that an organisation is under pressure, plans could be put in place earlier to minimise any negative effect on patients. We want to investigate whether we can use patterns in primary care data to measure pressure on services (e.g., if a practice is overwhelmed, it might carry out less routine/non-urgent tasks) and how those measures of pressure vary between different practices and over time. We also want to investigate whether patients experience worse outcomes (e.g., presenting at A&E) if a practice does look to be overwhelmed. Finally, we would like to develop and test algorithms that can identify service pressure early on and alert primary care staff so that they can respond accordingly.

During the pandemic there has been a substantial change to the pattern of circulating viruses and bacteria that cause illnesses [ref https://bjgp.org/content/72/714/e63, https://www.opensafely.org/approved-projects/#project-108, https://www.gov.uk/government/news/ukhsa-update-on-scarlet-fever-and-invasive-group-a-strep-1]. In order to support ongoing response and recovery of NHS service from the COVID-19 pandemic, it is useful to have detailed information on patterns of disease being reported by the NHS and treatments such as antibiotics.

In the winter of 22/23 UKHSA (December 8th) reported an unseasonal increase of scarlet fever and group A strep infections. Sadly, between 19th September 2022 and 1st January 2023 there have been 151 deaths across all age groups, including 29 children under 18 in England. UKHSA indicates that the increase is likely to reflect increased susceptibility to these infections in children due to low numbers of cases during the COVID-19 pandemic, along with current circulation of respiratory viruses.

We would like to use OpenSAFELY (traditional code based and interactive) to analyse presentations to NHS services, recording of symptoms and diagnostic tests in peoples NHS records as well as any treatments, such as antibiotics. We will evaluate data in the years preceding the pandemic up until the latest available data.

Our expected outputs will be charts, blogs and academic papers describing our findings.

Ursodeoxycholic acid (UDCA) is used routinely to treat a liver condition called primary biliary cirrhosis (PBC) and is taken long-term to prevent the disease progressing. There is some biological evidence UDCA may prevent serious COVID-19 outcomes. A small study using observational liver disease registry data showed less severe COVID-19 outcomes in those taking UDCA. We plan to test this association in a much larger population using OpenSAFELY data. The study population will be all people with PBC, we will compare COVID-19 deaths in those treated with UDCA to those not treated with UDCA initially in first year of the pandemic, and then further extended up to the end of 2022.

This is important for those at high-risk of COVID-19, for whom vaccinations are less effective, where treatments to prevent the disease may be important to reduce the risk of severe COVID-19.

We want to know whether anastrazole use has been impacted by COVID-19. Anastrazole is a type of hormone treatment, which works by lowering the levels of oestrogen hormones in your body. It is mainly prescribed for women who have been through the menopause and have hormone-dependent breast cancer. It is however also prescribed to prevent breast cancer if you are at high risk and have been through the menopause; or to treat breast cancer in men and younger women; or as fertility treatment if you have polycystic ovary syndrome.

It is taken as a tablet once a daily, and will usually be taken for five to ten years. Some people start taking it after a number of years of taking another hormone therapy drug called tamoxifen. It is estimated from primary care prescribing data that around 70,000 patients have been prescribed anastrazole in the year from February 2021 to January 2022. Issuance in secondary care was between 40-50,000 per month up to March 2020 when it fell to almost 30,000 per month, and has gradually recovered over the last two years. Important to know what impact COVID may have had on this treatment for cancer.

COVID-19, flu, and other infections are more common in winter, and the effects of non-COVID infections may be exacerbated in people co-infected with COVID-19. Some risk factors, for example asthma or high blood pressure, may make you more likely to be infected or to become severely ill. When you visit your doctor or go to a hospital, information on your medical conditions is recorded. Approved researchers can access this information. We will use this information to learn more about which factors may make you more likely to have COVID-19 and other common infections in winter, and to become severely ill because of these infections. We also want to learn about which risk factors are linked with just COVID-19 and which are linked with multiple infections. Knowing this will enable doctors and other health care professionals to take extra care of people at risk. This should mean that they are less likely to need to see a doctor or go to hospital because of COVID-19 and other infections. It may also reduce winter pressures on the national health service.

Using antimicrobial medicines carefully and only when needed, also known as antimicrobial stewardship, is an important step to reduce the spread of antibiotic resistance and a key national priority.

The COVID-19 pandemic has had an impact on antibiotic prescribing. For example using openly available medicines data we have previously found (see blog https://www.bennett.ox.ac.uk/blog/2020/05/impact-of-covid-19-on-prescribing-in-english-general-practice-march-2020/) that substantially more prescriptions were issued for antibiotics used to treat respiratory conditions while prescribing for antibiotics for conditions like urinary tract infections decreased.

We would like to use OpenSAFELY-Interactive to describe trends and variation in specific clinical conditions (e.g. COPD) and different patient demographic groups (e.g. children/those resident in a care home). General practice gives out the majority of antibiotics in England so we will focus on prescribing in primary care and our expected outputs will be charts, blogs and academic paper describing our findings.

Sickle cell disease is the name for a group of inherited health conditions that affect the red blood cells. It is a lifelong condition where severe pain episodes, lasting days-weeks, develop episodically. It also increases the risk of infections and anaemia that can cause tiredness and shortness of breath.

During the pandemic people with sickle cell disease were advised to shield and it is also possible that their care was affected due to the indirect impacts of the pandemic. Additionally we have found that people with sickle cell disease who were eligible for COVID-19 therapeutics in the community, got lower access to these treatments than others (see OpenSAFELY Reports).

We would like to use OpenSAFELY-Interactive to assess the impact of the COVID-19 on the care of people with sickle cell disease in primary care. For example we will describe the number of people captured in primary care records before and after the onset of the pandemic which may indicate any disruptions in the sickle cell screening programme, describe the trends in prescriptions for pain to the group and prescribing of antimicrobials to manage symptoms. Our expected outputs will be charts, blogs and academic papers describing our findings.

Examples of specific measures for people with sickle cell disease that we would like to assess using OpenSAFELY-Interactive include; Measure of coding for sickle cell (trait/disease); coding of sickle cell crisis, trends in certain prescriptions such as painkillers and antibiotics.

Rationale: Gout is by far the most common cause of inflamed joints worldwide, yet one of the worst managed. There are very effective medications that can prevent flares of pain and swelling in people with gout. However, previous studies have shown that patients with gout are often not prescribed these medications. National guidelines have been introduced with the aim of encouraging the optimal treatment of gout, but there remains very limited data on how care for people with gout has been affected by the COVID-19 pandemic or by the introduction of these guidelines.

Aim and Methods: Our goal is to use routinely collected data from electronic medical records through OpenSAFELY, to evaluate how the COVID-19 pandemic has impact on the quality of care for people with gout in England.

We will evaluate national and regional changes in several important aspects of care, which align with national recommendations, including: i) the number of new diagnoses of gout; ii) the prescription of guideline-recommended treatments; iii) the monitoring and attainment of treatment targets; iv) and the number of hospital admissions for gout flares.

We will investigate trends in these outcomes using interrupted time-series models, which will enable us to compare pre-pandemic, early pandemic and late pandemic phases.

Expected output of project: In the short term, this project will highlight areas for improvement at national and regional levels in England. In the long term, this routinely captured data will be used to monitor for improvements in care for patients with gout.

Individuals with compromised immune systems are particularly vulnerable to the effects of SARS-CoV-2, and have endured a disproportionate burden of COVID-19-related morbidity and mortality throughout the course of the pandemic. Although risk factors for severe disease in the general population have been explored in detail, few studies have focused specifically on immunocompromised individuals.

Our study aims to identify factors associated with disease progression among immunocompromised individuals infected during the Omicron wave of the pandemic. The population of interest includes immunocompromised individuals as defined by: solid organ transplant; bone marrow transplant; haematologic malignancy; primary or acquired immunosuppressive condition; recent immunosuppressive medication; or recent chemotherapy or radiotherapy.

Key risk factors of interest include vaccination status, prior infection, and clinical comorbidities, among others. A better understanding of how these factors shape the risk of severe disease in immunocompromised populations may help identify individuals who may benefit from access to emerging therapeutics and support targeted clinical trial recruitment.

General practice has been disrupted by the pandemic in many clinical areas (e.g., Curtis et al., 2021; Williams et al., 2020). We aim to assess the impact of the pandemic on the routine management of blood pressure and hypertension. High blood pressure is one of the leading risk factors for several diseases (e.g., cardiovascular disease, stroke) worldwide. Research suggests that delays in the management of high blood pressure are associated with worse clinical outcomes, for example acute cardiovascular events, or death (Xu et al., 2015).

The Quality and Outcomes Framework (QOF) outlines several indicators that focus on blood pressure (BP) and hypertension (HYP) targets. We aim to use OpenSAFELY to quantify the extent to which any of the relevant QOF indicators were disrupted during the pandemic but wont link our results to clinical outcomes.

Our expected outputs include: (1) clinical reports will be published at https://reports.opensafely.org/, (2) reusable analytic code and codelists will be available for other researchers to use and modify, and (2) peer-reviewed academic papers discussing our results including actionable suggestions for service improvement within the NHS. Our outputs will be published timely and updated continuously. Further, we aim to supplement the aggregated results available at https://qof.digital.nhs.uk/ with detailed breakdowns of important demographic categories to assess health inequalities.

Background:

Chronic kidney disease (CKD) is a public health problem facing unprecedented challenges due to the disruption caused by the COVID-19 pandemic. CKD patients were amongst those at the greatest risk of death and severe illness as a result of COVID-19, whilst restrictions and the subsequent backlog have impacted healthcare access. Monitoring of CKD by GP practices may have been impacted due to stretched resources and changes to remuneration, while kidney transplantation programmes had to be paused at times.

Aim:

In this project, we plan to make use of existing healthcare data from across England to determine how healthcare use and quality of life has varied year on year since 2017-2018 for people with CKD.

Proposed methods:

We will undertake analyses for each financial year from 2017-2018 to 2022-2023.

For each year, we will report the number of people with CKD at the beginning of each financial year, grouped into different CKD severity groups including dialysis patients and kidney transplant recipients. We will then describe various measures of healthcare needs by each CKD group for each quarter of each year.

For people not on dialysis or with kidney transplant, we will report how many people in each CKD group either progressed to a more advanced stage of CKD or died each year. For people on dialysis or with kidney transplant, we will describe the number of people who died.

From our counts of healthcare needs, we will estimate the costs across each financial year from 2017-2018 to 2022-2023 for each CKD stage. We will estimate the average healthcare cost per person for each CKD stage, investigating further for geographical and socioeconomic disparities. We will then estimate the quality of life impact across each year for each CKD group, again investigating further for geographical and socioeconomic disparities. Topic areas

We are interested in describing and measuring the effect of COVID disruption on GP appointments. We have access to a database of GP appointment data that contains appointment scheduling information for 40% of the population in England.

We would like to explore the data in more detail to look at patterns over time to assist with health service recovery from COVID-19. We are particularly interested in looking at appointment lead time, which means the time between requesting an appointment and the appointment date. We will explore whether this (and other measures of interest) change over time, particularly before, during and after COVID. In addition to looking at this for the whole population, we will look at changes in particular subgroups of patients; for example, in particular age groups or ethnic groups, or for groups of patients with the same diagnosis (e.g., mental health).

With a better idea of how access to GP appointments varied over time for these groups, we would like to explore the data in more detail by applying data science methodologies to identify any short- and long-term patterns of change. We will be asking questions like: were some practices able to reduce appointment lead times faster than others in the COVID recovery period? If so, do those practices have any characteristics in common? This is important as it may help us identify best practice and describe the most effective and efficient response to primary care disruption due to COVID-19 in the future.

Diabetes is a condition in which your blood sugar level is too high. It can happen when your body doesn’t produce enough insulin, the body’s own insulin is not effective, or if your body cannot produce any insulin at all. Prolonged high blood sugar levels can have negative health consequences, such as seriously damaging your heart, eyes, feet, or kidneys. The aim of diabetes treatment is to keep your blood glucose at a healthy level, to lower your chances of getting complications.

The COVID-19 pandemic caused significant disruption to healthcare services. A reduction in routine blood sugar monitoring and subsequent treatment adjustments in diabetes patients may increase the risk of complications for these individuals.

The purpose of this study is to describe any changes in routine diabetic management during the COVID-19 pandemic, signalled by changes to measures defined in NICE guidance. For example we plan to report the overall rate of diabetes within the study population and then describe variation in both diabetes monitoring and treatment. Additionally we plan to report differences between subgroups of interest, for example: age groups, geographical regions, other illnesses This may help to better understand inequalities in healthcare during the COVID-19 pandemic and highlight potential areas for further research and improvement.

Chronic lymphocytic leukaemia (CLL) is a cancer which affects white blood cells that mostly affects people over age 60. Individuals with CLL have weaker immune systems and therefore are at great risk from both contracting, and subsequently experiencing severe outcomes from, COVID-19. Vaccines within the CLL population are thought to be less effective than in the general population.

The purpose of this study is to describe trends over time in CLL measures, such as the mortality rate, during the COVID-19 pandemic. We intend to compare trends within the CLL subpopulation against a subpopulation without CLL. Additionally we plan to report differences between subgroups of interest, for example: age groups, geographical regions, other illnesses. There are many factors that could affect mortality risk, such as: behavioural interventions (e.g. lockdowns, mask wearing), vaccination status of individuals, or the dominant strain of SARS-CoV-2 circulating within the population. To provide context to our mortality rates, we will describe the temporal population-wide state of these factors.

This may help to better understand the impact of the COVID-19 pandemic on individuals with CLL.

During COVID-19, GPs have supported people to manage a decline in their mental wellbeing. Many services and community activities were not available during COVID and this may have resulted in changes to prescribing of drugs as a result of declining mental wellbeing. Certain medicines that are effective in the short term to elevate symptoms of poor mental wellbeing can cause dependence. This study will establish any changes to what dependence forming medicines have been prescribed, how alternatives to medication have been employed and the impact this has on people in terms of how often they turn to their GP for continuing support.

Analysis to assess the statistical variation in patient demographics receiving care incentivised via the QOF and how this has changed or been affected by the COVID-19 pandemic and subsequent recovery of Primary Care Services.

It is widely believed, although not previously analysed in detail, that there will be a difference in the demographic of patients who receive and those who do not receive care via QOF incentivised activities. We would be seeking to analyse the effect of COVID-19 on these differences by directly comparing data before the pandemic with data during the pandemic (including post-pandemic recovery).

The expectation is that the results of this are disseminated within NHSE as a policy development paper. With wider code and results published as per OpenSAFELY’s standard procedure.

Systemic lupus erythematosus (SLE) is a condition where the body attacks itself, so it can affect the skin, joints, kidneys, lungs, and the nervous system. It is expected that people with SLE will be at higher risk of severe COVID-19 because the medication used for the condition will reduce a person’s immunity and ability to fight infections. This study aims to investigate if SLE (systemic lupus erythematosus) and/or the treatments of SLE increase a person’s chance of getting COVID-19,if they get it more severely and if they are more likely to develop long COVID-19. We will also examine if having COVID increases a person’s chance of developing SLE. We will look back in individuals records to examine how many people with SLE developed COVID-19 and if they had more severe disease compared to people without SLE. We will use the records of people in the OpenSAFELY and the SAIL database in Wales between 1st March 2020 to 31st December 2021. OpenSafely covers 96% of the GP population in England and SAIL covers 98% of the GP population in Wales.

