NHS Commissioning Support

Text removed is struck through; text added is underlined. Unchanged paragraphs are summarised rather than repeated.

Objective for processing

i5 Health Ltd provides consultancy services to support Integrated Care Boards (ICBs), [44 words unchanged] Health Ltd (hereafter known as i5 Health) uses data from NHS England Data Access Request Services (NHSE DARS). Data. [13 paragraphs unchanged] - i5 Health has been evaluating the impact of Non-Medical Prescribing by [24 words unchanged] England. The economic impact of community nursing will be evidenced by using NHSE NHS England data data to review early discharge support, re-admission avoidance, service delivery and service [44 words unchanged] has locally or, if monitored more widely, has on different health economies. [17 paragraphs unchanged] - To support this work i5 Health used SUS, ECDS and HES data provided by NHSE NHS England on an annual basis to carry out its impact assessments. NHSE NHS England data contains costing which is essential for health economic analysis purposes but [87 words unchanged] the impact of inequalities and gaps in service provision for Lincolnshire ICS. [3 paragraphs unchanged] - The number, richness and extent of the data sets NHSE NHS England currently provides to i5 Health are essential for the training of the [138 words unchanged] Networks for Population Health Management to support Atrial Fibrillation Screening Programmes’ (https://www.researchgate.net/publication/325335156). [6 paragraphs unchanged] The lawful basis for this processing is Article 6(1)(f) - 'processing is necessary for the purposes of the legitimate interests ...' and Article 9(2)(j) 'processing is necessary for archiving purposes in the public interest, scientific or historical research purposes or statistical purposes in accordance with Article 89(1) based on Union or Member State law which shall be proportionate to the aim pursued, respect the essence of the right to data protection and provide for suitable and specific measures to safeguard the fundamental rights and the interests of the data subject. The lawful basis for processing personal data under the UK GDPR is: i5 Health have undertaken a Legitimate Interests Assessment and concluded that they can rely on legitimate interests for this processing. i5 Health’s legitimate interests are a necessary and lawful basis for processing and controlling SUS, ECDS and HES data, as they enable i5 to assist NHS customers with the access and use of i5 analytics products and services, and the development of the same. i5 considers access to its analytics to be in the public interest as well as of critical value to end-user NHS customers, so they can better understand their patient pools and make better health and social care decisions. Legitimate interests also include for i5 Health, as a commercial for-profit company, to continue as a provider of leading analytics in the UK, and as well to support compliance with the clinical and non-clinical performance expectations of the NHS. Article 6(1)(f) - the processing is necessary for the legitimate interests of the data controller or the legitimate interests of a third party, unless there is a good reason to protect the individual’s personal data which overrides those legitimate interests i5 Health as part of their Legitimate Interests, will use the data to provide commercial services to improve healthcare for citizen’s and provide support services to healthcare providers to enable those providers to deliver better healthcare to citizens. The lawful basis for processing special category data under the UK GDPR is: Article 9(2)(j) - processing is necessary for archiving purposes in the public interest, scientific or historical research purposes or statistical purposes in accordance with Article 89(1) based on Union or Member State law which shall be proportionate to the aim pursued, respect the essence of the right to data protection and provide for suitable and specific measures to safeguard the fundamental rights and the interests of the data subject. This processing is in the public interest because it adheres to the UK Policy Framework for Health and Social Care Research, which protects and promotes the interests of patients, service users and the public, and aims to produce generalisable and publicly available information to inform future decisions over patients’ treatments or care [5 paragraphs unchanged] i5 Health have received government funding but no commercial funding or sponsorship. As evidenced in this Agreement, NHSE NHS England data will not be used solely for commercial purposes. Given the nature [21 words unchanged] for the public are proportionate to the commercial advantages to i5 Health.

Processing activities

NHSE will provide i5 Health with one drop of the latest annual record level pseudonymised Community Service Data Set (CSDS), Emergency Care Data Set (ECDS) and Hospital Episode Statistics (HES), which will include 2 years of data for 21/22 and 22/23 bar Secondary Use Services (SUS) Payment by Results (PbR) which will only include 1 year of data for period 21/22, data via the Secure Electronic File Transfer (SEFT) system. There will be no requirement nor attempt to re-identify individuals within the data sets. No data will flow to NHS England for the purposes of this Data Sharing Agreement (DSA). NHS England will provide the relevant records from the Community Service Data Set (CSDS), Emergency Care Data Set (ECDS) and Hospital Episode Statistics (HES) datasets to i5 Health. The Data will contain no direct identifying data items. The Data will be pseudonymised and individuals cannot be reidentified through linkage with other data in the possession of the recipient. The Data will not be transferred to any other location. The Data will be stored on servers at i5 Health. [2 paragraphs unchanged] • Data is received from NHSE using secure transfer SEFT NHS England [12 paragraphs unchanged] - There will be no flow of data into NHSE from i5 Health. - The only data transfer under this agreement will be the dissemination of the pseudonymised data to i5 Health by NHS England. - The only data transfer under this agreement will be the dissemination of the pseudonymised data to i5 Health by NHSE . [1 paragraph unchanged] The Data will be accessed by authorised personnel via remote access. The Controller(s) must confirm and provide evidence upon audit by NHS England that access via any remote device complies with the data security obligations within this DSA and the Data Sharing Framework Contract. For remote access: - Remote access will only be from secure locations situated within the territory of use (as further restricted elsewhere within the DSA if so done) stated within this DSA; - Access controls granting users the minimum level of access required are in place; - Remote access is only via secure connections (e.g., VPNs or secure protocols) to protect data; - Multifactor authentication (MFA) is required for remote access; - Device security, including up-to-date software and operating systems, antivirus software, and enabled firewalls are utilised for the remote access; - All remote access is undertaken within the scope of the organisation’s DSPT (or other security arrangements as per this DSA) and complies with the organisation’s remote access policy. The above applies in addition to any condition set out elsewhere within the DSA (e.g. who may carry out processing, and for what purpose). Remote processing will be from secure locations within England and Wales. The data will not leave England and Wales at any time. Access is restricted to employees of i5 Health Limited who have authorisation from the Operations Director All personnel accessing the Data have been appropriately trained in data protection and confidentiality. The Data will not be linked with any other data. There will be no requirement and no attempt to reidentify individuals when using the Data. Analysts from the i5 Health Limited will process/analyse the Data for the purposes described above.

Expected output

i5 Health used the NHSE NHS England data to create aggregated reports and data models. All outputs used in [106 words unchanged] ICB, GP practice or any other suitable levels with small numbers supressed. [15 paragraphs unchanged] - The Coronavirus pandemic has accelerated the use of coronavirus risk stratification [88 words unchanged] continue improving the mortality risk stratification model on an annual basis using NHSE NHS England data due to the risk of a recurrence of the pandemic. This [5 words unchanged] Health (self-funded) will continue in 2022/23 to maintain a state of readiness. [1 paragraph unchanged] - A national exercise to inform NICE guidance based on evidence contained in the NHSE NHS England data has been carried out since 2019. This exercise, referred to as [43 words unchanged] NICE guidance and the categorisation of patient cohorts that had unintended consequences. [11 paragraphs unchanged]

Expected measurable benefits

[9 paragraphs unchanged] The following list the expected benefits for each of the categories i5 [58 words unchanged] be applied to future hospital activity data (HES or SUS) received from NHSE NHS England to provide evidence of efficacy. [14 paragraphs unchanged]

Unchanged: Benefits reported.

Text removed is struck through; text added is underlined. Unchanged paragraphs are summarised rather than repeated.

Unchanged: Objective for processing, Processing activities, Expected output, Expected measurable benefits, Benefits reported.

Text removed is struck through; text added is underlined. Unchanged paragraphs are summarised rather than repeated.

Datasets: + Community Services Data Set (CSDS)

