Impact case study database

5% of adults have asthma and 10% of these have severe asthma. 50% of patients with severe asthma have a type of inflammation characterised by high levels of eosinophils, called severe eosinophilic asthma. In the UK, 200,000 people have life-threatening, debilitating severe asthma, which cannot be controlled with the usual medicines. The World Health Organisation estimates that asthma causes loss of 15,000,000 disability-adjusted life-years and 250,000 deaths every year. COPD is one of the commonest chronic conditions, affecting an estimated 384,000,000 people globally. In the UK, COPD is diagnosed in 4.5% of people over 40; an estimated 1,200,000 people live with this diagnosis, and it is estimated that a further 2,000,000 people remain undiagnosed. It causes high levels of mortality and morbidity, being a common cause of emergency admissions and the third leading cause of mortality in the world.

In 2019, the Global Initiative for Chronic Obstructive Lung Disease (GOLD) Science Committee COPD guidelines [A] recommended that ICS are used only where directed by a blood eosinophil indication, citing the University of Oxford research [1,2,3] as evidence. The recommendations are based on the consistent effects of ICS seen at ≥100 eosinophils/μL, as found in the University of Oxford analyses investigating blood eosinophil counts as a continuous variable [2]. This was the first time in any airway disease that ICS were targeted to a biomarker-identified patient population. Similarly, Pavord and Bafadhel’s research influenced the European Respiratory Society guidelines, published June 2020, recommending withdrawal of ICS in patients with COPD without frequent exacerbations, but not withdrawing ICS if blood eosinophil count is ≥300 eosinophils/μL [B]. The importance of their research on both sets of guidelines was confirmed in a letter from the chair of the GOLD Science Committee [C], stating “[Pavord’s] work had implications on the GOLD report…and an official European Respiratory Society (ERS) guideline”. These international guidelines have been adopted into local practice, influenced by Bafadhel’s research, confirmed by the chair of the Respiratory Prescribing Group for an NHS Trust [D].

Approximately 75% of patients with COPD are treated with ICS [D], whereas only 20% have blood eosinophil counts of ≥ 300 eosinophils/μL. In England around 80,000 people are diagnosed with COPD each year. This suggests that implementation of these guidelines results in approximately 44,000 fewer patients receiving ICS in England alone. According to the chair of an NHS Respiratory Prescribing Group “ Implementation of the guideline supports cost effective prescribing … leading to a reduction in prescribing costs as well as the knowledge that the risk of adverse effects from ICS are being minimised” [D]. Therefore, the revised guidelines enable ICS to be used more economically and effectively, both improving treatment for severe disease and reducing adverse effects of treatment for other patients.

The University of Oxford-led analysis showed simple biomarkers identify which patients benefit most from life-changing anti-IL-5 monoclonal antibody therapy [4]. This has enabled life-changing, but expensive, treatments to reach the appropriate patient population. Use of these drugs in severe asthma is the only biomarker-directed use of biologics in a non-malignant condition.

Changes to international guidelines: In 2019, the Global Initiative on Asthma (GINA) recommended that anti-IL5 treatment for severe eosinophilic asthma should be for patients with blood eosinophils above a specified level and more than a specified number of exacerbations in the last year, and that high blood eosinophils and higher number of severe exacerbations are “ strongly predictive” of good asthma response (citing [4]) [E]. The Chair of GINA confirmed the impact of Pavord’s work on eosinophils: “ …[Pavord’s] *investigation of prognostic and predictive factors relevant to clinical management of patients with asthma…have led to changes in clinical recommendations for treatment of asthma in national guidelines and in the strategy report of the Global Initiative for Asthma (GINA)*” [F].

National approvals based on blood eosinophil thresholds: The UK National Institute for Health Care and Clinical Excellence (NICE) had issued interim guidance in April 2016 not to recommend the anti-IL5 biologic mepolizumab for use in severe asthma as it did not meet NICE cost-effectiveness criteria [Gi]. However, after the drug company GSK provided new modelling based on the specific severe asthma patient population with a blood eosinophil threshold of ≥300 cells/μL and a high number of exacerbations, which showed maximal benefit in the meta-analysis led by Pavord [Gii], (subsequently published in [4)]), and also revised the price for the NHS, NICE amended its initial guidance in Dec 2016 and recommended mepolizumab for this specific patient population [Giii, iv]. Feedback received in the appraisal process and committee response also confirmed the importance of Pavord’s analysis in this recommendation: for example, both NHS England and a Consultant Respiratory Physician referenced this research [4] with respect to choosing the appropriate patient population [Giv]. After the mepolizumab decision, NICE recommended two further anti-IL5 biologics (reslizumab in Oct 2017; benralizumab in Mar 2019), with stratification for blood eosinophil count and number of exacerbations as the criteria for selecting eligible patients [Gv, vi]. Subsequently, in Jan 2020, the recommendation for mepolizumab was aligned to benralizumab [Gvii], opening anti-IL5 biologic treatment to approximately 35,900 patients in England.

Blood eosinophil thresholds have thus become standard for enabling the NHS to provide biologics for severe asthma. The NHS England National Clinical Director for Respiratory Services stated in 2020 that as a consequence of Pavord’s research, “ measurement of peripheral blood eosinophil counts is now part of routine practice for severe asthma care and has proven integral to several NICE Health Technology Appraisals for biologics in severe asthma” [H]. According to Asthma UK:

“Pavord’s work to establish the use of eosinophil levels as a biomarker that predicts response to treatment has transformed our ability to match the right patient to the right drug and to create vital eligibility criteria for using the new biologic drugs” [I].

This was confirmed by the Clinical Lead of the UK Severe Asthma Registry (UKSAR) who stated that Pavord’s University of Oxford research “ has been crucial to the success of Mepolizumab and other biologics as it has allowed us to target treatment effectively and demonstrate to regulators that the drug can be used efficiently and economically” [J].

**Wider approval of biologics: The US Food and Drug Administration and European Medicines Agency each approved the anti-IL4/13 biologic dupilumab for patients with moderate-to-severe eosinophilic asthma in 2018 [Ki] and 2019 [Kii], respectively, based on the pivotal clinical trial programme which included the eosinophil and exhaled nitric oxide biomarker-directed analysis (6). A retrospective real-life cohort study in France (64 patients, published May 2020), showed that patients treated with dupilumab outside of the clinical trials also had significantly improved asthma control and lung function, reduced oral steroid use, and reduced exacerbations [Li].

Improvements in clinical care and quality of life: Suffering from severe asthma has severely debilitating effects on the daily lives of patients, with breathlessness, anxiety, frequent hospitalisations, and toxic side effects from long-term high-dose corticosteroids. Asthma UK estimated that in April 2019 more than 3,000 people in the UK were being treated with biologics, such as mepolizumab, and this number is increasing [I]. An Asthma UK survey in 2020 of more than 200 people receiving biologic treatment showed “ for 1 in 5 it has been completely life-changing and for two thirds it has reduced their asthma symptoms and frequency of asthma attacks” [I]. Asthma UK’s qualitative research in 2019 showed dramatic improvements in quality of life. For example, two participants stated:

“I just wish I had been put on this biologic a lot sooner. Because the period I was suffering…it was just so depressing that sometimes you think your life is just not worth living anymore” and

“What [the biologic] has also done is give me a sense of confidence…that extra dimension of freedom…That’s an invaluable thing” [I].

The transformative effect of the reduction of exacerbations for patients on mepolizumab is also illustrated by the testimonial of a carer submitted to NICE, describing the benefit to her daughter whilst being treated: “ she didn't have one episode of exacerbation of her asthma and finally felt that there was hope for her to have some kind of near normal life” [Hiv].

The Clinical Lead for UKSAR has stated that

“Mepolizumab and other biological drugs…have had a massive impact on the lives of patients with the most severe forms of asthma. For example, the use of regular oral corticosteroids to treat severe asthma has been substantially reduced, and which was the most common therapeutic intervention for severe asthma” [J].

The benefits to patients of reducing long-term oral corticosteroid use include reductions in risk of osteoporosis, diabetes, glaucoma, stomach ulcers and susceptibility to infection. In 2020, a global, prospective, observational cohort study also confirmed that benefits of mepolizumab to patients in real-world settings are consistent with those seen in the clinical trials, with significant reductions in exacerbations and doses of corticosteroids [Lii]. This study also showed a discontinuation rate of less than 20%, (with only 4% of patients reporting lack of efficacy as the reason) [Lii]; approximately half the discontinuation rate seen for biologics in other inflammatory conditions without biomarker-guided treatment.

Commercial success for pharmaceutical industry: Total respiratory sales of mepolizumab (Nucala) for GSK exceeded GBP174,000,000,000 from 2016-2019, increasing year-on-year including GBP768,000,000 in 2019, GBP206,000,000 of which was from Europe [M]. Approvals for mepolizumab have continued to increase, including for paediatric use in the EU, and for self-administration in 2019. Overall, biologic therapies are driving growth in the asthma therapeutics market, much of this benefiting the UK pharmaceutical industry.

Contact tracing is a key tool for enabling locally specific and relevant control measures, identification of emergence of dangerous clusters, enabling an informed public health response and promoting socially responsible behaviours for individuals. The Oxford research informed decisions by governments and development companies to pursue digital contact tracing for COVID-19 and detailed parameters for deployment to maximise the impact of this public health intervention to prevent infections, break chains of transmission, save lives and reduce the burden on health systems.

Within a week of sharing the preprint of [2] with the UK Government, Oxford researchers embarked on developing a digital contract tracing programme with NHS-X from 7 March 2020. NHS-X CEO confirms that [2] “ provided compelling evidence that digital contact tracing could play an important role in controlling the pandemic”, that it “played a pivotal role in ministers’ decision to pursue digital contact tracing”, and it “contributed significantly to the government’s response to Covid-19”. [A]. Fraser was called as a witness to the House of Commons Select Committee on Science and Technology on 28 April 2020 alongside the NHS-X CEO who confirmed the Oxford research provided the scientific justification for the NHS app stating “W e very quickly moved when Professor Fraser gave us the epidemiological basis for why it would be a powerful intervention”. [B].

This role of the Oxford research in the decision to pursue digital contact tracing is described by the Government Office for Science, who stated

“Prof Christophe Fraser was present from the outset and played the central role in providing evidence that mobile contact tracing – using an app – could play a powerful role in controlling the pandemic. His work, subsequently published in Science [2], played a decisive role in both convincing us that this was an important direction to pursue and also illuminating the directions that might be taken to implement the approach. Prof Fraser was scrupulous in helping us to understand strengths and limitations.” [C]

The Oxford team produced two reports for the UK Government [D, E] assessing options in detail, which led to the UK pilot of the mobile contact tracing app on the Isle of Wight in May 2020 as part of a pilot of the Test and Trace system, which was successful in reducing transmission and controlling the epidemic [5]. Based on their research including [2], the Oxford researchers laid out in May 2020 five epidemiological and public health requirements that any COVID-19 tracing app should satisfy [F]. These are

1. Sensitively and specifically quarantine infectious individuals

2. Higher user uptake and adherence

3. Rapid notification

4. Integration with local health policy

5. Ability to evaluate effectiveness transparently

The NHX-C CEO described that these requirements “summarised succinctly the guiding principles of the app’s development” [A], citing [2] and [D,E,F].

The Government Office for Science letter confirms “Prof Fraser and his team […], actively contributed to all aspects of the app development” and in describing [D, E] notes the “ major contributions to the setting of the parameters of the app and modelling of design alternatives. They also participated in extensive work on validation including preparation for multiple searching independent checks of their work.” [C]

The NHS Test and Trace Product Director gave the following example of how the research has impacted the app design:

“Their initial modelling work, in particular the assessment of the distribution over index case infectiousness, is used within the NHS COVID-19 app, and also referenced in recommendations around the world. Inclusion of this distribution within the app’s risk scoring helped to reduce the number of false notifications by over 30%.” [G]

Following the Isle of Wight pilot, the UK app moved to the Google/Apple platform. The Oxford group established a partnership with Google to develop the platform. Oxford’s five principles were used by the Google/Apple Exposure Notification API developers as requirements for their app to meet before launch.

The Google/Apple API drives the UK NHS COVID-19 contact tracing app and many other apps worldwide. In Europe in particular, countries have embraced the decentralised model provided by Google/Apple including UK, Germany, Italy, Austria, Switzerland and Ireland. The API is integrated into the operating system of Apple and Android phones and a risk scoring algorithm is used to decide how to sum up exposure times to arrive at an overall risk score which will lead to an exposure notification to the user if one or more of the contacts later test positive for COVID-19. The UK NHS app uses its own risk scoring algorithm which is based on research by the Fraser group and the Alan Turing Institute.

In a letter, the Vice President of Engineering at Google commented on impact on the press:

“The initial absence of evidence made health authorities reluctant to invest in the technology and users hesitant to adopt it. When the press misinterpreted findings to question the feasibility of digital contact tracing, Professor Fraser’s team proactively clarified the misunderstanding; changing the narrative in the mainstream press.” [H]

Describing the impact of [4] on approaches by governments and health authorities, Google wrote:

“It’s difficult to overstate how impactful this evidence base has been in enabling governments make informed decisions around adopting Exposure Notifications technology. The joint publication’s focus on low levels of app adoption and the added impact to traditional contact tracing was also instrumental in convincing uptake of the technology by many public health authorities, as part of collaboration with different third party entities, and even with uptake by the public.” [H]

On the topic of global impact, the Government Office for Science said:

“the app as it currently exists would not exist without the scientific work of Prof Fraser. Furthermore, the work of Prof Fraser informed the efforts of numerous other countries (including France, Norway, Ireland, Israel) each of which took different directions but nevertheless relied upon scientific and modelling direction established by Prof Fraser. I will be unequivocal: this is the finest example of true scientific impact that I have encountered in my career. It combines excellent scientific work, advocacy, engagement, ethics and selfless work in extremely difficult conditions and under great pressure. It has undoubtedly saved lives and had substantial economic impact.” [C]

Google describe the impact of the Oxford research on their decision to pursue digital contact tracing:

“Prior to this pandemic, there was little known about effectiveness of digital contact tracing technologies and factors key to their success. The Fraser group’s initial publication in Science [2] and report to the NHS provided scientifically credible and immensely valuable information in this area.” [H]

Google also describe specific examples of how the underpinning research influenced the app design, including the requirement set out in [F] for rapid notification:

“At Google and Apple, these results directly influenced API design decisions; they demonstrated user-reported symptoms were an important facilitator of timely notifications, leading us to support self-reported symptoms in the API. Findings also informed quantifiable benchmarks for system efficacy (e.g., the number of days by which contacts must be notified), which enabled strategic investment of time and resources (e.g., effort to help health authorities automate distribution of codes verifying a positive test).” [H]

Between launch on 24 September 2020 and 31 December 2020, NHS COVID-19 app was downloaded to over 21,000,000 phones, and used regularly by approximately 16,500,000 users in England and Wales, which is 49% of the eligible population with compatible phones, and 28% of the total population. During these three months, the app sent 1,700,000 exposure notifications to break chains of transmission and in an analysis by the Fraser group with collaborators [I], they estimated 600,000 cases had been prevented. The success of the app during 2020 was made public by the Government in February 2021 [J].