The effect of the COVID-19 pandemic, and of the measures introduced to control it, is not yet fully understood. To date, we know that mental health of the population was affected, and that primary care, secondary care and emergency department consultations related to mental health decreased dramatically in the first phase of the pandemic. We also know that the system has changed in response to the pandemic, for example more virtual consultations are now conducted with the reduction in face-to-face consultations as compared to before the pandemic. In this research, we will quantify the change during the COVID-19 pandemic in trends of psychiatric emergencies across different patient groups. We will investigate the following psychiatric emergencies: self-harm, suicide-related, alcohol-related, eating disorders-related, and home violence-related hospitalisations and emergency department visits, as well as mortality (due to above mentioned psychiatric emergencies). We will compare the trends (counts and rates) in psychiatric emergencies observed (that actually happened) during the COVID-19 pandemic to trends that would be expected if the pandemic had not happened (predicted using mathematical models). With this, we will investigate the effect of the COVID-19 pandemic on 1) mental health of the population, 2) the severity of psychiatric emergencies, 3) service utilisation by patients and 4) inequalities between different patient groups. This research is important to support patients, healthcare professionals, and the healthcare system in the recovery from the pandemic.

We are interested in describing and measuring the effect of COVID disruption on GP appointments. We have access to a database of GP appointment data that contains appointment scheduling information for 40% of the population in England. Because we haven’t used this dataset before, we will first check that the data within it looks sensible; for example: is the date on which an appointment is booked always before the appointment date? Do we always know when an appointment is booked? If this information is missing, can we find any patterns of where it is missing (e.g., do some practices tend not to record it)? Does it look like some appointments are repeated in the dataset? These kinds of checks will help us decide whether any additional data cleaning steps need to be implemented before we carry out any our analyses.

There is a theory that COVID-19 disease and vaccination could could lead to inflammation in the prostate, increasing levels in the blood of the prostate specific antigen (PSA). PSA test is a common prostate cancer screening offered to men. So an increase in PSA could lead to unnecessary and harmful false cancer flags.

This is study we will investigate the possible effect of the COVID-19 vaccination and infection on PSA levels. We will also investigate how long the effect is.

This research could help us understand more about the potential effect of COVID-19 on PSA and alert clinicians to be vigilant. It will help the patients and clinicians to use the PSA testing more effectively and interpret the results more appropriately.

Cancer services including cancer diagnosis were heavily disrupted by the COVID-19 pandemic. This study was designed to investigate the effect of the pandemic on prostate cancer diagnosis overall and by demographics such as age, ethnicity, deprivation, region. We will use primary care records to describe the trends in diagnosis (new and existing) over time from January 2015 to October 2022. We will look if different subgroups of patients (age, ethnicity, deprivation, region) were affected by the COVID-19 pandemic in a different way. This knowledge gained from this study will help NHS understand if there was an effect of the pandemic on prostate cancer diagnosis and if this was different for patient with different demographic characteristics. The knowledge could help the NHS improve the services post-pandemic and address the potential backlog that was created by the pandemic.

Prior to the pandemic the NHS had recognised that prescribing of antipsychotics was excessive in certain at-risk populations including in patients with dementia, learning disability, autism and those in care homes. Antipsychotics are licensed to treat serious mental illness, but they can also be used to manage behaviour where patients are severely distressed or at risk of harming themselves or others.

During the pandemic many of the social and charitable structures that supported people with disabilities were substantially affected and researchers have described the negative consequences in these at-risk groups, including increases in agitation, anxiety, depressive symptoms and negative behaviours such as aggressive and self harming behaviours.

The purpose of this study is to determine if the COVID-19 pandemic affected the rate of prescribing of Antipsychotic Prescribing in individuals with autism, dementia, learning disability, serious mental illness or those resident in a care home.

Medication reviews are an important aspect of primary care to ensure that a patient’s prescribed medications are safe, effective and appropriate. The COVID-19 pandemic has been shown to have caused significant disruption to healthcare services. A reduction in the frequency of medication reviews may increase the risk of harm from medications and/or result in suboptimal management of long term conditions.

There is no single definition of what constitutes a medication review. A medication review can ensure the appropriateness of a prescribed medication or ensure appropriate monitoring has taken place. They may be undertaken to review all of a patient’s current repeat prescriptions or to review medication for a single condition only. For example, the review may ensure that a patient prescribed medicine to lower their blood pressure, has had their blood pressure checked recently.

The purpose of this study is to observe changes in the number of medication reviews conducted in primary care during the COVID-19 pandemic.

The study will report the rate of medication reviews overall and then describe variation between key clinical, regional, and demographic subgroups (e.g., Index of Multiple Deprivation, ethnicity, or a record of learning disability). This may help to better understand inequalities in healthcare and highlight potential areas for further research and improvement.

BMI has been identified as a risk factor for clinical outcomes in patients with COVID-19. Electronic Health Records (EHR) hold significant potential for studying the association between BMI and COVID-19. This depends on the validity, completeness, and currency of BMI information that is available in or can be derived from EHR. Thus, this study aims to describe and compare methods for identifying BMI in EHR.

The study will check for consistency, completeness, and precision of the BMI variable, comparing the two methods of identifying BMI in OpenSAFELY: (1) calculation from height and weight and (2) GP-recorded values. It will evaluate the validity of each method (e.g., identify proportion of missing values or values out of range) and summarise any discrepancies across the two methods. Furthermore, it will aim to evaluate whether there are any patterns in missingness by relevant subgroups in the population.

The analysis will be completed using primary care EHRs of adults between 21 and 110 years registered with a GP-surgery using TPP and EMIS on 1 March 2020. Descriptive tables will be produced to describe missing and out of range values. Between 1 March 2015 to 1 March 2020, unique BMI for each patient will be considered to summarise how often BMI records are updated (average number of unique records over 5 years, average duration between BMI updates). Analyses will be performed for the overall population as well as by demographic categories, clinical conditions, and practice/region/STP.

Medication related harm describes any error related to prescribing that may cause harm to a patient (e.g., prescribing a medicine at the wrong dose or not carrying out a necessary blood test before prescribing a medicine). Some important indicators of medication related harm have been designed as part of the PINCER programme, which helps pharmacists find instances of “potentially hazardous prescribing” in the electronic health records of their practices, so that clinically important errors can be reduced.

We are interested in understanding whether these PINCER hazardous prescribing indicators were affected by the disruption of primary care services due to COVID-19 and to what extent those rates of medication related harm have recovered since primary care services were restored. Once we have an overall understanding of how rates have changed, we would like to explore the data in more detail by looking at the rates for each practice over time and applying data science methodologies to identify any short- and long-term patterns of change. We will be asking questions like: are there any practices that did not experience any COVID-19 disruption in their PINCER indicators? Were some practices able to recover faster than others? Should we find any practices that correspond to these patterns, we will explore whether these practices have any characteristics in common. This is important as it may help us identify best practice and describe the most effective and efficient response to primary care disruption in the future.

We do consider this project a “proof of concept”, in which we demonstrate the utility of our approach. The intention in the longer term is to use the same analysis pipeline across all clinical areas to identify variation in change over time and expose factors associated with positive change over time.

The COVID-19 pandemic and related public health measures have led to disruptions to the healthcare system, including prescribing behaviour. Internationally, increases in opioid prescribing to people in care homes and a shift away from non-pharmacological treatment to opioids for people with pain since the start of the COVID-19 pandemic have raised concerns, and in the UK there have been reports of increased opioid use in people on outpatient procedure waiting lists. Prior to COVID-19, there were concerns about high opioid prescribing rates in the UK, especially for non-cancer pain and in areas of greater social deprivation. Thus it is important to understand if and how prescribing behaviour has changed during the COVID-19 pandemic, and how this varies by sociodemographic factors.

We will use OpenSAFELY data to identify all registered adults prescribed opioids from Jan 2018 to Mar 2022. We will describe rates of opioid prescribing prior to the COVID-19 pandemic, and estimate any changes in prescribing patterns in the two years after the start of the COVID-19 pandemic, especially in relation to the start dates of the nationwide lockdowns. By focussing on key subgroups (people in care homes, people without cancer) and stratifying by relevant characteristics (age, Index of Multiple Deprivation, region) we will identify which groups were most affected which can help identify targets for improved prescribing.

An increased risk of Covid-19 infection and subsequent admission to hospital with COVID has been seen in groups already carrying a burden of biological, psychological and social risk factors (e.g. poverty, poor housing and unemployment) associated with poor mental and physical health outcomes. Many patients with COVID have required treatment on Intensive Care Units. There is a substantial literature on the emergence of psychiatric illness after discharge from ICU. We know that many patients who have had Covid-19 develop a psychiatric disorder, although the reasons for this are unclear. Our study aims to help understand better how poor mental health after admission to hospital with Covid-19 is related to patients’ experiences in hospital (e.g admission to ICU), and how much is explained by difficulties experienced by patients before they became ill. We will use data collected about patients’ hospital admissions, information they provide in a questionnaire about life before Covid, and some information from their GP records about overall health and functioning.

In our previous research using a study called ISARIC, we studied people who became unwell with Covid-19 and had complications related to their kidneys. We found which people were most likely to have kidney complications and what effect this had on their survival. In this new project, we will study whether kidney complications are unique to Covid-19 or whether similar problems develop from other infections. We will also study the long-term kidney effects of being unwell with Covid-19. These subjects are important because complications related to kidneys are the most common amongst people with Covid-19. We need to know more about what impact Covid-19 has on long-term kidney health.

According to a survey published in May 2022 by the Office for National Statistics, 1.8 million people, nearing 3% of the total UK population, had long COVID. The number of long COVID cases is enormous; however, the impact on the healthcare system has not yet been quantified. Our study aims to investigate the healthcare utilisation of people with long COVID and the risk factors for high users of healthcare resources.

We will identify people with long COVID using their electronic health records and analyse their healthcare contacts and costs, specifically primary care consultation, accident and emergency room visits, secondary care (inpatient and outpatient), and prescriptions. In addition, we will explore the potential factors associated with high healthcare utilisation among people with long COVID and compare their healthcare utilisation to historical records before the pandemic.

Alternatives to COVID vaccines are being developed for people whose immune system is compromised, and who therefore are unable to benefit from currently available vaccines. One example is Evusheld, a combination of two monoclonal antibodies, tixagevimab and cilgavimab. Once these treatments are launched, their safety will be closely monitored for example through spontaneous reporting of suspected adverse drug reactions by patients, carers and healthcare professionals. In order to contextualise potential safety signals arising from spontaneous reporting, a detailed understanding of the background rate of key clinical events in an immunocompromised population is required. This will help with the interpretation of potential signals, and to prioritise which studies should be done to further investigate signals. This study will therefore describe the underlying rate of key clinical outcomes in immunocompromised people in the UK.

Background: OpenSAFELY is a secure research platform that includes data from 58 million English patients. OpenSAFELY has been used to answer important questions about COVID-19, influencing policy decisions and increasing our understanding of the effects of the virus. OpenSAFELY holds data from national PCR and lateral flow testing. Importantly, the availability of testing has changed over the course of the pandemic, and testing was often driven by patient choice. This has likely led to important differences between people who had COVID-19 tests and those who did not. These differences mean that studies using the testing data might produce incorrect results, unless these differences can be accounted for in the study. This is important for COVID-19 and for future UK pandemics.

Aim: The aim of this project is to determine how we can account for differences in testing in OpenSAFELY to address important questions in COVID-19, and then how we can use this to inform policy for future pandemics.

Methods: We will use linked data from OpenSAFELY and the ONS COVID-19 Infection Survey (ONS CIS). ONS CIS is much smaller than OpenSAFELY but it randomly sampled households across the UK and regularly tested all participants, which means that the data were not reliant on access to testing or patient choice. We will compare people who tested positive in ONS CIS to those who tested positive in OpenSAFELY. We will then look at risk factors for hospitalisation and death among those with recorded COVID-19 infection in OpenSAFELY, and use methods to understand how using the national testing data (rather than regularly testing everyone) might have led to incorrect results. Finally, we will use various statistical methods to try and overcome any errors introduced by use of COVID-19 testing data.

During the COVID-19 pandemic, NHS England issued guidance around prescribing of anticoagulants (commonly known as blood thinners); recommending a switch from warfarin to a newer type of anticoagulant, commonly known as DOACs, in some groups of patients. This was partially due to the need for less frequent blood testing in patients prescribed a DOAC. However, DOACs are not recommended for patients with a mechanical heart valve. This study aims to carry on and refine previous work on DOAC prescribing in patients with a mechanical heart valve.

In December 2021, COVID-19 medicine delivery units were launched across England to offer antiviral medicines and neutralsing monoclonal antibodies to nonhospitalised COVID-19 patients thought to be at high risk of severe outcomes. Amid understandable concerns surrounding early regulatory authorisations, noninterventional evidence surrounding the effectiveness of these medications is needed to guide policy surrounding the use of these medications in routine clinical practice. We aim to emulate a randomised trial using observational data to estimate the effectiveness (prevention of hospitalisation and death) of either sotrovimab or molnupiravir versus no-treatment, amongst non-hospitalised COVID-19 patients in one of the high-risk groups.

Capturing ethnicity information in routine health records is recognized as a necessary pre-requisite to understanding and addressing inequalities in health service usage and outcomes. OpenSAFELY has been, and continues to be, used to support research that delivered urgent results related to the global COVID-19 emergency.

Assessing sample representativeness, consistency and completeness is a critical component of health research and provides context for researchers to understand the validity / applicability of their research.

Summary

This study profiles the representativeness, consistency and completeness of routinely collected primary care ethnicity data.

Intended Outputs:

• Short data report on ethnicity variables
• Build on data curation / validation framework to support future work

This study aims to evaluate whether the COVID-19 pandemic affected the completion of shared care drug monitoring. It also considers whether any such effect was associated with other variables, such as patient demographics or comorbidities.

The analysis will cover all patients registered with a GP practice aged 18 years and over, who are currently prescribed currently prescribed Azathioprine, Methotrexate, Mercaptopurine, Mycophenolate, Leflunomide or Levothyroxine.

Understanding the impact of COVID-19 on monitoring is important, as in the short-term this will lay the foundation for an analysis of the health impact and recovery planning, and in the longer-term an analysis of whether current monitoring recommendations remain appropriate. Furthermore, understanding factors which underpin variance in resilience to the impact of COVID-19 on monitoring adherence may be valuable learning points, with a view to increasing the resilience of the population overall, and reducing health inequalities between the least and most resilient segments.

Rheumatology services were directly affected during the pandemic, as rheumatologists were frequently seconded to treat COVID-19 patients, rheumatology appointments were moved to remote appointments and drug monitoring was reduced. The impacts of these changes has not been measured. Patients with rheumatoid arthritis are treated with immunosuppressive medications that require blood monitoring for toxicity and patients are monitored by a rheumatologist to manage the disease activity, thereby preventing further complications. This study will identify a rheumatoid arthritis population between 1st March 2018 to 30th April 2022 from the OpenSAFELY platform and will measure blood monitoring and associated outcomes, 2) rheumatology outpatient appointments, 3) hospital admissions and 4) deaths. The impact of the pandemic will be measured by comparing the pre-pandemic period to the pandemic period.