Objective for processing

i5 Health Ltd provides consultancy services to support to Clinical Commissioning Groups (CCG), Commissioning Support Units (CSUs), Sustainability and Transformation Plans (STP), Integrated Care Boards (ICBs), Integrated Care Partnerships (ICPs/STPs), Integrated Care Systems (ICSs/CSUs), Acute services, Trusts, District General Hospitals (DGHs), Community/District Nursing, NHS England and other government healthcare organisations as well as Local Authorities (LA) (LAs) (hereafter known as Government Healthcare Organisations) in their decision-making for commissioning purposes. For this, i5 Health Ltd (hereafter known in this section as i5 Health) requires an extension to the term of this Agreement, and a renewal of pseudonymised uses data from NHS Digital for the following purposes: England Data Access Request Services (NHSE DARS). Purpose 1) i5 Health require data for: i5 Health Ltd (i5 Health) evaluates - on behalf of the Health Education Board of NHS England, the economic impact of Non-Medical Prescribing (NMP) - the prescribing of drugs by health practitioners other than doctors. i5 Health analyses the relevant activity data in order to identify utilisation of NMP practitioners in various healthcare settings. In doing so, they can measure the impact NMP has or, if introduced more widely, will have on different health economies. (Academic Paper ID: WNC 48 'Nurse Prescribing' - Worldwide Nursing Conference, Singapore 2014; abstract Purpose 1 Commissioning Support http://www.citeulike.org/user/gstf/article/1324789). First full first report: (http://www.i5health.com/NMP/NMPEconomicEvaluation.pdf ) - Provision of consultancy services to support Integrated Care Boards (ICBs), Integrated Care Partnerships/Sustainability Transformation Partnership (ICPs/STPs), Integrated Care Systems/Commissioning Support Units (ICSs/CSUs), Acute Trusts, District General Hospitals (DGHs), Community/District Nursing, NHS England and other government healthcare organisations as well as Local Authorities (LAs) in their decision-making for commissioning purposes and to address Inequalities and the Health and Wellbeing Gap in populations. Purpose 2) Purpose 1.1 Out-Of-Hospital Service Identification i5 Health provides consultancy services to support to Clinical Commissioning Groups (CCGs), CSUs, Sustainability and Transformation Plans (STP), Acutes, NHS England and Local Authorities (LA) in their decision making for commissioning purposes. - To identify realistic NHS Quality, Innovation, Productivity and Prevention initiatives for government healthcare organisations to spot trends and to perform benchmarking that support commissioners in particular with their operational, strategic planning and commissioning. Current work includes identification of suitable initiatives for Out-of-Hospital services to reduce pressure on hospitals. It also includes provision of patient cohorts for Long Term Conditions (LTCs) to enable GPs to evaluate the quality of their Quality Outcome Framework (QOF) risk registers and devise appropriate actions to improve patient outcomes. Purpose 2.1) - This requires 5 years of historic data due to the length of commissioning plans and cycle, for example, the “Five year Cancer commissioning Strategy for London”, or the “NHS Southwark 5 year Commissioning Strategy Plan”. To identify realistic NHS Quality, Innovation, Productivity and Prevention (QIPP) QIPP initiatives for specific CCGs, Commissioning Support Units (CSU) and Providers in order to spot trends and to perform benchmarking that support commissioners in particular with their operational, strategic planning and co-commissioning. Current work includes with NHS England to identify suitable initiatives for Specialist Services like Cardiology and Cardiac Surgery. It also includes provision of patient counts for Long Term Conditions (LTC) to GPs to enable them to evaluate the quality of their Quality Outcome Framework (QOF) registers and devise appropriate actions (with small numbers suppressed). Purpose 1.2 Supporting Voluntary Sector Organisations Purpose 2.2) - i5 Health advises Voluntary Sector Organisations (VSOs) that have charitable status and exist to complement the work of the NHS in improving patient care. Such VSOs include, but are not limited to, Age UK, Asthma UK and Anaphylaxis UK. Only VSOs that are commissioned by the NHS are clients of this service. VSOs receive information from i5 Health to match VSO services to the needs of their local populations. All i5 Health reports for VSOs are aggregated with small numbers supressed, in line with HES Analysis Guidance, to streamline their own specific charitable works for NHS patients. i5 Health advises Voluntary Sector Organisations (VSOs) that have charitable status and exist to complement the work of the NHS in improving patient care. Such VSOs include Age UK and Asthma UK. Only voluntary organisations that are commissioned by the NHS will be clients of this service. - This requires 10 years due to the nonfrequent occurrence of anaphylactic shocks in patients, for example, the 10-year review found increasing incidence trends of emergency egg allergy reactions and food-induced anaphylaxis in children. Purpose 2.3) Purpose 1.3 Identification of Care and Quality Gaps - To measure standards of care and identify gaps in healthcare provision to inform commissioning strategy. A number of CCGs government healthcare organisations have been working with i5 Health in this respect to develop their strategies. Where NHS Digital Examples include the pre-COVID impact of digital health that has already given formal approval for led to the implementation of remote consultations in primary and secondary care and the sharing of digital pathology results. Those digital services were established pre-COVID and were used rapidly during the pandemic. Other work i5 Health is performing relates to analyse data (IG Ref DSCON066/Halton CCG), the outcome was described by evaluation of the Director of Transformation positive impact community nursing has on secondary care and gaps in services where patients are admitted due to adverse health events such as giving stroke, anaphylactic shock, epilepsy, cardiac conditions, mental health, etc. where fast first line support is beneficial. "…..Halton CCG a unique glance into what financial results could be made through our partnership approach. Unlike any other piece of consultancy, i5 and COP shone an economic light on what schemes are working well and what areas i5 Health could prioritise our energy on." - This requires 5 years of historic data due to the length of commissioning plans and cycle, for example, the “NHS Five Year Forward View”. Purpose 2.4) Purpose 1.4 Nursing / Community Nursing To provide Case Finding and Risk Stratification services. - i5 Health has been evaluating the impact of Non-Medical Prescribing by nurses with a view to extend this programme to include the economic impact of community nursing - on behalf of Community Nursing for NHS England. The economic impact of community nursing will be evidenced by using NHSE data to review early discharge support, re-admission avoidance, service delivery and service gaps. i5 Health analyses the relevant activity data from Hospital Episode Statistics (HES), Secondary Use Services (SUS), Community Services Data Set (CSDS) to assess the utilisation of community nurses in various healthcare settings. In doing so, i5 Health can measure the impact community nursing has locally or, if monitored more widely, has on different health economies. i5 Health has created, free of charge, for the NHS a Calculator that establishes the Health Risk of Coronavirus. Like in the contexts of the Diagnosis Stratification and Targeted Social Prescribing tools and the EBI work for NHS England, it uses past medical history of people to predict their risk levels – in this case if infected by a form of Coronavirus (ie not based on the data of SARS-CoV-2, which causes COVID-19). It is intended the Calculator be upgraded to take account of SARS-CoV-2 once the relevant data is made available. - In health economies where community nursing is understaffed due to funding gaps, sub optimal outcomes can be measured in avoidable readmissions, delayed discharges and high ambulance conveyances, etc. The i5 Coronavirus Health Risk Calculator ("Calculator") establishes a person’s health risk category as either low, medium, high or very high, in the event of being infected by Coronavirus. The Calculator has been developed from NHS hospital medical profiles of patients that had either Coronavirus prior to the emergence of COVID -19 or Influenza. - This analysis can identify pockets of success that community nursing delivers. That success, measurable through NHS Data, is up-scalable and transferrable to other areas to encourage more investment, promotion of best practices, staff retention and sustainability in community nursing. The Calculator may be used by: - This requires 10 years of data due to the time between implementation and realisation of benefits, for example, “Expanding the NHS community workforce: what will this mean for the future of district nursing?”, or the Royal College of Nursing (RCN) report on the urgent investment in District Nursing, as new figures show the number of District Nurses working in the NHS has dropped by almost 43% in England alone in the last 10 years. • Clinicians to support assessment of individuals and advising them on adjustments to their lifestyles to minimise the risk of infection Purpose 2 Prevention of Exacerbations and Illnesses • Clinicians and Hospital Management for prioritising care, improving bed management and ensuring right facilities in ICUs - To provide Case Finding and wide-reaching Risk Stratification services that support the NHS Long Term Plan in the “Treating and Preventing Ill Health”, “Aging Well” and “Personalised Care” areas of work. The key objectives of the purposes of processing data are to turn the NHS into a pro-active health support service away from a reactive illness service. i5 Health is building scientifically validated risk stratification models using pseudonymised data that are provided to the NHS to identify populations at risk in the context of the following sections. • Clinicians for telephone assessment or remote consultations Purpose 2.1 Early Identification of Risk of Long-Term Conditions • Public Health authorities and governments, for planning for disease control, levels of quarantine and targeted shielding - In many parts of the country, a number of Long-Term Condition (LTC) patients are sub-optimally treated because they fail to get on to the relevant registers at GP practices; additionally, there are many patients that have conditions which, if identified early enough, could receive treatment that reduces the risk of them progressing to a full LTC. i5 Health has developed algorithms for a service that identifies, at surgery level, the numbers of patients that fall into both these categories. The novel Coronavirus, SARS-CoV-2, is the pathogen responsible for the infectious respiratory disease COVID-19. NHS Digital collects patient data every month and processes it into a form data experts can use for analysis and advice to the NHS. By early September 2020, data on patients hospitalised to the end of July2020 July will be available, thus enabling an upgrading of the calculator. The tool is currently based on Secondary care data though it is expected that the upgrading will also include Primary Care data when that becomes available. The science underlying the Calculator is set out in the academic paper entitled ‘Predicting Health Risk in Patients with Coronavirus or Influenza using Artificial Intelligence‘ at: https://www.i5analytics.com/HealthRiskInPatientsWithCoronavirus.pdf - At the request of NHS England, i5 Health has provided data models in respect of GP practices on Merseyside (specifically, Southport and Formby (19 practices) and South Sefton (30 practices) which are part of the NHS Cheshire and Merseyside Integrated Care Board; similar case finding data models have been provided to the NHS in London and in the Midlands. The case finding and risk stratification data models of i5 Health has been further called on (at the specific request of the Board of NHS England) in the context of Social Prescribing for all of London; for the CSU support of COVID-19 shielding and vaccination programmes in London; for Evidence Based Interventions (EBI) work to improve NICE guidance of NHS England and the Academy of Medical Royal Colleges (AoMRC) across the country. In many parts of the country, a number of Long Term Condition (LTC) patients are sub-optimally treated because they fail to get on to the relevant registers at GP practices; additionally, there are many patients that have conditions which, if identified early enough, could receive treatment that reduces the risk of them progressing to a full LTC. i5 Health has developed algorithms for a service that identifies, at surgery level, the numbers of patients that fall into both these categories. At the request of NHS England, i5 Health has applied this system in respect of GP practices on Merseyside (specifically, 19 practices in Southport and Formby CCG and 30 practices in South Sefton CCG); similar case finding services using these algorithms have been provided to the NHS in London and in the Midlands. The case finding and risk stratification roles of i5 Health has been further called on (at the specific request of the Board of NHS England) in the context of Social Prescribing for all of London; for the CSU support of CCG shielding and vaccination programmes in London ; for Evidence Based Interventions (EBI) work of NHS England and the Academy of Medical Royal Colleges (AoMRC) across the country. - This requires 10 years of data to ensure there is sufficient training data for AI to avoid bias e.g. towards groups of patients that are under-represented or where the female/male ratios are unbalanced, etc. For example, the report: “Multifactorial 10-Year Prior Diagnosis Prediction Model of Dementia”. The AI development will be undertaken in England/Wales, and any ethical issues are reviewed on a regular basis guided by the Information Asset Owners Handbook. To support this work i5 Health needs SUS, ECDS and HES data on an annual basis to carry out its functions as each category on its own does not contain sufficient elements. SUS data contains costing which is essential for health economic analysis purposes but does not contain all diagnosis and procedure codes and procedure dates. HES does not contain all Payment by Results (PbR) related information contained in SUS though does contain all the necessary clinical information such as full medical history and treatments of a patients including procedure dates. Purpose 2.2 COVID-19 Mortality Risk Stratification Unfiltered historic data is essential to build full and accurate medical histories and monitor progression of diseases to achieve the most precise outcomes forecast that specific interventions can achieve (as demonstrated in the Evidence Based Interventions (EBI) exercise being carried out by i5 Health for NHS England). The availability of seven years of longitudinal data improves data quality for the purposes of creating algorithms because pre-existing conditions or interventions that have been coded seven years ago might otherwise be lost to essential research. Increasing the provision of data from five years to seven years reduces the clinical risk of omitting pre-existing conditions and interventions and should improve efficacy of future treatment. In other words, i5 Health has determined seven years of data will be more evidential when creating algorithms for the purpose of future case finding and risk stratification. Data will be destroyed on a rolling basis, and evidence of data destruction will be provided to NHS Digital on a yearly basis. - i5 Health has created a data model that establishes the Health Risk of Coronavirus for the NHS free of charge. Like in the contexts of the Diagnosis Stratification and Targeted Social Prescribing tools and the EBI work for NHS England, it uses past medical history of people to predict their risk levels – in this case prior to the outbreak in April 2020 an early form of the human coronavirus was used. It is intended the Calculator be upgraded to take into account the latest COVID-19 admissions to improve the data model for future epidemics - this has not yet been completed due to insufficient funding. In an NHS England and Academy of Medical Royal Colleges (AoMRC) exercise carried out by i5 Health over the past two years and continuing into 2022 on Evidence Based Interventions (EBI), the longevity of the records allows i5 Health to strengthen the theory that historical NHS data, if properly interrogated and learnt from, is rich enough to provide valuable supporting evidence for NICE guidance and clinicians addressing new patients’ needs. - This requires 10 years of data to ensure that all pre-existing conditions of patients that had poor outcomes are sufficiently represented in the AI training datasets so that any future COVID outbreaks can be supported with mortality risk models. For example, the reports: “Ten years of severe respiratory syncytial virus infections in a tertiary paediatric intensive care unit”, or “Performing risk stratification for COVID-19 when individual level data is not available”. The AI development will be undertaken in England/Wales, and any ethical issues are reviewed on a regular basis guided by the Information Asset Owners Handbook. i5 Health maintains a rolling seven full years of data and the oldest year is destroyed on receipt of the latest year. At the end of a retention period, i5 Health removes expired data and provides to NHS Digital appropriate destruction certificates. Data is retained only for so long as necessary for the purposes set out herein and as agreed with NHS customers and complies with data deletion requests. Purpose 2.3 Evidence Based Interventions Voluntary Sector Organisations (VSOs) already cooperate with i5 Health to improve the extent and quality of the information that i5 Health relies on to support, with data processing, clinical commissioning within the NHS. The VSOs have occasion to ask for i5 Health reports (aggregated with small numbers supressed, in line with HES Analysis Guidance), based on data analysis that can improve their own specific charitable works for NHS patients. - Historic de-identified clinical data is essential to build full and accurate medical histories and monitor progression of diseases to achieve the most precise outcomes forecast that specific interventions can achieve (as demonstrated in the Evidence Based Interventions (EBI) exercise being carried out by i5 Health for NHS England). The availability of 10 years of longitudinal data improves data quality for the purposes of creating algorithms because pre-existing conditions or interventions that have been coded 10 years ago might otherwise be lost to essential research. The number, richness and extent of the data sets NHS Digital currently provided to i5 Health are essential for the training of the organisation’s Neural Networks that are at the core of i5 Artificial Intelligence (AI) modules. What are being trained are prediction models that, when applied to data of patients, can forecast either the onset of a disease or the prognosis and likely outcomes of a disease when applied to patients’ data. The models use patients’ past medical histories to predict future outcomes and are trained on the outcomes of patients with similar medical histories and the outcomes they experienced. Once a model is trained, none of the original training data is retained within the model - only the cause and effects which can be applied to a single medical record for prediction. The training and underlying methodology are best illustrated in academic papers i5 Health produces for algorithmic tools. An example can be found on ReseachGate in respect of Atrial Fibrillation: ‘Artificial Neural Networks for Population Health Management to support Atrial Fibrillation Screening Programmes’ (https://www.researchgate.net/publication/325335156). - In an NHS England and Academy of Medical Royal Colleges (AoMRC) exercise carried out by i5 Health between 2020 and 2022, with a view to extend beyond 2024, relating to Evidence Based Interventions (EBI), the longevity of 10 years of records allows i5 Health to strengthen the theory that historical NHS data, if properly interrogated and learnt from, is reliable to provide valuable supporting evidence for NICE guidance and clinicians addressing patients’ needs. For example, the report: “House of Commons Science and Technology Committee Evidence-based early years intervention” which refers to the Choice for parents, the best start for children: a ten year strategy for childcare. i5 Health has reviewed the requirement for data being supplied through this agreement and has assessed the data requested as necessary for the purposes of their agreement. In carrying out the training of Neural Networks it is suboptimal to exclude or minimise data within the medical records being used for training. Doing so would detach the algorithms from reality and therefore make them less effective when applied to the data of actual patients whose care is being decided on. Because machine learning uses many input variables (features) for categorising patients for different treatment and risk bands, minimising the data reduces the categorisation ability for fields that are non-obvious. - This requires 10 years of data since some undesired outcomes occur many years after the procedure was performed such as insertion of implantable devices. For example, the report from the Health Foundation titled: “Learning from unintended consequences”. i5 Health does not, as a rule, seek to target children’s data or that of any other vulnerable group. However, there are occasions when analysis of such is specifically requested by the NHS (e.g. researching the variations in childhood asthma within the Brent CCG area) and i5 Health could not fulfil that sort of need without having the relevant historical data. In addition, i5 Health believe not having the data would limit the efficacity of their advice on Population Health Management generally. Providing less than complete data would detach the algorithms from reality, introduce bias and discrimination and therefore make them less effective when applied to the data of actual patients whose care is being decided on. Purpose 2.4 Health Economic Impact Analysis National data is required as i5 Health provide reports from local through Regional to National levels. Multiple years of data are required in order to produce time-series and predictive modelling; historic data is retained to enable this. The data cannot be minimised by applying filters to specific conditions of relevance as the full data is needed in order to produce the outputs as outlined within this Agreement. - To support this work i5 Health used SUS, ECDS and HES data provided by NHSE on an annual basis to carry out its impact assessments. NHSE data contains costing which is essential for health economic analysis purposes but does not contain all diagnosis and procedure codes and procedure dates. HES does not contain all costing - Payment by Results (PbR) - related information contained in SUS though does contain all the necessary clinical information such as full medical history and treatments of a patients including procedure dates. By providing commissioners with aggregated cost and clinical information, business plans can be evaluated before they are moved into practice. Two such examples are the impact of anaphylaxis on the healthcare system for Medway and Swale ICS and the impact of inequalities and gaps in service provision for Lincolnshire ICS. HES data itself has many diagnosis fields. This is critical for the development of i5 Health’s algorithms. A key case in point is the current major project for NHS England at Skipton House that i5 Health is centring. This is Evidence Based Interventions (EBI) referenced in the Yielded Benefits Purpose section under #2.1) Case Finding: “Develop criteria of adverse outcomes based on treatment received to support the separation of patient cohorts that should or should not have received treatment. Such definitions may include readmissions, complications, further treatment needs and death. These criteria may also be used to develop further NICE guidance”. The breadth of the information needed for the EBI project is important because of the number of potential and different pathways of analysis. i5 Health need the maximum number of fields to permit clustering methods based on common co-morbidities used to investigate if commonalities exist amongst patient groups. - Populations at risk of adverse periods of health include sepsis, anaphylaxis, epilepsy, etc, can have significantly better health outcomes if early interventions or prophylactic actions are implemented by ICSs or PCNs. Such actions may include support of charities such as Anaphylaxis UK with the provision of Adrenaline Auto-Injectors (AAI) to at risk patients or raising awareness in geographical areas e.g., high streets at national level. SUS PbR data provides a very different type of information. It is, by definition, costed data and grouped relating to finance – needing HIG codes and national tariffs. SUS PbR is the basis of i5 Health’s costing advice when it comes to using the i5 Commissioning Opportunities (i5 COP) set of algorithms that provide the value for CCGs of alternative (principally non-hospital) treatments. - This requires 5 years of historic data due to the length of commissioning plans and cycle. For example, the “NHS Five Year Forward View”. The number, richness and extent of the data sets NHS Digital currently provide to i5Health are essential for the training of the organisations Neural Networks that are at the core of i5 Artificial Intelligence (AI) modules. The result is the highest level of reliability when reports are created for commissioning and NHS England decision makers. i5 Health has reviewed the requirement for the amount of data being supplied under this Agreement and assessed the data requested as necessary for the purposes of this Agreement. Purpose 2.5 Preventing Ill Health The lawful basis for this processing is Article 6(1)(f) - 'processing is necessary for the purposes of the legitimate interests ...' and Article 9(2)(j) - The number, richness and extent of the data sets NHSE currently provides to i5 Health are essential for the training of the organisation’s data models that are at the core of i5 Artificial Intelligence (AI) modules. What are being trained are prediction models that, when applied to data of patients, can forecast either the onset of a disease or the prognosis and likely outcomes of a disease when applied to patients’ data. The models use patients’ past medical histories to predict future outcomes and are trained on the outcomes of patients with similar medical histories and the outcomes they experienced. Once a model is trained, none of the original training data is retained within the model - only the cause and effects which can be applied to a single medical record for prediction. The training and underlying methodology are best illustrated in academic papers i5 Health produces for algorithmic tools. An example can be found on Atrial Fibrillation: ‘Artificial Neural Networks for Population Health Management to support Atrial Fibrillation Screening Programmes’ (https://www.researchgate.net/publication/325335156). 'processing is necessary for archiving purposes in the public interest, scientific or historical research purposes or statistical purposes...'. - This requires 10 years of data to ensure there is sufficient training data for AI to avoid bias e.g. towards groups of patients that are under-represented or where the female/male ratios are unbalanced, etc. For example, The Department of Health & Social Care have released their report: “Prevention is better than cure" and from the ICO: "What do we need to know about accuracy and statistical accuracy?”. The AI development will be undertaken in England/Wales, and any ethical issues are reviewed on a regular basis guided by the Information Asset Owners Handbook. Purpose 2.6 Prioritisation of Waiting Lists - Waiting times have reached the highest in the history of the NHS in 2022 and many patients are admitted as emergencies whilst waiting instead of receiving elective admissions. Through prioritisation of patients at risk of emergency admission, deterioration of health status can be prevented as well as shorter periods of stay can be achieved. Based on prior conditions and current health status, patients on the waiting lists may also be referred to out of hospital services, social prescribing or community nursing services whilst waiting for treatment to improve pre-admission health, recovery process and outcomes. Waiting times can also be reduced through identification of cohorts that are likely not to attend a scheduled appointment. Those patients may receive an additional reminder from the NHS so that the appointment is not forgotten or cancelled no longer required. Patients at risk of health deterioration whilst on the waiting list may also be included in targeted Hospital at Home programmes or Virtual Wards where remote consultations are performed with or without the use of Digital Health sensor technologies. - Prioritisation of patients on the waiting list based on past medical history and likely outcomes will optimise use of existing capacity, inform commissioners on workforce needs, reduce inequality, and avoid adverse health outcomes for patients. - This requires 3 years of historic data due to the wait time of patients that on the waiting list to facilitate the point when a patient becomes an emergency admission. For example, House of Commons Committee of Public Accounts NHS backlogs and waiting times in England or UK Parliament NHS backlogs and waiting times in England. i5 Health Limited are the sole controller who also process the data for the purposes described within this Agreement. The lawful basis for this processing is Article 6(1)(f) - 'processing is necessary for the purposes of the legitimate interests ...' and Article 9(2)(j) 'processing is necessary for archiving purposes in the public interest, scientific or historical research purposes or statistical purposes in accordance with Article 89(1) based on Union or Member State law which shall be proportionate to the aim pursued, respect the essence of the right to data protection and provide for suitable and specific measures to safeguard the fundamental rights and the interests of the data subject. [1 paragraph unchanged] The purpose of i5 solutions includes, though not limited to, utilisation by NHS customers to produce baselines for innumerable outcome metrics derived from key data combinations, to support better use of resources, staff, and services, ultimately resulting in better health and social care outcomes. It is important to note that access to SUS and HES data by i5 Health’s NHS customers is only ever to aggregated data, with numbers suppressed in line with ICO standards. As part of i5 Health’s Legitimate Interests Assessment, a risk to benefit evaluation has been undertaken. For example, possible risks relating to the machine learning algorithms are mitigated as follows: i5 Health processing activities are a necessary, targeted, and proportionate means of achieving the purposes of the legitimate interests outlined above, and these interests are not overridden by moral or ethical issues and are balanced against any impact on individual rights. i5 Health uses the least intrusive means possible to achieve the analytics it provides to i5 Health customers, strictly within the UK only. The risks of any data breach are clearly understood by i5 Health, which is why i5 Health requests that the SUS, ECDS and HES data be pseudonymised before it is provided to the company and does not require access to the encryption key in order to process SUS, ECDS and HES information. 1) Using up to 10 years of historical medical data in the training and validation datasets, i5 Health aim to evaluate the presence of underrepresented groups in the population. i5 Health builds machine learning models based on large datasets to achieve statistical significance to ensure that any interventions suggested by a healthcare professional which are informed by the models are valid. Any predictive model that could lead to a change in clinical practice would be clinically verified by the relevant healthcare professional, to ensure that e.g. the intervention being offered to an individual is appropriate. The solutions that i5 Health delivers are limited to the NHS customers like CCGs, STPs, CSUs, hospital trusts, NHS England, care quality commission registered providers, public health departments, and similar health care providers within the UK. The objective for processing is to provide support for commissioning activities, operational and financial analytics, comparators and indicators, data quality validation, and other critical insights as requested and directed by NHS customers on the basis of the pseudonymised SUS and HES data. 2) i5 Health use the term ‘risk model’ to describe the outputs of the machine learning, to ensure that clinicians are clear that outputs are probabilistic rather than facts. Models are used to assist clinical decision-making, rather than making e.g. a diagnosis themselves. i5 Health provides these services only where there is benefit to the NHS and its patients. i5 Health Limited does not provide nor would provide these services to commercial sector health bodies 3) The machine learning models are updated when new training data improves the predictability of an undiagnosed condition. i5 Health Limited are the sole data controller who also process the data for the purposes described within this Agreement. Community Services Data Set (CSDS) data will be used to evaluate the benefits Community Nursing delivers to the NHS which is a deliverable to the Director of Community Nursing NHS England. This data request has been minimised to only include information required to perform the evaluation. The benefits evaluation with regards to nursing provides evidence to the community nursing teams that can be used for deriving positive messages where service improvements have been delivered. Such positive messages can improve morale and retention of nursing staff. The benefits evaluation also applies to patients where service gaps offer improvement opportunities including catheter care, wound care, falls prevention, and nutrition. i5 Health is currently supporting the Catheter Care Networks across England to reduce emergency admissions relating to Urinary Tract Infections (UTIs) which is one of the main reasons for non-elective admissions. By having access to CSDS, i5 Health are able to review the efficacy of care to provide improvement plans for service redesign, training, and education. i5 Health is a regular presenter at Future NHS Bitesize Webinars and the “Value of Community Nursing” events that are part of the “Annual Community Nursing Programme”. i5 Health have received government funding but no commercial funding or sponsorship. As evidenced in this Agreement, NHS Digital NHSE data will not be used solely for commercial purposes. Given the nature [21 words unchanged] for the public are proportionate to the commercial advantages to i5 Health. There are different commercial arrangements depending on the specific services/needs of the NHS organisation concerned. Examples include: • Analysis for NHS England of the likely effect of on the structure and finances of the NHS in London arising from the implementation of 44 digital initiatives. i5 Health invoiced on a per diem basis. • Application of i5 algorithms, at the request of NHS England, to data of Merseyside CCGs to establish whether there were critical gaps between the numbers on local GP LTC registers and the likely numbers of, as yet, undiagnosed sufferers. i5 Health invoiced on the basis of size of the CCG population. • At the request of NHS England, establishing whether administrative data can help to select inappropriate procedures for consideration for the Evidence Based Interventions (EBI) programme. It involves analysing longitudinal patient records to ascertain what criteria is present that presents an unacceptable outcome for a patient undergoing a specific procedure. Also establishing whether administrative data can help clarify the criteria for intervention to contribute to NICE guideline development. i5 Health invoices on the basis of monthly Work Packages agreed with NHS England. • Licensing of i5 Artificial Intelligence algorithms installed in the server of NHS Arden & GEM CSU with which i5 Health is a longstanding BI partner. A per annum license fee has been paid. An alternative commercial arrangement is one whereby, for a fee based on the size of a CCG, i5 Health is asked by the CSU to apply its algorithms to NHS data in order, e.g. to ascertain the financial value of introducing specific primary care initiatives that reduce the dependency on secondary care.