The NHS Test and Trace Product Director explains that

“The app has an estimated (lower bound) secondary attack rate of approximately 6%, which compares favourably to traditional contact tracing, yet has the advantage of being privacy preserving and extremely fast. The app has therefore been crucial in breaking chains of transmission and protecting users and their communities.” and goes on to confirm that the University of Oxford’s “contributions to its development have helped prevent many cases of COVID-19, and ultimately deaths.” [G]

The COVID-19 social network open source simulation OpenABM-Covid-19 [3] was the main tool used by NHS England and NHS Wales to predict COVID-19 infection rates and therefore plan key resources e.g. staffing, hospital beds, equipment, medications. The AI technology specialist company Faculty use Oxford’sOpenABM-Covid-19 as the underpinning simulation for their contact with NHS-England to provide data visualisations dashboard information for key central government decision-makers with a deeper level of information about the current and future coronavirus situation to help inform the response [K]. Similarly, the company Armakuni provide data visualisations for NHS-Wales. OpenABM-Covid-19 [3] has also been used by Singapore and France to predict infection rates and therefore plan key resources e.g. staffing, hospital beds, equipment, medications.

The University of Oxford research [2] showed that the Oxford ELISA met MHRA performance metrics for sensitivity and specificity with minor adjustments of assay thresholds. The work was published in the form of a government report [A]. On the basis of performance results [1,2] the Oxford team formed a partnership with Thermo Fisher in November 2020 to commercialise the Oxford ELISA as the OmniPATH Combi SARS-CoV-2 IgG ELISA Test [B(i)]. The test was CE marked by MHRA in November 2020 and was made commercially available in December 2020. The collaboration between Oxford and Thermo Fisher led to the rapid commercialisation of an accurate, widely used antibody test.

Thermo Fisher stated “ The last eight months has demonstrated the benefits of close collaboration between the academic knowledge and research capabilities of Oxford and the technology transfer and manufacturing expertise of Thermo Fisher. The commercialization of the assay in a new agile way of working has enabled accelerated time to market to serve in the pandemic response” [B(ii)]. As of 31 December 2020,the Oxford ELISA had been used in population surveys (ONS CIS [C]; 88,120 tests performed), the Oxford COVID-19 vaccine programme (843 tests performed), screening programmes (Oxford-OUH Staff Testing Programme; 21,332 tests performed), by the UK Biobank (106,540 tests performed) and by RECOVERY (5,760 tests performed) and REMAP-CAP (640 tests performed) clinical trials. In total, 223,235 Oxford ELISAs were used to inform seroprevalence across these groups.

Oxford research [2] also showed sensitivity and specificity, with minor adjustments of assay thresholds, for four commercial ELISAs (produced by Abbott, DiaSorin, Roche and Siemens) which allowed the companies to better understand the performance of their assays and provided the insight required for them to modify to meet MHRA performance requirements. The Head of Medical Affairs at Roche stated that “ Research conducted and published by Oxford University [1, 2] provided important benchmarking to assess the performance of our COVID-19 antibody testing platforms” [D].

In early 2020, Oxford research [1], led directly to the withdrawal of the early LFIA testing kits nationally in the UK, in the US, Spain and Italy [E], preventing inaccurate tests being used to inform pandemic control. The former head of the UK DHSC Antibody Testing and Prevalence Studies (which went on to become NHS Test and Trace) stated

“Without this information, ministers would not have had the data to make informed decisions about which tests to procure and how relevant they were to policy ideas such as immunity certificates. It was the studies led by Derrick and team into the performance of the tests that resulted in decisions to delay mass deployment of antibody tests and the return of underperforming tests to their manufacturers” [E].

The LFD evaluation study was published in the form of a report to government in November 2020 [4] and was used to inform test choice (Innova SARS-CoV-2 Antigen Rapid Qualitative Test) for a pilot COVID-19 community testing study of 300,000 people in Liverpool (6-16 November, alongside PCR testing. [F]). This pilot resulted in Liverpool moving from tier 3 to tier 2 restrictions and the subsequent deployment of 1,000,000 LFDs to 50 directors of public health on 16 November. The Community Testing Programme, including every local authority, was rolled out nationwide in December 2020 and offered local areas the opportunity to deploy large-scale testing to asymptomatic individuals in a way that best suited them and the needs of their communities. The evaluation of LFDs by the Oxford team and PHE led to the government decision to roll out asymptomatic testing to the 18,000,000 people who were required to leave home to work [G], including NHS staff (approximately 1,500,000 employees) and care home staff, reducing the risk of the virus being passed on to hospital patients, care home residents, colleagues and family members. The government also used rapid LFDs to offer asymptomatic testing at 126 universities across the UK, reaching 75% of the student population, providing assurance that they could travel home safely during the travel window (3-10 December) for the Christmas break and minimise the risk of transmission [H]. Oxford and PHE’s research led to a successful application to the MHRA in December 2020 for using LFDs as a self-test, increasing deployment to settings that could not support supervised testing [G].

According to NHS Test and Trace data, between 26 November and 31 December 2020, 967,525 LFD tests were taken across England identifying 13,755 positive cases (number does not include healthcare workers). Subsequent confirmation and isolation of these cases contributed to infection control. Lord Bethell, Parliamentary Under Secretary of State (Minister for Innovation), who leads COVID-19 policy, confirmed that “ Oxford research enabled the development of mass community testing. These tests have successfully identified a significant number of individuals with SARS-CoV-2 infections who may otherwise not have been identified” [I]. DHSC Deputy Director Testing Policy summarised; “ research conducted by Oxford has supported (and enabled) the widest deployment of asymptomatic testing in the western world” [G].

In April 2020, Walker led the ONS CIS nationwide survey to track COVID-19 in the population [3]. Since May 2020, the ONS CIS weekly reports using information from 400,000 individuals across England, Wales, Scotland, and Northern Ireland have been used by scientists and the UK government to manage the pandemic. Data from the ONS CIS directly feeds into the government’s COVID-19 Task Force twice weekly meetings, providing real-time information to decision-makers [J]. ONS CIS data feeds into every Scientific Advisory Group for Emergencies (SAGE) meeting and helps shape its recommendations to the government [J]. Data also feeds into the SAGE sub-group Scientific Pandemic Influenza Group on Modelling (SPI-M) to model national and regional infection rates and estimate R (reproduction number).

Sir Patrick Vallance, Government Chief Scientific Advisor summarised the effects of this data:

“Several SPI-M groups use data from the ONS infection survey to inform their modelling, including estimating R… *the Survey has been useful in informing and checking calculations of R and other key metrics used to understand the state of the epidemic, and therefore forms part of the critical evidence-base needed to make the best-informed policy decisions of national importance.*” [J].

Data from the ONS CIS has been used to inform tiering and national intervention decisions, to assess the effectiveness of non-pharmaceutical interventions such as face coverings and social distancing, and to advise key scientific groups such as SAGE and NERVTAG on epidemiological questions including new variants of SARS-CoV-2 [K]. Several models have been developed to estimate the number of lives saved as a result of national lockdown in the UK. Using data from the ONS CIS and comparing to data from Sweden where no lockdown was imposed, researchers at University College London estimated that lockdown on 23 March 2020 saved 17,700 lives in England and Wales to 7 August 2020 [Li]. Using data from the ONS CIS, researchers at London School of Health and Tropical Medicine modelled the effect of tiered restrictions in England introduced in October 2020 and showed that they had some effect in slowing transmission but that a temporary lockdown (4 Nov – 5 Dec) provided the strongest effect in reducing COVID-19 deaths, saving approximately 20,000 lives between 1 October and 31 Dec 2020 [Lii].

Professor Sir Ian Diamond, National Statistician, highlighted the advancement in analytical capability of staff at the ONS through the collaboration with the Oxford team,

“The involvement of Professor Sarah Walker and Dr Koen Pouwels ... has significantly broadened the statistical infrastructure and capability in surveillance studies at ONS. ... The partnership has proven the success of flexible and innovative ways of working across sectors to provide vital evidence for key policy decisions” [K].

As the major cause of bacterial meningitis, septicaemia and pneumonia, Streptococcus pneumoniae is responsible for approximately 3,700,000 cases of severe disease and 500,000 deaths globally each year in young children. The schedule of immunisations is essential for effective and sustainable vaccination programmes, as the exact number and timing of doses influences efficacy, public acceptance, costs, and long-term sustainability.

World Health Organisation, and low and middle income countries: In Nepal, in Nov 2013 the National Committee on Immunisation Practices (NCIP) recommended the introduction of the PCV10 pneumococcal vaccine into the national immunisation schedule for infants [A], based on the OVG’s research [1, 2, 4]. Specifically, a 2+1 dosing schedule was chosen instead of 3+0, based on the OVG trial data [2]. Based on the NCIP recommendation, from 2015, PCV10 was implemented in the Nepalese infant immunisation schedule, available to more than 500,000 infants per year. Further, the OVG’s research [3] validated the choice in Nepal to use an ‘accelerated’ 2+1 dosing schedule, which was more acceptable to the public and the research showed was as effective as the standard 2+1 schedule.

The OVG research demonstrating benefits of using a 2+1 versus a 3+0 dosing schedule in Nepal [2] was extensively referenced in a 2017 systematic review of the impact of pneumococcal conjugate vaccines, which was prepared for the WHO Special Advisory Group of Experts (SAGE) for immunisation [B]. In Feb 2019, the WHO recommended a 2+1 schedule for infant immunisation against pneumococcal disease [C], informed by this systematic review and the OVG research [2, 3] [D]. Based on these WHO recommendations, as of Nov 2020, 2+1 schedules for PCVs had been adopted for routine infant immunisations by 7 LMICs receiving support from the Global Alliance for Vaccines and Immunisations (GAVI), in addition to Nepal: India, Indonesia, Uzbekistan, Georgia, Mongolia, Kyrgyzstan and the Republic of Moldova [E]. This implementation of the 2+1 immunisation schedule applies to approximately 30,700,000 infants each year across these 8 countries. Notably, India had the world’s highest number of pneumococcal deaths in 2015 and introduced routine immunisation following the 2+1 schedule in 2017 [E], prioritising states with the highest pneumonia burden.

UK vaccination schedules: In October 2017, the UK Joint Committee for Vaccines and Immunisation (JCVI) reviewed data from the OVG-led trial comparing a 2+1 to a 1+1 schedule for PCV13 [6] and, based on this research, made a recommendation to the UK Department of Health and Social Care that the UK infant immunisation schedule should be changed from the 2+1 schedule (immunisations at 2, 4 and 12 months) to a 1+1 schedule (immunisations at 3 and 12 months) [Fi]. Public Health England announced in April 2019 that the new 1+1 schedule should be offered to all infants born on or after 1 Jan 2020 [Fii]. This was the first implementation of a 1+1 PCV13 schedule in the world. This trial and change in the UK policy is a landmark that provided credibility and reassurance to other countries that reduced dosing is safe and has cost benefits, instigating further studies in India, Vietnam, South Africa and the Gambia, funded by the Bill and Melinda Gates Foundation, to see if this reduced dose schedule is similarly immunogenic in LMICs [Gi]. Indeed, results from the study in South Africa (referencing [6]), confirmed that the 1+1 schedule was not inferior to 2+1 in a LMIC with an established PCV immunisation programme, providing the opportunity to reduce the cost of PCV procurement [Gii].

Decreased invasive pneumococcal disease and fewer injections for infants in the UK: The switch from PCV7 to PCV13 in the UK vaccination schedule in 2010, which was influenced by OVG research [5] showing that PCV13 was highly immunogenic and well tolerated in the UK schedule [H], resulting in large decreases in cases of invasive pneumococcal disease, including since August 2013. Specifically, a PHE study showed that in 2013/14 the incidence of invasive pneumococcal disease in England and Wales had decreased by 32% compared to the level before the introduction of PCV13 immunisations; this resulted in at least 1,800 fewer cases of these serious, life-threatening infections in 2013/14 [Ii]. A further PHE study showed that, in 2016/17, invasive pneumococcal disease cases resulting from PV13-type serotypes had decreased by 64% since the introduction of PCV13 [Iii]. Overall, PHE data from England and Wales, shows that pneumococcal vaccination has prevented an estimated 9,000 cases of invasive pneumococcal disease in children under 5 between 2013-2017 [Iii,iii].

Implementation of the 1+1 immunisation schedule in the UK resulted in approximately 750,000 fewer doses of pneumococcal vaccine being administered annually. Fewer injections means less distress and discomfort for infants and their parents. Mathematical modelling by PHE, based on the immunogenicity data from the trial led by the OVG [6], indicated that this reduction in discomfort through reducing the number of injections will not cause any significant loss of control of pneumococcal disease, with an estimate of only 2 additional cases of invasive pneumococcal disease in children under 2 years, over 5 a year period [J].

Global decreases in pneumococcal deaths, severe disease and economic impacts: The global mortality rate for pneumococcus in 2015 was estimated to be 45 deaths (uncertainty range 29–56) per 100,000 children under the age of 5, with uncertainty in part due to the pathogen not being identified in many cases of pneumonia [K]. Overall, pneumococcal vaccine programs, including those based on OVG research, have resulted in a mean annual reduction of global paediatric deaths from invasive pneumococcal disease of 47,400 between 2010 and 2015, and an estimation that this reduction will have been at least equalled for every year since 2013 [K]. As of November 2020, 138 countries, including 58 LMICs, had introduced a pneumococcal conjugate vaccine into their national immunisation schedule [E], and a minimal estimate is that 22,800,000 children across the world have been immunised against pneumococcal infections [E]. Nepal is an illustrative example of the impact, where the vaccine programme was heavily influenced by the OVG research: in 2018 there was a 34% reduction in cases (at least 15,500 fewer cases per year) of bacterial pneumonia in children compared to the pre-vaccination period (2014-2015)[Li]; and by 2019, carriage of vaccine-serotype pneumococcus had decreased by 74% among healthy infants [Li]. Pneumococcal disease has a major economic impact on poor families, through costs of healthcare and loss of earnings; for example, in Nepal, out-of-pocket payments for pneumonia hospitalizations exceeded 40% of the monthly household spending for more than 70% of households in the three poorest quintiles [Lii]. Analysis of the impact of introduction of the PCV10 vaccination in Nepal estimated that vaccination averted 85% of catastrophic health expenditures among the poorest quintiles of the population [Liii].

The reduction in doses of PCV13 administered in the UK, through introduction of the 1+1 schedule from 1 Jan – 31 Dec 2020, has resulted savings to the NHS of GBP7,500,000 in vaccine administration costs alone, with additional savings through reduced vaccine purchasing costs. The ‘shelf price’ of this vaccine is GBP49 per dose and, although NHS purchase price is not publicly available, the total annual savings to the NHS for 2020 are estimated as at least GBP20,000,000.