OpenSAFELY is a powerful platform for analysing patient records in English primary care and other linked healthcare datasets. After rapid development at the start of the COVID-19 pandemic (March 2020), we launched a pilot programme to allow external research/informatics groups to use the platform for COVID-19-related questions. This has been highly successful; however, such project applications require a certain technical skillset, and there is a relatively long lead-time to start performing analyses; although feedback from users is that the onboarding time is relatively short compared with many platforms (typically 5-10 weeks), it is not possible to conduct any rapid analysis. While we are creating dashboards that show activity in selected clinical areas, we would like to permit some simple counts to be accessible more flexibly and widely, for example, for policy experts and other individuals involved in NHS service planning, commissioning, monitoring and best practice guideline development.

For the pilot phase, access will be provided to a defined group of NHS England policy makers. Once approved, individuals will be able to submit simple analyses, by selecting a clinical codelist (a list of clinical codes used to capture all the ways an activity or condition may have been recorded, e.g. a blood test or asthma) from the OpenCodelists tool. For the pilot phase, only codelists which have been used for existing OpenSAFELY COVID-19 related projects on OpenCodelists.org will be incorporated into the tool. Users also select a time period and a small range of other options.

For the pilot phase, we propose the following category of analyses, returned to the requester within a simple PDF/html. The number of occurrences of the chosen codelist are counted over the selected time period and figures may include: a table of the most common codes within the codelist, variation in code use between practices per week/month, rates broken down by sex/age group/ethnicity/region. Patient counts will be rounded and small numbers suppressed to protect patient privacy. We will apply automated statistical disclosure controls where appropriate, and where necessary results will be reviewed prior to release by trained output-checkers. Practices will not be identified; except in cases where an appropriate governance approval has been granted, such as a practice requesting to identify only its own data within a graph of other practices.

Using this tool users will be able to: preview codelists; see which codes in a particular codelist are most often used; assess the change over time in a particular clinical activity; get an indication of the level of coding of conditions/activities of interest; or ensure an activity/outcome of interest is recorded in sufficient numbers to permit further analysis prior to submitting a full project application.

In the pilot phase, results are not publishable - but used to inform development of OpenSAFELY-Interactive.

Chronic kidney disease (CKD) is one of several factors that makes a person substantially more vulnerable to COVID-19. Not only are individuals with CKD at greater risk of severe disease if infected, but they are less likely to respond adequately to current COVID-19 vaccines. Accordingly, individuals with CKD were among the earliest to be offered COVID-19 vaccines in the UK, and have subsequently been offered up to two further doses with the aim of providing stronger and more durable protection against severe disease.

This project will use the OpenSAFELY platform to answer two key questions. First, what levels of vaccine coverage have been achieved in CKD patients? Second, how does vaccine effectiveness (VE) vary among CKD patients who have received different types of vaccine? Where possible, the project will distinguish between key subgroups of CKD patients, such as those who have received organ transplants or dialysis. For the analyses of VE, key objectives will be to compare vectored (ChAdOx1-S) versus mRNA (BNT162b2) vaccines, and heterologous (“mix-and-match”) versus homologous (single type) schedules.

The study will thus identify how effectively CKD patients have been reached during different phases of vaccine roll-out, and how effective different vaccination strategies have been in this vulnerable population.

UKHSA has recently detected higher than usual rates of liver inflammation (hepatitis) in children in England (see news story).

Possible causes are being investigated including COVID-19 and other causes such adenovirus ; it is thought that changes in patterns of circulating virus due to COVID-19 pandemic regulations could have also contributed to this outbreak.

In order to support the response teams investigation to prevent harm to children the OpenSAFELY team is conducting an analysis to provide data to inform the response. We will look at data on the recording of clinical codes on children’s electronic health records and provide charts on the trends over time. For example we will look at how many children have records that indicate they may have a gastrointestinal illness, a key sign of liver inflammation on children presenting in general practice for symptoms such as gastrointestinal illness and look for changes in the number of liver function tests in children which GPs may order to further investigate and diagnose hepatitis.

In the past year a body of research has emerged highlighting a potential link between treatment with statins and COVID-19 related outcomes. Bergqvist el al (2021), for example, find that individuals taking statins were at a lower risk of death attributable to COVID-19. This is potentially explained by the role of statins in reducing inflammation and blood clotting, features frequently observed in severe COVID-19 cases.

In this study we aim to further explore the extent to which a causal relationship between statin use and COVID-19 outcomes exists using a regression discontinuity study design, as well as instrumental variable (IV) methods (using QRISK as an IV for example). Such approaches have the potential to adjust for sources of bias common in observational studies, which have currently been employed by the majority of studies investigating the relationship between statins and COVID-19 outcomes. In this way, this study will potentially provide greater insight into the mechanisms underpinning COVID-19 related health outcomes.

Applying such methods will provide a different perspective which can be incorporated within an evidence triangulation framework to better inform health policy with respect to the use of statins in preventing adverse COVID-19 health outcomes. We anticipate the results of this study will be presented in the form of a peer-reviewed paper, and potentially a subsequent paper formally triangulating evidence from multiple studies of differing design.

References

Bergqvist R, Ahlqvist VH, Lundberg M, Hergens M-P, Sundström J, Bell M, et al. (2021) HMG-CoA reductase inhibitors and COVID-19 mortality in Stockholm, Sweden: A registry-based cohort study. PLoS Med 18(10): e1003820. https://doi.org/10.1371/journal.pmed.1003820

On 16th December 2021, COVID-19 Medicine Delivery Units (CMDUs) were launched across England to provide antiviral medicines and neutralising monoclonal antibodies (nMAbs) to treat COVID-19 patients in community settings who are at high risk of severe outcomes. The approval and early clinical use of these medications have been mainly supported by evidence from several randomised controlled trials of limited size. It is crucial to evaluate the effectiveness and safety of these medications on preventing COVID-19 disease progression in real-world settings with wider populations.

This study aims to examine and compare the effectiveness of sotrovimab and molnupiravir, the two most frequently prescribed medications by CMDUs so far, in non-hospitalised high-risk adult patients with COVID-19 across England. The main comparator groups are patients who received sotrovimab vs. those who received molnupiravir, between whom the risk of COVID-19 related hospitalisation/death within 29 days after treatment initiation (and other efficacy/safety outcomes of interest) will be compared.

The STOMP initiative launched in 2016 sought to reduce inappropriate prescribing of psychotropics among those with learning disability and autism (LD&A). Many non-pharmacological options to managing anxiety, distress and behavioural problems in LD&A require access to specialist staff, communal spaces, specialist learning environments and much of this was restricted during the COVID-19 lockdown.

In the Quality and Outcomes Framework, 2020-2021 report, NHS Digital found depression to have the greatest increase in QOF recorded prevalence, an increase of 0.7 percentage points from 2019-2020. The QOF framework also includes an indicator for depression diagnosis review.

This study will describe how the diagnosis and review of depression and the use of psychotropic medicines in the vulnerable LD&A population may have changed before and during the COVID-19 pandemic.

OpenSAFELY-TPP and OpenSAFELY-EMIS will be used to identify patients of any age with a diagnosis of learning disability or autism. For the study period of January 2018 to January 2022 monthly rates of current and new depression diagnosis and current and new antidepressant prescription will be described, looking separately at Selective serotonin reuptake inhibitors (SSRIs) and other antidepressants (Tricyclics, Monoamine oxidase inhibitors (MAOIs) and others). Based on the QOF metric DEP003, the monthly rates of individuals 18 and older who received a review for a new diagnosis of depression within 10 to 65 days will also be described.

SARS-CoV-2 infections are often asymptomatic. Asymptomatic infections are challenging to pandemic control because they are a substantial source of transmission but are difficult to identify (and therefore target directly) within the resource constraints of health systems. New knowledge about characteristics of individuals that are associated with asymptomatic (versus symptomatic) infection may inform public health policy and clinical understanding of COVID-19. In this study, we will use national household survey data linked to patients’ general practice records to investigate the symptoms, if any, reported by individuals with a SARS-CoV-2 infection in England. We will examine factors that may be associated with whether an infection results in symptoms or not, such as socio-demographic characteristics, pre-existing health conditions, vaccination and prior infection, and the SARS-CoV-2 variant and load.

We are investigating the characteristics of those with long covid within the Covid Symptom Study Biobank longitudinal cohort study. We would like to use OpenSAFELY Electronic Health Records to gather information such as past diagnoses of long-covid, other comorbidities and vaccination status that would otherwise be unavailable; this will help to enhance our understanding of the factors which lead to long covid and determine its severity.

The number of infections that have become resistant to antibiotic treatment has increased over the past years. This antibiotic resistance can be driven by increased antibiotic use. Using antibiotics carefully and only when needed, also known as antimicrobial stewardship, is therefore important to reduce the spread of antibiotic resistance.

The COVID-19 pandemic has had an impact on antibiotic prescribing. For example, it has resulted in more prescriptions being given for dental infections. General Practitioners also say that they were more likely to give antibiotics for throat infections during the pandemic, hence increasing antibiotic use.

General practice gives out the majority of antibiotics in England. This project therefore will investigate the impact that the COVID-19 pandemic has had on long-term and repeated antimicrobial prescribing in general practice. This includes, for example, antibiotics prescribed regularly for infections of the lung, bladder and skin.

Aims

• Evaluate the impact of COVID-19 on long-term and repeated prescriptions of antimicrobials in general practice
• Produce guidance tools to support healthcare colleagues in general practice to review patients on long-term and/or repeated antimicrobials
• Share knowledge with scientific colleagues to encourage wise use of antibiotics and help reduce antibiotic resistance

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is a viral infection; while many people experience mild symptoms and recover quickly, others go on to develop severe coronavirus disease (COVID-19) resulting in hospitalization or death. Many studies have shown that severe COVID-19 is more common in people who are older, overweight, or with existing health conditions. In countries such as the UK and USA, COVID-19 has had a greater impact on ethnic minority populations, including those of Indian ethnicity.

At the start of the pandemic, India has the lowest number of COVID-19 deaths worldwide, leading to the belief that severe COVID-19 may be less common in people residing in India than in Indian people in other countries. There are several plausible explanations for this; Firstly, people residing in India may have greater environmental exposure to sunlight than Indians residing in other countries, which may result in higher levels of vitamin D. Secondly, people living in India may have greater immunity to COVID-19 resulting from greater exposure to other infections. This is related to socioeconomic status as people living in overcrowded households with poor sanitation may be more likely to contract infections than Indian populations overseas. Thirdly, people residing in India may also be less obese and have fewer pre-existing conditions than Indian people living in higher-income countries. Lower BMI in Indian residents may result from lower calorie intake or higher physical activity - both of which may lower the risk of severe COVID-19.

Since 2021, the COVID-19 pandemic has significantly worsened in India, and the comparison between Indian people residing in India vs. overseas has reversed. While it is possible that real biological differences may explain the initial lower risk of severe COVID-19 in Indian residents, there may be alternative explanations such as differences in the quality of COVID-19 data and age structure of the populations. For example, people living in India may be less likely to visit the hospital with COVID-19 due to stigma around the virus, or they may be more likely to die at home instead of in a medical facility. This means that the cause of death may be less well captured in India than in other countries, and the true number of people with severe COVID-19 may be underestimated. To date, no study has fully answered whether the severity of COVID-19 truly differs between Indian populations in India and in other countries.

In this study, we will compare the risk of severe COVID-19 between Indians in the UK and India and identify possible explanations. We use the anonymised medical records of Indians in the UK and interview data from 220,000 people living in urban and rural India. We will first compare the proportion of the population who have ever had severe COVID-19, defined as either going to the hospital for or dying from COVID-19.

We will then explore whether differences in the prevalence of severe COVID-19 can be explained by differences in data quality, the age structure of the two populations, or differences in the prevalence of key risk factors.

We expect that the impacts of our research will include the development of COVID- 19 prevention and management strategies which take into account the specific needs of Indian populations worldwide, and a deeper understanding of factors which lead to the the development of severe COVID-19 in people of Indian origin.

Rationale

The coronavirus disease 2019 (COVID-19) caused by the novel severe acute respiratory syndrome coronavirus 2 (SARS-Cov-2) is the biggest challenge the National Health Service (NHS) has faced in living memory. For rheumatology services, the impact of the pandemic has been enormous, with appointments delayed, consultations shifted to telephone, and procedures cancelled. Variation across the country is inevitable and we have an opportunity to learn from each other.

The COVID-19 pandemic offers a unique opportunity to change the way we measure care quality in the NHS, reducing data burden and improving completeness. Care quality for rheumatoid arthritis (RA) services is currently assessed by the Health Quality Improvement Partnership (HQIP) National Early Inflammatory Arthritis Audit (NEIAA). The audit collects information across the first 12 months after diagnosis capturing performance against the NICE quality guidelines. However, data collection is burdensome for Trusts, often incomplete, and was paused during the COVID-19 pandemic.

Aim and Methods

Our goal is to use routinely collected data from electronic medical records through OpenSAFELY to describe the impact of the COVID-19 pandemic on the delivery of services to patients with newly diagnosed rheumatoid arthritis, identifying variations in care and in the speed of recovery across England.

We will describe: i) delays in the diagnosis of rheumatoid arthritis; ii) delays in the shared prescribing of treatments for rheumatoid arthritis with primary care; and iii) delays in escalating treatment for rheumatoid arthritis. We will analyse this data using an interrupted time-series model, which allows us to compare pre-COVID-19 and post COVID-19 periods from March 2020, with the post COVID-19 periods further split at the start of subsequent pandemic waves.

Expected output of project

We anticipate that the output of this project will contribute to a reduction in variation in care across NHS service providers. In the long term, this research will set a precedent for the monitoring of early rheumatoid arthritis care to be undertaken using routine captured data, reducing the burden of clinician level data entry for every patient seen.

Prevention of suicide deaths is a national priority for NHS England and Improvement. Analysing data on suicide deaths to understand what contact with healthcare took place in the days before they completed suicide will help us find opportunities to improve suicide prevention. Around a quarter of people who die by suicide are known to specialist mental health services, but many others may have been in contact with their GP or other primary care services.

Studies show that survivors of COVID-19 infection are at increased risk of developing mental health disorders, but additionally, the COVID-19 pandemic has led to changes to daily living and social restrictions, contributing to decline in mental health and well-being of the general population. COVID-19 also affected how people accessed healthcare.

This analysis is part of wider analysis on suicide deaths following contact with healthcare services, which will help to shape future suicide prevention improvement programmes. OpenSafely data will be used to examine suicide deaths following contact with GP or primary care services at a national level.

Background

Primary care records offer an opportunity to ascertain cases of COVID-19 which do not necessarily result in hospital admission or death. This could be useful for studying the burden of COVID-19 in the community, risk factors for SARS-CoV-2 infection separately to risk of severe COVID-19, risk factors for mortality and case fatality ratios among those infected, and post-viral effects in people who had COVID-19 that did not require hospitalisation.

There are over 100 primary care (CTV3) codes with terms related to COVID-19 used by TPP and available for selection in studies performed in the OpenSAFELY platform (https://opensafely.org/). The majority of these codes have been newly created for use in the current pandemic. The aim of this work was to assign these codes into categories related to the identification of COVID-19 in primary care, and to provide advice for studies using the OpenSAFELY platform that require people to be classified by their COVID-19 case status as defined in primary care records (either as an exposure or as an outcome).

Methods

An initial list of TPP primary care codes related to COVID-19 was obtained by searching the TPP database for terms containing “COV-2”, “Coronavirus”, or “COVID”. The returned terms were cross-checked against the NHS Digital COVID-19 SNOMED CT codes and CTV3 codes for any missing terms which were added to the list when found. The resulting list of terms was then reviewed by a team of clinicians, epidemiologists and statisticians in order to identify distinct categories of terms and assign terms into one of these distinct categories.