Processing activities

Under this version of the Agreement NHS Digital NHSE will provide i5 Health with one drop of the latest annual record level pseudonymised SUS PbR, ECDS Community Service Data Set (CSDS), Emergency Care Data Set (ECDS) and HES Hospital Episode Statistics (HES), which will include 2 years of data for 21/22 and 22/23 bar Secondary Use Services (SUS) Payment by Results (PbR) which will only include 1 year of data for period 21/22, data via the Secure Electronic File Transfer (SEFT) system. There will be no requirement nor attempt to re-identify individuals within the data sets. A Database Analyst (DBA) from i5 Health will load the record level pseudonymised data into an i5 Health secure database. Data Flow The database will be managed locally and secured by the DBA with user access control. All data processing is done within the England and Wales where a Data Base Analyst (DBA) from i5 Health will load the record level pseudonymised data into an i5 Health secure database using the following process: Record-level pseudonymised data will only be accessed by individuals within the Analytics Team, who have the authorisation from the Operations Director (who is also Caldicott Guardian), to access the data for the purpose (s) described within this DSA, all of whom are substantive employees of i5 Health. All those accessing data under this Agreement are substantive employees of i5 Health, and have received appropriate training in data protection and confidentiality. Data will only be accessed at the named processing location as set out in this Agreement. • Data is received from NHSE using secure transfer SEFT The additional SUS PbR, ECDS and HES data being provided under this Agreement will be linked to SUS PbR data already held across the data sets (e.g. SUS PbR Episode data with SUS PbR A&E data); from National Level to GP Practice Level. • Data is uploaded by the IT lead onto an encrypted secure server and imported into an encrypted database. There will be no requirement nor attempt to re-identify individuals within the data sets. • Any statistical analysis, data modelling and verification is performed on the encrypted server. Data will not be made available to any third parties, including VSO’s, except in the form of aggregated outputs with small numbers suppressed in line with the HES Analysis Guide which also covers suppression rules for SUS. For example, a report containing aggregate data for cohorts from localities to large geographical areas will be produced for NHS England. • Aggregate data with small numbers suppressed is used in reports and presentations to improve patient care in the NHS There will be no data linkage undertaken with NHS Digital data provided under this agreement that is not already noted in the agreement. • Data Models are built to support predictive healthcare that only contain cause and effect relationships All organisations party to this agreement must comply with the Data Sharing Framework Contract requirements, including those regarding the use (and purposes of that use) by “Personnel” (as defined within the Data Sharing Framework Contract - i.e. employees, agents and contractors of the Data Recipient who may have access to that data). • After the data retention period, data is securely deleted on the server and backups i5 Health are requesting to use 10 years’ worth of data to produce reports and data models which will allow sufficient longitudinal medical history to discover patterns that can help to predict future adverse health events in populations. Data Access - The database will be managed locally and secured by the DBA with user access control and record-level pseudonymised data will only be accessed by individuals within the Analytics Team, who have the authorisation from the Operations Director (who is also Caldicott Guardian), to access the data for the purpose (s) described within this agreement, all of whom are substantive employees of i5 Health and have received appropriate training in data protection and confidentiality. Data access is audited frequently by i5 Health ltd. - Patient level data will not be made available to any third parties. Data in aggregated form with small numbers suppressed in line with the HES Analysis Guide will be used for reporting. Data Processing - Data processing will only be carried out by substantive employees of i5 Health who have been appropriately trained in data protection and confidentiality. - No data will be shared with third parties. - There will be no flow of data into NHSE from i5 Health. - The only data transfer under this agreement will be the dissemination of the pseudonymised data to i5 Health by NHSE . - The data requested is pseudonymised and will not be linked with other datasets which could allow re-identification. The data will remain on a local secured dedicated server at the i5 Health offices with restricted access by a restricted number of substantive employees trained in Information Governance and Data Security.