The ‘end-to-end’ electronic transfusion process was implemented initially in all acute hospitals in Oxfordshire prior to 2013. Benefits described below have been realised from 2013 to 2020, quantified by follow-up studies from the Oxford team. The team has long supported adoption in other settings, for example specification for an electronic process for transfusion was developed for the National Patient Safety Agency as early as 2006, and engagement with national schemes such as Serious Hazards of Transfusion (SHOT). Through this engagement and the dissemination of the primary research and follow up studies, the take-up of these processes has been recommended in national guidelines from 2015 onwards. The Oxford team also developed a long-standing collaboration with Haemonetics Ltd (Boston), enabling incorporation of research findings into Haemonetics’ commercially available software and products, including the BloodTrack® system, which are now marketed internationally.

The development of a process for electronic blood product ordering with an integrated clinical decision support system (CDSS) allows collection of data regarding the patient’s clinical condition and the justification for the blood product order. An alert is generated in real time if the order is outside agreed guidelines. Feedback is provided to clinicians in several ways including contact with the prescribing clinician if an alert is generated, to ask for further information and if necessary, provide education; an online dashboard allowing all clinicians to visualise their blood ordering practice; a summary report circulated on a quarterly basis to senior clinical staff for cascade to their teams; and monthly review meetings with the junior haematology doctors. These processes have multiple benefits:

Improvements to Patient Care and Safety: A study with Serious Hazards of Transfusion (SHOT) analysed data from 57 wrong component transfused (WCT) events from approximately 2,000,000 blood component transfusions administered in the UK in 2015 and 2016 and showed that patient safety has been improved by reduced wrong transfusion events and avoidance of unnecessary transfusion. In cases that related to sample labelling, blood collection and administration, no WCTs occurred with sampling or administration using electronic processes, whereas 17 WCTs occurred with manual processes [A **(**i)].

Improvements in compliance with good practice. Local data has found much greater compliance with correct procedures and fewer errors [1,2,5]. Murphy and colleagues found that significant improvements were made following the introduction of the electronic system, including a rise from 8% to 100% in checking that the blood group and unit number on the blood pack matched the compatibility label and the pack was in date (p≤0.0001). Similar significant improvements were found in blood sample collection, the collection of blood from blood refrigerators, and the documentation of transfusion. Improved compliance with guidelines for transfusion and avoidance of unnecessary transfusion following the implementation of electronic blood ordering; for example, in the haematology ward in an Oxford hospital, compliance with RBC transfusion guidelines increased from 74% in 2014 to 95% in 2019 [A(ii)].

**Improvements In Staff Time and Efficiency. The Oxford team also found that staff found the system easy to operate and preferred it to standard procedures [2]. The process involves one nurse rather than two, and there is a 50% reduction in the time for a nurse to undertake pre-transfusion patient/blood identification at the bedside [1]. The speed of delivery of blood for urgent transfusions through electronic remote blood issue has also been increased [3,6]. The electronic remote issue process avoids the need for multiple blood transfusion laboratories in centres with several hospitals as blood can be provided safely for urgent cases at remote sites. This has meant that the transfusion service in Oxfordshire operates from a central laboratory at the John Radcliffe with satellites at the Churchill and Horton Hospitals rather than with 3-4 stand-alone blood transfusion laboratories.

In the **NICE guideline on blood transfusion (2015) using an electronic patient identification system when undertaking blood transfusion is recommended in the guideline [B] largely based on the Oxford research [5]. The research was also endorsed nationally as an NHS QIPP (Quality, Innovation, Productivity and Prevention) case study in 2016, entitiled: “ Electronic blood transfusion: improving safety and efficiency of transfusion systems” [C].

The British Society for Haematology guidelines on the administration of blood components (2017) state that “…Electronic patient ID: Electronic systems are available to help improve patient ID procedures…” and “the systems employed must be robustly designed and implemented to ensure that patient safety is enhanced” [D] , in both cases citing [5].

The research provided the evidence for a recommendation in the 2017 Annual Report of the Serious Hazards of Transfusion (SHOT) scheme [E]:

“All available information technology (IT) systems to support transfusion practice should be considered and these systems implemented to their full functionality. Electronic blood management systems should be considered in all clinical settings where transfusion takes place. This is no longer an innovative approach to safe transfusion practice, it is the standard that all should aim for. Action: Hospital Chief Executives, Hospital Risk Managers and Hospital Transfusion Teams.” [E].

The Medical Director for SHOT confirmed in October 2020,

“The use of barcode scanning technology…has proved to be a simple but effective way to reduce errors at the point of sample collection. Professor Murphy’s seminal studies on this practical solution…were highly successful at his home institution at Oxford. The resulting publications demonstrating a clear safety benefit to patients has helped to stimulate the adoption of electronic patient identification in the UK and internationally. This was one of the key SHOT recommendations in 2017 and SHOT continue to stress the importance of incorporating electronic blood management in all clinical settings where transfusion takes place.” [L(ii)]

Professor Murphy provided information, advice and guidance to the Healthcare Safety Investigation Branch (HSIB) during their 2019 investigation into solutions to prevent wrongly labelled blood samples [F]. In its report entitled “ Wrong Patient Details On Blood Sample - Healthcare Safety Investigation I2019/003” the HSIB Recommendation 2019/46 was that “ … NHSX should take steps to ensure the adoption and ongoing use of electronic systems for identification, blood sample collection and labelling”.

Internationally, the 2016 haemovigilance report from Australia’s National Blood Authority highlighted the importance of technology, including portable barcode readers and/or radio-frequency identification scanners, to reduce transfusion process-related errors, and the value of decision support tools to improve clinical prescribing, based on the Oxford research [G,L(iii)].

The Chief Medical Officer, in her 2014/ 2015 NIHR Annual Report, highlighted the savings resulting from the system:

“... a new blood management system trialled and tested by our Oxford NIHR Biomedical Research Centre saved Oxford University NHS trust half a million pounds [GBP500,000] last year. It uses barcode patient identification syste guaranteeing each and every patient reciveds the right blood in the right amount. This system, if implemented across the NHS could create savings of more that £50m [GBP50,000,000] each year and is a fool-proof way of ensuring patients’ safety.” [H]

Longer-term analysis showed the first hospital-wide implementation of electronic blood ordering and decision support in Oxford hospitals resulted in a 26% reduction (USD1,340,000; GBP1,000,000 per year) in blood costs over the period 2013-2019 [A(ii)].

A survey of UK hospitals participating in SHOT published in 2019 found that over 50% of UK hospitals were using an electronic system for the bedside transfusion process and/or blood collection from blood refrigerators. [A(i)]. The NHS Blood and Transport 2018 national comparative audit of blood transfusion demonstrated a widespread adoption of the remote blood allocation systems, with 67.5% of hospital now utilising these [I].

The Co-Director of the Massachusetts General Hospital confirmed that “ As a direct consequence of the work at Oxford, my hospital began in 2012 to develop a similar system for patient identification and safe transfusion” [L(i)]. A recently reported study [J] shows a step-change in compliance with basic patient identification in theatre at the time of transfusion upon its implementation in 2014. Adoption has also been confirmed in Toronto [L(i)] and a leading Australian hospital [L(iii)].

As at October 2020, the Haemonetics BloodTrack® system has been implemented in the NHS in [text removed for publication]; in Ireland in [text removed for publication]; and in many other countries including in the United States, where it is used in [text removed for publication]. In Italy, where market entry commenced in 2018, it has been implemented in [text removed for publication] and during 2020 the software has been translated into French, Spanish and German [K]. The The Co-Director of the Massachusetts General Hospital corroborates that “ *commercial systems such as Haemosafe® and Bloodtrack® have recently been introduced in an effort to reduce barriers to implementation. These commercial systems are fundamentally based on the original Oxford work.*” [L(i)]. The General Manager (Hospital, W. Europe) for Haemonetics confirmed: “ It’s quite clear that the relationship we have enjoyed with yourself and the EBTS team in Oxford has had far reaching impact to transfusion safety very broadly in the UK and with increasing presence around the Globe.” [K]

The Oxford team is continuing to work on enhancing the electronic transfusion process, and supporting other hospitals to implement it including further documentation of the clinical and economic benefits in the largest implementation to date at Barts Health (4 acute hospital sites serving 2,500,000 people in east London) [L(iv)].

The University of Oxford-led research has influenced international strategy that is being implemented in the GMS. Specifically, in 2014, the WHO Drug Resistance and Containment Technical Expert Group for malaria (chaired by A. Dondorp, F Smithuis on committee) provided evidence to the Malaria Policy Advisory Committee of the WHO (MPAC; N. White on committee). This evidence included University of Oxford-led research on the prevalence of artemisinin resistance in the GMS [1], pilot MDA in the GMS [5, 6], and community health worker programmes in Myanmar (see below, led by F. Smithuis) [A]. On the basis of evaluation of this evidence, MPAC recommended the elimination of P. falciparum in the GMS by 2030 [B], to counter the threat of multidrug resistance – as advocated by the University of Oxford research (e.g. [3]) – and this goal was adopted as WHO strategy in 2015 [C].

Also, several tools for malaria elimination that were developed and assessed through University of Oxford research have been endorsed by the WHO. MDA has been controversial, and University of Oxford researchers and their findings (e.g. [3, 5, 6]) have had a major influence on the debate [D]. Since 2015, the WHO has endorsed MDA as a strategy for malaria elimination in low transmission areas [E], influenced by evidence from University of Oxford research projects (e.g. [5, 6]) [A, B]. Specifically for the GMS, the 2015 WHO recommendations state “ given the threat of multidrug resistance and the WHO call for malaria elimination in the Greater Mekong subregion (GMS), MDA may be considered as a component of accelerated malaria elimination efforts in areas of the GMS with good access to treatment, vector control and surveillance” [E]. Also, the assessments of drug risk from University of Oxford research (e.g. [4]), contributed to the recommendations of drug choices for MDA, with the MPAC noting in 2017 that “ the cardiovascular risk associated with the antimalarial drugs piperaquine, amodiaquine, chloroquine is considered very low and these medicines can be used in mass drug administration for malaria” [F], and the WHO 2017 Malaria MDA Field Manual recommending dihydroartemisinin-piperaquine as part of MDA [G].

Myanmar has the greatest malarial burden in the GMS, with 79% of all malaria cases in the GMS in 2012. Reducing malaria transmission in high-transmission areas in Myanmar is one of the priorities of the WHO GMS strategy [C]. Much of the burden of malaria in Myanmar is in poor and extremely remote communities, particularly in the border regions; the villages may be accessible only on foot or by off-road vehicles, and inaccessible during monsoon rains. Before 2013, healthcare systems in these communities had been weakened by long-term under-investment and conflict, leaving poor infrastructure and most villagers with no access to healthcare professionals. The University of Oxford research units in the region applied the knowledge and tools they had developed (including [1-6]) to implement targeted malaria elimination strategies that have caused dramatic declines in cases of malaria, transforming the health of local communities with a combined population of more than 1,000,000 people and contributing substantially to the WHO goal of malaria elimination by 2030.

Between May 2014 and April 2017, the Shoklo Malaria Research Unit (SMRU) established 1,222 malarial ‘posts’ providing free access to early malaria diagnosis and treatment to remote communities in Karen State, along the Myanmar-Thai border [H]. In this time, they diagnosed and treated approximately 365,000 people [H]. Using the ultrasensitive PCR methods (e.g. [2]), they identified villages that were malarial hotspots, and 50 hotspot villages were treated with MDA drug regimes - dihydroartemisinin–piperaquine plus single-dose primaquine once per month for 3 consecutive months - guided by University of Oxford research on using MDA as an elimination strategy and the choice of drugs (e.g. [3]). In the MDA-treated villages, there was a 5-fold decease in P. falciparum incidence. By April 2017, these interventions had resulted in 965 (79%) of 1,222 villages with malarial posts being free from P falciparum malaria for at least 6 months [H]. According to SMRU, in 2020, there were 1,222 malarial posts in operation and 1,060 of these (87%) had reported zero cases of P. falciparum malaria during 2020, showing sustained near-elimination of P. falciparum malaria. In the HpaPun district (the most remote and malarious area of Karen State) the yearly incidence of P. falciparum malaria has decreased by 96% since 2014, when the programme began, and by 83% since 2016, when the majority of the malarial post network was established. Importantly, the prevalence of artemisinin resistance molecular markers was stable over 3 years (2014-2017), indicating success in preventing spread of drug resistance [H].

This elimination of malaria has transformed health on both sides of the Myanmar-Thai border. Importantly, removing the burden of malaria from the health care systems in these remote areas enables their limited resources to be redirected to treating other diseases. The director of a hospital in Tak Province in Thailand, close to the border with Myanmar, credited the interventions by SMRU and MORU for a dramatic change numbers of patients with malaria, and stated: “ Malaria (P. falciparium) was one of the major causes of illness and hospital admission along the Thai-Myanmar border region until the last five years. Most of the malaria came from the Myanmar side of the border…Now it has almost gone” [I]. The director of the Department of Medical Science in the Ministry of Public Health for Thailand confirmed that, due to the SMRU work in Karen State, “ by 2020, P. falciparum malaria had been almost completely eliminated, thus preventing the spread of artemisinin resistance…The impact on the health of the population has been remarkable”, with cases in Tak Province (population approximately 600,000) being very rare in 2020 compared to being “ a major cause of morbidity and mortality in the past” [J]. Refugees and migrant workers in the border region have particularly poor access to healthcare [D], and the director of a charity providing health services to these people confirmed that SMRU and MORU’s malaria eradication campaigns have made “significant changes” in the region [K].

Since 2013, the Myanmar Oxford Clinical Research Unit (MOCRU) has worked together with the medical aid charity Medical Action Myanmar (MAM) to achieve malaria elimination through community health worker (CHW) programmes. These CHWs in remote border areas of Myanmar, adjoining Thailand, India, Bangladesh and China, use rapid diagnostic tests for malaria and provide high-quality treatment. The rationale for implementing these CHW programmes in these remote regions was based on University of Oxford research demonstrating the need for malaria elimination in these key areas with emerging drug resistance (including [1, 2, 3, 6]).

MAM initially started the programme in 2011 with 103 CHWs in Mon State, but working with MOCRU from 2013 the programme expanded to 1,326 CHWs across eight regions and covering approximately 728,057 people (1.3% of population of Myanmar) [L]. They diagnosed and treated more than 21,000 cases of malaria. For each year that the CHW programme operated, there was a 70% (95% CI 66–73%) decrease in P. falciparum incidence and a 64% (95% CI 59–68%) decrease in P. vivax incidence. Overall, this project led to a dramatic decrease in malaria: 97% decrease in falciparum malaria and 95% decrease of vivax malaria within 3 years (up to 2016) [L].

Crucially, from 2013, the CHWs began to provide a basic health care package in addition to malaria diagnosis and treatment. This included treatment for malnutrition, diarrhoea and respiratory tract infections, and it brought two strands of benefits. Firstly, local people with other health conditions received diagnosis and treatment, including, in 2016, diagnosis of 14,509 cases of pneumonia and identification and referral to hospital of 6,278 patients with suspected tuberculosis [L]. Secondly, this provision of services beyond malaria by the CHWs ensured that people continued to seek their help even after there were large reductions in malaria, ensuring the few remaining cases of malaria continued to be detected and treated early, which is essential for elimination [L]. In all 167 communities in Mon state where MAM had introduced CHWs, falciparum malaria was eliminated, with only a single - imported - positive case detected from 55,000 rapid diagnostic tests during 2017-2018.