An initial analysis of (probable case and suspected case) sub-categories was then performed by plotting the following using OpenSAFELY data from between February 2020 - November 2021 (1) the frequency of codes entered into TPP software by GPs over time and (2) the proportion of people dying due to (a) COVID-19 and (b) causes other than COVID-19 (using ONS cause of death data) in the 80 days after a record of a positive test in either primary care TPP data or in SGSS data.

This study sets out to describe subgroup differences in COVID-19 related death rates over time (February 2020 - December 2021) and describe COVID-19 related death risk in different subgroups of the population in the three pandemic waves. Subgroups are based on age category, sex, body mass index category, ethnicity, deprivation and comorbidities.

This study will investigate the negative impacts of the pandemic on a range mental and physical health conditions not directly related to coronavirus and identify how these impacts differ by ethnicity, and socio-economic status. Understanding these negative impacts will inform future healthcare policy by identifying population groups at high risk and identifying health conditions most negatively affected by the social restrictions and healthcare burden caused by the pandemic.

The COVID-19 pandemic has caused substantial illness and death, with the impacts most serious for older adults, ethnic minority groups, and socially deprived populations. In addition to illness and death caused by COVID-19 itself, the pandemic has had effects on wider mental and physical health. Reasons for this include reduced resources for the prevention, monitoring and treatment of illnesses un-related to COVID-19, and individual concerns around fear of exposure to the virus and burdening of already stretched health care systems. Mental health and wellbeing may be negatively affected by anxiety related to social-isolation, limits on household and social mixing, loss of employment or income, and increased burden of childcare.

This study will compare clinical monitoring, hospital admissions and death for diabetes, lung disease, heart disease, and mental health between the UK, Denmark, and Sweden. The study will explore 1) whether these adverse outcomes differ by ethnicity, and social-deprivation, and 2) whether risks have changed over time as population restrictions have been introduced and lifted. Comparisons between countries will help determine whether certain public health approaches are associated with better health outcomes.

People need to attend hospital for investigations and procedures to diagnose and treat illnesses. There is evidence that fear of COVID-19 kept them away in the pandemic, and that hospital acquired COVID-19 occurred and adversely affected some patients who attended. Patients need to know that they can come to hospital safely, and patients and doctors need to understand the risk and impact of acquiring COVID-19 and which factors make its post-operative outcomes worse. Then safe policies can be implemented and monitored, and patients better informed of the risks and benefits of attending hospital for routine care.

Within our hospital we use routinely collected data to measure these risks and monitor impacts of policy changes. However, data in a single trust cannot provide the power to fully characterise the risks pertaining to different patients and procedures. This proposal using national health data will provide more precise, generalisable, repeatable risk estimates appropriately adjusted for differences between patients and procedures. We will use the OpenSAFELY Trusted Research Environment to provide hospital data on procedures and investigations, readmissions, and critical care admissions from secondary care data, deaths, COVID-19 vaccination and testing data, and primary care diagnoses and prescriptions to incorporate pre-existing patient factors and the varying community levels of COVID-19 infections.

This work will provide reusable tools in real-world data for patients attending hospital for care; identifying high risk patients that might need pathway improvements and low risk patients and procedures

The Covid-19 restrictions and lockdowns in England had a dramatic effect on the way people use health services. These effects may have been more pronounced for vulnerable groups, such as people exposed to domestic violence, people with mental health or addiction problems, or children with safeguarding concerns. We aim to analyse the frequency of GP visits over the period before, during and after the lockdowns to see how health care access changed for vulnerable groups and how that compares to the rest of the population. This will help us understand who experienced barriers to GP health care through the pandemic, if those barriers persisted after the lockdowns and how we can improve them.

General practice has been disrupted by the pandemic in many clinical areas, including in the delivery of asthma related care. We aim to assess how the COVID-19 pandemic has impacted the delivery of services for people with asthma and their care. The first step will be to ascertain people with asthma in the electronic health records in the OpenSAFELY platform. Then we will analyse the diagnoses, tests, and how the patient’s care and condition were managed, with reference to national standards of care during the COVID-19 pandemic. We will look to assess the impact of the pandemic on different groups for example those in certain regions, ethnicities or deprivation groups. Our analyses will help inform the NHS recovery from the COVID-19 pandemic, in various ways, such as identifying best practice others can learn from and prioritising high value care activities.

While vaccines remain the best strategy to prevent COVID-19, recent evidence suggests monoclonal antibodies (nMABs) or antivirals could potentially benefit certain vulnerable populations before or after exposure to SARS-CoV-2, such as the unvaccinated or recently vaccinated high-risk patients.

With the recent roll out of nMABs and antivirals, there is an urgent need to for knowledge and understanding around the use of nMABs and antivirals in the treatment of patients with COVID-19, such as factors of relevance in determining nMAB and antiviral treatment and the impact of nMAB and antiviral treatment in real word settings.

This project will cover three main areas relating to the roll out of nMABs and antivirals for non-hospitalised patients with COVID-19; (1) eligibility and uptake, (2) effectiveness and (3) safety.

Pain management is one of the central elements of primary healthcare. Prior research suggests that between one third and one half of all UK adults may be suffering from noncancer-related painful conditions that last more than three months (i.e., chronic pain). The high prevalence of chronic pain (CP) and associated costs of around £12 billion per year to the UK economy warrant a closer look at the impact of COVID-19 on CP management. Longitudinal data from March 2019 to March 2022 are analysed to determine if patients who were treated for CP in March 2020 experienced changes in long-term prescribing (4 or more within 12 months) for controlled analgesics, a shift to stronger/weaker pain medication, as well as treatments for dug dependence or misuse management in acute hospital settings. CP patients are a potentially vulnerable group and it is important to understand if and how the COVID-19 pandemic has disrupted pain management.

A previous analysis of breakthrough infections has found that people on renal replacement therapy (dialysis and transplantation) appear to be more susceptible to breakthrough infections with COVID19 even if they were previously vaccinated.

The aim of this project is to use combined data on kidney function and the requirement of renal replacement therapy from both primary care and UK Renal Registry data as well as hospital codes to quantify how impaired kidney function and modalities of renal replacement therapy modify the risk of breakthrough infections after vaccination.

We will also quantify the protective impact using third as well as booster-vaccinations compared to not having received booster vaccines.

The output will be a research paper and inform national vaccination policy.

We aim to ensure that the most accurate data are used to identify people with advanced kidney disease and those who are undergoing kidney replacement therapy (dialysis and kidney transplantation). This is important as for example people with dialysis may have different risks of COVID19 compared to people who have advanced kidney disease but not yet started dialysis, even if both groups have been already vaccinated. Current data held by GPs will have inaccuracies as GPs may only know about dialysis start in a patient after some delay (in receiving letters), and not all hospitals code dialysis as a hospitalisation, therefore existing data held in OpenSAFELY may be inaccurate.

The COVID-19 pandemic has placed large pressures on the NHS. To help make space to treat patients with COVID-19, the NHS has cancelled, postponed, or changed health care. Some people were discouraged from using services for fear of catching COVID-19 or not wanting to take up resources during a pandemic.

We will study the impacts of health care disruption (e.g., delays and confusion over service delivery) in England. We will study how hospital admissions that could have been prevented have varied during the pandemic. We will look at trends over time, including which types of people were more affected than others. We will share the findings with policy makers, NHS care managers, and politicians to help inform:

• Which services to invest in or prioritise first
• Which types of people need additional help

Our study will help benefit patients and minimise any impacts of this disruption.

The use of triage tools during surge and mass casualty events is recommended in national guidance. The need for such tools was identified again in preparation for the recent Olympic and Paralympic Games.

The evidence base for this guidance was challenged by GPs at the RCGP Pandemic Flu Summit in March 2010. GPs questioned the extent to which they as a group had been involved in the development of the tools. GPs were also concerned that there is little data supporting use of such tools in primary care.

This study aims to validate a previously developed triage tool (for pandemic influenza) called FLUCATs (https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0075384) on patients with COVID-19 . This study will test how well the FLUCATs test how parts of a general practitioner’s questions and assessment of children and adults with COVID-19 like illness can predict who: can safely be kept at home; need hospital admission; need high dependency or intensive care; are most at-risk of dying.

Fit notes (sometimes called “sick notes”) are issued by GPs in the UK where a person needs to demonstrate to their employer that they are unable to work for health reasons for more than 7 days. Given the large number of patients living after being diagnosed with COVID-19, it is important to know the degree to which recovering from COVID affects people’s ability to work. Coding of fit notes by GPs should provide a window on the scale of the continuing public health and economic burden for those recovering from COVID-19. This in turn can assist policy makers in prioritising resources for COVID recovery services.

In parallel with this, measuring the diagnoses that are associated with fit note coding would provide important insights into the types of conditions that are causing substantial continued problems for people recovering from COVID-19. This could provide important information for the continued care of these patients.

This study examines to what extent (rate and duration) fit notes are given to patients with COVID (identified through three routes—those testing positive for SARS-CoV-2, those with a probably diagnostic code in primary care, and those hospitalised with a diagnostic code for COVID) and compare against general populations from 2019 and 2020, as well as an active comparator (patients hospitalised with pneumonia in 2019 for comparison with hospitalised COVID patients). It also evaluates what diagnoses are commonly coded alongside the recording of a fit note. The study also analyses what type of patients get a fit note, according to demographic, clinical, and organisational factors.

Care home residents have restricted mixing with the general population but staff provide a route of infection introduction from the surrounding community. The aim of this analysis was to investigate the timing of first evidence of introduction of COVID-19 to care homes across England, to understand care home characteristics associated with earlier introduction and the relationship between risk of introduction and the surrounding local epidemic.

A rolling analysis over time was planned with the primary outcome being occurrence of a covid-related event (test positive, clinical diagnosis, admission or death) among residents of a care home within the next 14 days from the current time step. Explanatory factors for 14-day risk considered were care home size, whether residential/nursing, proportion of residents with dementia, deprivation index and rural/urban classification of the home’s postcode, and daily incidence of probable COVID-19 infections within the local population of the home’s Middle Super Output Area (MSOA).

The analysis was ultimately retired due to complications with identifying individuals residing in the same household and defining whether or not each household was a care home or private home.

The SARS-CoV-2 Delta variant first emerged in India in late 2020. The Delta variant became the dominant circulating form of SARS-CoV-2 in the UK by summer 2021.

This study considered the feasibility of an analysis to estimate the relative severity of the SARS-CoV-2 Delta variant compared to the previously circulating forms of the virus.

The SARS-CoV-2 Alpha variant was first identified in Kent, UK in autumn 2020. The variant was more transmissible than the prevalent wild type virus and quickly became the dominant circulating form of the virus.

This study aimed to estimate the relative severity of the SARS-CoV-2 Alpha variant compared to the previously circulating wild type virus.

The study population was defined by all people testing positive for SARS-CoV-2 in the window 16th November 2020 to 11th January 2021, with data available on the gene marker that can differentiate between the two virus types (SGTF).

The exposure of interest was infection with the Alpha variant. The outcomes of interest were hospital admission, admission to the intensive care unit (ICU), and all-cause mortality.

The key finding of this research is that infection with the Alpha variant is associated with a 70% increased risk of all-cause death compared to wild type virus.

There is some thought that viruses that cause respiratory illness may compete with one another so that only one can become established in the host at a time. Further, it may be possible that respiratory illnesses other than COVID-19 may induce some immune protection against COVID-19.

This study aimed to estimate the impact of non-COVID-19 respiratory infections on the incidence of COVID-19 infection and its severity among the general population.

The exposure of interest was infection with non-COVID-19 respiratory infections. The outcomes of interest were infection with COVID-19 and the resulting outcome of COVID-19 infection.

Around a third of people admitted to hospital with COVID-19 develop kidney damage (known as acute kidney injury or AKI). While AKI is often recoverable, it can sometimes lead to worsening of long-term kidney function and in some cases, permanent kidney failure requiring long-term dialysis or a kidney transplant. The complications caused by long-term kidney damage exert considerable stress on the NHS and it is important that we know what to expect to be able to ensure that services are planned with sufficient capacity to cope. For example, growing numbers of people requiring dialysis will require investment in the workforce and infrastructure. We plan to investigate what happens to long-term kidney function after COVID-19 using electronically coded GP records combined with hospital records routinely-collected from day-to-day patient care inEngland.

Our study will include adults without kidney failure to begin with (i.e. those not on dialysis or a kidney transplant). We will compare people with and without COVID-19 for differences in the rates of the following kidney outcomes:

1. Newly-diagnosed kidney failure,
2. 50% or more reduction in kidney function from before,
3. Further AKI.

We will be able to determine whether outcomes vary based on the severity of COVID-19, based on time (e.g. during peaks, after widespread use of effective anti-COVID-19 drugs such as steroids), and based on levels of vaccination. We will account for factors such as age, sex, ethnicity, socioeconomic deprivation, smoking, BMI, and pre-existing medical problems such as diabetes.

As kidney failure is relatively rare in the general population, very large studies are needed to be able to investigate whether or not a group is at increased risk compared to other groups. Because of the size of the population in OpenSAFELY, we will be uniquely placed to be able to detect differences in this outcome due to COVID-19.

The pandemic put pressure on routine healthcare services such as family practices, who are usually responsible for giving a range of childhood vaccines, up to the age when children start school. We set out to investigate the impact of the pandemic on childhood immunisations, for example to explore whether they were more likely to be given later or missed altogether. We aimed to produce a prototype of a dashboard which would show, nationally and at local levels, the impacts of the pandemic on childhood immunisation rates and any “catch-up” action which may be required.

Patients with COVID-19 are thought to be at higher risk of cardiometabolic and pulmonary complications, but quantification of that risk is limited. We aimed to describe the overall burden of these complications in survivors of severe COVID-19.

Working on behalf of NHS England, we used linked primary care records, death certificate and hospital data from the OpenSAFELY platform. We constructed three cohorts: patients discharged following hospitalisation with COVID-19, patients discharged following hospitalisation with pneumonia in 2019, and a frequency-matched cohort from the general population in 2019. We studied eight cardiometabolic and pulmonary outcomes. Absolute rates were measured in each cohort and Cox regression models were fitted to estimate age/sex adjusted hazard ratios comparing outcome rates between discharged COVID-19 patients and the two comparator cohorts.

Cardiometabolic and pulmonary adverse outcomes were markedly raised following hospitalisation for COVID-19 compared to the general population. However, the excess risks were more comparable to those seen following hospitalisation with pneumonia. Identifying patients at particularly high risk of outcomes would inform targeted preventive measures.

Some people who have had COVID-19 (the disease caused by the coronavirus responsible for the 2020 pandemic) complain of ongoing symptoms, sometimes for many months, after they recover from their initial infection. While we know that ongoing post-COVID-19 symptoms occur, we do not yet know how often this happens, or the full spectrum of possible symptoms. We want to know whether doctors decide that the symptoms are due to COVID-19 or another disease. We would like to find out which, if, any, symptoms resolve over time and how quickly or slowly this happens. Finally, we also do not know whether there are any factors that can predict which groups of people are more likely to have continuing symptoms after they have COVID-19, specifically in relation to age, sex and vaccination status.

It is vital that we know the answers to these questions so that we are better able to support those struggling with ongoing symptoms and plan the health services they need.

Therefore, we plan to investigate these questions using routinely recorded health information from very large numbers of the UK general population. This routine healthcare information is recorded every time someone has any sort of contact with a general practitioner or a hospital, making it a large and powerful source of information for answering these questions. We access this information in a very safe way to ensure that we never see any individual’s health records, we only ever see the results of our studies in groups.