Expected output

All outputs will be aggregated analysis with small numbers suppressed (in line with HES Analysis Guidance, and SUS suppression rules) for inclusion within economic evaluation and Clinical Commissioning Group (CCG) strategy. All outputs are solely provided to the NHS customers and will be aggregated outputs with small numbers suppressed in line with the HES Analysis Guide. No service/product/data will be supplied to any commercial organisation by i5 Health except in so far as is permitted for Voluntary Sector Organisations (VSO). i5 Health used the NHSE data to create aggregated reports and data models. All outputs used in reports will be aggregated with small numbers suppressed (in line with HES Analysis Guidance). All data models retain the cause-and-effect relationships of the data and are scientifically validated and non-reversible. The format of reports that i5 Health is using include static reports such as Word, PowerPoint, Excel, PDF, etc and interactive reports such as Excel or Power BI. The format of data models includes equations and number matrixes which are commonly used in number theory, algebra and machine learning. All reports and data models would be to run in the context of the whole of England to facilitate comparisons and may be broken down to Trust, ICB, GP practice or any other suitable levels with small numbers supressed. The data provided will be used solely for the purposes identified above. Purpose 1 Commissioning Support The outputs i5 Health have provided (according to each purpose): Purpose 1.1 Out-Of-Hospital Service Identification Purpose #1) Non-Medical Prescribing - Outputs include building of a data model that supports the new ICBs commissioning model for contracting services. This includes new models using a more sustainable solution that involves out of hospital services which is a key system deliverable for the NHS Long Term Plan. Out of hospital services free secondary care resources needed by more acute patients as well as optimise pathways through better collaboration between providers that may result in financial savings and also improve patient experience. This is an ongoing project internally referred to as Commissioning Opportunity (COP) that spans across various ICBs since 2016. Initially first national report for Health Education Board of NHS England on the economic value of Non-Medical Prescribing (NMP) called on three categories of NHS Digital information (latest HES data in respect of long term health conditions (LTC)); Nurses currently in the workforce, and Nurses using FP10 Prescription forms). Thereafter continued to update value of NMP to the country. The input from i5 Health formed the basis of a decision by NHS England to increase considerably the funding of NMP. Purpose 1.2 Supporting Voluntary Sector Organisations The i5 exercise and its findings continue to be quoted in reports by decision makers in support of expanding the practice of NMP. i5 Health maintains readiness to provide Health Education England (HEE) with further updates, based on the NHS data, of the workforce, patient care and financial benefits continuing to arise from Non-Medical Prescribing. - Voluntary Sector Organisations (VSOs) receive performance and service gaps information that correspond with the services provided or planned to be provided by the VSO. Illustrative examples of the information provided may include counts of patients with LTCs, adverse outcomes, in need of discharge support, in need of self-management support, etc. Purpose 2.1) Purpose 1.3 Identification of Care Gaps i5 Health is currently working on a project with Lincolnshire CCG to ascertain if information about patient behaviour, conditions, and events, captured from wearables, monitors and other smart technologies can predict demand for services, then providing these technologies to patients and using the data generated will enable providers to pre-empt and redirect demand or design new services. The key outcomes provided by i5 Health are the creation of cohorts of patients for the exercise, in cooperation with the Data Services for Commissioners Regional Office (DSCRO) of Arden & GEM CSU, and the evaluation of the data provided by the wearables. (Creation of cohorts would not include identifying specific individuals using NHS Digital data.) The project is expected to continue into 2022 and beyond. - In many parts of the country, a number of Long-Term Condition (LTC) patients are sub-optimally treated because they fail to get on to the relevant registers at GP practices; additionally, there are many patients that have conditions which, if identified early enough, could receive treatment that reduces the risk of them progressing to a full LTC. i5 Health has developed algorithms that identify, at surgery level, the numbers of patients that fall into both these categories. NHS England have requested for i5 Health to carry out a study in respect of the GP practices in Southport and Formby. This NHS England initiative is continuing. The case-finding role of i5 Health was further called on in the context of Social Prescribing for all of London (at the specific request of the Board of NHS England). Internally, i5 Health refers to those projects as Diagnosis Stratification (DST) and Targeted Social Prescribing (TSP). Purpose 2.3) Purpose 1.4 Community Nursing Coming into this category is the major national exercise being focused by i5 Health on Evidence Based Interventions (EBI) – on the instructions of NHS England and the Academy of Medical Royal Colleges (AoMRC). A member of the of i5 Health team is to become a member of the Expert Advisory Committee of AoMRC which has oversight of progress of the project. The project has characteristics that brings it also within Purpose #2.4 below and is being renewed through to 2022 at least. - i5 Health has a long-standing history supporting the recognition of nursing from an evidence perspective to support uptake, funding and education of nurses. Outputs to support nursing include health economic reports e.g. Non-Medical Prescribing or Service Reviews which may be static or dynamic in nature. Purpose 2.4) - A localised report for East of England relating to the achievements of community nursing has been completed and presented at the Value of Community Nursing Event sponsored by NHS England. This event highlighted the importance of Catheter Care and will be extended in Q4 2023 nationally. In many parts of the country, a number of Long Term Condition (LTC) patients are sub-optimally treated because they fail to get on to the relevant registers at GP practices; additionally, there are many patients that have conditions which, if identified early enough, could receive treatment that reduces the risk of them progressing to a full LTC. i5 Health has developed algorithms that identify, at surgery level, the numbers of patients that fall into both these categories. NHS England have requested for i5 Health to carry out a study in respect of the GP practices in Southport and Formby. This NHS England initiative is continuing. The case –finding role of i5 Health was further called on in the context of Social Prescribing for all of London (at the specific request of the Board of NHS England). - The i5 Health’s outputs and findings continue to be quoted in reports by decision makers in support of expanding the practice of community nursing. i5 Health maintains readiness to provide decision makers within nursing teams with further updates, based on the NHS data, of the workforce, patient care and financial benefits continuing to arise from nursing. In respect of Arden & GEM CSU, it is expected that the Case Finding work carried out for its clients, Milton Keynes CCG, will be extended to more of the 70 CCGs within the CSUs ‘footprint’. It is also expected that work started in early 2020 for the CSU’s clients, South Lincolnshire CCG, on an Internet of Things (IoT) initiative will continue through into 2021. In 2019, discussions started with NEL CSU, after the success of the Haringey CCG Atrial Fibrillation exercise, to integrate the Case Finding algorithms into the CSU’s server or at least partner with the CSU by providing an API link. The Coronavirus crisis has accelerated these considerations and NEL CSU is now benefitting from accessing the Risk Stratification service provided by the i5 Health Coronavirus Health Risk Calculator - up to 22 million uses through the DSA year (see explanation in 5 d iii) below). I5 Health has received funds from Innovate UK to commence the Calculator Work Packages on 15th July 2020. The 90 FTE days allocated would give an end date of the middle of October 2020 but i5 Health will endeavour to accelerate the process as time is of the essence. Purpose 2 Prevention of Exacerbations and Illnesses Within this category are, currently, in respect of Kent and Medway CCG: Purpose 2.1 Early Identification of Risk of Long-Term Conditions - Facilitating, through i5 Activation Score algorithms based on the analysis of secondary care data and in cooperation with the DSCRO at NEL CSU and Health Diagnostics, the creation of cohorts of patients that are hard-to-reach for Health Check purposes. - Arden & GEM CSU has been performing case finding work for its client Bedfordshire, Luton and Milton Keynes Integrated Care Board (ICB) using the Diagnosis Stratification (DST) tool. This exercise is planned to be extended to all the health and care organisations within the ICB. Work has started to identify patients at risk of developing LTCs in early 2022 for the CSU's client South Lincolnshire CCG. The project is a co-operation with several organisations in which i5 Health is providing data analytics. This project will continue through into Q2 2023. NEL CSU is also interested in joining the DST project for case finding of patients with undiagnosed conditions to improve the results of NHS Diagnostic Hubs which is one of the opportunities for sharing good practice. - Application of diagnosis stratification algorithms, in cooperation with the DSCRO at Arden & GEM CSU, to establish those cohorts at risk of diabetes induced leg ulcers. This exercise is in the context of the CCG’s Social Prescribing programme. Purpose 2.2 COVID-19 Mortality Risk Stratification It is envisaged that these exercises will continue into 2022 and beyond. - The Coronavirus pandemic has accelerated the use of coronavirus risk stratification to avoid hospitalisation of at-risk patients. Since the integration of the i5 Health Coronavirus risk model at NEL CSU, the population of London is now benefitting from a system that has been tried and tested in case of a resurgence of the pandemic. During the 2020 pandemic the Risk Stratification model was used for free for over 9 million times in London and over 50 million times around the world. This free global service was made possible through funding received by Innovate UK. i5 Health will endeavour to continue improving the mortality risk stratification model on an annual basis using NHSE data due to the risk of a recurrence of the pandemic. This work is funded by i5 Health (self-funded) will continue in 2022/23 to maintain a state of readiness. Planning is also being undertaken on a project with South West London CCG and NEL CSU to address the problem of hard-to-reach members of the Black, Asian and minority ethnic (BAME) community for cardiovascular problems. Purpose 2.3 Evidence Based Interventions The Evidence Based Initiatives (EBI) programme and its need for analysing secondary care data over several years has enabled i5 Health to create an AI tool to facilitate the exercise. This has ensured a great deal many more intervention can be spotlighted over the next 12-24 months. - A national exercise to inform NICE guidance based on evidence contained in the NHSE data has been carried out since 2019. This exercise, referred to as Evidence Based Interventions (EBI) within the NHS, is supported by i5 Health on the instructions of NHS England and the Academy of Medical Royal Colleges (AoMRC). Some of the key outputs are outcomes evaluations for specific interventions that result in better use of NICE guidance and the categorisation of patient cohorts that had unintended consequences. i5 Health has developed Artificial Intelligence (AI) algorithms that interrogate, in various ways, secondary care data provided by NHS Digital. The results of those interrogations, when combined with other data (e.g., workforce, size of population…etc), contribute to reports requested by and for NHS decision makers. These include reports on: - i5 Health have developed criteria of adverse outcomes based on treatment received to support the separation of patient cohorts that should or should not have received treatment. Such definitions include readmissions, complications, further treatment needs and death. These criteria are used to develop further NICE guidance, which has included guidance on the clinical coding inclusion/exclusion criteria of surgical intervention for chronic sinusitis, removal of adenoids, and cystoscopy for men with uncomplicated lower urinary tract symptoms. - risk stratification Purpose 2.4 Health Economic Impact Analysis - outcome prediction - i5 Health is currently engaged in a clinical trial with Lincolnshire CCG to ascertain if information about patient behaviour, conditions, and events, captured from wearables, monitors and other smart technologies can predict demand for services. The view is that providing these technologies to patients and using the data generated will enable providers to pre-empt and redirect demand or design new services. The key outcomes provided by i5 Health are the review of cohorts of patients to evidence the impact of the exercise in cooperation with Arden & GEM CSU and the evaluation of the data provided by the wearables. The project was delayed due to COVID-19 and is expected to complete in Q2 2023. - service recommendations Purpose 2.5 Preventing Ill Health - health economy planning - The AI based data models for DST are forecasting either the onset of a disease or the prognosis and likely outcomes of a disease is based on patients past medical histories to predict future outcomes and are trained on the outcomes of patients with similar medical histories and the outcomes they experienced. These models are used to count the number of people that may have undiagnosed conditions to support the scaling and type of screening programmes for a region. The models are also provided to the CSUs within the NHS to facilitate case finding. - invoice validation Purpose 2.6 Prioritisation of Waiting Lists - reports on the patient care and financial benefits of specific activities eg Non-Medical Prescribing, digital initiatives, etc. - NHS South West London CSU is concerned about its waiting lists and has embarked on a project with i5 Health to prioritise patients that otherwise would deteriorate and be admitted as emergencies. The key output is to count the number of patients that require prioritisation to right-size the service provision and timing. Depending on the scientific evaluation and accreditations, the CSU may consider using the data model to prioritise patients on the waiting list to prevent deterioration of their conditions. - It is a radical but rational step to take this concept of intense research into historical patient records into the field of Waiting Lists to find patterns that, translated into algorithms, allow prediction of whether an individual is transitioning from the Elective category into the Non-Elective category. If this were possible, reprioritising that patient could not only save a life or prevent serious distress but also avoid expensive secondary care costs for the NHS. - It is also thought possible that prediction of some patients on Waiting Lists simply not attending treatment Did Not Attend (DNA) could be arrived at because of natural improvement without treatment; if correctly assessed, those patients might be reprioritised or put on a less acute pathways not using Waiting List spaces. Purpose 2.7 Staying Healthy for Longer by predicting exacerbations - Predicting the likelihood of unmanaged or unmanageable respiratory exacerbations reduces respiratory distress of patients and reduction in use of healthcare resources. The prediction model output will be used to risk stratify populations that are living in deprived or polluted areas, do not receive sufficient support for self-management or do not adhere to recommendations for inhaler use. Such populations can be supported with smart sensor technology or self-management advice to reduce non-adherence and the risk of exacerbations.