Global uptake of INTERGROWTH-21^st Preterm Postnatal Growth StandardsIn 2016, the team began the INTERPRACTICE-21^st Project to scale up global implementation of the INTERGROWTH-21^st Preterm Postnatal Growth Standards [5]. WHO and the US Centers for Disease Control and Prevention (CDC) have both recommended use of the INTERGROWTH-21^st standards and feeding protocols [A]. As of 31 July 2020, 370 neonatal units around the world have implemented the standards and protocols in countries including Dubai, Haiti, Saudi Arabia, Sri Lanka, Thailand, and the USA. The standards and protocols have also been adopted by national paediatric and neonatal societies and Ministries of Health, for example the National Paediatric Society of Argentina and the Italian Society of Neonatology. The standards seek to reduce the overdiagnosis of extrauterine growth restriction, which results in overfeeding that can lead to obesity and cardiometabolic disease in adulthood. Monitoring preterm infants with the INTERGROWTH-21^st standards, which describe optimal postnatal growth, reduces the risk of overfeeding. The feeding protocols detail how to feed a preterm infant, emphasising the use of human milk.

In 2018, following an analysis of evidence to prevent stillbirths, the Scottish Government recommended the use of the INTERGROWTH-21^st Symphyseal-fundal Height (SFH) Standards and Fetal Growth Standards for monitoring pregnancies [3]. A letter in November 2018 from the Children and Families Directorate of the Scottish Government, to Chief Executives, Medical Directors and Nurse Directors [B] states that “The expert Scottish Government led Stillbirth Group has reviewed current evidence around accuracy of charts and recommends the use of Intergrowth 21^st SFH and ultrasound growth charts which are free to use and readily available….”.

The INTERGROWTH-21^st standards, describing how a preterm infant should grow, have been integrated into BadgerNet, the pregnancy and neonatal unit data management system widely used across the NHS for the collection, storage and reporting of perinatal patient data. The raw data available in the BadgerNet system are being used by the Neonatology GIRFT Project to benchmark the extent to which preterm infants are growing optimally and to provide feedback to neonatal units on how well their patients are growing against the standards. The GIRFT Clinical Lead for Neonatology highlights how the data are being used: “Thank you so much for providing GIRFT Neonatology with Z scores from both the INTERGROWTH-21^st Newborn Size Standards and Preterm Postnatal Growth Standards for all premature babies admitted to neonatal units in England for the three year period 2016-19. I have used the data to develop some benchmarking metrics to review growth between admission and discharge for babies at different gestational ages” [C] .

During the 2016 Zika virus (ZIKV) outbreak in Brazil the use of the INTERGROWTH-21^st standards redefined the definition of microcephaly and led to changes in global practice and guidelines. On 8 December 2015, the Brazilian Ministry of Health changed its definition of microcephaly from head circumference at birth ≤33cm to ≤32cm, irrespective of the baby’s sex or gestational age at birth. In a letter to the Lancet in February 2016 a group of Brazilian epidemiologists argued to use the international INTERGROWTH-21^st Newborn Size Standards as these are sex and gestational age specific standards for optimal size at birth. Their use, they stated, could reduce the number of suspected cases in Brazil from an anticipated 600,000 to 63,000 or 3,000 for cut-offs of -2SD or -3SD, respectively [D].

WHO published an interim Rapid Advice Guideline in February 2016 and an update in August 2016 [E], which included the recommendation to use the INTERGROWTH-21^st Preterm Postnatal Growth Standards for pre-term neonates (and for term neonates if accurate gestational age is known) to interpret postnatal changes in head circumference until 64 weeks postmenstrual age. The CDC’s ZIKV management guidelines also recommend the INTERGROWTH-21^st Fetal Growth Standards to assess head circumference at three points during pregnancy [F]. The use of a more specific definition of microcephaly focuses screening efforts, reduces the need for further investigations, and would also reduce the burden and costs on the health system. Furthermore, it alleviates the emotional effects on parents of healthy infants who might have otherwise been given a false-positive result in the screening assessment [D]. The Brazilian Government subsequently adopted the standards for all pre-term babies born in the country.

In a 2015 JAMA paper, the Child Health Epidemiology Reference Group at Johns Hopkins University concluded that using the INTERGROWTH-21^st Newborn Size Standards rather than the US Birth Weight References reduced the prevalence of small for gestational age (SGA) in 16 LMIC cohorts without any effect on neonatal mortality [G], i.e. using the INTERGROWTH-21^st standards resulted in a decrease in the percentage of neonatal deaths attributable to SGA. In a 2017 BMJ paper, the same group selected the INTERGROWTH-21^st Newborn Size Standards [4] as the most reliable tool for estimating the true prevalence of SGA in LMICs worldwide [H]. The revised 2012 data, using the INTERGROWTH-21^st definition, indicated that 23,300,000 infants were born SGA, with an estimated 606,500 attributable neonatal deaths, i.e. 21.9% of all neonatal deaths worldwide.

For implementation purposes, the study team made the INTERGROWTH-21^st clinical tools, publications, tables, charts, calculators, apps, and training resources freely downloadable [Ii]. Clinicians and health practioners are further supported through a number of training resources [Iii]. Between July 2014 and May 2020 the resources have been downloaded from The Global Health Network website 231,726 times by users in 195 countries and territories, and 23,580 healthcare professionals have been trained using INTERGROWTH-21^st e-learning modules [J].

In 2019, The Global Health Network conducted an online survey of website users (hospitals/health centres 57.0%; universities 19.9%; NGOs 15.1%; Health Ministries/ others 8.0%). The following percentages of these users had introduced the INTERGROWTH-21^st standards into clinical practice in their institutions: pregnancy dating 43.2%; symphyseal-fundal height 31.6%; fetal growth by ultrasound 40.1%; newborn size 56.4%; gestational weight gain 34.9%, and preterm postnatal growth 50.6%. More than 100 health care professionals have been trained to deliver the INTERGROWTH-21^st Neurodevelopment Package (INTER-NDA), which has been used to assess the neurodevelopment of over 15,000 2-year old children in 14 countries. The operation manuals and protocols for the INTER-NDA are freely available [Iiii].

TB is serious respiratory infection caused by M. tuberculosis, and is the leading cause of death from a bacterial infection. In 2019, there were approximately 10,000,000 cases of TB and 1,400,000 fatalities, and tackling TB is a priority of the World Health Organisation (WHO). A major problem is the rapid rise in drug-resistant strains; the WHO estimates that in 2018 there were 484,000 new cases with resistance to the most effective first-line drug (rifampicin), 78% of which had multidrug resistance. Clinicians require accurate diagnosis of whether an infection is M. tuberculosis or another Mycobacterium, and whether the infection is susceptible or resistant to specific antibiotics. Public health bodies need to be able to identify the relatedness of different TB cases to act to control transmission. The University of Oxford research showed that WGS can yield all this information, and provided the evidence and tools that have enabled improved comprehensive mycobacterial diagnostics at national and international levels.

England has one of the highest rates of TB in western Europe. Public Health England (PHE) and the University of Oxford collaborated to develop, validate and implement an end-to-end mycobacterial diagnostic service, from the clinical sample through to the results report [6]. Implementation of this service [A] was completed in Jan 2018 and achieved full UKAS accreditation in late 2019. As announced by the UK government in March 2017, this was the first use of WGS in diagnostics for any disease at this scale [B]. The PHE pipeline incorporates University of Oxford approaches for DNA extraction [1], and analysis of sequencing data using the University of Oxford-developed COMPASS pipeline [6] to identify species and drug susceptibility [2, 3, 4], and relatedness [5] [A]. With phased introduction starting in Dec 2016, since Jan 2018 all positive mycobacterial cultures in England have been analysed through this WGS pipeline at reference laboratories in London and Birmingham [A]. In 2019, a total of 14,605 mycobacterial cultures were processed for WGS by PHE. Based on University of Oxford-led research [4], since 2018 phenotypic susceptibility testing for M. tuberculosis was stopped for 80% of isolates [A].

WGS of most or all mycobacterial samples has been implemented by several public health bodies including in the USA, Netherlands, Italy, Australia, Germany and Canada. The University of Oxford research, and pipeline adopted by PHE, influenced these decisions and the approaches used.

New York State: The New York State Department of Health, Wadsworth Center, have been developing their pipeline for mycobacterial WGS with input from University of Oxford researchers since 2015 [C]. This includes adopting laboratory assays [1], which contributed to the successful Wadsworth application for regulatory approval, and incorporating methods into an analysis algorithm [C]. These WGS approaches became routine from Oct 2018 and enabled replacement of phenotypic drug susceptibility testing for 80% of M. tuberculosis isolates [C]. Since 2016, Wadsworth have used WGS for 3500 TB isolates, with approximately 800 TB cases identified per year [C]. Wadsworth shared the pipeline with other US states and the US Association of Public Health Laboratories, to provide a centralised platform [C].

Netherlands: In the Netherlands, WGS is now used for all M. tuberculosis complex isolates (approximately 600 per year) and, on the basis of University of Oxford research [4], phenotypic resistance testing has been discontinued for the 90% of isolates that have no resistance mutations in WGS [D] . According to the Head of the Netherlands Tuberculosis Reference Laboratory at the National Institute for Public Health and the Environment, “ The impact of the research at the Oxford University regarding the introduction of WGS is major…This formed the basis to a true revolution in the diagnosis of this disease” [D] .

Italy, Australia, and others: The Italian supranational reference laboratory for TB now offers routine WGS for patients at high risk of drug-resistant TB. The head of this laboratory stated that “ The Oxford University research contribution was a major breakthrough in support for the implementation of WGS” and that interpretation of TB WGS results in Italy is based on evidence from University of Oxford research [E]. Based directly on input from Crook, T. Walker and colleagues, the lead of the Mycobacterium Reference Laboratory in New South Wales (NSW), Australia, established WGS for all clinical isolates (over 1,200 isolate) of M. tuberculosis (piloted from Oct 2016, routine service from July 2019) [F]. The University of Oxford research [e.g. 4] was critical for the design of the genomics service, and has resulted in more accurate and rapid detection of drug-resistant TB and improvement in the resolution and timeliness of detection of clusters of recent community transmission [F]. The NSW implementation led to roll-out of similar programmes in other Australian states in 2019 [F]. Routine WGS for TB is also now used in other countries including Germany and Canada.

World Health Organisation: Based on their research on sequencing-based drug susceptibility prediction [4], the University of Oxford team was contracted by the Foundation for Innovative Diagnostics (FIND) to provide a catalogue of genetic variants associated with drug susceptibility for the WHO [Gi,ii]. In Dec 2020, the WHO confirmed that the catalogue, based on WGS of 40,000 isolates, will form the basis of a new reference standard for national and international TB programmes and industry [Gi,ii].

The implementation of WGS for Mycobacteria has led to benefits for public health laboratories, clinical care for patients, and in public health responses.

**Decreased turnaround times, costs, and risks in public health laboratories: Because WGS yields all the required diagnostic and epidemiological information, the reports are available faster than by separate conventional methods. For example: PHE report drug susceptibility within 5-7 days, compared to 3-4 weeks previously [A]; and Wadsworth have a 7-day average turnaround time for reporting susceptibility, which is 7 days earlier than previously for first-line drugs and several weeks faster for second-line drugs [C].

Although WGS is an expensive technology, the end-to-end pipeline removes the need for numerous laboratory assays, and PHE estimated saving approximately 7% per sample by using WGS [6]. The Netherlands implementation has been “ highly cost effective” because of the reduced need for other expensive and labour-intensive methods [D], and Wadsworth report “ savings in staff time and cost savings”, which are covering the costs of implementing WGS and freeing up staff to develop further testing improvements [C]. The reduction in the number of laboratory assays enabled by WGS also improves safety as technicians are less frequently exposed to the hazards of working with viable bacteria in biosafety level 3 laboratories, reducing the risk of laboratory infections [D].

**Faster, tailored clinical care: The rapid return to clinicians of drug susceptibility results from WGS has decreased the time to treat patients with appropriate antibiotics. For cases where the isolate is shown to be susceptible to first-line drugs, clinicians are given rapid reassurance that they are using the appropriate regimen [A, C, D]. Crucially, patients with drug-resistant infections previously had to wait several weeks for effective treatment and they now receive it in 7 days, transforming their clinical care. For example, in 2017-2018 in England, an effective treatment regime was identified 3-16 weeks earlier for 700 patients with drug-resistant TB [A]. When TB is treated in hospital, it is expensive to healthcare providers – for example, approximately GBP900 per day in an airflow-controlled ward – and rapid identification of an effective treatment regime speeds up discharge, yielding further savings. Choice of drugs for a patient is also now personalized when the WGS shows identical resistance mutations to an infection in a previously treated patient [A]. Additionally, WGS reveals low-level resistance to the first-line drug rifampicin, which was previously very difficult to detect, so clinicians are now able to tailor drug treatment for patients infected with these strains [C, D].

Improved precision in public health interventions: According to PHE, the speed and increased resolution of SNP typing has transformed cluster investigation to prevent further cases [A]. For example, PHE has been able to: pinpoint transmission between specific individuals, avoiding unnecessary extensive contact tracing efforts; focus a targeted public health intervention in a healthcare setting where rapid transmission through brief contact was discovered to have occurred; identify and address transmission due to breaches in infection control practice in a healthcare setting, avoiding further risk to patients; and contribute to stopping the spread of international clusters of extensively drug resistant TB [A].

Bovine TB, caused by Mycobacterium bovis, is a major problem in the UK, leading to compulsory slaughter of approximately 50,000 cattle annually and costs of approximately GBP150,000,000 to UK taxpayers in eradication efforts, plus additional costs for the cattle industry. A 2018 government review determined it is “ feasible and cost-efficient to move to whole-genome sequencing… [which] allows disease transmission pathways to be identified with greater accuracy” [Hi]. This has been enabled by the UK Animal and Plant Health Agency (APHA) adopting a modified version of the University of Oxford-PHE human TB pipeline in 2018. APHA has validated and accredited WGS, and since 2019 it is in routine use for surveillance and research into the control of TB [Hii]. APHA uses a web portal called ‘ViewBovine’ developed with Peto and University of Oxford colleagues to utilize the WGS data [Hii]. These University of Oxford tools have been used by APHA to identify infection pathways and sources more rapidly, accurately and with greater certainty than standard methods [I].

Bash the Bug (set up by Fowler and University of Oxford colleagues in April 2017) is the most popular biomedical citizen science project on the Zooniverse platform. Volunteers help to classify images of bacteria growing on a microtitre plate developed by the CRyPTIC consortium to gather data on the drug susceptibility of approximately 20,000 M. tuberculosis samples, which are subject to WGS. More than 45,000 users from at least 19 countries have contributed more than 4,500,000 image classifications up to August 2020. A survey of Zooniverse volunteers by the University of Nottingham (May 2020) found that 83% of 92 respondents who listed Bash the Bug as their favourite project had sought out ways to learn about scientific topics since taking part [Ji]. Similarly, in response to a single-question online poll (Oct 2020), 43 of 49 respondents (88%) agreed with the statement “ BashtheBug has increased my interest in medical research” [Jii]. Comments received from volunteers showed Bash the Bug has provided benefits to some people in empowering them to help medical research, particularly during the COVID-19 pandemic, when notably there was a huge spike in classifications in April-May 2020 [Jii].