This study is investigating the use of sodium glucose cotransporter 2 inhibitors (SGLT2is) on the risk of COVID-19 outcomes in the general population.

To describe quality annd completeness of ethnicity recording in TPP, EMIS linked data sources and validate against 2021 Census ethnicity from ONS.

This study aimed to look at whether COVID-19 increases the risk of incident type 1 diabetes in children and adults.

We want to understand the impacts of long COVID on the quality-of-life of people who have developed this condition. For many health conditions, researchers ask people to fill in standardised surveys which measure quality-of-life, which help them understand how much the condition is affecting people. These kinds of measures are used in planning the resources for health systems, because they allow a standard way of comparing different health conditions, and, if people fill them in repeatedly, tracking changes in wellbeing over time.

These quality-of-life measurements are not currently available for people with long COVID who did not need to be hospitalised during their COVID-19 episode. Our study aims to collect these measurements, by using a smartphone app linked to a patient’s GP records. We will ask people to voluntarily participate in the study and use the app, and if they do so to give their consent for filling the questionnaires and for linkage to their health records.

These results will tell us if long COVID has different quality-of-life impacts among different age groups, ethnicities, geographic regions, or because of any underlying health conditions.

We will convert the results into standardised measurements that are used in the NHS to assess the impact of illnesses. This means they will be useful for working out how much the effects of long COVID cost the health service, and how we should prioritise interventions such as booster vaccines to avoid more people getting long COVID.

Together with other researchers studying long COVID we will provide results to support long-term care. At all stages we will work with people affected by long COVID to ensure their input is central to designing the study plans and interpreting the outputs.

Loneliness and living alone have been linked with poor mental and physical health. Evidence suggests that the COVID pandemic exacerbated feelings of loneliness, social isolation and connectivity, due to national lockdowns, social distancing rules and shielding. There is also evidence that the pandemic increased psychological distress including feelings of anxiety and depression, and perhaps particularly for those living alone.

This study will use a large cohort of adults to see how diagnoses of mental health conditions changed before and during the COVID pandemic, and how this differed among people living alone and not living alone.

To provide background rates of adverse events in populations similar to those recruited in COVID-19 vaccine trials. This helps to contextualise safety outcomes in vaccine trials.

To describe the occurrence and timings of key COVID-19 related clinical outcomes, such as positive tests, hospital admissions, and death.

The national COVID-19 vaccination effort in the United Kingdom continued in September 2021 with third doses administered to the most vulnerable, and with eligibility extended to the entire adult population by the end of the year.

This study will assess effectiveness of a third dose of COVID-19 vaccine compared with a complete 2-dose vaccination course.

PINCER is a pharmacist-led information technology intervention that aims to reduce the risk of patient harm due to medication errors as captured by 13 indicators of hazardous prescribing or delayed medication monitoring. The COVID-19 pandemic has been shown to have caused significant disruption to healthcare services. This study examines the rate of hazardous prescribing or delayed monitoring as indicated by each individual PINCER indicator between September 2019 and September 2021 to assess the impact of the pandemic.

The HbA1c test, used to measure blood glucose level, is a key primary care service that has been disrupted by the COVID-19 pandemic. Getting an accurate read of blood glucose level is necessary to make new diabetes diagnoses and detect increased risks of complications for existing diabetics. Furthermore, diabetes has been linked to greater COVID-19 severity and mortality, making regular monitoring and management of blood glucose levels more crucial.

This study aims to describe the impact of service disruption on HbA1c test results across 23.8 million patients’ records in England. It presents cohort-based observational analyses to examine metrics such as the number of HbA1c tests recorded, percentage of abnormal results, and mean of test values between January 2019 and June 2021, focusing on breakdowns by healthcare inequalities, which include age group, sex, ethnicity, region, IMD, diabetes, learning disability, and mental health conditions.

Safe and effective vaccination is vital for the public health response to the COVID-19 pandemic. The COVID-19 pandemic prompted rapid vaccine development and roll-out. Each vaccine demonstrated a reassuring safety profile in large clinical trials before licensing for use. However, post-licensing vaccine safety surveillance remains important for vaccine safety and public confidence. On March 11 2021, the Danish authorities temporarily paused delivery of COVID-19 vaccination while investigating a report of a fatal blood clot in a person who had received the AstraZeneca COVID-19 adenovirus vector vaccine. A new syndrome of thrombosis with thrombocytopaenia has been described, prompting further investigation of venous thromboembolism events following vaccination.

This study aims to investigate the associations of each of Pfizer BioNTech and AstraZeneca COVID-19 vaccines with any venous thromboembolism (VTE) in England. The study will also look at the specific outcomes of pulmonary embolism and cerebral venous thrombosis.

Primary analysis will use a self-controlled case series (SCCS) design, a case-only design which compares incidence of events within the same individual across different time periods. A separate analysis will be conducted for each vaccine brand. Vaccination status will be obtained from the primary care record. We will use primary care and hospital admission records to define episodes of venous thromboembolism.

A secondary objective is to investigate the extent to which self-controlled case series (SCCS) and cohort study designs differ in their estimations of any associations of COVID-19 vaccination with venous thromboembolic events, to aid interpretation of findings from other observational vaccine safety studies. A secondary analysis will use a cohort study design with both historical and concurrent comparator groups. A cohort study design can be a useful complement to investigate the robustness of findings from SCCS analysis in safety studies, since they have a different profile of biases. We will use this analysis to explore the effect of study design on estimates of vaccine safety.

This study is a partnership between the OpenSAFELY collaboration, the NIHR Health Protection Research Unit (HPRU) in Vaccines and Immunisation (a partnership between the London School of Hygiene and the UK Health Security Agency), the NIHR HPRU in Emerging and Zoonotic Infections at University of Liverpool, and the Medicines and Healthcare products Regulatory Agency (MHRA).

Safe and effective vaccination is vital for the public health response to the COVID-19 pandemic. The COVID-19 pandemic prompted rapid vaccine development and roll-out. Each vaccine demonstrated a reassuring safety profile in large clinical trials before licensing for use. However, post-licensing vaccine safety surveillance remains important for vaccine safety and public confidence. This study aims to investigate for any association of COVID-19 vaccination with each of Guillain-Barré syndrome, transverse myelitis and Bell’s palsy.

Primary analysis will use a self-controlled case series (SCCS) design, a case-only design which compares incidence of events within the same individual across different time periods. A separate analysis will be conducted for each combination of vaccine brand and outcome. Vaccination status will be obtained from the primary care record. We will use primary care and hospital admission records to define episodes of Guillain-Barré syndrome, transverse myelitis and Bell’s palsy, and will additionally use emergency care attendance data to enhance ascertainment of Bell’s palsy.

We also plan to follow with a secondary analysis using a cohort study design with both historical and concurrent comparator groups. A cohort study design can be a useful complement to investigate the robustness of findings from SCCS analysis in safety studies, since they have a different profile of biases. We will use this analysis to explore the effect of study design on estimates of vaccine safety.

This study is a partnership between the OpenSAFELY collaboration, the NIHR Health Protection Research Unit (HPRU) in Vaccines and Immunisation (a partnership between the London School of Hygiene and the UK Health Security Agency),the NIHR HPRU in Emerging and Zoonotic Infections at University of Liverpool, and the Medicines and Healthcare products Regulatory Agency (MHRA).

Many studies have been conducted that estimate the effect of COVID-19 vaccines in reducing infection or disease compared with unvaccinated people. However, there have been no randomised controlled trials that have made direct head-to-head comparisons. Such comparisons are difficult in observational data without the rollout of both vaccines across the same population at the same time. The concurrent rollout of the Pfizer-BioNTech vaccine and the Oxford-AstraZeneca vaccine in health and social care workers in England provides a rare opportunity to do this.

We compared effectiveness in around ⅓ million health and social care workers in England, vaccinated between 4 January and 28 February 2021, and followed up for 20 weeks. Primary outcomes were recorded SARS-CoV-2 infection, COVID-19-related accident and emergency attendance, and COVID-19-related hospital admission.

We found very little difference in between the vaccines on the studied outcomes. Whilst some material differences in the protective effect may exist, both are extremely effective for reducing COVID-19 disease and there were too few events to reliably detect such differences up to 20 weeks after vaccination.

The COVID-19 vaccines were developed and rigorously evaluated in randomized trials within a year of the first reports of COVID-19 in Wuhan, China. However, many questions about vaccine effectiveness were not answered by randomized trials, for example protection for sick or elderly people, or against severe outcomes such as hospitalisation. The UK started its general population COVID-19 vaccination programme in early December 2020, prioritising vaccination in the most vulnerable people, in particular elderly people, who were not well-represented in trials.

We assessed vaccine effectiveness amongst >2.6 million people aged over 70 in England, using linked routine clinical data from the OpenSAFELY TPP database. We looked at the vaccine’s ability to reduce infection, hospitalisation, and death compared with those who were not vaccinated. We studied both the Pfizer-BioNTech and the Oxford-AstraZeneca vaccines.

We found that the vaccines were highly effective. However, we also found evidence of biases in these estimates. For example, the estimated protective effect of vaccination was very high in the first two weeks after vaccination, yet we know from trials that immunity does not develop until after the first week.

We discuss the challenges of estimated vaccine effectiveness in observational data, using our study as an exemplar.

The vaccination programme for COVID-19 in the UK started on 8th December 2020. As with any vaccine, despite being vaccinated, COVID-19 vaccine breakthrough is likely and a small number of people will still fall ill, be hospitalised, or die from COVID-19.

As of 30th June 2021, a total of 10,782,870 individuals were identified as being fully vaccinated against COVID-19, with a median follow-up time of 43 days (IQR: 23-64). From within this population, the total number of COVID-19 vaccine breakthrough cases was 17,030 (0.2%). The majority of COVID-19 vaccine breakthrough cases were mild with relatively few fully vaccinated individuals being hospitalised or dying as a result. However, some concerning differences in rates of breakthrough cases were identified in several clinical groups; comorbidities with the highest rates of breakthrough COVID-19 included renal replacement therapy, organ transplant, haematological malignancy, and immunocompromised

The continued increase in numbers of positive SARS-CoV-2 tests in England are concerning and, as numbers of fully vaccinated individuals increases and follow-up time lengthens, so too will the number of COVID-19 breakthrough cases.

To help inform roll-out of the third and booster dose vaccination, and to support the redesignation of dialysis patients into the 3rd dose group, with eligibility for subsequent vaccination, this analysis was rerun in November 2021 (with the previous CKD group split out into categories). As expected, there had been a rise in all rates (calculated per 1000 person-years);

• Positive SARS-CoV-2 test from 12.33 (12.14-12.51) to 99.44 (99.31-99.57)
• COVID related hospitalisation from 0.70 (0.65-0.74) to 2.31 (2.29-2.34)
• COVID related death from 0.12 (0.1-0.14) to 0.64 (0.63-0.66)

Despite these increases, reassuringly, breakthrough cases remain relatively mild; 312, 530 Positive SARS-CoV-2 tests vs 3840 COVID-19 related deaths (with the majority (80%) of deaths occurring in those over 70 years of age).

This project was initiated to quantify care home mortality on the care home level as so called excess mortality. After initial feasibility work, this was paused due to the lack of a time-updated household identifier.

This short data report compared different strategies for identifying care home residents in electronic health records in England. Historically, GPs do not enter whether or not a person is living in a care home in their medical record. There is also no central registry of care home residents. Therefore, when we want to study people in care homes using electronic health records from GP surgeries, we need to rely on other methods to identify these people.

In this report we compared using medical codes, an address linkage, and household information to identify care home residents. We describe how these methods are implemented in OpenSAFELY, and compared the overlap between them.

The COVID-19 pandemic has had a significant impact on the delivery of care within the NHS. We have developed the OpenSAFELY Service Restoration Observatory to describe this impact on primary care activity and monitor the restoration of service. We produced 11 measures reflective of clinical activity in general practice. A substantial drop in activity was observed in all measures at the outset of the COVID-19 pandemic. By April 2021, the median rate had recovered to within 15% of the median rate in April 2019. The remaining measures showed a sustained drop, ranging from a 18.5% reduction in medication reviews to a 42.0% reduction in blood pressure monitoring. Three measures continued to show a sustained drop by December 2021. Since December 2021 activity rates have recovered to (or in some cases above) the pre-pandemic baseline for measures involving blood tests including cholesterol, liver function, thyroid, full blood counts, HbA1c, and renal function testing as well as for medication review activity.

This study aimed to quantify the mortality risk among care home residents compared to residents of private homes during the first and second wave of the pandemic. We used a technique called standardisation to account for age differences between care home residents and residents of private homes, and described how the mortality risk changed between February 2019 and March 2021.

We found that care home residents had an approximately 10-fold higher risk of death compared to residents of private homes prior to the first pandemic wave; this increased to an approximately 17 times higher risk during the first wave. Despite absolute increases in the risk of death during the second wave, the relative mortality risk for care home residents remained stable at approximately 10-fold higher compared to residents of private homes.

These analyses were largely descriptive, and as such we cannot say why there was a spike in the risk of death for care home residents during the first but not the second wave. Plausible explanations include improved protections in care homes such as testing and better use of personal protective equipment (PPE), pre-existing immunity from the first wave among residents and staff, or an element of mortality displacement. The peak of the second wave occurred prior to the point where the UK vaccination program is anticipated to have an effect. This study quantifies the impact of the pandemic on care home residents in England, and suggests a number of important questions for future research to ensure that this vulnerable population is protected for the next pandemic.

National guidance was issued during the COVID-19 pandemic to switch patients on warfarin to direct oral anticoagulants (DOACs) where appropriate as these require less frequent blood testing. DOACs are not recommended for patients with mechanical heart valves. We conducted a retrospective cohort study of DOAC prescribing in people with a record of a mechanical heart valve between September 2019 and May 2021, and describe the characteristics of this population. We identified 15,457 individuals with a mechanical heart valve recorded in their records, of whom 1058 (6.8%) had been prescribed a DOAC during the study period. 767 individuals with a record of a mechanical heart valve were currently prescribed a DOAC as of May 31st 2021.

COVID-19 disease in children and young people is typically mild or asymptomatic, however the role of children and young people in transmission of SARs-CoV-2 remains unclear. At the start of the COVID-19 pandemic, schools were shut, on the premise that children may be vectors of COVID-19, as they are for other respiratory diseases, such as seasonal influenza. Evidence on the absolute impacts of educational settings being closed on community transmission is mixed; quantifying the impact is challenging, as closures have been implemented alongside other public health measures, infection control measures vary across educational settings and school opening may be associated with other behavioural changes. The negative impacts of school closures on the educational, developmental and mental health of children is undisputed.

In our previous analysis we showed during the first wave of the UK pandemic, when schools were largely closed, for adults 65 years and under living with children there was no evidence of a markedly increased risk of SARS-CoV-2 infection or severe COVID-19 outcomes. However, during wave 2 (1st September to 19th December 2020) when schools were largely open, there was evidence of increased risk of infection and hospitalisation with COVID-19, although the risk of death from COVID-19 was not increased in either wave.

In this analysis, we examine the risk of SARS-CoV-2 infection and outcomes from COVID-19 associated with living with children compared to those without during 2021, and the impact that adult vaccination had on that risk.

This was an early look at work which was later written up for publication (see Risks of COVID-19 hospital admission and death for people with learning disabilities - a cohort study.)