Expected measurable benefits

The Non-Medical Prescribing (NMP) has been in existence for 26 years. Over time there are good reasons to believe that the returns from it, across the board, have been very positive: from cost effectiveness, through staff development to patient satisfaction. Benefits Type: Activities carried out with regards to NMP analysis: 1) Staying Healthy for Longer by predicting of exacerbations based on past medical history. - Global research of NMP experience 2) Independent Living including improved Self-Management - National research through personal interviews and group sessions 3) Improvements in early disease detection to support Healthy Aging - Data analysis and modelling to establish: 4) Digital Health to facilitate Hospital at Home, Virtual Wards and Telemedicine - Local Impact 5) Financial savings due to better use of existing resources and use of non-acute resources where the patient need is non-medical - National Impact 6) Increased funding resulting in retention and growing of workforce, in particular nurses and community nurses. - Potential cost savings 7) Optimisation of limited resources such as operating theatres through prioritisation of waiting lists based on predicted exacerbation timeframes - Activity and cost outcomes 8) Supporting a thriving non-clinical support service economy creating jobs and expertise in Social Prescribing services that meet patients’ non-clinical needs Not least of all, the clinicians Audit, in growing use since 2009, has elicited important data supportive of that contention. However, greater evidence of the performance and effect of NMP is necessary. i5 Health is reviewing existing HES and studies with a view to drawing out relevant information, proposing new methodology, refining existing audits and promoting new ones to provide a comprehensive analysis of NMP to assist in decisions on whether and to what extent NMP should be adopted more widely in England. The following list the expected benefits for each of the categories i5 Health is working in to improve patient care. With projects at various stages of maturity, it is difficult to precisely quantify benefits since i5 Health is not involved in direct patient care nor requests data on service provision from commissioners to avoid creating additional reporting burden. All metrics listed represent a strategy of measure for quantification that may be applied to future hospital activity data (HES or SUS) received from NHSE to provide evidence of efficacy. Voluntary Sector Organisations (VSO) cooperate with i5 Health to improve the extent and quality of the information that i5 Health relies on to support, with data processing, clinical commissioning within the NHS. Strategy of measures for Purpose 1: Commissioning Support The benefits of the Evidence-Based Interventions (EBI) programme are the prevention of avoidable harm to patients, avoidance of unnecessary operations and freeing up clinical time by only offering interventions on the NHS that are evidence-based and appropriate. Phase 1 of the programme targeted 17 such interventions. The 17 interventions in question were snoring surgery in the absence of Obstructive Sleep Apnoea (OSA), dilatation and curettage (D&C) for heavy menstrual bleeding in women, knee arthroscopy for patients with osteoarthritis and injections for non-specific low back pain without sciatica, breast reduction, removal of benign skin lesions, grommets for glue ear in children, tonsillectomy for recurrent tonsillitis, haemorrhoid surgery, hysterectomy for heavy menstrual bleeding, chalazia removal, arthroscopic shoulder decompression for subacromial shoulder pain, carpal tunnel syndrome release, Dupuytren’s contracture release, ganglion excision, Trigger finger release, varicose vein surgery. The current Phase 2 of the programme is selecting further procedures and Phase 3, commencing during Q1 2021, will select more. • Purpose 1.1 Out-Of-Hospital Service Identification includes commissioning of additional services that match the needs of the population and reduction of hospital admissions that can be deemed avoidable. This is an ongoing project spanning several years. The initial EBI work of i5 Health was to answer two key questions: • Purpose 1.2 Supporting Voluntary Sector Organisations includes new services in the VSO space to support vulnerable patients and reduced admissions attributable to those services. Anaphylaxis project to end Q4 2023. Question 1: 'Can administrative data help to select inappropriate procedures for consideration for the EBI programme?' Essentially this is about establishing, from longitudinal patient records, what criteria is present that presents an unacceptable outcome for a patient undergoing a specific procedure. • Purpose 1.3 Identification of Care Gaps includes development of Social Prescribing services that support patients with non-clinical needs and patients on the waiting list. This is an ongoing project spanning several years. Question 2: 'Can administrative data help us to clarify the criteria for intervention to contribute to our guideline development?' • Purpose 1.4 Community Nursing includes recruitment of more nurses, higher level of retention, more services across the country and additional funding. This project is scheduled to end Q2 2023. Having provided positive answers to both, i5 Health have developed criteria of adverse outcomes based on treatment received to support the separation of patient cohorts that should or should not have received treatment. Such definitions include readmissions, complications, further treatment needs and death. These criteria are used to develop further NICE guidance, which has included guidance on the clinical coding inclusion/exclusion criteria of surgical intervention for chronic sinusitis, removal of adenoids, and cystoscopy for men with uncomplicated lower urinary tract symptoms. Strategy of measures for Purpose 2: Prevention of Exacerbations and Illnesses For the purposes of Phases 2 and 3 of EBI, i5 Health are doing the following, using the data supplied under this agreement: • Purpose 2.1 Early Identification of Risk of Long-Term Conditions includes lower rates of undiagnosed patients that are otherwise diagnosed in A&E resulting in lower A&E attendance rates as well as lesser complications during admissions. This is an ongoing project spanning several years. - Initially volumes and variations of elective procedures performed in the NHS are being analysed. The procedures are evaluated in order of volume and widest variation. • Purpose 2.2 COVID-19 Mortality Risk Stratification includes reduced mortality rates in the event of another COVID-19 epidemic and prioritised immunisation programmes based on prior conditions. This is an ongoing project spanning several years. - The evidence for most frequent procedures or those with most variation is being examined by the clinical team working with i5 Health to assess appropriateness. • Purpose 2.3 Evidence Based Interventions includes improved NICE guidance resulting in appropriate referrals with less adverse outcomes. This is an ongoing project spanning several years. - Procedures that are deemed inappropriate in certain circumstances will then be further examined to describe the criteria that make the procedure inappropriate (e.g., low rate of diagnosis confirmed, high rate of complications). • Purpose 2.4 Health Economic Impact Analysis includes informed commissioning decisions based on local population needs reducing inequality and hospital activities. This is an ongoing project spanning several years. The Coronavirus Calculator tool is intended to benefit: • Purpose 2.5 Preventing Ill Health through prediction models that enable patients to self-manage. This is an ongoing project spanning several years. • Individuals - to adjust their lifestyle and minimise the risk of infection. • Purpose 2.6 Prioritisation of Waiting Lists includes prioritisation of patients on waiting list to reduce transition from the elective pathway to the non-elective pathway. This is an ongoing project spanning several years. • Clinicians and Hospital Management for prioritising care, improving bed management and ensuring right facilities in ICUs • Other benefits to patients, the NHS, and the wider population have included: - Reduction in the cases and prevention of LTCs - Reduction in the suffering of patients - Avoidance in the decline of quality of life - Reduction in the mortality risk from LTCs - Early treatment for identified patients, probably in a primary care or social prescribing setting - Society benefitting from less pressure on the patient and family members and the greater availability of individuals to work - Reduction of acute hospital admissions for patients - Improvements in operational efficiency - Efficacy improvements for screening in primary care and community settings by pre-selecting for clinical review only that cohort identified by the tool as having or likely in the next 12 to 24 months to have an LTC (e.g., Atrial Fibrillation) - Reduction of pressure on secondary care and related costs through fewer hospital admissions • Clinicians for telephone assessment or remote consultations • The benefits of the Evidence-Based Interventions (EBI) programme are the prevention of avoidable harm to patients, avoidance of unnecessary operations, and freeing up clinical time by only offering interventions on the NHS that are evidence-based and appropriate. Phase 1 of the programme targeted 17 such interventions. The 17 interventions in question were snoring surgery in the absence of Obstructive Sleep Apnoea (OSA), dilatation and curettage (D&C) for heavy menstrual bleeding in women, knee arthroscopy for patients with osteoarthritis and injections for non-specific low back pain without sciatica, breast reduction, removal of benign skin lesions, grommets for glue ear in children, tonsillectomy for recurrent tonsillitis, haemorrhoid surgery, hysterectomy for heavy menstrual bleeding, chalazia removal, arthroscopic shoulder decompression for subacromial shoulder pain, carpal tunnel syndrome release, Dupuytren’s contracture release, ganglion excision, Trigger finger release, varicose vein surgery. The current Phase 2 of the programme is selecting further procedures and Phase 3, which commenced during Q1 2021 and was completed in Q1 2022. The continuation of the programme into Phase 4 is expected in Q2 2023 and may run for 6 months. The work which is envisaged to continue in 2022/23: Consequent on delivery of Preliminary EBI Evidence Pack, working with Getting It Right First Time (GIRFT) Clinical Coding Team, Expert Working Groups, Clinical Classifiers, Clinical Leads and Price Assessors to enrich the material in Preliminary EBI Evidence Pack. • Public Health authorities and governments, for planning for disease control, levels of quarantine and targeted shielding South West London CCG have used the COVID-19 tool for outcome risk stratification, identifying patients for shielding and subsequently vaccinations in six boroughs. There are two particular points about this exercise that relate to Long Term Conditions: - The Calculator risk assessments can lead to persons, clinicians and government authorities taking more informed decisions that could reduce the effect of the condition Covid-19 itself (eg the screening exercise carried out for London) - The focus given by the NHS to Covid-19 during this pandemic has a negative effect on the treatment of Long Term Conditions in general; this might be because of delays in screening or of treatment itself. The Calculator can contribute to authorities carrying out more selective screening of populations which means more optimal use of resources and better availability of resources for other Long Term Conditions. Identification of patients that have (or risk having) an Long Term Condition (LTC) but do not appear on the GP's risk register can lead to better management of their health requirements Evidence Based Interventions (EBI) work which will continue into 2021: Consequent on delivery of Preliminary EBI Evidence Pack, working with Getting It Right First Time (GIRFT) Clinical Coding Team, Expert Working Groups, Clinical Classifiers , Clinical Leads and Price Assessors to enrich the material in Preliminary EBI Evidence Pack. Continue working on answers to following two key questions : Question 1: 'Can administrative data help to select inappropriate procedures for consideration for the EBI programme?' Essentially this is about establishing, from longitudinal patient records, what criteria is present that presents an unacceptable outcome for a patient undergoing a specific procedure. Question 2: 'Can administrative data help us to clarify the criteria for intervention to contribute to our guideline development?' Develop criteria of adverse outcomes based on treatment received to support the separation of patient cohorts that should or should not have received treatment. Such definitions may include readmissions, complications, further treatment needs and death. These criteria may also be used to develop further NICE guidance. i5 Health have received government funding but no commercial funding or sponsorship. As evidenced in this Agreement, NHS Digital data will not be used solely for commercial purposes. Given the nature of the services/outputs of i5 Health for the NHS, as described within this Agreement, i5 Health believe the benefits of such for the public are proportionate to the commercial advantages to i5 Health. There are different commercial arrangements depending on the specific services/needs of the NHS organisation concerned. Examples include: • Analysis for NHS England of the likely effect of on the structure and finances of the NHS in London arising from the implementation of 44 digital initiatives. i5 Health invoiced on a per diem basis. • Application of i5 algorithms, at the request of NHS England, to data of Merseyside CCGs to establish whether there were critical gaps between the numbers on local GP LTC registers and the likely numbers of, as yet, undiagnosed sufferers. i5 Health invoiced on the basis of size of the CCG population. • At the request of NHS England, establishing whether administrative data can help to select inappropriate procedures for consideration for the Evidence Based Interventions (EBI) programme. It involves analysing longitudinal patient records to ascertain what criteria is present that presents an unacceptable outcome for a patient undergoing a specific procedure. Also establishing whether administrative data can help clarify the criteria for intervention to contribute to NICE guideline development. i5 Health invoices on the basis of monthly Work Packages agreed with NHS England. • Licensing of i5 Artificial Intelligence algorithms installed in the server of NHS Arden & GEM CSU with which i5 Health is a longstanding BI partner. A per annum license fee has been paid. An alternative commercial arrangement is one whereby, for a fee based on the size of a CCG, i5 Health is asked by the CSU to apply its algorithms to NHS data in order, e.g. to ascertain the financial value of introducing specific primary care initiatives that reduce the dependency on secondary care.

Benefits reported

Purpose 1.) Non-Medical Prescribing (NMP) Ongoing projects where measurable outputs were achieved: - • The Non-Medical Prescribing (NMP) has been in existence for 26 years. Over time there are good reasons to believe that the returns from it across the board have been very positive, including cost effectiveness, staff development, and patient satisfaction. The immediate benefits benefit of the NMP analysis carried out by i5 Health were was the increase in investment of £500,000 by the NHS in nurse prescribing training. The increased number of , in particular, nurses who are able to prescribe will accelerate patient treatment, freeing trained as prescribers accelerated primary care visits and freed up doctors to spend more time for patients who with complex patients. Additionally, health visitors have continued to visit families in their own homes throughout the COVID-19 pandemic. Health visitor prescribing has meant that families do not need to be seen attend GP surgeries to seek medication, thus saving GP time, and reducing the risk of infection for GPs and families by a doctor. reducing face-to-face contact, see: Clinical and cost-effectiveness of non-medical prescribing: A systematic review of randomised controlled trials. Purpose 2.4) Case Finding and Risk Stratification • Voluntary Sector Organisations (VSOs) cooperate with i5 Health to improve the extent and quality of the information that i5 Health relies on to support, with data processing, clinical commissioning within the NHS, particularly that information relating to the provision of social prescribing services as evidenced by the Greater London Authority: VCSE Sector Engagement and Social Prescribing and National Academy for Social Prescribing: Supporting Voluntary and Community Organisations. - Haringey CCG missioned NEL CSU to apply i5 Health algorithms to its population to establish the identities of individuals (within their own identifiable patient care datasets) with as yet undiagnosed Atrial Fibrillation (AF). Because Haringey’s population of over 270,000 had a lower recorded AF prevalence level (0.9%) than the national average prevalence (1.5%), i5 Health was missioned by the CCG, along with North East London CSU, to apply their tool to find the ‘missing’ patients. As a result of this audit a minimum of 76 patients were identified from the list provided by secondary care data and coded on the Quality Outcomes Framework (QOF) register with a confirmed diagnosis of AF, an increase in the diagnosed prevalence of AF in Haringey CCG from 0.9% and closer to the estimated prevalence of 1.5%. 46 out of the 76 patients (68%) added to the QOF AF register as AF patients who were anticoagulated for stroke prevention. • Following the initial results of the Evidence Based Interventions (EBI) programme of NHS England and the Academy of Medical Royal Colleges (AoMRC), it was decided that not enough input had been achieved through the analysis of historical patient data. i5 Health developed algorithms to interrogate data and to analyse intervention results. The positive benefits have been the reinforcement of NICE guidelines on interventions and an extension of the programme using i5 input from 17 interventions to circa 60. Further evidence can be found here: https://ebi.aomrc.org.uk/. - i5 Health co-operated with Arden & GEM CSU on the application of i5 algorithms to the population of the NHS Milton Keynes CCG area (300,000) in respect of: Hypertension, Atrial Fibrillation, Heart Failure, Diabetes Mellitus, Cancer, Asthma, COPD, Coronary Heart Disease, Palliative Care, Dementia, Depression (18+), Obesity (16+), Epilepsy (18+), Chronic Kidney Disease (16+), Stroke and TIA, Rheumatoid Arthritis (16+), Peripheral Arterial Disease, Osteoporosis (50+). The results enabled the GP practices to strengthen the relevant registers and treat patients more effectively. • The COVID-19 Calculator tool benefits: - The on-line process created by i5 allowed each of the 32 CCGs and 5 STPs in London to establish the value of introducing of Social Prescribing within their areas. CCGs accessing the site have been able to introduce the findings into their planning processes during the year 2019/20 and 2020/21. - Individuals to adjust their lifestyles and minimise the risk of infection. - It was decided, following the initial results of the Evidence Based Interventions (EBI) programme of NHS England and the Academy of Medical Royal Colleges (AoMRC), that not enough input had been achieved through the analysis of historical patient data. i5 Health developed algorithms to interrogate data and to analyse intervention results. The positive benefits have been the reinforcement of NICE guidelines on interventions and an extension of the programme using i5 input from 17 interventions to circa 60. - Clinicians and Hospital Management in prioritising care, improving bed management, and ensuring right facilities in ICUs. - The NHS in London selected the medical records of circa 4,100,000 Londoners and successfully processed them using i5 risk stratification algorithms, along with other tools, to create heat maps of South West London to the establish the districts of greatest risks from Coronavirus and to support GPs in their shielding and vaccinations exercises. - Clinicians carrying out telephone assessments or remote consultations. Other benefits to patients, the NHS and the wider population have included: - Public Health authorities and governments in planning for disease control, levels of quarantine and targeted shielding. · Reduction in the cases and progression of LTC Further evidence can be found here: https://digital.nhs.uk/coronavirus/risk-assessment. · Reduction in the suffering of patients • NEL CSU have used the COVID-19 tool for outcome risk stratification, identifying patients for shielding and subsequently vaccinations in all London boroughs. It is envisaged that the same methodology could be used in the event of future COVID outbreaks and i5 Health is planning to keep the predictive tool up to date at its own expense and at no cost to the NHS - as it did in response to the April 2020 outbreak. · Avoidance of a decline in the quality of life • The COVID-19 Calculator risk assessments can lead to persons, clinicians and government authorities taking more informed decisions that could reduce the effect of the condition Covid-19 itself (e.g., the screening exercise carried out for London). · Reduction in the mortality risk from LTC • The focus given by the NHS to COVID-19 during the last pandemic has had a negative effect on the treatment of Long-Term Conditions in general because of consequent delays in screening or of treatment. The COVID-19 Calculator can contribute to authorities carrying out more selective screening of populations which means more optimal use of resources and better availability of resources for other Long-Term Conditions. · Early treatment for identified patients, probably in a primary care or social prescribing setting • Identification of patients that have (or risk having) a Long-Term Condition (LTC) but do not appear on the GP's risk register can lead to better management of their health requirements. Haringey CCG commissioned NEL CSU to apply i5 Health algorithms to its population to establish the identities of individuals (within their own identifiable patient care datasets) with as yet undiagnosed Atrial Fibrillation (AF). Because Haringey’s population of over 270,000 had a lower recorded AF prevalence level (0.9%) than the national average prevalence (1.5%), i5 Health was missioned by the CCG, along with NEL CSU, to apply their tool to find the missing patients. As a result of this audit a minimum of 76 patients were identified from the list provided by secondary care data and coded on the Quality Outcomes Framework (QOF) register with a confirmed diagnosis of AF, an increase in the diagnosed prevalence of AF in Haringey CCG from 0.9% and closer to the estimated prevalence of 1.5%. 46 out of the 76 patients (68%) added to the QOF AF register as AF patients who were anticoagulated for stroke prevention. Further evidence from the Stroke Association can be found here: https://www.stroke.org.uk/sites/default/files/af-data_2018_haringey-ccg_08d_1.pdf. · Society benefitting from less pressure on the patient and family members and the greater availability of individuals to work • i5 Health co-operated with Arden & GEM CSU on the application of i5 algorithms to the population of the NHS Milton Keynes CCG area (300,000) in respect of: Hypertension, Atrial Fibrillation, Heart Failure, Diabetes Mellitus, Cancer, Asthma, COPD, Coronary Heart Disease, Palliative Care, Dementia, Depression (18+), Obesity (16+), Epilepsy (18+), Chronic Kidney Disease (16+), Stroke and TIA, Rheumatoid Arthritis (16+), Peripheral Arterial Disease, Osteoporosis (50+). The results enabled the GP practices to strengthen the relevant QoF registers and treat patients more effectively. Further evidence from AGEM CSU can be found here: The Complete Care Community Programme - Evaluating The Early Development And Progress Of The Programme https://www.ardengemcsu.nhs.uk/media/2869/ccc-evaluation-report_v2.pdf. · Reduction in acute hospital admissions for patients • Reports created by i5 allowed each of the 32 CCGs and 5 STPs in London to establish the value of introducing of Social Prescribing within their areas. CCGs were able to introduce the findings into their planning processes during the year 2019/20 and 2020/21 and raise significant funding for Social Prescribing interventions. Further evidence can be found here: Embedding social prescribing across the five London STP footprints, https://www.kingsfund.org.uk/sites/default/files/media/Shaun_Crowe.pdf. · Improvements in operational efficiency · Avoidance of mass screening in primary care and community settings by pre-selecting for clinical review only that cohort identified by the tool as having or likely in the next 12 to 24 months to have an LTC (eg Atrial Fibrillation) · Reduction of pressure on secondary care and related costs through fewer hospital admissions.

Text removed is struck through; text added is underlined. Unchanged paragraphs are summarised rather than repeated.