Liver transplantation is a highly effective treatment for end-stage liver failure, but is heavily constrained by the limited pool of donor organs. In the UK in the year to March 2018, there were 1,574 deceased donors, 1,149 liver retrievals and 975 liver transplants, indicating 62% overall liver utilisation. The highest-risk donor organs, including from donors after circulatory death rather than donors after brain death, were particularly poorly utilised (34%). As a result of this under-utilisation, in 2018 some 30% of listed patients were still waiting for a transplant for six months after listing, and 12% of listed patients died whilst still on the waiting list within two years [A] .

The research undertaken by Friend demonstrated the potential to increase the number of livers available for transplantation through increased utilisation of donor livers by warm perfusion, and the collaboration with Coussios made this technically possible through the development of the OrganOx metra. The first patient to benefit from this new approach was successfully transplanted in February 2013, and following successful completion of this Phase I study [5] the OrganOx metra received its CE mark as a medical device in 2016. Based on evidence from [6], machine perfusion was endorsed by NICE in January 2019 [B]. By December 2020, the OrganOx metra was being used routinely across all seven UK transplant centres in the UK (Birmingham, Cambridge, King’s, Leeds, Edinburgh, Royal Free and Newcastle). It total it has been used to enable over 850 liver transplants in 4 continents and 13 countries (UK, Austria, Italy, Spain, Germany, Belgium, France, the Netherlands, India, UAE, Australia, Canada, and USA) [C].

The OrganOx metra has significant benefits for both liver transplant recipients and patients with end stage liver disease who are awaiting a liver transplant. Improved patient outcomes are due to a reduction in damage to the perfused liver after it has been removed from the donor, and the ability to assess the donated liver’s function before it is transplanted, which result in more livers available for transplantation by using organs which would previously have been discarded. By increasing how long the liver can be stored before a transplant, clinicians can also choose to undertake a transplant at the optimal time point.

In the large randomised trial [6], the Oxford team halved the organ discard rate from 24% to 12%. Even though more organs (including more marginal organs) were utilised and were preserved for 50% longer, the post-transplant outcomes were still significantly better for normothermically preserved organs, as evidenced by a 50% decrease in post-transplant graft injury. The Professor of Transplantation at Addenbrooke’s Hospital, Cambridge, the first UK centre to fully adopt the OrganOx metra for routine clinical use, confirmed in October 2020:

“Since its introduction into our routine clinical practice in February 2018 we have increasingly used the machine in three settings, to assess donor livers where doubt existed over viability, to enable us to tackle difficult recipient procedures without the pressure of accumulating cold ischaemia, and to overcome logistical issues such as where we accepted two livers simultaneously. By far the biggest use (70%) has been in the assessment of donor livers. Since we started using the normothermic liver perfusion our transplant activity has increased around 30%...As you would imagine this has had an impact on our waiting times and waiting list mortality, as well as enabling us to successfully use livers that would otherwise not have been used by any centre in the UK.” [D]

NICE issued a press release in January 2019 announcing their recommendation of the procedure and highlighting the benefits to patients:

“Every year hundreds of people with advanced liver disease die whilst waiting for a transplant. This new device offers real hope as it may improve transplant outcomes and allow livers that were previously thought to be unsuitable to be used and also increase the time that livers are able to be kept. It is an exciting development that has the potential to shorten waiting list times and reduce mortality rates from advanced liver disease. After transplant, the vast majority of people go on to lead full and healthy lives and it is truly amazing to see the transformation.” Director of Policy at the British Liver Trust [E].

In addition to increasing donated organ utilisation, the introduction of the OrganOx metra in to the seven UK transplant centres has had a significant impact on clinical practice in the field of liver transplantation. An independent study (VITTAL: Viability Testing and Transplantation of Marginal Livers, NCT02740608), led by the University of Birmingham between November 2016 – February 2018, investigated the impact of normothermic machine perfusion on utilisation of liver grafts that had been discarded by all seven liver transplant centres in the UK. Of the 31 liver grafts perfused on the OrganOx metra, 22 were subsequently transplanted with 100% 90-day survival, demonstrating that normothermic machine perfusion can enable safe utilisation of over 70% of presently discarded grafts [F]. Chief Investigator of the VITTAL study, transplant surgeon at University Hospital's Birmingham NHS Foundation Trust, said:

“In the 30 years I've been involved with transplantation, there have been three or four events which have been game changers and I'm absolutely certain we are looking at a game changer that will change the way we practice organ storage and transplantation. It is already changing practice at the centres that have been able to use this technology either within clinical trials or within an expansion of service evaluation” [E].

The OrganOx metra has also had a significant impact on transplantation logistics. By increasing preservation times from the conventional maximum of 12 hours to well over 30 hours [G], clinicians have more flexibility about the timing of surgery and are able to plan more transplants as day cases, as noted in [6]. A study on the introduction of normothermic machine perfusion (NMP) at the Medical University of Innsbruck, Austria, found that “NMP in a multidisciplinary approach enables a safe prolongation of liver preservation and overnight organ care. A first field test of NMP indicates safety and benefit of this approach” [G]. This experience led to the omission of night-time procedures and parallel transplantations in Innsbruck [G]. In a separate study, use of the OrganOx metra device allowed 84% of transplants to take place during the day versus 65% with static cold storage [H]. This may help transplant surgical teams avoid ‘burn-out’ of key personnel. There is good evidence that daytime operations also have better outcomes for patients; for example, Yang et al [I] found that after-hours surgery was associated with significantly increased postoperative mortality and morbidity, and that: “ The timing of surgery plays a role in the outcome after the procedure and that both mortality and morbidities increase in surgeries performed outside of the regular working hours”.

A UK health economic study conducted by the University of Southampton in 2019 [J] concluded that, even though normothermic machine perfusion is more costly than static cold storage, it is also significantly more cost-effective by virtue of enabling additional transplants, with patients experiencing lower rates of early allograft dysfunction and adverse events: its incremental cost-effectiveness ratio was found to be GBP7,876 per Quality-Adjusted Life Year (QALY) gained.

In the first three years since CE marking in 2016, the OrganOx metra has seen over GBP7,000,000 of cumulative direct sales since product CE marking in 2016. As of October 2020, the company employs 29 staff (headcount): 23 in the UK, 4 in the US, 1 based in France and 1 based in Germany. Revenue for the period 1st May 2019 – 30th April 2020 was GBP2,200,000 which represents a 30% increase in revenue from the previous year [ C]. 40 devices are in use worldwide across Europe, North America, Asia and Australia. In January 2020, the company received a GBP4,600,000 investment from BGF, a UK investment company, to enable expansion in the US, where a subsidiary has been incorporated [C].

The OrganOx metra was recognised in the 2013 IET Innovation Awards, receiving 1^st prize in 3 of 15 categories, including: ‘Best Healthcare Technology’, ‘Best Intelligent System’ and ‘Best Emerging Technology Design’. The OrganOx metra was further shortlisted as one of 4 finalists for the 2019 50^th anniversary MacRobert award of the Royal Academy of Engineering. In September 2019, OrganOx won the ‘Best Proof-of-Value of an Innovation’ category at the US Medtech Insight Awards in Boston, attracting the following citation:

“This is a major advancement in organ preservation. There is well-documented unmet medical need in this area, and it is a true achievement to have a method to help secure desperately needed organs for patients. The judges commented that OrganOx represents the true aspects of what it takes to be best value and best for patients. The need for better preservation for transported organs has been a true roadblock for patients desperately waiting for organ transplants. The perseverance of the team is obvious, and they overcame huge roadblocks to achieve success.” [K].

The University of Oxford research showed the utility and validity of generating, analysing and interpreting data from pathogen WGS to uncover the underlying drivers of outbreaks, enabling effective interventions [e.g. 1, 2, 4, 6]. Their methods now underpin routine pathogen WGS in healthcare organisations and public health agencies. Since their original demonstration of the power of WGS to address healthcare-associated infections [1], pathogen WGS has expanded rapidly, with more than 20,000 C. difficile and 78,000 S. aureus genomes sequenced (deposited in NCBI short-read archive, Nov 2020).

Crook was PHE Director of the National Infection Service 2015-2019 and the University of Oxford team worked closely with PHE to address specific outbreaks [e.g. 5, 6], and to develop national strategies, bioinformatic tools and methods [building on 4 and 6] for routine analysis of pathogen sequence data to reduce outbreaks, particularly in healthcare settings (for example, for C.difficile, [A]). Embedding WGS in PHE labs and optimizing the use of WGS-based information is now Strategic Priority 7 of the PHE Infectious Diseases Strategy 2020-25 (published Sep 2019, [B]).

This research has also contributed to implementation of WGS for outbreak control by public health bodies internationally. For example, an Associate Director for Science at the US Centers for Disease Control and Prevention (CDC) stated that the University of Oxford research informed CDC policy and procedures, guided the field in how to establish and implement WGS bioinformatics, and “their greatest impact is in leading the application of advanced genomics for HAIs [healthcare acquired infections]” [Ci]. Further, the Associate Director of Clinical Microbiology at the University of Virginia, US, stated: “ research and researchers at University of Oxford…have been integral to our ability to assist the entire United States with [nosocomial] outbreak investigation” [Cii]. An example from Virginia is that a WGS investigation based on University of Oxford methods informed redesign of hospital plumbing, dramatically reducing spread of antibiotic resistant bacteria, reducing new patient infections with carbapenem resistant isolates from 70 to 15 per year, since 2016 [Cii]. In Australia, research by Crook and colleagues “ strongly influenced the implementation of genomics into routine diagnostic testing and public health laboratory surveillance in NSW [New South Wales],…reflected in the development and implementation of the NSW Health Genomics Strategy (June 2017) and the Australian National Microbial Genomics Framework 2019-2022”, according to a lead for public health microbiology in NSW [D].

C. difficile is the biggest cause of infectious diarrhoea in hospitalised patients and can lead to serious infections, with mortality of up to 25% in frail patients. It presents a worldwide problem, causing an estimated 500,000 infections and 29,000 deaths in the US in 2012, and more than 170,000 cases in Europe. The estimated annual attributable cost in the US is USD6,300,000,000.

**Redesign of national surveillance for C. difficile: Following the University of Oxford discovery [2] that sources other than symptomatic patients are the source of most C. difficile infections in endemic settings with routine infection control, PHE designed a new WGS-based national surveillance system for C. difficile, substantially based on University of Oxford research [A]. This is the world’s first structured national surveillance programme for C. difficile based on WGS. Sentinel hospitals were chosen based on University of Oxford analysis of national patient admission and transfer data. The system was approved and scheduled for implementation in 2020, but delayed until 2021 due to the COVID-19 pandemic [A]. The system is in the business plan of PHE for Healthcare Associated Infections (current, Dec 2020) [A], to provide all acute NHS trusts in England (approximately 150) with access to WGS and process approximately 12,000 C. difficile samples per year. According to the PHE lead on C. difficile the new system is “ a more cost-effective and future-resilient surveillance system to track known and provide an early warning of emergent C. difficile clones and types” [A], as well as extending global knowledge of how to diagnose, report and characterise a key infection that is a health service quality performance indicator in many countries, including the UK and US [A].

Widespread adoption of antimicrobial restriction to control C. difficile infection: The demonstration that fluoroquinolone restriction was associated with marked falls in incidence of C. difficile infection [3] has led to widespread adoption of polices to restrict fluoroquinolone use internationally. This research [3] influenced US public health and clinical leaders to focus on fluoroquinolones [Ci], and the 2017 update by the Infectious Diseases Society of America and Society for Healthcare Epidemiology of America of the Clinical Practice Guidelines for Clostridium difficile Infection in Adults and Children cites several studies by Crook and colleagues and recommends restriction of fluoroquinolones [E]. The 2018 guidelines of the European Society of Clinical Microbiology and Infectious Disease recommend “ Restriction of antibiotic agents/classes is effective in reducing CDI rates” [F], and cites University of Oxford research (including [2, 3]). C.difficile infections in the US reduced by approximately 22% from 2017 to 2019, which is likely due to a combination of infection control measures and altered antibiotic prescribing [Ci].

An illustrative example of benefits to healthcare providers is Betsi Cadwaladr University Health Board in Wales, which had one of the highest rates of C. difficile infection in the UK. A 2016 WGS investigation by Public Health Wales [G], collaborating with Eyre and colleagues [based on 3], found evidence of antibiotic use being a key driver in patterns of C. difficile infection. The report recommended prioritisation of antimicrobial stewardship across primary and secondary care [G]. A spokesperson from the health board was quoted by The Pharmaceutical Journal in 2019 [H] as stating that the Board had achieved a 38.7% decrease in C. difficile infections comparing 2017/18 to 2018/19, which they partly attributed to a decrease in the antibiotic prescription rate (12.6% decrease in primary care, April 2016-March 2019). The median cost of treating a patient with C. difficile infection is GBP7,500, compared to GBP2,800 for patients with other medical conditions, so a decrease in cases provides immediate cost-savings to the healthcare provider, as well as decreasing the high risk of mortality in frail patients.

Better infection control measures for C. auris: C. auris is an emerging multidrug resistant fungus, with healthcare-associated outbreaks detected in at least 30 countries. C. auris infection occurs predominantly in patients in high-dependency settings, with high mortality rates – up to 50% – and presenting major treatment challenges due to antifungal resistance and transmission within healthcare units. The University of Oxford discovery that re-useable equipment, including temperature probes, can be a major route for spread and persistence of this pathogen [4] had the immediate impact of stopping the Oxford hospital outbreak in August 2017. Previous infection control measures had no significant effects until removal of the temperature probes from the affected neurosciences intensive care unit. This was the first time an outbreak of C. auris had been ended with a clear understanding of the cause, and resulted in the immediate benefit of no further cases from an outbreak that had infected 70 highly vulnerable patients.

This discovery of the infection risk associated with multi-use equipment led to changes in healthcare guidelines. PHE updated C. auris guidelines in August 2017 (Ii), specifically referring to reusable equipment such as temperature probes, and there have been no major outbreaks since these changes in practice (Iii). The cost of controlling one previous outbreak was reported as more than GBP1,000,000 (Iii), so preventing outbreaks provides major cost-savings for healthcare and public health bodies. In the US, the University of Oxford research [4] informed CDC guidance on environmental cleaning for C. auris prevention [Ci]. CDC guidelines were updated in 2018 to specifically mention disinfection to avoid the risk of transmission via mobile equipment including temperature probes [Ji]. In South Africa, C. auris has been detected in approximately 100 hospitals and has caused large outbreaks. To address this, the Federation of Infectious Diseases Societies of Southern Africa issued new guidelines in 2019 [Jii], extensively recommending single-use equipment and disinfection of reusable equipment, citing [4].