Two cohorts were assembled, reflecting the two first waves of COVID-19 experienced in the UK, comprising males and females aged up to 105 years: (1) wave 1, from 1st March 2020 until 31st August 2020; (2) wave 2, 1st September 2020 until 31st December 2020 (for admissions) or 8th February 2021 (for deaths). The main outcomes were COVID-19 related death and COVID-19 related hospitalisation. Various risk factors were included, with particular focus on whether or not the person was on the learning disability register.

Generally, age was the strongest factor associated with both outcomes, but much stronger for COVID-19 death than mortality. Various comorbidities were associated with increased absolute risks, aligning with the earlier work on risk factors (Nature, 2020). Learning disabilities was noted as being associated with increased risks (leading to subsequent publication focused on that question).

This study aimed to assess the association between learning disability and risk of hospitalisation and mortality from COVID-19 in England among adults and children.

Two cohorts were assembled, reflecting the two first waves of COVID-19 experienced in the UK, comprising males and females aged up to 105 years: (1) wave 1, from 1st March 2020 until 31st August 2020; (2) wave 2, 1st September 2020 until 31st December 2020 (for admissions) or 8th February 2021 (for deaths). The main exposure group was people included on a general practice learning disability register (LDR), with a subgroup of people classified as having profound or severe learning disability. We also identified patients with Down syndrome and cerebral palsy. The main outcomes were COVID-19 related death and COVID-19 related hospitalisation. Non-COVID-19 related death was also explored.

In wave 1, of 14,301,415 included individuals aged 16 and over, 90,095 (0.63%) were identified as being on the LDR. 30,173 COVID-related hospital admissions, 13,919 COVID-19 related deaths and 69,803 non-COVID deaths occurred; of which 538 (1.8%), 221 (1.6%) and 596 (0.85%) were among individuals on the LDR, respectively. Wave 1 hazard ratios for individuals on the LDR, adjusted for age, sex, ethnicity and geographical location, were 5.3 (95% confidence interval (CI) 4.9, 5.8) for COVID-19 related hospital admissions and 8.2 (95% CI: 7.1, 9.4) for COVID-19 related death. Wave 2 produced similar estimates.

Associations were stronger among those classed as severe-profound and among those in residential care. Down syndrome and cerebral palsy were associated with increased hazard of both events in both waves; Down syndrome to a much greater extent. Hazards of non-COVID-19 related death followed similar patterns with weaker associations.

People with learning disabilities have markedly increased risks of hospitalisation and mortality from COVID-19. This raised risk is over and above that seen for non-COVID causes of death. Ensuring prompt access to Covid-19 testing and health care and consideration of prioritisation for COVID-19 vaccination and other targeted preventive measures are warranted.

Obtaining accurate estimates of risk is challenging in the context of changing levels of circulating infection, which standard risk prediction models commonly ignore.

We adopt a landmarking approach for risk prediction to dynamically incorporate time- and region-dependent information on infection prevalence. We compare this approach with more commonly used Cox models that do not take infection prevalence into account, in terms of discriminaton and calibration. We also compare simpler models to more richly specified ones.

The target population was adults registered at a general practice in England in March 2020. The outcome was 28-day COVID-19-related-death. Predictors included demographic characteristics and comorbidities. Three proxies of local infection prevalence were used: model-based estimates; rate of COVID-19-related attendances in emergency care; and rate of suspected COVID-19 cases in primary care.

Prediction models were developed on 11,972,947 individuals, of whom 7,999 experienced COVID-19-related-death. All models discriminated well between individuals who did and did not experience COVID-19 mortality, including simple models adjusting only for basic demographics and number of comorbidities: C-statistics 0.92-0.94. However, absolute risk estimates were substantially miscalibrated when infection prevalence was not explicitly modelled.

Our study suggests that models that ignore the infection prevalence provide poorly calibrated estimates of absolute risk; models that include time-varying measures of the infection prevalence can provide more accurate estimates. Simple models based only on number of comorbidities and basic demographics performed almost as well as more complex risk prediction models, both within models including infection prevalence and models that ignore this, suggesting that policies targeting population level reduction of COVID-19 mortality risk may not need to distinguish between all comorbidities in detail.

The COVID-19 vaccination programme was launched in England on 8th December 2020. Limited information was available on vaccination coverage in detailed demographic and clinical subgroups.

We deployed a data monitoring framework using publicly auditable methods and a secure, in-situ processing model, using linked but pseudonymised patient-level NHS data on 57.9 million patients. We describe vaccine coverage and time trends across a range of subgroups of Joint Committee on Vaccination and Immunisation (JCVI) priority cohorts.

Our study highlights lower vaccination coverage among certain key groups: ethnic minorities, those living in areas of higher deprivation, and individuals living with severe mental illness or learning disabilities; where targeted activity may be needed to ensure equitable protection against COVID-19.

Ethnic differences in the risk of severe COVID-19 may be linked to ethnic differences in household composition. In this study our aim was to assess whether household composition (= the number of age-based generations in a household) was associated with risk of severe COVID-19 for older individuals (people aged 67 or older) after taking account of other risk factors for severe COVID-19, and whether this association differed by ethnicity.

We created a cohort that included all people over the age of 67 in OpenSAFELY, and analysed whether the number of younger generations that the 67+ year old lived with was associated with their risk of being hospitalised or dying from COVID-19. We performed separate analyses for wave 1 (01/02/2020-31/08/2020) and wave 2 (01/09/2020-31/01/2021) in the UK, and analysed the effect of the following living arrangements: 67+ year old living alone, 67+ year old living with other 67+ year olds, 67+ year old living with people from 1 younger generation, 67+ year old living with people from 2 younger generations, 67+ year olds living with people from 3 younger generations (with distinct younger generations defined as 0-18, 18-29, 30-66). For all analyses, we assessed whether observed effects were different by ethnic group (defined as White, Black, South Asian, Mixed and Other).

Initial results showed that in Wave 1 there was no association between household composition and severe COVID-19 for any ethnicity, but that in wave 2 living with increasing numbers of younger generations was associated with an increased risk of severe COVID-19 for White and for South Asian people. We also found that older South Asian people are over-represented within multi-generational household in the most deprived settings, and older people in these settings have a particularly high risk of severe COVID-19.

All patients in England within vaccine priority groups were offered a COVID-19 vaccine by mid-April 2021. Clinical record systems contain codes to denote when such an offer has been declined by a patient (although these can in some cases be entered for a variety of other reasons including vaccination delay, or other administrative issues). We set out to describe the patterns of usage of codes for COVID-19 vaccines being declined.

We identified all patients in key vaccine priority groups: aged over 50, or over 16 and at increased risk from COVID-19 (Clinically Extremely Vulnerable [CEV] or otherwise “at risk”). We describe the proportion of patients recorded as declining a COVID-19 vaccination for each priority group, and by other clinical and demographic factors; whether patients recorded as “declined” subsequently went on to receive a vaccination; and the distribution of code usage across GP practices.

We found that clinical codes indicative of COVID-19 vaccinations being declined are widely used in English general practice. Of 24.5 million patients in priority groups as of May 25th 2021, 89.2% had received a vaccine, 8.8% had neither a vaccination nor a decline recorded, and 2.7% had a decline code recorded. Of patients with a recorded decline, 18.9% were subsequently vaccinated.

Declines were substantially more commonly recorded among Black and South Asian patients, and patients from more deprived areas. There is a need for more detailed survey and/or qualitative research with patients and clinicians to determine the most common reasons for these recorded declines.

The COVID-19 pandemic has disrupted healthcare activity. The NHS stopped non-urgent work in March 2020, later recommending services be restored to near-normal levels before winter where possible.

We aimed to describe the volume and variation of coded clinical activity in general practice across the whole population in OpenSAFELY-TPP between January 2019-September 2020, taking respiratory disease and laboratory procedures as examples.

We found that activity recorded in general practice declined during the pandemic, but largely recovered by September 2020. There was a large drop in coded activity for laboratory tests, with broad recovery to pre-pandemic levels by September. One exception was International Normalised Ratio test, with a smaller reduction. The pattern of recording for respiratory symptoms was less affected, following an expected seasonal pattern and classified as “no change”. Respiratory infections exhibited a sustained drop not returning to pre-pandemic levels by September. Asthma reviews experienced a small drop but recovered, while COPD reviews remained below baseline.

In conclusion: We delivered an open source software framework to describe trends and variation in clinical activity across an unprecedented scale of primary care data. The COVID-19 pandemic led to a substantial change in healthcare activity.

From the beginning of the COVID-19 pandemic, the NHS recommended putting measures in place to reduce face-to-face contact with patients where possible to reduce the risk of transmission. This included guidance to consider switching warfarin patients to a direct acting oral anticoagulant (DOAC) where suitable. This aimed to reduce the number of times patients had to attend for blood tests. However, it was not known how widespread this practice was, whether it resulted in potentially dangerous co-administration of both drugs in some patients, or whether those patients who remained on warfarin were tested frequently and maintained good clotting results.

We identified patients who were taking warfarin prior to the pandemic and the factors associated with being switched to DOACs.

This is the first widespread study on the impact of the NHS England guidance. We showed that a substantial number of patients in general practice in England had their anticoagulant switched from warfarin to a DOAC. We also showed that there was a small number of patients who potentially received both warfarin and a DOAC. There did not appear to be a widespread reduction in INR control during the pandemic.

This study highlights that national prescribing guidance can be implemented in primary care in England, both on a large scale and at a fast pace when necessary to improve patient care. It also highlights the importance of good clinical governance processes when switching medications, in order to reduce the risk of co-prescribing of both agents.

There is concern about medium to long-term effects of COVID-19. The aim of this study was to investigate whether people have a higher risk of hospital admission and death after being discharged from hospital following illness with COVID-19.

We used anonymised electronic health records from English hospitals to identify people who had been admitted to hospital with COVID-19 and then later discharged. We also gathered further details about these individuals from linked GP records. We then looked at whether these patients were re-admitted to hospital later, or died (using national death registration data). For comparison, we also used data from before the pandemic and looked at hospitalisations and deaths in people who had previously been in hospital with influenza, and in people who had not been in hospital at all.

24,673 people who had been discharged from hospital following illness with COVID-19 were identified. These individuals had higher risks of returning to hospital and dying compared with people who had never been in hospital. They also had higher risks of dying than people who had been in hospital with influenza prior to the pandemic, but similar risks of being readmitted to hospital.

We concluded that the higher risks for rehospitalisation and death in people discharged after a COVID-19 hospital admission compared with the general population might create a substantial extra burden on healthcare.

(1) It was reported that people with severe COVID-19 are at risk of blood clots. Warfarin is a commonly used blood thinning treatment. A previous study suggested that it might increase the risk of severe COVID-19 related outcomes due to the process of how it produces an effect in the human body to prevent blood clots, which is different from its alternative treatment (direct oral anticoagulants, DOACs). However, little is known about the impact of routine use of warfarin on COVID-19 related blood clots.

Therefore, we studied a large number of people with atrial fibrillation (a medical condition of an irregular heart rhythm that can lead to blood clots in the heart) who were prescribed warfarin or DOACs as part of normal clinical care and followed them up to compare the effects of warfarin on several COVID-19 related outcomes with DOACs.

Overall, this study shows no evidence of harmful effects of warfarin on severe COVID-19 disease, compared with DOACs. This study provides reassurance about the safety of warfarin use among patients with atrial fibrillation in the context of the COVID-19 pandemic.

(2) Early studies showed that blood thinning treatment (anticoagulants) reduce the risk of thrombotic events amongst those infected. However, current evidence of the role of routinely prescribed oral anticoagulants (OACs) in COVID-19 outcomes is unclear.

Therefore, we studied a large number of people with atrial fibrillation (a medical condition of an irregular heart rhythm that can lead to blood clots in the heart) who had a low risk of stroke. Among this study cohort, the decision of prescribing OAC would typically be borderline. We followed them up to compare the effects of those who prescribed oral anticoagulants on several COVID-19 related outcomes with those who did not .

In conclusion, this study shows that people receiving OACs had a lower risk of testing positive for SARS-CoV-2 and severe COVID-19 outcomes than non-users. This might be explained by the potential effect of OACs in preventing severe COVID-19 outcomes or more cautious behaviours leading to reduced infection risk.

COVID-19 has disproportionately affected minority ethnic populations in the UK. Our aim was to quantify ethnic differences in SARS-CoV-2 infection and COVID-19 outcomes during the first and second waves of the COVID-19 pandemic in England.

GP data were linked to data from other sources on the outcomes of interest: SARS-CoV-2 testing and positive test results and COVID-19-related hospital admissions, intensive care unit (ICU) admissions, and death. Self-reported ethnicity as captured on the GP record was used, grouped into five main categories (White, South Asian, Black, other, and mixed) and 16 subcategories.

Our findings show that minority ethnic groups have higher risks of COVID-19-related hospitalisation, ICU admission, and death even after accounting for other medical conditions individuals may be suffering from; in particular, the risks of COVID-19-related hospitalisation and death in wave 2 were increased for South Asian groups and reduced in all other ethnic minority groups relative to the White group.

The causes of the worse outcomes for minority ethnic groups is likely to involve many factors and require action across many fronts, including reducing structural inequalities, addressing barriers to equitable care, and improving uptake of testing and vaccination.

This study aimed to investigate whether people with HIV had different risks of dying from COVID-19 compared with people without HIV.

We used anonymised electronic health records from GP practices and information from national death registrations to compare COVID-19 death rates between people who had a record of HIV in the GP record and people with no record of HIV. We also took account of other factors, like age and other illnesses, that might affect risk of COVID-19 death.

Over 17 million people were included, and 27 480 had a record of HIV. People living with HIV had higher risk of COVID-19 death than those without HIV after taking account of other factors and this was particularly pronounced in people of black ethnicity.

We concluded that people with HIV in the UK seem to be at increased risk of dying from COVID-19. Targeted policies should be considered to address this raised risk as the pandemic response evolves.

The aim of this study was to answer the question: to what extent is a person’s risk of dying from COVID-19 governed by their broader risk of death, based on their age, medical history, and other characteristics?

We used anonymised routinely collected GP records data from over 17 million people and information on deaths and causes of death from national death registrations. We looked at all deaths after February 2020 and analysed how individual characteristics like age, ethnicity, and medical history were linked to both COVID and non-COVID deaths.

We found that a person’s individual characteristics broadly affected their risk of COVID-19 death and non-COVID deaths in similar ways. But there were exceptions, most notably ethnicity: non-white groups had higher risks of dying from COVID-19 death, despite generally lower risks of dying from non-COVID causes.

We concluded that COVID-19 largely multiplies existing risks faced by patients, with some notable exceptions. Identifying the unique factors contributing to the higher COVID-19 mortality risk among non-white groups is a priority to inform efforts to reduce deaths from COVID-19.

In the early pandemic, there have been concerns that non-steroidal anti-inflammatory drugs (NSAIDs, such as ibuprofen, naproxen, diclofenac, etc.) may increase the risk of COVID-19 related death, but the evidence for this was lacking. This study investigated the routine use of NSAIDs and risk of COVID-19 related death using data from patients in a database covering 40% of the population of England.

In the first study we identified people with an NSAID prescription in the last 3 years from the general population. We compared the risk of COVID-19 related death among these people with those who were not prescribed NSAIDs.

In the second study we identified people with rheumatoid arthritis/osteoarthritis and divided them into two groups: those who were prescribed NSAIDs and those who were not prescribed NSAIDS. We compared the risk of COVID-19 related death between these two groups.