Datasets: + HES-ID to MPS-ID HES Admitted Patient Care; + HES-ID to MPS-ID HES Outpatients

Objective for processing

i5 Health Ltd provides consultancy services to support to Clinical Commissioning Groups [24 words unchanged] i5 Health Ltd (hereafter known in this section as i5 Health) requires an extension to the term of this Agreement, and a renewal of pseudonymised data from NHS Digital for the following purposes: [1 paragraph unchanged] i5 Health Ltd (i5 Health) evaluates - on behalf of the Health [63 words unchanged] (Academic Paper ID: WNC 48 'Nurse Prescribing' - Worldwide Nursing Conference, Singapore 2014;abstract 2014; abstract [2 paragraphs unchanged] i5 Health provides consultancy services to support to Clinical Commissioning Groups (CCG), (CCGs), CSUs, Sustainability and Transformation Plans (STP), Acutes, NHS England and Local Authorities (LA) in their decision making for commissioning purposes. Under Purpose 2 there are a number of specific sub-purposes: [17 paragraphs unchanged] In many parts of the country, a number of LTC Long Term Condition (LTC) patients are sub-optimally treated because they fail to get on to the [25 words unchanged] of them progressing to a full LTC. i5 Health has developed algorithms for a service that identify, identifies, at surgery level, the numbers of patients that fall into both these categories. At the request of NHS England requires England, i5 Health to carry out a study has applied this system in respect of the GP practices on Merseyside (specifically, 19 practices in Southport and Formby. This Formby CCG and 30 practices in South Sefton CCG); similar case finding services using these algorithms have been provided to the NHS England initiative is continuing. in London and in the Midlands. The Case Finding role case finding and risk stratification roles of i5 Health was has been further called on (at the specific request of the Board of NHS England) in the context of Social Prescribing for all of London; for the CSU support of CCG shielding and vaccination programmes in London (at the specific request of the Board ; for Evidence Based Interventions (EBI) work of NHS England). England and the Academy of Medical Royal Colleges (AoMRC) across the country. To support this work i5 Health needs both SUS data SUS, ECDS and HES data on an annual basis to carry out its functions [58 words unchanged] as full medical history and treatments of a patients including procedure dates. Unfiltered historic data is essential to build full and accurate medical histories [12 words unchanged] that specific interventions can achieve (as demonstrated in the Evidence Based Interventions (EBI) exercise being carried out by i5 Health for NHS England). The availability [78 words unchanged] creating algorithms for the purpose of future case finding and risk stratification. In an NHS England exercise carried out by i5 Health last year and continuing for the rest of this year on Evidence Based Interventions (EBI), the longevity of the records allows i5 Health to strengthen the theory that historical NHS data, if properly interrogated and learnt from, is rich enough to provide valuable supporting evidence for NICE guidance and clinicians addressing new patients’ needs. It is not intended that more than seven years data Data will be sought. destroyed on a rolling basis, and evidence of data destruction will be provided to NHS Digital on a yearly basis. In an NHS England and Academy of Medical Royal Colleges (AoMRC) exercise carried out by i5 Health over the past two years and continuing into 2022 on Evidence Based Interventions (EBI), the longevity of the records allows i5 Health to strengthen the theory that historical NHS data, if properly interrogated and learnt from, is rich enough to provide valuable supporting evidence for NICE guidance and clinicians addressing new patients’ needs. [1 paragraph unchanged] Voluntary Sector Organisations (VSOs) already cooperate with i5 Health to improve the [18 words unchanged] within the NHS. The VSOs have occasion to ask for i5 Health reports, reports (aggregated with small numbers supressed, in line with HES Analysis Guidance), based on data analysis that can improve their own specific charitable works for NHS patients. [1 paragraph unchanged] i5 Health has reviewed the requirement for the amount of data being supplied through this renewal agreement and has assessed the data requested as necessary for the purposes of their agreement. [69 words unchanged] minimising the data reduces the categorisation ability for fields that are non-obvious. [1 paragraph unchanged] LEGAL BASIS FOR PROCESSING National data is required as i5 Health provide reports from local through Regional to National levels. Multiple years of data are required in order to produce time-series and predictive modelling; historic data is retained to enable this. The data cannot be minimised by applying filters to specific conditions of relevance as the full data is needed in order to produce the outputs as outlined within this Agreement. Article 6(1)(f) - 'processing is necessary for the purposes of the legitimate interests ...' and Article 9(2)(j) HES data itself has many diagnosis fields. This is critical for the development of i5 Health’s algorithms. A key case in point is the current major project for NHS England at Skipton House that i5 Health is centring. This is Evidence Based Interventions (EBI) referenced in the Yielded Benefits Purpose section under #2.1) Case Finding: “Develop criteria of adverse outcomes based on treatment received to support the separation of patient cohorts that should or should not have received treatment. Such definitions may include readmissions, complications, further treatment needs and death. These criteria may also be used to develop further NICE guidance”. The breadth of the information needed for the EBI project is important because of the number of potential and different pathways of analysis. i5 Health need the maximum number of fields to permit clustering methods based on common co-morbidities used to investigate if commonalities exist amongst patient groups. 'processing is necessary for archiving purposes in the public interest, scientific or historical research purposes or statistical purposes...' SUS PbR data provides a very different type of information. It is, by definition, costed data and grouped relating to finance – needing HIG codes and national tariffs. SUS PbR is the basis of i5 Health’s costing advice when it comes to using the i5 Commissioning Opportunities (i5 COP) set of algorithms that provide the value for CCGs of alternative (principally non-hospital) treatments. i5 Health have undertaken a Legitimate Interests Assessment and concluded that they can rely on legitimate interests for this processing. The number, richness and extent of the data sets NHS Digital currently provide to i5Health are essential for the training of the organisations Neural Networks that are at the core of i5 Artificial Intelligence (AI) modules. The result is the highest level of reliability when reports are created for commissioning and NHS England decision makers. i5 Health has reviewed the requirement for the amount of data being supplied under this Agreement and assessed the data requested as necessary for the purposes of this Agreement. i5 Health’s legitimate interests are a necessary and lawful basis for processing and controlling SUS and HES data, as they enable i5 to assist NHS customers with the access and use of i5 analytics products and services, and the development of the same. i5 considers access to its analytics to be in the public interest as well as of critical value to end-user NHS customers, so they can better understand their patient pools and make better health and social care decisions. Legitimate interests also include for i5 Health, as a commercial for-profit company, to continue as a provider of leading analytics in the UK, and as well to support compliance with the clinical and non-clinical performance expectations of the NHS. The lawful basis for this processing is Article 6(1)(f) - 'processing is necessary for the purposes of the legitimate interests ...' and Article 9(2)(j) The purpose of i5 solutions includes, though not limited to, utilisation by NHS customers to produce baselines for innumerable outcome metrics derived from key data combinations, to support better use of resources, staff, and services, ultimately resulting in better health and social care outcomes. It is important to note that access to SUS and HES data by i5 Health’s NHS customers is only ever to aggregated data, with numbers suppressed in line with ICO standards, and pseudonymisation always occurring prior to transfer of SUS and HES data to i5 Health. 'processing is necessary for archiving purposes in the public interest, scientific or historical research purposes or statistical purposes...'. i5 Health processing activities are a necessary, targeted, and proportionate means of achieving the purposes of the legitimate interests outlined above, and these interests are not overridden by moral or ethical issues, and are balanced against any impact on individual rights. i5 Health uses the least intrusive means possible to achieve the analytics it provides to i5 Health customers, strictly within the UK only. The risks of any data breach are clearly understood by i5 Health, which is why i5 Health requires SUS and HES data to be pseudonymised before it is provided to the company and does not require access to the encryption key in order to process SUS and HES information. i5 Health have undertaken a Legitimate Interests Assessment and concluded that they can rely on legitimate interests for this processing. i5 Health’s legitimate interests are a necessary and lawful basis for processing and controlling SUS, ECDS and HES data, as they enable i5 to assist NHS customers with the access and use of i5 analytics products and services, and the development of the same. i5 considers access to its analytics to be in the public interest as well as of critical value to end-user NHS customers, so they can better understand their patient pools and make better health and social care decisions. Legitimate interests also include for i5 Health, as a commercial for-profit company, to continue as a provider of leading analytics in the UK, and as well to support compliance with the clinical and non-clinical performance expectations of the NHS. The purpose of i5 solutions includes, though not limited to, utilisation by NHS customers to produce baselines for innumerable outcome metrics derived from key data combinations, to support better use of resources, staff, and services, ultimately resulting in better health and social care outcomes. It is important to note that access to SUS and HES data by i5 Health’s NHS customers is only ever to aggregated data, with numbers suppressed in line with ICO standards. i5 Health processing activities are a necessary, targeted, and proportionate means of achieving the purposes of the legitimate interests outlined above, and these interests are not overridden by moral or ethical issues and are balanced against any impact on individual rights. i5 Health uses the least intrusive means possible to achieve the analytics it provides to i5 Health customers, strictly within the UK only. The risks of any data breach are clearly understood by i5 Health, which is why i5 Health requests that the SUS, ECDS and HES data be pseudonymised before it is provided to the company and does not require access to the encryption key in order to process SUS, ECDS and HES information. [1 paragraph unchanged] i5 Health provides the these services solely for the only where there is benefit of to the NHS and its patients. i5 Health Limited does not provide nor would provide these services to commercial sector health bodies i5 Health has developed Artificial Intelligence (AI) algorithms that interrogate, in various ways, secondary care data provided by NHS Digital. The results of those interrogations, when combined with other data (e.g. workforce, size of population…etc), contribute to reports requested by and for NHS decision makers. These include reports on: i5 Health Limited are the sole data controller who also process the data for the purposes described within this Agreement. - risk stratification i5 Health have received government funding but no commercial funding or sponsorship. As evidenced in this Agreement, NHS Digital data will not be used solely for commercial purposes. Given the nature of the services/outputs of i5 Health for the NHS, as described within this Agreement, i5 Health believe the benefits of such for the public are proportionate to the commercial advantages to i5 Health. - outcome prediction There are different commercial arrangements depending on the specific services/needs of the NHS organisation concerned. Examples include: - service recommendations • Analysis for NHS England of the likely effect of on the structure and finances of the NHS in London arising from the implementation of 44 digital initiatives. i5 Health invoiced on a per diem basis. - health economy planning • Application of i5 algorithms, at the request of NHS England, to data of Merseyside CCGs to establish whether there were critical gaps between the numbers on local GP LTC registers and the likely numbers of, as yet, undiagnosed sufferers. i5 Health invoiced on the basis of size of the CCG population. - invoice validation • At the request of NHS England, establishing whether administrative data can help to select inappropriate procedures for consideration for the Evidence Based Interventions (EBI) programme. It involves analysing longitudinal patient records to ascertain what criteria is present that presents an unacceptable outcome for a patient undergoing a specific procedure. Also establishing whether administrative data can help clarify the criteria for intervention to contribute to NICE guideline development. i5 Health invoices on the basis of monthly Work Packages agreed with NHS England. - reports on the patient care and financial benefits of specific activities e.g. Non-Medical Prescribing, digital initiatives, etc. • Licensing of i5 Artificial Intelligence algorithms installed in the server of NHS Arden & GEM CSU with which i5 Health is a longstanding BI partner. A per annum license fee has been paid. An alternative commercial arrangement is one whereby, for a fee based on the size of a CCG, i5 Health is asked by the CSU to apply its algorithms to NHS data in order, e.g. to ascertain the financial value of introducing specific primary care initiatives that reduce the dependency on secondary care. There will never be any requirement or attempt by i5 Health to re-identify individuals.

Processing activities

Under this version of the Agreement NHS Digital will provide i5 Health with one drop of the latest annual record level pseudonymised SUS PbR PbR, ECDS and HES data via the Secure Electronic File Transfer (SEFT) system. A Database Analyst (DBA) from i5 Health will load the record level pseudonymised data into an i5 Health secure database. database. The database will be managed locally and secured by the DBA with user access control. Record-level pseudonymised data will only be accessed by individuals within the Analytics [23 words unchanged] within this DSA, all of whom are substantive employees of i5 Health. All those accessing data under this Agreement are substantive employees of i5 Health, and have received appropriate training in data protection and confidentiality. Data will only be accessed at the named processing location as set out in this Agreement. Data will only be accessed at the named processing location as set out in this Agreement. The additional SUS PbR, ECDS and HES data being provided under this Agreement will be linked to SUS PbR data already held across the data sets (e.g. SUS PbR Episode data with SUS PbR A&E data); from National Level to GP Practice Level. The additional SUS PbR and HES data being provided will be linked to SUS PbR data already held, across the data sets (e.g. SUS PbR Episode data with SUS PbR A&E data); from National Level to GP Practice Level. There will be no requirement nor attempt to re-identify individuals within the data sets. National data is required as i5 Health provide reports from local through Regional to National levels. Multiple years of data are required in order to produce time-series and predictive modelling, historic data is retained to enable this. The data cannot be minimised by applying filters to specific conditions of relevance as the full data is needed in order to produce the outputs as outlined within this Agreement. There will be no requirement nor attempt to re-identify individuals within the data sets. HES data itself has many diagnosis fields. This is critical for the development of i5 Health’s algorithms. A key case in point is the current major project for NHS England at Skipton House that i5 Health is centring. This is Evidence Based Interventions (EBI) referenced in the Yielded Benefits Purpose section under #2.1) Case Finding: “ Develop criteria of adverse outcomes based on treatment received to support the separation of patient cohorts that should or should not have received treatment. Such definitions may include readmissions, complications, further treatment needs and death. These criteria may also be used to develop further NICE guidance”. The breadth of the information needed for the EBI project is important because of the number of potential and different pathways of analysis. I5 Health need the maximum number of fields to permit clustering methods based on common co-morbidities used to investigate if commonalities exist amongst patient groups. Data will not be made available to any third parties, including VSO’s, except in the form of aggregated outputs with small numbers suppressed in line with the HES Analysis Guide which also covers suppression rules for SUS. For example, a report containing aggregate data for cohorts from localities to large geographical areas will be produced for NHS England. SUS PbR data provides a very different type of information. It is, by definition, costed data and grouped relating to finance – needing HIG codes and national tariffs. SUS PbR is the basis of i5 Health’s costing advice when it comes to using the i5 Commissioning Opportunities (i5 COP) set of algorithms that provide the value for CCGs of alternative (principally non-hospital) treatments. Processing specific to the i5 Coronavirus Health Risk Calculator: The pseudonymised medical records of patients admitted to hospital with Covid-19 will be analysed to establish the characteristics in their history that might be additional to/different from those identified in the earlier analysis of Influenza and historical (ie non-Covid-19) Coronavirus patients. The proposed Work Packages are as follows: WP1 - Collect/categorise data (15 days) • Three years of medical records with over 200 fields including 21 diagnosis and 21 procedures and any diagnosis present prior to Covid-19 • Categorise into balanced cohorts (Low, Medium, High and Very High) for optimal separation between Sensitivity and Specificity WP2 - Principal Component Analysis (PCA) (10 days) • PCA carried out on diagnosis codes to ensure only conditions used that are pertinent for prediction WP3 - Data split (15 days) • Create separate data sets for AI training, monitoring of overfitting and performance testing • Balancing sets to contain patients with various outcomes and avoid over-representation WP4 - Coding and Training (25 days) • Short term -- risk level prediction; Medium/Long term -- Healthcare Planning information • Train various AI models and topologies to evaluate which combinations works best WP5 - Test beds (15 days) • Testing of revised calculator with several Integrated Care Systems nominated by CSUs • Address issues and develop any additional functionality The number, richness and extent of the data sets NHS Digital currently provide to i5Health are essential for the training of the organisations Neural Networks that are at the core of i5 Artificial Intelligence (AI) modules. The result is the highest level of reliability when reports are created for commissioning and NHS England decision makers. i5Health has reviewed the requirement for the amount of data being supplied under this Agreement and assessed the data requested as necessary for the purposes of this Agreement. Data will not be made available to any third parties except in the form of aggregated outputs with small numbers suppressed in line with the HES Analysis Guide which also covers suppression rules for SUS. For example, a report containing aggregate data for cohorts from localities to large geographical areas will be produced for NHS England. The inclusion of Voluntary Sector Organisations (VSO) as recipients in no way changes the processing activities set out in this paragraph, but will be limited to aggregate data with small numbers suppressed in line with the HES analysis guide which covers suppression rules for SUS. [1 paragraph unchanged] Data will only be accessed and processed by substantive employees of i5 Health and will not be accessed or processed by any other third parties not mentioned in this agreement. All organisations party to this agreement must comply with the Data Sharing Framework Contract requirements, including those regarding the use (and purposes of that use) by “Personnel” (as defined within the Data Sharing Framework Contract - i.e. employees, agents and contractors of the Data Recipient who may have access to that data). There will never be any requirement or attempt by i5 Health to re-identify individuals. All organisations party to this agreement must comply with the Data Sharing Framework Contract requirements, including those regarding the use (and purposes of that use) by “Personnel” (as defined within the Data Sharing Framework Contract - i.e. employees, agents and contractors of the Data Recipient who may have access to that data.