Safer cardiac bypass surgery: From 2013 to 2017 in Europe, Australia and the US, more than 100 cases of severe M.chimaera-linked infections were reported, associated with cardiac surgery using HCUs, with high morbidity and mortality (9 of 18 patients (50%) in the UK died). Infection was estimated to occur in 1 in 100-1000 patients undergoing surgery with HCUs. Research [including 5] provided certainty about the source, which empowered the case for redesign of the equipment. Robust interventions led to the termination of the outbreak, through withdrawal of the implicated HCU and major improvements in HCU design by the company [Ki]. The International Society for Cardiovascular Infectious Diseases issued guidelines (Nov 2019) on prevention of M. chimera infection following cardiac surgery [Kii], citing this research [5]. No cases of M. chimera infection have been found in patients who had open heart surgery in the UK since 2016, showing that accurate identification of the outbreak source led to elimination of this dangerous infection.

The World Health Organisation (WHO) reports there were approximately 87,000,000 new cases of gonorrhoea in 2016, and it results in substantial morbidity and economic cost worldwide. Spread of drug-resistant gonorrhoea strains is a major problem and can occur quickly in high-risk populations. Since there are few treatment options, rapid investigation and prevention of outbreaks is essential. In 2018, the WGS methods developed by Eyre and colleagues [6] were adopted by PHE and public health bodies in Australia to conduct rapid investigations into extensively drug-resistant gonorrhoea, enabling tracing of sources and transmission and targeted public health intervention [Li, ii]. For example, two UK cases were traced to Ibiza, Spain, prompting a rapid local and international response to prevent spread [Lii, iii]. According to the WHO [Liv], the rapid contact tracing and public health messaging undertaken by PHE based on this work, resulted in low risk of further transmission in the UK, thus preventing further near-untreatable infections. Also, based on this research, in 2019 the ECDC called on EU/EEA member states to strengthen surveillance, testing and tracing (Liii), and the University of Oxford research (e.g. [6]) has contributed to the implementation of WGS for gonorrhoea surveillance by the WHO [M].

Shoulder pain is a common musculoskeletal condition associated with a high socioeconomic burden. Subacromial shoulder pain, which accounts for up to 70% of shoulder pain, has increasingly been treated with surgery over the past two decades, despite a lack of compelling evidence for clinical effectiveness or cost effectiveness. The University of Oxford team collaborated with the University of Bristol on a subsequent longitudinal study of use and cost of subacromial decompression surgery [A]; this investigated the use and cost of subacromial decompression in England over the last decade compared with other countries and explored how this related to the conduct and outcomes of randomised, placebo-controlled clinical trials e.g. [4]. They found that in England, the use of arthroscopic subacromial decompression (ASAD) increased by 91% from just over 15,000 procedures in 2007/8 to nearly 29,000 procedures in 2016/17 at a cost of approximately GBP125,000,000 per annum [A]. Despite this increase, publication of the CSAW trial results in November 2017 showed no difference in outcome between two surgical groups undergoing ASAD or placebo surgery for subacromial shoulder pain at any time point [3].

NHS England responded rapidly to the publication of the CSAW trial results. In 2018, the NHS England Evidence Based Intervention programme, which aims to reduce the number of inappropriate interventions on the NHS, placed ASAD for subacromial shoulder pain on a ‘procedures of limited value’ list for national Clinical Commissioning Groups, meaning ASAD would be performed only when specific criteria are met. The guidance for CCGs states “Recent research has indicated that in patients with pure subacromial impingement (with no other associated diagnoses such as rotator cuff tears, calcific tendinopathy and acromio-clavicular joint pain), non-operative management with a combination of exercise and physiotherapy is effective in the majority of cases.” [B p.35]. The CSAW trial [3] is cited as the primary source in the rationale for this recommendation. An updated patient information leaflet was produced (C).

Reduction of surgery rates and cost saving

Most of the increase in ASAD in England over the past two decades took place before 2011/12 [A]. The subsequent plateau in rates between 2011/12 to 2016/17, illustrated in the graph, could be attributed to raised awareness of the issue of treatment uncertainty and correlates with the CSAW trial registration in 2012, publication of the trial protocol in 2015 [3], presentation of the trial by Beard and Carr with the clinical community, and media attention [D].

Following publication of the CSAW trial results in 2017 and subsequent changes to NHS guidelines in 2018, annual rates of surgery in England have fallen 80% from 28,000 in 2016/17 to 5,720 in 2019/20 [A, E]. In 2016/2017, the median cost of an elective admission for ASAD alone was GBP4,476 [4]. This places ASAD costs at an estimated GBP129,000,000 in 2016/17 compared with GBP26,000,000 in 2019/20. The fall in ASAD rates following publication of the CSAW trial therefore represents a cost saving to the NHS of GBP103,000,000 per annum. A study published in 2020 found that surgery rates also fell by 29% in Scotland between 2014 and 2018 [F], and the authors identified this decline as “ *corresponding to the publication of epidemiological studies demonstrating a rise in [ASAD], and awareness of studies which questioned the benefit of [ASAD]*”, citing [1] and [4]. Meanwhile, the FIMPACT trial also found no clear evidence that ASAD is cost-effective [5].

The use of surgery has also been in decline in other countries, and this has been linked to CSAW trial data. For example, in Sweden rates of ASAD increased from approximately 80 per 100,000 in 2008/09 to approximately 150 per 100,000 in 2014/15 and subsequently decreased to 120 per 100,000 in 2018/19. According to the Professor of Orthopaedic Surgery at Lund University, Sweden: “This decrease was without doubt influenced by the awareness raised in the orthopaedic community of the limited evidence to support for efficacy of the procedure by first the publication of your trial protocol, and then the outcomes of the trial in The Lancet.” He continues: “ The Publication of the CSAW trial, and associated papers, has initiated a review within the Swedish Governmental Institutes of ‘National Board of Health and Welfare’, and the ‘Swedish Agency for Health Technology Assessment’…The result will be an update of the Swedish national clinical guidelines for patients with shoulder pain” [Gi]. The Director of the Whitlam Orthopaedic Research Centre reports changes in practice in response to the CSAW trial:

“I have been pleased to see the practice change that has occurred in sub-acromial decompression surgery worldwide. In my own country, rates of sub-acromial decompression have significantly declined and this has been seen in many countries worldwide, including the UK” [Gii].

Arthroscopic surgery is low risk, but some patients can experience infection, stiff shoulder, damage to blood vessels or nerves and ongoing pain [G]. Furthermore, analysis of mixed shoulder arthroscopic procedures suggests around 5 in 1,000 patients will suffer a significant or life threatening complication such as deep infection, pulmonary embolus or death. This small incidence of serious harms has been reduced as a result in the fall in surgery fates following publication of the CSAW trial. Indeed, a BMJ Rapid Response Review on Subacromial Decompression for Shoulder Pain published in 2019 in response to the CSAW and FIMPACT trials concluded, “ almost all informed patients would choose to avoid surgery because there is no benefit but there are harms and it is burdensome. Subacromial decompression surgery should not be routinely offered to patients with sub acromial shoulder pain” [H]. The review also highlights the burden of surgery on patients including a 2–10 hour outpatient stay following surgery, approximately two weeks off work during recovery and avoidance of activities such as overhead carrying for three months.

As noted by the Director of the Whitlam Orthopaedic Research Centre: “The impact on health that has largely stemmed from the C-SAW paper is not only felt in cost savings by avoiding unnecessary care, but it has also decreased the risks associated with unnecessary care” [Gii]. Similarly the President of the Canadian Orthopaedic Association and Research Chair in Evidence-Based Orthopaedics noted, “ *I firmly believe that the CSAW trial has decreased the risks associated with sub-acromial decompression, allowing patients to return to their usual activities with greater confidence and ease.*” [Giii]

In addition to the direct effect on shoulder pain care, the philosophy and approach to CSAW has resulted in many wider impacts on the scientific and public perception of surgical evaluation which remain ongoing. The CSAW trial led to significant media coverage, sparking public debate on the role of placebo surgery [I]. Together with epidemiological studies [1] it has also contributed to a step change in the orthopaedic community’s acceptance of evidence-based decisions for costly based interventions thereby maximising the effectiveness of hospital care budgets, for example in one account from Australia:

“I have found the C-SAW publication to be a recent example of one that has had a major impact not only in changing clinical practice, but in changing it for the better in that clinical practice is now much more reliant on the evidence available”

Director, Whitlam Orthopaedic Research Centre, Australia [Gii].

From being a rare and pharmacologically influenced design, the 3-way placebo surgical control trial pioneered in CSAW is now championed as a rigorous and appropriate method for fundamental assessment of surgical efficacy and treatment mechanisms in all surgical subspecialities. In 2018, the MRC and NIHR Methodology Research Programme jointly commissioned a workshop to set guidelines for use of placebo controls in surgical trials, led by Beard, leading to the ASPIRE guidelines to assist commissioners and researchers at a global level to design and conduct high quality placebo surgical trials [J]. Several surgical placebo control trials are now underway both in the UK and abroad, including the ACCURATE trial recruiting from Finland, Sweden and Norway. The Director of the Whitlam Centre wrote that “…the impact of the C-SAW study has been felt worldwide … in the change it has made to orthopaedic thinking by pushing it towards a greater understanding of and respect for the evidence” [Gii]. The President of the Canadian Orthopaedic Association summarised the impact of the University of Oxford research:

“Without doubt, the CSAW trial has had a global impact on research and clinical practice. This trial has empowered health care professionals and clinician scientists to pursue similar trials and sparked many discussions on the topic of placebo surgery within the orthopaedic field” [Giii].

Screening and management of PSA-detected prostate cancer has been a controversial healthcare topic globally. Cancer Research UK estimates that there are approximately 48,500 new cases of prostate cancer each year in the UK, and the American Cancer Society estimates that there are approximately 191,930 new cases each year in the US.

Opportunistic PSA testing in asymptomatic men has led to early detection and drives subsequent radical treatments that were intended to improve quantity and quality of life of these men, but are associated with a risk of significant adverse side-effects. Prior to the initiation of the ProtecT trial, studies in the UK (British Association of Urological Surgeons) and US (Department of Veterans Affairs) showed that over 90% of men diagnosed with localised prostate cancer received radical treatment, mostly in the form of radical prostatectomy.

The Oxford research [1 – 5] provided a powerful evidence base for the effectiveness of screening and treatments, highlighting how the harms of screening and treatment can outweigh the benefits. This has informed practice changes globally, reflected in global NICE, US and European guidelines recommending more active monitoring to reduce harmful non-beneficial interventions allowing clinicians and individuals to benefit by making better decisions about care.

The NICE guideline on Prostate cancer: diagnosis and management published in May 2019 [Ai] describes and references the ProtecT trial in detail and appends a presentation given by Hamdy and Donovan (Oct 2018), which cites [1, 2 and 5] to the evidence considered by NICE [Aii]. The research informed the NICE guidelines as follows:

• Citing [2], NICE recommends active surveillance be offered as an option to people with low-risk localised prostate cancer: “ Based on the evidence from the ProtecT trial (Donovan et al., 2016), the choice of active surveillance, prostatectomy or radiotherapy appears to be a trade-off between the benefits offered by prostatectomy and radiotherapy against their potential risk of side effects…….. Based on this evidence, the committee decided that all three treatment options may be suitable for different people and therefore agreed to keep the existing recommendation to offer active surveillance as an option to people with low-risk localised prostate cancer” [Aii]

• Based on [2], NICE kept the recommendation that the progression to radical treatment should be based on the man’s personal preferences. Side-effect profiles extracted from and referring to ProtecT analysis are listed in the guidance [Ai].

The US Preventive Services Task Force changed its policy for prostate cancer screening from a D (“Discourage the use of this service”) recommendation in 2011, to a C recommendation (“Offer or provide this service for selected patients depending on individual Circumstances”) in 2018 for men aged 55 - 69 years, largely based on outcomes from the ProtecT and CAP trials. The cited Oxford research that showed no significant improvement in all-cause or prostate cancer mortality in any of the treatment groups [1, 5] and the analysis of surgery and radical treatment vs. active monitoring [3] is cited in the policy. The “ change in recommendation grade further reflects new evidence about and increased use of active surveillance of low-risk prostate cancer, which may reduce the risk of subsequent harms from screening.” The guidelines go on to elaborate that “ the decision to be screened for prostate cancer should be an individual one” and “ before deciding whether to be screened, men aged 55 to 69 years should have an opportunity to discuss the potential benefits and harms of screening with their clinician and to incorporate their values and preferences in the decision”. [B]

The European Association of Urology prostate cancer guidelines published in 2019, refer to the ProtecT Study research [1] on describing Active Surveillance (AS) / Active Monitoring (AM) thus: “ …, the ProtecT study has reinforced the role of deferred active treatment (i.e. either AS or some form of initial AM) as a feasible alternative to active curative interventions for patients with low-grade and low-stage disease” and that “AS should be the default management strategy in patients with low-risk disease and a life expectancy > 10 years” [C].

The NHS and other healthcare providers have benefitted from costs and resources avoided by not implementing a regular PSA test-based screening platform. A UK options appraisal [D] found that annual screening would result in almost 10,000,000 more PSA tests per year and 1,400,000 biopsies and, therefore, a large increase in many resources would be required (e.g. GP nurse sessions, PSA tests, radical treatments, hormone treatment, outpatient appointments). Total additional lifetime costs for a cohort of men aged 50 of a screen-once policy at 50 were estimated as GBP50,000,000, rising to almost GBP1,000,000,000 for an annual screening policy in the UK [D]. The clinical costs of screening in the UK, without administrative costs, were estimated to be GBP600,000,000 to GBP1,700,000,000 per year [E].

In November 2020, the UK National Screening Committee (UK NSC) completed its first periodic review since 2016 and concluded that a screening programme for prostate cancer in the UK was not recommended, for reasons including the harms caused and no single better treatment for early-stage disease [H]. The evidence document included numerical data from the ProtecT trial and cited [1] and [2] as part of the 2019 NICE review for NG131 [A].

The high-quality evidence from ProtecT enables men to make properly-informed choices about testing and, for those with clinically localised prostate cancer, to decide on the balance between the benefits and side-effect profiles of radical treatments.

The cohort study of men undergoing prostate biopsy in the ProtecT study showed that 1.3% required hospital admission and 10.4% consultation with a doctor because of post-biopsy adverse symptoms including pain, fever, and blood in urine, faeces and ejaculate [4]. Among the two-thirds of men who received a negative or inconclusive biopsy result as part of ProtecT, around 20% reported high distress persisting up to 12 weeks [F]. By contributing to the evidence base used by policy makers to decide not to put screening in place, the Oxford research has helped avoid 1,400,000 biopsies per year [D] and subsequent harms avoided e.g. an estimated 18,200 men requiring hospital admission, 145,600 men suffering post-biopsy adverse symptoms and 184,800 men suffering high distress, per year.