In both studies we found no evidence of a harmful effect of routinely prescribed NSAIDs on COVID-19 related deaths. We concluded that becoming infected by COVID-19 should not influence decisions about the routine prescribing of NSAIDs.

This is the first study, to our knowledge, to investigate the ongoing routine use of hydroxychloroquine and risk of COVID-19 mortality in a general population. In this study, which uses a database that includes around 40% of the population of England, we investigated whether routine use of hydroxychloroquine before the COVID-19 outbreak (ie, pre-exposure) prevented COVID-19 death.

We found no evidence to support a substantial benefit of hydroxychloroquine in preventing COVID-19 death. At the same time, we have shown no evidence of harm. The absence of clear harms or benefits would therefore support ongoing clinical trials using hydroxychloroquine.

This study aimed to evaluate how regular use of a type of drug used to treat respiratory diseases, inhaled corticosteroids (ICS), might have influenced the risk of dying from COVID-19 during the first wave of the pandemic. We studied two groups of people - those with asthma, and those with chronic obstructive pulmonary disease (COPD) - and compared their risk of death from COVID-19. We used a number of techniques to try and account for the fact that people who receive treatments such as ICS are different from those who do not.

We found that:

• For people with COPD, those who were prescribed ICS were at higher risk of dying from COVID-19 compared to those not receiving ICS treatments.
• For people with asthma, those receiving high-dose (but not low-dose) ICS were at higher risk of dying compared to those individuals receiving a non-steroid based reliever inhaler (called salbutamol).

However, our study results also suggest that the higher risk of death in individuals taking ICS might not be fully related to the use of the ICS medication, and might also be related to other factors, for example, people using ICS typically have more severe lung disease, which might explain the reason why they have a higher risk of COVID-19 related death.

We therefore concluded that our study provided no convincing evidence that use of ICS was associated with an increased risk of dying from COVID-19 during the first wave of the pandemic, which should be seen as reassuring for people with these respiratory diseases who rely on ICS to manage their respiratory symptoms.

We used the newly created OpenSAFELY platform to examine factors associated with COVID-19-related death. General practice records of 17,278,392 adults were linked to 10,926 COVID-19-related deaths. A range of factors, including demographic characteristics and comorbidities were explored to determine their relationship with COVID-19 related death.

COVID-19-related death was associated with: being male; greater age and deprivation; diabetes; severe asthma; and various other medical conditions. Compared with people of white ethnicity, Black and South Asian people were at higher risk, even after adjustment for other factors.

We measured a range of factors associated with COVID-19-related death in one of the largest studies of COVID-19 on this topic at this point in time, very early in the pandemic.

Atrial fibrillation is a heart condition characterised by an irregular heartbeat. Patients with atrial fibrillation have an increased risk of stroke and NICE recommends they are treated with drugs that reduce that risk of stroke (blood thinners or anticoagulants). Warfarin used to be the drug of choice but new anticoagulants that are safer and more convenient to take have been preferred for the past years. This is because warfarin requires adjusting the dose on a regular basis based on blood tests, which are a burden for patients and doctors. For safety reasons, during the COVID-19 pandemic, doctors were encouraged to switch patients from warfarin to the new anticoagulants. However, it is uncertain how many patients were actually switched to the new anticoagulants over time and which proportion of patients are currently prescribed the new anticoagulants as per NICE guidelines. Therefore, this study aims to assess the impact of the pandemic on compliance with NICE guidelines regarding anticoagulation for atrial fibrillation.

Cardiovascular disease is the main cause of death in England, accounting for one in four deaths. Risk factors for cardiovascular disease include non-modifiable factors (e.g., age, sex, family history, ethnic background) and modifiable risk factors (e.g., smoking, raised blood pressure and cholesterol). Assessing risk of cardiovascular disease and treating risk factors is very important to prevent heart attacks and strokes. However, the COVID-19 pandemic caused substantial disruption to usual care in GP practices, particularly during lockdowns. It is possible that many patients with risk factors for cardiovascular disease have missed opportunities for GPs to assess their risk and take the necessary steps, including prescribing drugs, to reduce that risk. Therefore, the aim of this study is to investigate the impact of the COVID-19 pandemic on assessment and treatment of cardiovascular risk factors, particularly raised blood pressure and cholesterol.

The Covid-19 pandemic has seen a huge shift in where patients are dying, in addition to increased mortality overall. While the majority of patients who died from Covid-19 did so in hospitals or care homes, between March 2020 and July 2021 a third more people than usual died at home, largely from causes other than Covid-19. At the same time, services for people at the end of life were severely disrupted, with many services shifting to remote working as far as possible. This impacted how much support patients and families received to care for people in their own homes.

This study will use data from GP and hospital records, and data on registered deaths to:

• Understand what services were provided to patients who died over the course of the pandemic, compared with pre-pandemic, and how did service use vary between sub-groups of patients? This will include considering place and cause of death, demographic, social and clinical factors.
• Develop measures of quality of end of life care from the data, taking account of the quality of the data available. We will use published standards on quality of care, for example, guidance from NICE, to identify potential measures of quality which could be identified from the data.
• Develop questions for future research, once we have identified which measures of quality of care are feasible to assess from the data.

The shift in place of death during the pandemic, and the exacerbation of existing health inequalities and pressures on health and care services, make this an even more important policy area to address now.

Patients who need surgery for cancer must balance the risks of undergoing and not undergoing surgery. The decision to undergo surgery is complicated by other existing conditions and diseases that the patient has. For example, there is some evidence that patients with an indication of SARS-CoV-2 infection might be at greater risk of death shortly after surgery, compared with patients who did not have an indication of infection. This increased risk might persist up to 7 weeks after surgery, in some cases.

In light of the increased risk of death, NHS England and NHS Improvement issued guidance to delay non-urgent, non-emergency operations for three months. The findings of research studies around that time varied in their support of such delays. For example, one global study suggested that non-emergency surgeries should be delayed for more than 4 weeks, while another study suggested the increase in the risk of death related to a SARS-CoV-2 infection was very low and might not outweigh the risks of delaying. The conflicting evidence makes it difficult for hospitals to safely plan surgeries, especially because it is not clear why there is such a wide range of advice given.

Our study attempts to review the relationship between surgery timing and cancer patient outcomes after surgery in light of the NHS England and NHS Improvement guidance. In addition to patient outcomes - e.g. death and post-operative complications - we will study the effect of surgery timing and patients’ SARS-CoV-2 status on outcomes related to hospital planning (e.g. patients’ length of stay in hospital and readmissions after discharge).

We will describe the incidence of SARS-CoV-2 Omicron infections in England in comparison to Delta infections. Should Omicron become established in England, we will compare the risk of re-infection, hospitalisation and death between the Omicron and Delta variants.

These data will be vital to understand the impact that the emergence of Omicron will have on the course of the pandemic in England and around the world.

Reports of people developing long-lasting post COVID symptoms are rising, with a latest estimate of 1.1 million people living in private households in the UK (1.7% of the population), according to the Office of National Statistics. However, little is known about which groups of individuals with a history of COVID-19 are more likely to have symptoms over a long period (e.g. more than 12 weeks). Recent research showed that not all patients who self-report long COVID have been diagnosed clinically or referred to recovery services by their doctors.

This is a population based cohort study from the electronic health records using the OpenSAFELY platform. In this study, we aim to identify risk factors associated with long COVID diagnosis in the primary care records. We will investigate the following questions:

• What are the risk factors associated with long COVID diagnosis?
• What are the differences between risk factors for acute COVID and long COVID?
• Can we predict or discriminate between these two types of COVID?
• How the clinical coding of long COVID diagnosis impacts on the above analyses?

The identified risk factors will help increase the awareness of long COVID, improve the diagnosis coverage and subsequently improve the access/referral to recovery services and inform a risk prediction model to identify and prioritise care, education and rehabilitation for patients at risk of long COVID.

In this study we will investigate how blood test results can be interpreted in OpenSAFELY. Blood test results give us important information about someone’s health. However, for some tests, the values can be reported as “less than” or “more than” a threshold, typically for values outside the accurate range that the machine can detect. Therefore a test result may appear as “<10”, and for some tests this type of result is very common. However, when conducting research, researchers may wish to group patients together and if the comparators (< and >) are not taken into account we may group the values incorrectly, e.g. the “<10” above would be taken as “10” and incorrectly grouped with values “>=10”. Here we will identify which tests involve comparators most often, and which values they are typically associated with. For example, if a certain test often has values “>5.5” then when creating groups researchers might need to use 5.5 as their highest value. The findings of this audit will be useful in various context, for example (a) assessing patients’ severity of preexisting disease, which may impact their risk from COVID, (b) assessing the impact on patients’ health after contracting COVID or (c) assessing indirect impacts of COVID on healthcare provision (e.g. whether people with diabetes generally experienced worsening disease during the pandemic).

Health care and the way we consult doctors has changed as a result of the COVID-19 pandemic and lockdown. Many more appointments with doctors (GPs) have been over the telephone instead of in person. Many routine appointments, blood tests, cancer screening and non-urgent treatment have been cancelled in both GP practices and in hospitals. Urgent cancer referrals reduced dramatically.

However, it is unlikely that patients have stopped having symptoms. It is more likely that people have been avoiding going to the doctors or to the hospital. This may be because they do not want to bother doctors or nurses at this time of higher workload, or they are not sure that GPs are available to deal with non-COVID problems or they are worried about catching the virus. This could mean that there may be a build-up of medical problems in the community. Delays in the diagnosis or treatment of conditions such as cancer may cause longer-term health problems for some groups of patients. As social distancing rules are relaxed, and life starts to return to normal, doctors will need to ‘catch-up’ on clinical care for a wide range of conditions.

This study will track changes in the number of people visiting their GP for symptoms of cancer, the number of tests done, numbers of hospital appointments and longer-term outcomes focussing on cancer. We will present the data so that GP and hospital staff will be able to see which disease areas or groups of patients that will be most in need.

Androgen deprivation therapy (ADT) for prostate cancer is used to shrink the cancer before or after the surgery or radiotherapy. ADT is also used to manage advanced cancers that cannot be cured with surgery and radiotherapy. Patients receive injections monthly or every three months long-term. In this research we are seeking to investigate if and how this was affected during the pandemic. There is evidence that the pandemic affected access to healthcare, especially when face-to-face appointments were required for the provision of services. We will access primary care and hospital (SUS) data to assess the impact on ADT delivery.

In this study we will extract and analyse data on the prescribing of injectable forms of ADT for prostate cancer between pre-pandemic (2015-2019) and pandemic years (2020-2021) to investigate any potential effect of the COVID-19 on the delivery of ADT for prostate cancer. This information may help to identify where services require more support as well as assist government and policymakers plan for future lockdowns and pandemics.

We will also analyse by age, ethnicity, deprivation, region of the country and care-home residency to identify possible patient groups with difficulties in accessing medication. Prostate cancer, comparatively, is one of the cancers more strongly affected by demographic, socioeconomic and ethnic inequalities.

Pancreatic cancer is relatively rare (10,000 new cases per year) and a difficult cancer to diagnose early. Due to the difficulty in early diagnosis it has a high mortality rate, with 90% of people dying within a year from diagnosis.

It is therefore possible that the COVID-19 pandemic has resulted in delayed diagnoses and worse outcomes in pancreatic cancer care.

In this service evaluation we will investigate if there was any detrimental effect of COVID-19 on pancreatic cancer including stage at diagnosis (are people diagnosed at more advanced stages), and the mechanisms that lead to diagnosis and healthcare provided.

This information may help to identify where services require more support as well as assist government and policymakers plan for future lockdowns and pandemics.

The use of NHS@Home interventions as a means of patients better self-managing health and care at home has been a key part of NHS England’s response to the indirect impact of the COVID-19 pandemic. This analysis aims to establish the breadth and depth of coverage - geographically and among the target populations. This will not be used in performance management, but rather as an initial assessment of uptake and usage of the services to inform modelling and help shape future analysis requirements.

The OpenSAFELY framework allows analysis of the full pseudonymised patient primary care records for 57 million people registered at a general practice in England. This comprises 24m patients from practices that use TPP and 33m patients from practices that use EMIS. Use of the TPP and EMIS systems is known to be geographically clustered, however little is currently known about the extent to which the separate populations within these systems is representative of the population at large.

OpenSAFELY has been, and continues to be, used to support research that delivered urgent results related to the global COVID-19 emergency.

Assessing sample representativeness is a critical component of health research and provides context for researchers to understand the validity / applicability of their research. Understanding differences in geographic spread, sex, age, IMD and ethnicity distributions between TPP patients and non-TPP patients is important for the interpretation of COVID-19 research analysis and generalization to the background population.

In order to understand the representativeness of TPP data we will compare broad demographic data to ONS estimates for England for Index of Multiple Deprivation (IMD); age and sex and population per MSOA / STP. We will also compare ethnic breakdowns of TPP to the 2011 ONS Census and compare the 5 most common underlying causes of death in England in 2020 ONS to the equivalent in TPP.

Mental health conditions such as anxiety and depression have become more prevalent throughout the population during the coronavirus COVID-19 pandemic, but the existing evidence as to whether people infected with SARS-CoV-2 are at greater risk is mixed. This study will therefore examine rates of common mental disorders, serious mental illness and mental health behaviours (such as self-harm) in people aged 16 years or over infected with SARS-CoV-2, for up to a year after the first testing positive for the virus. Rates of these conditions will be compared with those in several non-infected control groups (people who tested negative for SARS-CoV-2, and people diagnosed with influenza before and during the pandemic) to ascertain the extent to which risk is elevated after SARS-CoV-2 infection. The findings of this study may identify groups of the population particularly at risk of mental health conditions following infection, and inform public health policies and service provision around mental health.

In a recent study we found that people with conditions that are related to problems with the immune system, such as Cohn’s disease, psoriasis, and arthritis, and people taking treatments affecting the immune system might be at an increased risk of severe COVID-19 infection. Atopic eczema, a common skin condition presenting with itching and rash, is also thought to be related to problems with the immune system, and some individuals with more severe eczema are treated with dupilumab, a drug that affects the immune system. In this study we aim to explore whether people with eczema and people with eczema who take dupilumab are at an increased risk of severe COVID-19 infection. Clinicians and public health experts need to know this, in order to develop strategies to prevent severe COVID-19 in people with atopic eczema.

All adults in the UK are eligible to receive coronavirus (COVID-19) vaccines to protect them against COVID-19. The vaccines currently available are given in two doses. Most people who have had a 2nd dose received it 8 to 12 weeks after their 1st dose. Although clinical trials have shown that the vaccines are safe and effective at protecting against COVID-19, there are still many questions that need to be answered in order to understand longer term effects of vaccination and keep people protected against COVID-19. Our work aims to understand the long-term effects by answering questions such as:

• Which groups are choosing not to get their COVID-19 vaccines?
• Does the protection given by the vaccines against COVID-19 infection, hospitalization and death reduce over time?
• Does the time between the 1st and 2nd doses affect the efficacy of the vaccines against COVID-19 infection, hospitalization and death?
• What is the effect of additional (more than 2) doses of the vaccine?

During COVID-19, GPs have been encouraged to switch patients from Warfarin to alternative anticoagulants called Direct Oral Anticoagulants (DOACs) as they require less frequent monitoring and avoid additional visits to the GP practice. DOACs do however require renal function tests to ensure the correct dose is prescribed to avoid harm in patients. This evaluation wants to identify the proportion of patients who are at risk of harm whilst prescribed the incorrect dose of DOAC.