Expected output

The All outputs will be aggregated analysis with small numbers suppressed (in line with HES Analysis Guidance, and SUS suppression rules) for inclusion within economic evaluation and Clinical Commissioning Group (CCG) strategy. All [36 words unchanged] except in so far as is permitted for Voluntary Sector Organisations (VSO). [1 paragraph unchanged] The outputs i5 Health have provided over the last 24 months are (according to each purpose); purpose): [1 paragraph unchanged] • Initially first national report for Health Education Board of NHS England on [56 words unchanged] a decision by NHS England to increase considerably the funding of NMP. It is expected that, at The i5 exercise and its findings continue to be quoted in reports by decision makers in support of expanding the request practice of NMP. i5 Health maintains readiness to provide Health Education England (HEE), i5 Health will provide during the year (HEE) with further updates, based on the NHS data, of the workforce, patient care and financial benefits continuing to arise from Non-Medical Prescribing. Purpose 2) Consultancy Service Purpose 2.1) A number of commissioning support reports are made for CCGs within the footprint of Business Intelligence partners, Arden & GEM CSU; this has been further facilitated by the installation within the server of Arden & GEM CSU of some of the i5 Health algorithms that are trained on the NHS data provided under this agreement. The work carried out at the request of Arden & GEM CSU within the NHS Milton Keynes CCG area on Commissioning Opportunities is expected to expand to other CCG clients of the CSU through to the end of the DSA contract year. NHS England has commissioned in 2018 i5 Health to provide commissioning reports for 32 CCGs and 5 Sustainability and Transformation Partnerships (STPs) in London as well as analysis of the effect on London over the next five years of the introduction of Digital technology. i5 Health continues to analyse the data received in order, on request, to update and add to the commissioning reports and the dashboards provided to the CCGs and STPs and the reports on the value of Digital technology. i5 Health is currently working on a project with Lincolnshire CCG to ascertain if information about patient behaviour, conditions, and events, captured from wearables, monitors and other smart technologies can predict demand for services, then providing these technologies to patients and using the data generated will enable providers to pre-empt and redirect demand or design new services. The key outcomes provided by i5 Health are the creation of cohorts of patients for the exercise, in cooperation with the Data Services for Commissioners Regional Office (DSCRO) of Arden & GEM CSU, and the evaluation of the data provided by the wearables. (Creation of cohorts would not include identifying specific individuals using NHS Digital data.) The project is expected to continue into 2022 and beyond. Purpose #2.1) End of Life Purpose 2.3) As part of NHS England’s Electronic Palliative Care Co-ordination System (EPaCCS) programme, i5 Health carried out an evaluation of data in respect of End of Life and its related costs. This is a continuing project – the outcomes being dependent on changes in trends of data. Coming into this category is the major national exercise being focused by i5 Health on Evidence Based Interventions (EBI) – on the instructions of NHS England and the Academy of Medical Royal Colleges (AoMRC). A member of the of i5 Health team is to become a member of the Expert Advisory Committee of AoMRC which has oversight of progress of the project. The project has characteristics that brings it also within Purpose #2.4 below and is being renewed through to 2022 at least. Purpose #2.4) Case Finding Purpose 2.4) In many parts of the country, a number of LTC Long Term Condition (LTC) patients are sub-optimally treated because they fail to get on to the [42 words unchanged] the numbers of patients that fall into both these categories. NHS England requires have requested for i5 Health to carry out a study in respect of the GP [30 words unchanged] of London (at the specific request of the Board of NHS England). In respect of Arden & GEM CSU, it is expected that the Case Finding work carried out for its clients, Milton Keynes CCG, will be extended during the next to more of the 70 CCGs within the CSUs ‘footprint’. It is also expected that work started in Q1 early 2020 for the CSU’s clients, South Lincolnshire CCG, on an Internet of [129 words unchanged] will endeavour to accelerate the process as time is of the essence. All outputs will contain only data that is aggregated with small numbers suppressed in line with the HES Analysis Guide which covers suppression rules for SUS. Within this category are, currently, in respect of Kent and Medway CCG: - Facilitating, through i5 Activation Score algorithms based on the analysis of secondary care data and in cooperation with the DSCRO at NEL CSU and Health Diagnostics, the creation of cohorts of patients that are hard-to-reach for Health Check purposes. - Application of diagnosis stratification algorithms, in cooperation with the DSCRO at Arden & GEM CSU, to establish those cohorts at risk of diabetes induced leg ulcers. This exercise is in the context of the CCG’s Social Prescribing programme. It is envisaged that these exercises will continue into 2022 and beyond. Planning is also being undertaken on a project with South West London CCG and NEL CSU to address the problem of hard-to-reach members of the Black, Asian and minority ethnic (BAME) community for cardiovascular problems. The Evidence Based Initiatives (EBI) programme and its need for analysing secondary care data over several years has enabled i5 Health to create an AI tool to facilitate the exercise. This has ensured a great deal many more intervention can be spotlighted over the next 12-24 months. i5 Health has developed Artificial Intelligence (AI) algorithms that interrogate, in various ways, secondary care data provided by NHS Digital. The results of those interrogations, when combined with other data (e.g., workforce, size of population…etc), contribute to reports requested by and for NHS decision makers. These include reports on: - risk stratification - outcome prediction - service recommendations - health economy planning - invoice validation - reports on the patient care and financial benefits of specific activities eg Non-Medical Prescribing, digital initiatives, etc.

Expected measurable benefits

[1 paragraph unchanged] The immediate benefits of the NMP analysis carried out by i5 Health were the increase in investment of £500,000 by the NHS in nurse prescribing training. More nurses who are able to prescribe will accelerate patient treatment, freeing up doctors to spend more hands on time with patients who need to be seen by a doctor. [8 paragraphs unchanged] Not least of all, the clinicians Audit, in growing use since 2009, [11 words unchanged] of the performance and effect of NMP is necessary. i5 Health is therefore being asked to review reviewing existing HES and studies with a view to draw drawing out relevant information, propose proposing new methodology, refine refining existing audits and promote promoting new ones to provide a comprehensive analysis of NMP to assist in decisions on whether and to what extent NMP should be adopted more widely in England. Voluntary Sector Organisations (VSO) cooperate with i5 Health to improve the extent [8 words unchanged] relies on to support, with data processing, clinical commissioning within the NHS. The VSOs have occasion to ask for i5 Health reports, based on data analysis that can improve their own specific charitable works for NHS patients. The benefits to the UK Health and Social Care system are better strategic planning and commissioning decisions, and subsequently improved care for patients due to better planning and strategy. The benefits of the Evidence-Based Interventions (EBI) programme are the prevention of avoidable harm to patients, avoidance of unnecessary operations and freeing up clinical time by only offering interventions on the NHS that are evidence-based and appropriate. Phase 1 of the programme targeted 17 such interventions. The 17 interventions in question were snoring surgery in the absence of Obstructive Sleep Apnoea (OSA), dilatation and curettage (D&C) for heavy menstrual bleeding in women, knee arthroscopy for patients with osteoarthritis and injections for non-specific low back pain without sciatica, breast reduction, removal of benign skin lesions, grommets for glue ear in children, tonsillectomy for recurrent tonsillitis, haemorrhoid surgery, hysterectomy for heavy menstrual bleeding, chalazia removal, arthroscopic shoulder decompression for subacromial shoulder pain, carpal tunnel syndrome release, Dupuytren’s contracture release, ganglion excision, Trigger finger release, varicose vein surgery. The current Phase 2 of the programme is selecting further procedures and Phase 3, commencing during Q1 2021, will select more. The financial benefit for the healthcare system varies from case to case. By way of example, analysis by i5 Health for the Sussex health economy two years ago resulted in a 10% reduction in Non-Elective admissions (NEL). The initial EBI work of i5 Health was to answer two key questions: i5 Health are currently instructed by NHS England on the following: Question 1: 'Can administrative data help to select inappropriate procedures for consideration for the EBI programme?' Essentially this is about establishing, from longitudinal patient records, what criteria is present that presents an unacceptable outcome for a patient undergoing a specific procedure. The aim of the Evidence-Based Interventions (EBI) programme is to prevent avoidable harm to patients, avoid unnecessary operations and free up clinical time by only offering interventions on the NHS that are evidence-based and appropriate. Phase 1 of the programme targeted 17 such interventions. The Phase 2 of the programme will select further procedures and is the subject of this proposal. Question 2: 'Can administrative data help us to clarify the criteria for intervention to contribute to our guideline development?' For the purposes of Phase 2, NHS England wishes to explore the feasibility of the i5 approaches and tools to detect and select inappropriate procedures as well as information that characterises the inappropriate procedures. To this end, i5 Health is doing the following using the data supplied under this agreement: Having provided positive answers to both, i5 Health have developed criteria of adverse outcomes based on treatment received to support the separation of patient cohorts that should or should not have received treatment. Such definitions include readmissions, complications, further treatment needs and death. These criteria are used to develop further NICE guidance, which has included guidance on the clinical coding inclusion/exclusion criteria of surgical intervention for chronic sinusitis, removal of adenoids, and cystoscopy for men with uncomplicated lower urinary tract symptoms. For the purposes of Phases 2 and 3 of EBI, i5 Health are doing the following, using the data supplied under this agreement: [1 paragraph unchanged] - The evidence for most frequent procedures or those with most variation are is being examined by the clinical team working with i5 Health to assess appropriateness. - Procedures that are deemed inappropriate in certain circumstances will then be further examined to describe the criteria that make the procedure inappropriate (e.g. (e.g., low rate of diagnosis confirmed, high rate of complications). [5 paragraphs unchanged] South West London CCG have used the COVID-19 tool for outcome risk stratification, identifying patients for shielding and subsequently vaccinations in six boroughs. [3 paragraphs unchanged] Consequent on delivery of Preliminary EBI Evidence Pack, working with GIRFT Getting It Right First Time (GIRFT) Clinical Coding Team, Expert Working Groups, Clinical Classifiers , Clinical Leads and Price Assessors to enrich the material in Preliminary EBI Evidence Pack. [4 paragraphs unchanged] i5 Health have received government funding but no commercial funding or sponsorship. As evidenced in this Agreement, NHS Digital data will not be used solely for commercial purposes. Given the nature of the services/outputs of i5 Health for the NHS, as described within this Agreement, i5 Health believe the benefits of such for the public are proportionate to the commercial advantages to i5 Health. There are different commercial arrangements depending on the specific services/needs of the NHS organisation concerned. Examples include: • Analysis for NHS England of the likely effect of on the structure and finances of the NHS in London arising from the implementation of 44 digital initiatives. i5 Health invoiced on a per diem basis. • Application of i5 algorithms, at the request of NHS England, to data of Merseyside CCGs to establish whether there were critical gaps between the numbers on local GP LTC registers and the likely numbers of, as yet, undiagnosed sufferers. i5 Health invoiced on the basis of size of the CCG population. • At the request of NHS England, establishing whether administrative data can help to select inappropriate procedures for consideration for the Evidence Based Interventions (EBI) programme. It involves analysing longitudinal patient records to ascertain what criteria is present that presents an unacceptable outcome for a patient undergoing a specific procedure. Also establishing whether administrative data can help clarify the criteria for intervention to contribute to NICE guideline development. i5 Health invoices on the basis of monthly Work Packages agreed with NHS England. • Licensing of i5 Artificial Intelligence algorithms installed in the server of NHS Arden & GEM CSU with which i5 Health is a longstanding BI partner. A per annum license fee has been paid. An alternative commercial arrangement is one whereby, for a fee based on the size of a CCG, i5 Health is asked by the CSU to apply its algorithms to NHS data in order, e.g. to ascertain the financial value of introducing specific primary care initiatives that reduce the dependency on secondary care.