The National Prostate Cancer Audit (NPCA) assesses the quality of services in England and Wales, and aimed to reduce over-treatment of men with low-risk prostate cancer, by auditing the percentage of men with low-risk prostate cancer receiving radical treatment. In 2016, the rate of over-treatment was 12%, and NPCA authors cited [1] and commented: “the proportion of men with low-risk disease being potentially ‘over-treated’ is an area of concern, especially given the recent publication of the ProtecT study” [Gi]. In the 2 years after the publication of the main ProtecT study clinical outcomes, over-treatment reduced by two-thirds (from 12% to 8% in 2017 [Gii], and to 4% in 2018 [Giii]). In 2017, the NPCA noted that “the trend seen towards a reduction in men with low-risk disease being “potentially over-treated” is encouraging” and suggested that the results of studies such as [1] were being disseminated into national practice [Gii].

Between 31 December 2019 and 31 December 2020 there were more than 83,207,000 confirmed cases of COVID-19 worldwide. The case fatality rate has been estimated at approximately 1% in high income countries, or 20-25% of all hospitalised patients. By performing the largest clinical trial of COVID-19 treatments, robustly and at speed, the RECOVERY trial led by the University of Oxford has achieved worldwide impact in guiding treatment, saving lives, and demonstrating the power of evidence-based medicine. For the first time, treatment of an epidemic disease was changed during a pandemic.

The streamlined trial design achieved unprecedented speed in initiation and recruitment, catching the first peak of infections in the UK, essential in a global pandemic with no known effective treatments. On 16 March 2020, the Chief Medical Officer and NHS England Medical Director endorsed RECOVERY and urged all NHS Trusts to adopt the trial, emphasising that it was crucial research and had been kept extremely simple [Ai]. Within 16 days, 1,000 patients had been randomised, 10,000 by 14 May, and 20,000 by 8 December 2020. Participants were recruited at 176 NHS hospital organisations and, during 2020, 10% of hospitalised UK COVID-19 patients were recruited. In August 2020, RECOVERY was selected to be the UK national platform for phase II as well as phase III COVID-19 trials, based on its unique national coverage and recruitment success [Aii].

In March 2020, Horby and colleagues reviewed COVID-19 treatment guidelines finding that corticosteroids were specifically not recommended in most COVID-19 treatment guidelines [B]. Less than 3 months after the trial started, RECOVERY proved that the corticosteroid dexamethasone reduces COVID-19 mortality, by approximately one third in ventilated patients and one fifth in oxygen-treated patients [5]. In December 2020, corticosteroids remained the only globally-available drugs proven to reduce mortality in severe and critical COVID-19.

Impact on national and international policy and clinical guidelines: It is unprecedented that research results are announced at lunchtime, become policy and practice by evening, and save lives by the weekend. The results were announced on 16 June 2020, a day on which 993 people in the UK died from COVID-19. Within 4 hours, dexamethasone – the world’s first coronavirus treatment proven to reduce the risk of death – was recommended for use across the NHS, and the Chief Medical Officer instructed hospitals to act immediately [Ci], stating “ dexamethasone has been demonstrated to have a clear place in the management of hospitalised patients” and urging clinicians to use it for patients who require oxygen or ventilation.

In the US, on 17 July 2020 the National Institutes of Health (NIH) [Cii] changed its guidance: “On the basis of the…(RECOVERY) trial,…(the Panel) recommends using dexamethasone”. COVID management protocols were revised globally to add recommend dexamethasone, including Saudi Arabia on 17^th June [Ciii], South Africa on 20 June [Civ], and India on 27 June [Cv]. On 22 June 2020, the WHO reviewed their guidance on corticosteroids for COVID-19 “ triggered…by the publication of the preliminary report of the RECOVERY trial” [Cvi]. The WHO conducted a meta-analysis of corticosteroids trials, of which RECOVERY was by far the largest, and the revised guidance, published on 2 Sep 2020, recommended “ *systemic corticosteroids rather than no corticosteroids for the treatment of patients with severe and critical COVID-19 (strong recommendation, based on moderate certainty evidence)*” based on the RECOVERY result [Cvi]. On 18 Sep 2020, the European Medicines Authority (EMA), unusually, provided a template for manufacturers to accelerate submission of amendments to their dexamethasone drug licenses to include the new indication, based on the RECOVERY results [Cvii].

Benefits to patients and global clinical care: Patients in the dexamethasone group in the trial benefited both from increased chance of survival and shorter duration of hospitalisation [3]. Health economics analysis [D] estimates that in the UK 12,000 lives (90% confidence interval, 4,250 - 27,000) were saved between 1 July 2020 and 31 December 2020. If dexamethasone has a similar effect size in settings where access to oxygen therapies is limited, in the same period this would translate into approximately 650,000 lives (90% confidence interval 240,000 - 1,400,000) saved globally [D].

Dexamethasone is off-patent, affordably available in most countries, on the WHO essential medicines list since 1977, and can be taken by everyone: for less than GBP50, eight patients can be treated and one life saved. It rapidly became standard of care for the sickest patients across the world. Six days after the RECOVERY result, the drug purchaser Vizient, which supplies approximately half of US hospitals, reported a 610% increase in demand for dexamethasone [Ei]. Independent analysis of US prescribing rates by health care technology company Aetion shows dexamethasone use for COVID-19 in hospital rising from 28% on 14 June 2020 to 52% on 28 June [Eii]. Clinical data reports from International Severe Acute Respiratory and emerging Infections Consortium (ISARIC), gathered from more than 550 sites across 42 countries, show high levels of steroid use globally since the RECOVERY result [F]: for patients admitted since 16 June (until 9 Nov 2020), 70% of those ventilated and 43% of those on oxygen received steroids.

Adoption of dexamethasone, based on RECOVERY, is widely credited with contributing to the decline in COVID-19 mortality; decreases of 18% in death rates for hospitalised COVID-19 patients have been reported between March and August 2020. For example, clinicians in the US [Gi] and India [Gii] are quoted in scientific news articles as attributing decreased mortality to steroids alongside other improvements in patient care.

Benefits to patients and healthcare providers: Learning a treatment is not effective is important as it protects patients from potential harm and avoids wasting resources. Early in the pandemic, in March 2020, both hydroxychloroquine and lopinavir-ritonavir were widely recommended [B], and hydroxychloroquine was championed by US President Donald Trump. RECOVERY announced on 5 June and 28 June 2020 that hydroxychloroquine and lopinavir-ritonavir, respectively, are ineffective for COVID-19. The large scale of RECOVERY allowed a definitive conclusion – and certainty for clinicians – on the lack of benefit of both treatments among hospitalised patients, which had been initially suggested by smaller trials (e.g. [2]) and non-randomised studies. As a direct result of RECOVERY, other clinical trials of both treatments were rapidly halted, including these arms of the WHO’s large, international SOLIDARITY trial [Hi,ii]. The hydroxychloroquine and lopinavir-ritonavir interim results from SOLIDARITY were consistent with RECOVERY [Hiii]. Both treatments have can have serious side-effects, including potentially fatal heart arrhythmias associated with hydroxychloroquine, so preventing unnecessary and ineffective prescribing reduced risks to patients, as well as avoiding raising false expectations. Proving that these drugs do not work avoided wasted resources for healthcare providers.

Impact on policy and clinical guidelines: The US Food and Drug Administration (FDA) had granted emergency use of hydroxychloroquine and chloroquine for COVID-19 on 28 March 2020, and the US government distributed millions of doses to treat patients not enrolled in clinical trials. As a direct result of the RECOVERY finding, the FDA revoked the emergency approval on 15 July 2020 [Ii], stating “Only randomized controlled trials can answer the question of whether HCQ or CQ is of clinical benefit in hospitalized patients with COVID-19, and the RECOVERY Trial results offer persuasive evidence of a lack of benefit of HCQ”. Hydroxychloroquine is not recommended by the WHO or EMA for COVID-19, with the EMA citing RECOVERY and SOLIDARITY [Iii]. The lopinavir-ritonavir drug regime is no longer recommended for COVID-19 by any international guidelines. Therefore, three RECOVERY results in little more than three weeks turned COVID clinical guidelines on their head: from widespread use of hydroxychloroquine and lopinavir-ritonavir and low use of dexamethasone in March, to the opposite pattern in July 2020. Subsequently, RECOVERY also found no benefit from azithromycin in patients hospitalised with COVID-19; this was announced on 14 Dec 2020 [Ji] and on 15 Dec the NHS recommended that azithromycin should not be used for these patients [Jii]. This change avoids inappropriate antibiotic use, which can increase antibiotic resistance.

RECOVERY has played a critical role, through media coverage, in changing the public perception of the importance of evidence-based medicine. During 2020, RECOVERY was covered 15,203 times in the media (online, print and broadcast), and #RECOVERYtrial was mentioned 19,000 times on social media. In particular, RECOVERY’s power to counter vocal claims of the beneficial effects of hydroxychloroquine has been an influential tool against fake news. The most prominent example is Twitter limiting the account of Donald Trump Jr and ordering him to delete a misleading tweet containing a video on 28 July 2020 after he made claims about the utility of hydroxychloroquine, which RECOVERY had already proved to be false [K]. Twitter also deleted several tweets shared by US President Donald Trump that contained the false claims, and added a note to its trending topics warning about the potential risks of hydroxychloroquine use [K]. In an article about RECOVERY and SOLIDARITY, expert authors including the Director of the Institute for Evidence-Based Healthcare, Bond University, Australia, commented “ it has been refreshing to see how perfectly such weakly founded claims [of efficacy] can be swept aside by evidence from properly conducted, large-scale, randomized trials” [L].

ChAdOx1 was patented in 2012 (PCT/GB2012/000467) and has been used in 11 vaccine programmes, including for malaria, influenza, Chikungunya, and MERS, showing a good safety profile and high immunogenicity. Recognising the potential for vaccine development using ChAdOx1, Gilbert and Hill founded Vaccitech in January 2016 to develop vaccine and T cell therapeutic products to improve global health. Vaccitech was spun out from the University of Oxford, and licensed intellectual property rights to the ChAdOx1 technology [1], raising GBP10,000,000 investment from Oxford Sciences Innovation (OSI) [A]. Vaccitech has subsequently attracted additional investment of GBP57,000,000 between Nov 2017 and Dec 2020 enabling programme development [A]. Vaccitech has grown from 1 employee in May 2017 to almost 50 employees in Dec 2020 [A], contributing to business development in the local area and has subsidiaries in Australia and the US to support global operations [A]. HAV Vaccines in London licensed ChAdOx technology from Oxford University Innovation (OUI) to develop a vaccine to treat and prevent infection by Mycobacterium avium subspecies paratuberculosis (MAP) which is responsible for the chronic gut inflammation in patients with Crohn’s Disease. A phase 1 study on healthy volunteers [Bi] showed no safety concerns and patients with Crohn’s Disease are being recruited into a follow-on phase 1 trial [Bii]. Aelix Therapeutics in Spain has also licensed the ChAdOx technology from OUI and is using the technology to develop a therapeutic vaccine for the treatment of HIV. Aelix Therapeutics launched a phase 1, randomised, double-blind, placebo-controlled safety, tolerability, and immunogenicity clinical trial in 45 individuals with HIV in July 2017 [C].

Between 23 April and 4 November 2020, Oxford led clinical trials in geographically and ethnically diverse populations over three continents. 11,636 individuals were vaccinated with ChAdOx1 nCoV-19 in clinical trials in the UK, Brazil, and South Africa. Based on interim efficacy results [6], these individuals now have a 70.4% reduced risk of developing COVID-19 and vaccinated individuals are protected against hospitalisation with COVID-19. In August 2020, AstraZeneca began a Phase III clinical trial in the US to determine the safety, efficacy, and immunogenicity of AZD1222 [D], enrolling 32,449 participants, with two thirds receiving the COVID-19 vaccine and one third the placebo. Data from the US trial showed that the vaccine reduced risk of severe COVID-19 by 100%, of clinical COVID-19 by 76% and was 85% effective in older adults.

On 30 December 2020, following review of the Oxford clinical trial data by the Commission on Human Medicines, ChAdOx1 nCoV-19/AZD1222 was authorised for emergency use supply in the UK by the Medicines and Healthcare Products Regulatory Agency (MHRA) [E], the first viral vectored vaccine against COVID-19 to be granted approval in Europe. Oxford research [6] was used to inform the Joint Committee on Vaccination and Immunisation (JCVI) decision on dosing regimen for the Oxford vaccine, and for the Pfizer-BioNTech mRNA vaccine that had received regulatory approval in the UK on 2 December 2020 and whose clinical trial data used a three-week dosing interval. Oxford research [6] showed that a longer interval between the first and second vaccine dose promoted a stronger immune response. On 31 December 2020, the JCVI recommended a two-dose schedule for the Pfizer-BioNTech and Oxford-AZ vaccines, with a maximal interval of 12 weeks between the first and second dose for both vaccines, based on the likelihood that immune responses to both vaccines were similar [F]. The Committee recommended prioritising delivery of the first vaccine dose as this was highly likely to have a greater public health impact, allowing more people to be vaccinated, and reduce the number of preventable deaths from COVID-19 [F]. Emergency use authorisation of ChAdOx1 nCoV-19/AZD1222 was granted in Argentina and El Salvador on 30 December 2020, and in Dominican Republic on 31 December 2020.

The UK, through the UK Vaccine Taskforce (VTF), was the first country in the world to secure access to the Oxford vaccine, securing 100,000,000 doses (enough for 50,000,000 people) in May 2020, 7 months before the vaccine received regulatory approval. The former Chair of the VTF wrote that,

“The portfolio of Oxford’s research on ChAdOx viral vectored vaccines was a critical factor for the VTF’s decision to secure 100 million doses of the Oxford/AZ vaccine before it had received regulatory approval” [G].

Guaranteeing doses of the Oxford vaccine helped the UK government meet its goal of securing access to COVID-19 vaccines for the UK population as quickly as possible [G].

On 30 April 2020, the University of Oxford and AstraZeneca signed a landmark agreement for the global development, manufacture, and distribution of ChAdOx1 nCoV-19/AZD1222 [Hi]. Of crucial importance to the Oxford/AstraZeneca partnership was the joint commitment to ensure broad and equitable access to the vaccine on a not-for-profit basis for the duration of the pandemic and in perpetuity to low and middle income countries (LMICs) [Hii]. In June 2020, AstraZeneca was the first global pharmaceutical company to join COVAX (COVID-19 Vaccines Global Access), a global initiative led by CEPI, GAVI and WHO, aiming to distribute 2 billion vaccine doses to 92 LMICs at no more than USD3 per dose. As of 31 December 2020, agreements were in place for the supply of 2,678,900,000 doses of the Oxford vaccine worldwide [I], almost twice as many as any other vaccine developer globally, and notably with approximately half committed to LMICs via COVAX [J]. The Oxford vaccine has the significant benefit of being easy to manufacture, store in a fridge and distribute, allowing it to be rapidly deployed in LMICs.

AstraZeneca confirmed,

“The ChAdOx platform was particularly attractive for us as it enabled the rapid development and deployment of novel vaccines with a manufacturing process and supply chain that would enable the equitable and global rollout we all envisaged, without some of the limitations of mRNA vaccines” [K].

Alongside expanding manufacturing capacity at AstraZeneca, the collaboration with Oxford also resulted in growth of

“a significant preclinical, clinical capability to support the vaccine programme. This now provides the platform for next generation vaccine development, not just for SARS CoV-2 but potentially for other important viral infections" [K].