This study will explore the impact of COVID-19 on those of ethnic minority backgrounds in London throughout the pandemic and determine whether the vaccination programme has made a difference in ethnic disparities in health outcomes. We will also look at potential mechanisms through which COVID-19 has an adverse effect on people of minority ethnic backgrounds, these include whether there is a greater impact from certain health conditions, living environments, certain health behaviours such as vaccination uptake.

Lockdowns, stay at home orders and social distancing guidelines have had an impact on the lifestyles of people living in the UK. Studies have shown a reduction in physical activity during the pandemic, using a variety of data sources such as smartphone apps, recording from implantable cardiac devices and surveys. Some international surveys have reported that lifestyle changes during the pandemic have resulted in weight gain and increased Body Mass Index (BMI) (a measure of if you are a healthy weight for your height). In the UK people joining the NHS Diabetes Prevention Program, a targeted intervention for people at high risk of diabetes, in the year following the pandemic were on average heavier than those who joined the programme in the years before the pandemic. But there is a lack of data on how the weight (and consequently BMI) of the general population of adults living in the UK has changed since the onset of the pandemic.

Even prior to the pandemic, obesity was common in the UK. The 2019 Health Survey for England demonstrated that 64% of adults were overweight or obese. Obesity has long been known to be a risk factor for the development of a range of conditions such as diabetes, heart disease and stroke. This can be through the effect of obesity on established markers of metabolic risk such as: Blood Pressure (BP), Blood Cholesterol and Lipids Profiles (Lipid Profile), and blood sugar levels measured by Haemoglobin A1c (HbA1c). Furthermore, obesity, and conditions linked to obesity such as diabetes, have been shown to increase the risk of hospitalisation and/or death from COVID-19.

In response to these challenges the NHS has invested in a range of interventions to support weight loss. Such as the NHS England Digital Weight Management Programme (DWMP), which is currently available to help overweight and obese people living with diabetes and/or hypertension manage their weight. This analysis will look at how the number of people meeting the eligibility criteria for the DWMP has changed since the onset of the pandemic through assessing BMI changes in these groups. We will also describe, amongst all adults living in the UK, how BMI and metabolic markers have changed since the onset of the pandemic, in order to assess the need for weight loss interventions in the overall population.

We will further investigate whether some groups of people have, since the onset of the pandemic, experienced more weight gain (and resultant increase in BMI) and/or worsening of their metabolic markers than others. These groups will be identified based on personal characteristics (such as age, sex, ethnicity, deprivation and/or region of residence) and whether they have another co-existent medical condition, such as asthma or chronic kidney disease. This will demonstrate specific population groups that may most benefit from targeted weight loss interventions. We will specifically look at co-existing medical conditions that have been deemed common and important by the NHS Qualify Outcome Framework (QOF) - a tool that is used to assess the quality of care provided by GPs.

The findings of this study will help inform NHS England health care policy around the extent of need for specific weight loss interventions and which specific groups are at greatest need of these interventions. Furthermore, we hope the findings of the analysis will inform future focussed research.

We need to know to what extent the actual treatment of people with ongoing symptoms of COVID-19 matches published NICE guidance. This corresponds to carrying out an audit. If differences are found, NICE could use the information either to improve the presentation of guidelines or improve its implementation work.

This study uses OpenSAFELY data to compare summarised information from GP records against NICE’s guidelines for long-term effects of COVID-19.

Since the outbreak of COVID-19 in the UK, primary care has changed quite dramatically, with many general practices restricting physical consultations and instead providing phone and video consultations. This has affected the monitoring and management of a range of long-term conditions and is of particular importance in people with diabetes who would ordinarily have their diabetes and associated comorbidities frequently monitored through diabetic reviews at their GP practice and at diabetes clinics within the NHS. It is not yet understood what impact the change in access to services has had on these patients in terms of their diabetes control, cardiovascular risk factors, diabetes-related complications, and mortality.

In previous work, we sought to compare the frequency of primary care interactions, services, monitoring, prescribing, and diagnoses before and after the start of the pandemic, with a view to quantifying the level of unmet need in this patient group and highlighting the anticipated backlog that primary care services would face as restrictions were gradually eased. In this proposed phase of the project, we aim to examine rates of hospitalisation, ICU admissions, diabetes-related complications and associated adverse events by combining data from primary and secondary care, and to investigate specific causes of mortality in people with diabetes.

The Covid-19 pandemic has resulted in major changes in how and whether patients can consult their General Practitioner. This has an impact on how their common infections (such as sore throat or bladder) are being diagnosed and treated. The number of telephone or web appointments has increased and there are concerns, in addition to patients not consulting their GP, that more antibiotics may be given inappropriately as it is difficult to diagnose remotely.

Prior to Covid-19, improvements in antibiotic prescribing in primary care (also known as antibiotic stewardship) were a major priority for the NHS. This was related to increased rates of infections in which antibiotics no longer work (known as antibiotic resistance), often related to overuse of antibiotics.

This project will provide up-to-date information to antibiotic stewardship activities in primary care.

Specific aims are to evaluate

• the impact of Covid-19 on the primary care treatment pathways for common infections
• the effects on antibiotic prescribing
• the effects of prior antibiotic exposure (types and extent) on severity of outcomes in Covid-19 infections (including during and after hospital admission)

We will investigate how the prescribing of groups of medicines used to treat the mental health of residents of care homes has changed over the COVID-19 period. Residents have been more socially isolated given restrictions on visits from family and friends and trips outside the care home setting. There have also been high rates of death in care homes due to COVID-19, which may have affected the mental health of surviving residents and staff.

Given rates of depression, anxiety and stress are considered to have increased in the general UK population over the COVID-19 pandemic, we are interested in looking at how the prescribing of anti-psychotic and anti-depressant medicines to residents in care homes has changed over the COVID-19 pandemic and relative to the general population.

The NHS Cancer Programme is working to identify more people with prostate cancer as part of the pandemic recovery. To aid that work, we would like to know:

• What was the impact of the pandemic on the number of patients who received prostate-specific antigen (PSA) tests?
• If there was an impact on PSA test undertaken, which patients did this affect (by patient characteristic, geography, symptoms at presentation)?
• Was the change in the quantity of PSA tests undertaken followed by a corresponding drop in the number of patients receiving onwards tests and treatment for prostate cancer?

We want to understand how many people are diagnosed with Long COVID in general practice, who is diagnosed with Long COVID, and how long after getting COVID-19 people are diagnosed with Long COVID. We will also find out how this changes over time; and how different comorbidities and other factors impact the above.

COVID-19 infection may increase the chances of certain conditions, including strokes and heart attacks. We will study people alive on 1 January 2020 with a record in OpenSAFELY to investigate whether this is the case. Specifically, we will compare the number of people with COVID-19 infection who also have a record of one of these conditions against the number of people without COVID-19 infection who also have a record of one of these conditions. This will allow us to estimate how much COVID-19 infection increases the risk of these conditions. Our work may inform whether people with COVID-19 infection should be offered preventative treatments to protect against these conditions.

We will assess the impact of COVID-19 on antidepressant prescribing; one of five medicine classes identified in the PHE Review which NHSE/I are prioritising and leading on. As described by the OpenSafely paper in the Lancet, addressing the effect that the pandemic has had on mental health is important and we will use primary care data to measure and mitigate the indirect health impacts of COVID-19 on the NHS.

We will investigate whether potential barriers to accessing mental health services over the course of the pandemic have led to specific (or a change in) indications driving the already increasing trend in antidepressant prescribing by observing:

• the rate of new initiations of antidepressants
• the discontinuation of antidepressants

We will investigate the the impact of COVID-19 on dependency forming medication use by age, gender, Index of Multiple Deprivation (IMD), ethnicity, region (and other system level geographical cuts)

Concern raised by the national clinical director of cardiovascular disease (CVD) about inequalities in the prevalence of CVD. The purpose for requesting GP record data in OpenSAFELY is to monitor service use for CVD pathways, blood pressure tests taken, corresponding reading, and medical prescription across geographical deprivation and ethnic groups. These metrics will serve as potential indicators for access and outcomes (for people of different ethnicities/people living in areas with different levels of deprivation) in CVD related primary care activity. We will particularly be monitoring the variations in these indicators that occurred over the course of the COVID-19 pandemic and assessing the relative success or failure in recovering primary care services.

Aims

• To better understand and be able to interpret disparities in access and outcomes related to CVD to narrow health inequalities in this area.
• To use CVD and this analysis as a test case in the processes necessary for developing primary care recovery of services indicators with OpenSAFELY

This project will link ISARIC-4C and OpenSAFELY data in order to follow up hospitalised patients with COVID-19 following discharge. ISARIC-4C is a UK-wide consortium responsible for prospectively collecting data and biological samples from hospitalised patients with COVID-19.

The aims of the project are to investigate children and young people under the age of 19 years with covid-19 to:

• Assess the feasibility of combining a detailed clinical dataset of patients hospitalised with SARS-CoV-2 (ISARIC-4C dataset) with wide-ranging primary care data (OpenSAFELY dataset)
• Investigate risk factors for hospital admission and disease severity
• Investigate outcomes for children hospitalised with covid-19 stratified by severity, compared to those cared for in the community

Recent research into coronavirus has shown that members of the public can play a crucial role in controlling infection outbreaks, by adopting simple behaviours to curb the spread of infection in the home such as handwashing, cleaning surfaces, wearing of face coverings and social distancing. Despite public health advice, evidence shows most people need to change their behaviour to help prevent infection.

Germ Defence is an interactive website employing behaviour change techniques to supplement public health advice. Germ Defence was originally developed during the H1N1 swine flu pandemic using theory, evidence and extensive feedback from members of the public. It was then trialled in over 20,000 patients and shown to reduce the number and severity of infections of users and members of their household. Germ Defence has recently been updated for the COVID-19 pandemic. The website aims to help users with pre-planning about effective isolation of an infected household member; personalised goal setting for increasing a range of infection control behaviours; changing the home environment to support new habits and problem solving to overcome barriers.

What is the aim of the project?

This project aims to examine the effects of randomising dissemination of the Germ Defence website via GP practices on rates of respiratory infection, including COVID-19 and seasonal flu.

What are we doing?

We will contact every GP practice in England and ask them to share the link to Germ Defence with their patients. Half of the practices will be randomly chosen and asked to immediately send out the Germ Defence link to their patients in the Autumn of 2020. These practices will be known as the intervention arm or immediate implementation group. The other half of the practices will be contacted to send out Germ Defence in March 2021 and will be known as the usual care arm or delayed implementation group. We will assess usage of the Germ Defence link from anonymous data produced by the website. We will then use anonymised national data that is collected as part of routine care to compare whether infection rates are lower in practices that sent Germ Defence information to their patients immediately, compared to those that didn’t send the information until later.

The aim of this work is to assess NHS England guidance issued at the peak of the COVID pandemic recommending FIT (stool) testing for all patients suspected of colo-rectal cancer which could be undertaken in primary or secondary care. The extent to which this advice has been implemented across the system, particularly in primary care, is unknown, since routine operational reporting from pathology systems is not a mandatory requirement.

In addition FIT testing in patients aged <50 (who are at very low risk of colo-rectal cancer) may be adding to the demand for the cancer two week waits, Lower GI referrals and subsequent colonoscopy.

Analysis of the data will allow us to determine if particular variables (e.g. age, gender, ethnicity, locality, FIT level, etc.) are related to an increased risk of colo-rectal cancer diagnosis, and importantly, if patients can be categorised into a very low risk of colo-rectal cancer. NHS England would then ensure that these outputs were included in further guidance to support more appropriate and targeted referral for colonoscopy when investigating suspected colo-rectal cancer.

The effectiveness of COVID-19 vaccines against clinical outcomes: a test negative case control study using OpenSAFELY

Background

As soon as SARS-CoV2 was identified as the virus causing the ongoing COVID-19 pandemic, scientists and vaccine manufacturers worked rapidly to develop vaccines against it. Since late 2020, the UK has been rolling out a phased mass vaccination programme in the adult population, starting with people at greatest risk of severe infection. In clinical trials with volunteers, all three of the UK’s currently authorised vaccines were good at protecting recipients against infection but now that the vaccination programme is underway, it is important to look at how protective the vaccines are in the real world.

Aims of the study

• To determine the type-specific vaccine effectiveness of individual deployed COVID-19 vaccines in preventing laboratory-confirmed SARS-CoV2 infection in specified subgroups of the population.
• To determine the type-specific vaccine effectiveness of individual COVID-19 vaccines used to vaccinate the UK population in preventing hospitalisation due to COVID-19 related disease in specified subgroups of the population.

Methods

In this study we will evaluate how good each vaccine is at protecting people against infection with SARS-CoV2, and also at preventing serious outcomes of COVID-19 such as hospitalisation. We will use a test-negative case control design study to compare COVID-19 vaccination status in people with symptoms of COVID19 who have a positive SARS-CoV2 test (cases) with people who also have symptoms but who have a negative test (controls). Using this study design, we will also look at the vaccines’ effectiveness in particular subgroups, such as the elderly and people with poorly functioning immune systems, to see if the vaccines work less well in some groups.

We will estimate the vaccines’ effectiveness in preventing infection and serious outcomes at different time points after the first and second doses of the vaccine, and after a booster dose if this is recommended. Logistic regression will be used to determine adjusted type-specific vaccine effectiveness estimates and we will repeat analyses at intervals for up to 3 years after the start of the vaccine rollout to detect waning VE.

We will use a secure analytic platform, OpenSAFELY, to run analysis codes against approximately 24 million routinely collected electronic health records (EHRs) for adult patients who are registered at GP practices. This study will complement other studies using different data sources or study design, and will add to the body of knowledge to inform the government’s expert independent advisory body, the JCVI, when it make recommendations on future vaccination policy.

The aim of the study is to estimate the effectiveness of the vaccines for COVID-19 using the age thresholds in the JCVI COVID-19 vaccination first phase priority groups under a modelling approach known as a regression discontinuity design.

An online consultation enables a patient to contact their GP, other health professional or GP practice staff over the internet, typically by using a smartphone, tablet or computer. This is a fairly new technology in terms of adoption, though its use had been accelerated by the pandemic. NHSE/I, through its Digital First Primary Care (DFPC) programme, has been working on enabling and stimulating the use of online consultation systems as a means to improve access to primary care and make best use of clinicians’ time.

OpenSAFELY was identified as a way of providing intelligence on online consultations activity prior and alongside the pandemic, particularly in the way events are recorded or not in primary care records via dedicated codes. In this initial discovery piece, we looked at:

• understanding how certain online consultation-relevant codes of were used in terms of adoption and volume
• understanding broad demographics and past clinical history of those associated with online consultation-relevant codes

These insights are being triangulated further with supplier data and qualitative findings as part of a wider research piece. Wider efforts on improving intelligence for operational, research and evaluation purposes are also underway.

We attempted to calculate the counts of eligible patients for the Covid vaccine based on the priorities set out in the Green Book and GOV advice for clinically extremely vulnerable.

The approach used primary care data mostly based on age and a list of high/moderate risk health conditions (without health care workers considered).

This work was conducted between October 2020 and January 2021. In January 2021 this analysis was superseded by the “SARS-CoV2 (COVID-19) Vaccine Uptake Reporting Specification” — opensafely/primis-covid-vacc-uptake-spec.

The work approved in this application is published here as a public record of analysis carried out on OpenSAFELY, but must not to be used for onward development, decision making or further analysis.

A vaccine uptake report has since been pre-printed doi.org/10.1101/2021.01.25.21250356