The former Chair of the VTF said “ The development of the Oxford vaccine helped drive the growth of the UK’s vaccine capability” [G]. In May 2020, the UK government announced GBP131,000,000 investment to accelerate construction of the Vaccines Manufacturing and Innovation Centre (VMIC) at Harwell, Oxfordshire, and to create a virtual VMIC which could produce vaccine doses whilst VMIC was under construction [Li]. As of 31 December 2020, three of the seven most advanced COVID-19 vaccines were being manufactured in the UK, including the Oxford vaccine [Lii].

The former Chair of the VTF stated,

“The development of the Oxford vaccine has stimulated new investment in skills and training and dramatically expanded the UK’s capacity to manufacture and fill vials across a range of bioprocessing and fill finish companies throughout the UK” [G].

Typhoid fever is a systemic infection caused by Salmonella Typhi, usually through ingestion of contaminated food or water. Between 11,000,000 and 21,000,000 cases and 128,000 to 161,000 typhoid-related deaths occur annually worldwide. While provision of clean water and improved sanitation could control the disease, the engineering works required to do this across low- and middle-income countries cannot be achieved quickly. Vaccines therefore offer early impact on disease while definitive investment in infrastructure is awaited. However, lack of licensed vaccines suitable for administration to children younger than five years had stalled progress towards introduction and there was no investment in the expensive field trials to help build confidence by demonstrating impact. University of Oxford researchers, under the leadership of Pollard, Angus and Basnyat, undertook key research to address this challenge. They then joined a partnership between the University of Maryland School of Medicine and PATH (an international global health non-profit organisation) to aid Gavi (Global Alliance for Vaccines and Immunisation) eligible countries in accelerating introduction of typhoid vaccines for children between 6 months and 15 years old.

The new typhoid conjugate vaccines had the potential to overcome many of the challenges that impeded uptake of earlier vaccines, including longer-lasting protection, fewer doses, and suitability for children younger than two years of age, allowing for inclusion in routine childhood immunisation programmes. However, they needed to be tested in a safe and ethical way. Conducting human challenge studies, where volunteers received an infection with a live pathogen, required the team to overcome many challenges to meet modern manufacturing, regulatory, safety and ethical standards. The development of the human challenge model of typhoid infection informed the WHO guidelines on development of new typhoid conjugate vaccines, and was prioritised by the 64th meeting of the WHO Expert Committee on Biological Standardization, 21-25 October 2013 [A]. These guidelines were made available for all manufacturers developing new vaccines for the disease.

The efficacy data on the new typhoid conjugate vaccine Typbar-TCV from the human challenge study conducted by the University of Oxford in 2016 proving a protective efficacy of 54%-87% against typhoid [3], were requested by the WHO as evidence in the prequalification of Typbar-TCV. Prequalification enabled procurement by UNICEF, Pan-American Health Organisation and Gavi. The WHO’s SAGE recommendations and the prequalification of Typbar-TCV directly supported the decision by Gavi to release USD85,000,000 to support countries roll out typhoid vaccine, announced in early 2018 [B].

Learning that an approach is unlikely to be effective is important as it avoids wasted resources. Observations on the role of typhoid toxin in invasion study [2] following development of the human challenge model [1] halted further development of therapeutic monoclonal antibodies using typhoid toxin as a target and experimental vaccines based on the toxin.

Having established the human challenge model, Pollard then took the opportunity to use these volunteers’ donated blood to develop a serum standard for typhoid conjugate vaccines [6], which was approved by WHOs Expert Committee on Biological Standardisation (ECBS) for all manufacturers developing typhoid conjugate vaccines [C]. This is a biological reference material vital for international standardisation of vaccine development comparisons and immunogenicity studies. As of 14 December 2020, the NIBSC have sold 232 vials of the standard.

Field trial participants: Field trials of the Typbar-TCV were conducted in Malawi (28,000 children), Bangladesh (32,500 children) and Nepal (20,000 children). In August 2018, interim study results from the 20,000 participants aged 9 months to 15 years vaccinated in Nepal demonstrated an efficacy rate of 82% [5], and thus it can be estimated from the data that in a trial population of 80,000 vaccinated children, this should result in a reduction of 200 cases of typhoid fever per year.

Mass adoption: 2018 was the first year in which countries could apply for Gavi support for purchase and delivery of typhoid vaccines for children between 6 months and 15 years old, with the first routine immunisations taking place in 2019 [D]. Zimbabwe carried out a Gavi-funded mass Typbar-TCV vaccination campaign between February and March 2019 that targeted 318,000 children in communities affected by an ongoing typhoid outbreak in Harare [E]. A two-week vaccination campaign, targeting 10,000,000 children of 9 months to 15 years old in urban areas of the Sindh province in Pakistan with Typbar-TCV, took place in late 2019 [F], the beginning of Pakistan’s introduction of the vaccine into their routine immunisation programme.

Routine immunisation: Though previously thought to be at low risk of typhoid, when it was discovered that under 3 year olds were at risk, there were concerns about how to best protect them, specifically as this age group respond less well to the previous vaccines and more to the newer conjugate vaccines. This age group is difficult to study as testing vaccines for young children requires prior testing in adults. Developing a way to test the conjugate vaccine for efficacy quickly, using the human typhoid challenge model [1, 3] was a crucial step. Following on from Pakistan’s introduction of the vaccine in to their routine immunisation programme at the end of 2019, Liberia and Zimbabwe also approved the introduction of TCV into their routine childhood immunisation programmes (implementation was due to take place in 2020 but was delayed to 2021 due to COVID) [F].

Vaccine rollout was given additional impetus due to the increasing problem of antimicrobial resistance, meaning in some areas typhoid had become untreatable with antibiotics. This was particularly a problem in the Sindh province of Pakistan where the Pakistan Health Authorities documented the outbreak of typhoid fever cases citing 5,274 cases of extremely drug resistant typhoid of a total of 8,188 typhoid fever cases (2016 - 2018). The vaccine efficacy data [3], amongst other University of Oxford research, is cited in the WHO position and background papers on the use of typhoid vaccines to combat resistance [Gi]. Following presentation of the interim field trial results in Nepal [5], the administration of Typbar-TCV to 10,000,000 children (detailed above) in Pakistan was approved on an emergency basis in 2019. The vaccine was specifically used here to contain the disease, due to widespread antimicrobial resistance of S. Typhi against ceftriaxone [H], one of the most commonly used antibiotics against typhoid fever. The impact of this effective vaccine in the control of antimicrobial resistance has been shown to be crucial, as already evidenced by the number of cases halving in areas of Pakistan following vaccination [I] (formal evidence from Pakistan delayed publishing due to COVID).

Vaccine efficacy and safety data generated in the human challenge model were used to support the 2018 WHO Strategic Advisory Group of Experts on Immunisation (SAGE) global policy recommendations for TCVs [Gi]. The policy references utilising the human challenge model ([3] and two additional University of Oxford publications), in the section on Immunogenicity, efficacy and effectiveness of the TCV. This policy recommended that the vaccine is used in all high burden areas from 9 months to 15 years of age. University of Oxford research is extensively referenced in the background papers [Gii] and guidance [1, 3 and three additional University of Oxford publications] as it provided the only efficacy data available at the time the recommendations were made. Furthermore, the safety data from Oxford-led field trials in Nepal [5] were referenced in WHO Global Advisory Committee on Vaccine Safety’s review of typhoid vaccines [J].

Age-stratified data on clinical and sub-clinical disease burden is important for countries to decide whether to prioritise vaccine introduction. Pollard led large-scale studies of the burden of typhoid in Bangladesh, Malawi and Nepal to inform the respective Governments on the priority of typhoid control [4]. For example, Pollard has been working in Nepal with Gavi and PATH [Ki] and has engaged with the Nepal Government to support introduction of TCV in Nepal’s routine immunisation. Following discussion of the surveillance study [4] and human challenge model data [2], in 2017 the National Committee on Immunization Practices recommended “ the need for typhoid vaccine studies to be conducted in Nepal, and in the view of the huge burden of disease, Nepal should be an early adopter of typhoid conjugate vaccine once funding is available” [Kii].

The fatality rate for Ebola is 25-90%. The 2013 - 2015 West Africa Ebola virus outbreak was unprecedented in scale, with 28,639 reported cases and 11,316 reported deaths. The outbreak cost the economies of Guinea, Liberia, and Sierra Leone USD2,200,000,000 and impacted their healthcare systems through loss of health care workers (Liberia lost 8% of its doctors, nurses, and midwives to Ebola – a total of 513 deaths). Furthermore, the outbreak set back the

treatment and control of other diseases (an estimated additional 10,600 lives were lost to HIV, tuberculosis and malaria during the epidemic), greatly affected children (17,300 lost one or both parents, 33-39 weeks schooling lost, gaps in vaccination schedules) and restricted travel [A].

The rVSV-ZEBOV-GP vaccine (also known as rVSVΔG-ZEBOV-GP, or rVSV-ZEBOV) progressed from the first in human Phase I study in 2014 [1] to being tested in Guinea in the continuing West African outbreak in 2014/15 in a vaccination trial. In 2015, 7,651 vaccinees in Guinea (along with tens of thousands of potential contacts of these vaccinees), benefitted from protection against Ebola, estimated at nearing 100% [B]. Ultimately this vaccine was used under “expanded access” or what is also known as “compassionate use” for 16,000 people in Guinea in 2015. As a result of the University of Oxford led trials and successful use during the outbreak, the European Medicines Agency announced conditional authorisation of the rVSV-ZEBOV-GP vaccine in October 2019 [C] and WHO announced pre-qualification approval in November 2019 [D]. The regional representative for Africa for Epicentre, the research arm of Médecins Sans Frontières (MSF), who was in charge of Laboratory Coordination for the MSF/Epicentre Phase 3 trial for the rVSV-ZEBOV-GP vaccine in Guinea during this outbreak, described the importance of the vaccine: “ The Ebola vaccine appeared as a game changer” and he highlighted “ its positive impact on limiting the transmission of the disease or reducing the number of deaths” [E].

The 10^th Ebola outbreak in the North Kivu and Ituri provinces of the DRC was the second largest in the world with 3,481 cases and 2,299 deaths. In examining the evidence required to make a recommendation for use of a vaccine in Ebola outbreaks, the October 2018 background paper for the WHO Strategic Advisory Group of Experts (SAGE) deliberations [F] referred to the University of Oxford study [1] showing the rVSV-ZEBOV-GP vaccine was immunogenic, with higher titres of neutralising antibodies produced at higher vaccine doses. Preliminary vaccination efficacy results followed in April 2019 that concluded " …the rVSV-ZEBOV vaccine should contribute to bringing the current Ebola outbreak in the DRC to an end, and to controlling future outbreaks as effectively and rapidly as possible" [G]. In a benefit to medical practice and policy, interim recommendations by WHO SAGE in May 2019 [H] indicated that rVSV-ZEBOV-GP should be offered as a priority to vaccinate those at high risk. Over 303,000 people received the rVSV-ZEBOV-GP vaccine between August 2018 and May 2020, [Ii]. An academic study lead by the Yale School of Public Health estimated that the deployment of rVSV-ZEBOV-GP during this outbreak reduced both geographical spread and risk by up to 70%, compared with projections without any vaccination campaign [J]. Whereas in the earlier West African outbreak approximately 900 health-care workers were infected and 513 died due to Ebola, in this outbreak in the DRC nearly 30,000 frontline healthcare workers, laboratory workers, surveillance teams and burial teams were offered the single dose vaccine with an efficacy calculated as 97.5% [G].

WHO Ebola outbreak response capabilities were further improved through the results of the first in human clinical trials [3, 4] on the Ad26.ZEBOV/MVA-BN-Filo Ebola vaccine as these led to this vaccine becoming the second WHO recommended vaccine for use for control of Ebola outbreaks. Following the initial use of rVSV-ZEBOV-GP vaccine in 2014/2015, there were two reasons that further vaccine development was essential. One major concern was the limited supply of the (then) currently employed vaccine (rVSV-ZEBOV-GP) referred to in [H]. The other was that this vaccine was quite reactogenic giving a high proportion of vaccinees a marked febrile reaction. Given that vaccinees are often contacts of people with Ebola, there were concerns of public uptake of the vaccine and that there may be diagnostic confusion and unnecessary containment and screening for Ebola. The prompt results of the Ad26.ZEBOV/MVA-BN-Filo Phase I vaccine trial [3] and Medicines and Healthcare products Regulatory Agency review of vaccine manufacturing, meant it was possible to move at pace from the Phase 1 first in human (3) to subsequent African trials, which were vital if patient benefit from a second Ebola vaccine was to be realised during this DRC epidemic. Though still at the investigational vaccine development stage, in May 2019 the WHO SAGE advised that individuals with indirect exposure to Ebola in the DRC be immunised with the Ad26.ZEBOV/MVA-BN-Filo vaccine [H] which was subsequently deployed in large scale population intervention study for outbreak control in the DRC. Between October 2019 and April 2020, 20,339 people received the first dose of the Ad26.ZEBOV/MVA-BN-Filo vaccine, and 9,560 of them received the second dose [I(i)]. The Ad26.ZEBOV/MVA-BN-Filo schedule was given European Medicines Agency licensure in July 2020 [Ki and Kii].

The progression of the Ad26.ZEBOV/MVA-BN-Filo vaccine to licensure and effect on WHO guidance was built on numerous reports and recommendations that refer to the University of Oxford-led research, including:

• the WHO SAGE October 2018 background paper [F] refers directly to the evidence on duration of the antibody response in [4] “ the information on the duration of protection for… candidate Ebola vaccines is up to 360 days post vaccination for the … Ad26.ZEBOV/MVA-BN-Filo vaccines” and summarises the results from all three University of Oxford-led Ebola vaccine trials [1, 3 and 5].

• the May 2019 WHO SAGE Interim Ebola Recommendations [H], which recommend that lower risk populations be vaccinated with the Ad26.ZEBOV-GP vaccine, referred to the WHO Ebola Vaccines Decision framework April 12, 2019 [L] which described, reviewed and assessed the safety and immunogenicity of the Phase I and II trial data for the Ad26.ZEBOV/MVA-BN-Filo vaccine [3, 5] as being stronger than that for an alternate vaccine being considered for testing in the field.

The work of Snape in leading the Phase I clinical trial of the Ad26.ZEBOV/MVA-BN-Filo vaccines in a dramatically expedited study [3], followed by extended safety and immunogenicity studies [4, 5] led directly to the availability of a second crucial tool in the fight against Ebola. The Director of the Department of Research and Production, National Laboratory for Public Health of Congo, Brazzaville confirmed that “ The double use of …[both]…. vaccines rapidly brought the epidemic to an end…..the development of two successful vaccines against EVD is highly significant for global health” [M].

Days before the 10^th outbreak was declared over, the 11^th Ebola outbreak was declared, which resulted in 119 cases and 55 deaths. Despite occurring during the COVID-pandemic, vaccination efforts began just four days after the outbreak occurred and more than 40,000 people at high risk were vaccinated with the rVSV-ZEBOV-GP vaccine [I(ii)].