Benefits reported

Purpose #2.1) Case Finding Purpose 1.) Non-Medical Prescribing (NMP) Coronavirus - The immediate benefits of the NMP analysis carried out by i5 Health were the increase in investment of £500,000 by the NHS in nurse prescribing training. The increased number of , in particular, nurses who are able to prescribe will accelerate patient treatment, freeing up doctors to spend more time for patients who need to be seen by a doctor. i5 Health has created for the NHS and for individuals a Calculator that establishes the Health Risk of Coronavirus (accessible at https://coronavirusrisk.org). It uses past medical history of people to predict their risk levels if infected by the virus 2019-nCoV leading to Covid-19. It categorises people into low, medium, high and very high risk in order to facilitate shielding and protection strategies for escalation and de-escalation, to support hospitals in their management – particularly of ICUs – and to provide information for decision making by clinicians and individuals. Purpose 2.4) Case Finding and Risk Stratification North East London (NEL) CSU has selected the medical records of circa 4,100,000 Londoners and successfully processed them through the i5 Health system in an exercise to identify those most at risk of Coronavirus. As data starts to come in about the new strain Covid-19, i5 Health will upgrade the Calculator accordingly and work with NEL CSU and Arden & GEM CSU to test the upgraded system. - Haringey CCG missioned NEL CSU to apply i5 Health algorithms to its population to establish the identities of individuals (within their own identifiable patient care datasets) with as yet undiagnosed Atrial Fibrillation (AF). Because Haringey’s population of over 270,000 had a lower recorded AF prevalence level (0.9%) than the national average prevalence (1.5%), i5 Health was missioned by the CCG, along with North East London CSU, to apply their tool to find the ‘missing’ patients. As a result of this audit a minimum of 76 patients were identified from the list provided by secondary care data and coded on the Quality Outcomes Framework (QOF) register with a confirmed diagnosis of AF, an increase in the diagnosed prevalence of AF in Haringey CCG from 0.9% and closer to the estimated prevalence of 1.5%. 46 out of the 76 patients (68%) added to the QOF AF register as AF patients who were anticoagulated for stroke prevention. NEL CSU has used the i5 Coronavirus Health Risk Calculator to provide the following: - i5 Health co-operated with Arden & GEM CSU on the application of i5 algorithms to the population of the NHS Milton Keynes CCG area (300,000) in respect of: Hypertension, Atrial Fibrillation, Heart Failure, Diabetes Mellitus, Cancer, Asthma, COPD, Coronary Heart Disease, Palliative Care, Dementia, Depression (18+), Obesity (16+), Epilepsy (18+), Chronic Kidney Disease (16+), Stroke and TIA, Rheumatoid Arthritis (16+), Peripheral Arterial Disease, Osteoporosis (50+). The results enabled the GP practices to strengthen the relevant registers and treat patients more effectively. • Mapping Covid 19 – the calculator enabled NEL CSU to create heat map of South West London to the establish the districts of greatest risks. - The on-line process created by i5 allowed each of the 32 CCGs and 5 STPs in London to establish the value of introducing of Social Prescribing within their areas. CCGs accessing the site have been able to introduce the findings into their planning processes during the year 2019/20 and 2020/21. • SHIELD – the calculator supported NEL CSU in the SHIELD List Segmentation exercise as requested by the government. - It was decided, following the initial results of the Evidence Based Interventions (EBI) programme of NHS England and the Academy of Medical Royal Colleges (AoMRC), that not enough input had been achieved through the analysis of historical patient data. i5 Health developed algorithms to interrogate data and to analyse intervention results. The positive benefits have been the reinforcement of NICE guidelines on interventions and an extension of the programme using i5 input from 17 interventions to circa 60. • Specific Patient Targeting – GP’s who were provided re-identified lists by NHS Digital of own patients with their risk scores. - The NHS in London selected the medical records of circa 4,100,000 Londoners and successfully processed them using i5 risk stratification algorithms, along with other tools, to create heat maps of South West London to the establish the districts of greatest risks from Coronavirus and to support GPs in their shielding and vaccinations exercises. By enabling NEL CSU to support the local health authorities in targeting efforts on specific districts at greater risk and deciding on increased shielding activities, with the potential to drill down to individual patients, it will allow the GPs to make decisions about their patients using the risk score calculated using the i5 Coronavirus Health Risk Calculator. Other benefits to patients, the NHS and the wider population have included: Diagnosis Stratification · Reduction in the cases and progression of LTC - i5 Health have finalised an exercise with NEL CSU and NHS Haringey CCG which started in Quarter 4 of 2017 and led to the successful application of i5 Health case finding algorithms to the medical records of the population of the NHS Haringey CCG area in respect of Atrial Fibrillation. The results reported in Quarter 2 of 2019 confirmed the power of the tool. · Reduction in the suffering of patients - i5 Health co-operated, in periods Quarter 2 to Quarter 4 of 2020 with Arden & GEM CSU and NHS Milton Keynes CCG on the application of the Case Finding Tool to the population of the NHS Milton Keynes CCG area in respect of: Hypertension, Atrial Fibrillation, Heart Failure, Diabetes Mellitus, Cancer, Asthma, COPD, Coronary Heart Disease, Palliative Care, Dementia, Depression (18+), Obesity (16+), Epilepsy (18+), Chronic Kidney Disease (16+), Stroke and TIA, Rheumatoid Arthritis (16+), Peripheral Arterial Disease, Osteoporosis (50+). · Avoidance of a decline in the quality of life Social Prescribing · Reduction in the mortality risk from LTC As noted last year, for all of London there is now an on-line process that allows each of the 32 CCGs and 5 STPs to establish the value of introducing of Social Prescribing within their areas. CCGs accessing the site have been able to introduce the findings into their planning processes during the year 2019/20. Patients that, with the cooperation of NHS Digital, were able to be recognised as 'missing' from the LTC registers or, if not treated, would suffer from an LTC continue to be contacted by their GPs for screening purposes, i5 Health advised at several Social Prescribing events during the year. Building on the above work, i5 Health carried out, in Quarter 3 of 2019, preliminary analysis in respect of the population within the NHS West Kent CCG area on Leg Ulcers and Diabetes and developed a proposal to be actioned post the NHS Kent and Medway CCGs merger by mid-2020 to the entire region. · Early treatment for identified patients, probably in a primary care or social prescribing setting A proposal was made in Quarter 2 of 2019 to Public Health England to apply the Social Prescribing algorithms to the rest of the South of England which proposal is still awaiting budget approval in 2020. · Society benefitting from less pressure on the patient and family members and the greater availability of individuals to work Evidence Based Interventions (EBI). · Reduction in acute hospital admissions for patients In Quarter 1 of 2019, the Department of Data and Analytics at NHS England and NHS Improvement commissioned i5 Health to use its knowledge and understanding in Case Finding combined with other of its analytical skills to centre a study on Evidence Based Interventions (EBI). · Improvements in operational efficiency EBI work done during 2019 comprised: · Avoidance of mass screening in primary care and community settings by pre-selecting for clinical review only that cohort identified by the tool as having or likely in the next 12 to 24 months to have an LTC (eg Atrial Fibrillation) a) Discovery Phase · Reduction of pressure on secondary care and related costs through fewer hospital admissions. - Team approach: top down (Clinical) and bottom up (Data Analysis) with support from Leads experts - A focus on CCG/Provider areas for numbers of elective interventions; establishing clusters of OPCS Classification of - Interventions and Procedures to check variability; increasing the criteria to include patient history and outcomes. b) Review Phase – 17 Original Procedures revisited by i5 Health and adjusted in the light of findings from Discovery Phase. c) 45 Additional Procedures Phase - Research and Identification on: - Using i5 Health's AI tools to find and address 45 additional procedures of limited value. - Prioritising them using various criteria including the most challenging in terms of costs, clinicians’ concerns, available evidence. - Reviewing data requirements for coding of clinical criteria for the 45 procedures. - Production of Preliminary EBI Evidence Pack for these procedures.

Text removed is struck through; text added is underlined. Unchanged paragraphs are summarised rather than repeated.

Datasets: + Emergency Care Data Set (ECDS)

Objective for processing

i5 Health Limited (i5 Health) requires the date from NHS Digital for the following purposes: i5 Health Ltd provides consultancy services to support to Clinical Commissioning Groups (CCG), Commissioning Support Units (CSUs), Sustainability and Transformation Plans (STP), Acute services, NHS England and Local Authorities (LA) in their decision-making for commissioning purposes. i5 Health Ltd (hereafter known in this section as i5 Health) requires a renewal of data from NHS Digital for the following purposes: Purpose #1) 1) i5 Health Limited Ltd (i5 Health) evaluates, evaluates - on behalf of the Health Education Board of NHS England, the economic [52 words unchanged] on different health economies. (Academic Paper ID: WNC 48 'Nurse Prescribing' - WorldwideNursingConference,Singapore2014;abstract Worldwide Nursing Conference, Singapore 2014;abstract http://www.citeulike.org/user/gstf/article/13247895 ) http://www.citeulike.org/user/gstf/article/1324789). First full first report: (http://www.i5health.com/NMP/NMPEconomicEvaluation.pdf ) First full first report (http://www.i5health.com/NMP/NMPEconomicEvaluation.pdf ) Purpose 2) Purpose #2) [1 paragraph unchanged] The specific purposes are:- Under Purpose 2 there are a number of specific sub-purposes: Purpose #2.1) 2.1) [1 paragraph unchanged] Purpose #2.2) 2.2) [1 paragraph unchanged] Purpose #2.3) 2.3) To measure standards of care and identify gaps in provision to inform commissioning strategy. A number of CCGs including NHS Halton CCG, C4G CCG, Brent CCG, Ashford CCG, have been working with i5 Health in this respect to develop their [17 words unchanged] DSCON066/Halton CCG), the outcome was described by the Director of Transformation as giving; giving [1 paragraph unchanged] i5 Health requires SUS PBR spells & episode at patient level, including procedure and diagnosis codes, in order to evaluate the applicability of a particular QIPP initiative for a group of patients. Data on PBR spells and episodes is essential in i5 Health establishing the nature and size of specific patient cohorts in a given acute provider setting. Such identification allows i5 Health to calculate accurately the effect of any proposed, specific initiative including the financial impact of that change (e.g. provision of certain alternatives in the primary care sector to hospital treatment). Purpose 2.4) To provide Case Finding and Risk Stratification services. i5 Health has created, free of charge, for the NHS a Calculator that establishes the Health Risk of Coronavirus. Like in the contexts of the Diagnosis Stratification and Targeted Social Prescribing tools and the EBI work for NHS England, it uses past medical history of people to predict their risk levels – in this case if infected by a form of Coronavirus (ie not based on the data of SARS-CoV-2, which causes COVID-19). It is intended the Calculator be upgraded to take account of SARS-CoV-2 once the relevant data is made available. The i5 Coronavirus Health Risk Calculator ("Calculator") establishes a person’s health risk category as either low, medium, high or very high, in the event of being infected by Coronavirus. The Calculator has been developed from NHS hospital medical profiles of patients that had either Coronavirus prior to the emergence of COVID -19 or Influenza. The Calculator may be used by: • Clinicians to support assessment of individuals and advising them on adjustments to their lifestyles to minimise the risk of infection • Clinicians and Hospital Management for prioritising care, improving bed management and ensuring right facilities in ICUs • Clinicians for telephone assessment or remote consultations • Public Health authorities and governments, for planning for disease control, levels of quarantine and targeted shielding The novel Coronavirus, SARS-CoV-2, is the pathogen responsible for the infectious respiratory disease COVID-19. NHS Digital collects patient data every month and processes it into a form data experts can use for analysis and advice to the NHS. By early September 2020, data on patients hospitalised to the end of July2020 July will be available, thus enabling an upgrading of the calculator. The tool is currently based on Secondary care data though it is expected that the upgrading will also include Primary Care data when that becomes available. The science underlying the Calculator is set out in the academic paper entitled ‘Predicting Health Risk in Patients with Coronavirus or Influenza using Artificial Intelligence‘ at: https://www.i5analytics.com/HealthRiskInPatientsWithCoronavirus.pdf In many parts of the country, a number of LTC patients are sub-optimally treated because they fail to get on to the relevant registers at GP practices; additionally, there are many patients that have conditions which, if identified early enough, could receive treatment that reduces the risk of them progressing to a full LTC. i5 Health has developed algorithms that identify, at surgery level, the numbers of patients that fall into both these categories. NHS England requires i5 Health to carry out a study in respect of the GP practices in Southport and Formby. This NHS England initiative is continuing. The Case Finding role of i5 Health was further called on in the context of Social Prescribing for all of London (at the specific request of the Board of NHS England). i5 Health needs both SUS data and HES data on an annual basis to carry out its functions as each category on its own does not contain sufficient elements. SUS data contains costing which is essential for health economic analysis purposes but does not contain all diagnosis and procedure codes and procedure dates. HES does not contain all Payment by Results (PbR) related information contained in SUS though does contain all the necessary clinical information such as full medical history and treatments of a patients including procedure dates. Unfiltered historic data is essential to build full and accurate medical histories and monitor progression of diseases to achieve the most precise outcomes forecast that specific interventions can achieve (as demonstrated in the Evidence Based Interventions exercise being carried out by i5 Health for NHS England). The availability of seven years of longitudinal data improves data quality for the purposes of creating algorithms because pre-existing conditions or interventions that have been coded seven years ago might otherwise be lost to essential research. Increasing the provision of data from five years to seven years reduces the clinical risk of omitting pre-existing conditions and interventions and should improve efficacy of future treatment. In other words, i5 Health has determined seven years of data will be more evidential when creating algorithms for the purpose of future case finding and risk stratification. In an NHS England exercise carried out by i5 Health last year and continuing for the rest of this year on Evidence Based Interventions (EBI), the longevity of the records allows i5 Health to strengthen the theory that historical NHS data, if properly interrogated and learnt from, is rich enough to provide valuable supporting evidence for NICE guidance and clinicians addressing new patients’ needs. It is not intended that more than seven years data will be sought. i5 Health maintains a rolling seven full years of data and the oldest year is destroyed on receipt of the latest year. At the end of a retention period, i5 Health removes expired data and provides to NHS Digital appropriate destruction certificates. Data is retained only for so long as necessary for the purposes set out herein and as agreed with NHS customers and complies with data deletion requests. [1 paragraph unchanged] LEGITIMATE INTERESTS The number, richness and extent of the data sets NHS Digital currently provided to i5 Health are essential for the training of the organisation’s Neural Networks that are at the core of i5 Artificial Intelligence (AI) modules. What are being trained are prediction models that, when applied to data of patients, can forecast either the onset of a disease or the prognosis and likely outcomes of a disease when applied to patients’ data. The models use patients’ past medical histories to predict future outcomes and are trained on the outcomes of patients with similar medical histories and the outcomes they experienced. Once a model is trained, none of the original training data is retained within the model - only the cause and effects which can be applied to a single medical record for prediction. The training and underlying methodology are best illustrated in academic papers i5 Health produces for algorithmic tools. An example can be found on ReseachGate in respect of Atrial Fibrillation: ‘Artificial Neural Networks for Population Health Management to support Atrial Fibrillation Screening Programmes’ (https://www.researchgate.net/publication/325335156). The legitimate interest which i5 Health serves is the furtherance of patient care as carried out by the NHS. i5 Health provides the service solely at the request of third parties ( ie NHS England, Acute Trusts, CCGs, CSUs and STPs) for the benefit of the NHS and its patients. i5 Health believe the objective would also be considered as contributing to broader societal benefits. i5 Health has reviewed the requirement for the amount of data being supplied through this renewal and assessed the data requested as necessary for the purposes of their agreement. In carrying out the training of Neural Networks it is suboptimal to exclude or minimise data within the medical records being used for training. Doing so would detach the algorithms from reality and therefore make them less effective when applied to the data of actual patients whose care is being decided on. Because machine learning uses many input variables (features) for categorising patients for different treatment and risk bands, minimising the data reduces the categorisation ability for fields that are non-obvious. i5 has developed algorithms that interrogate, in various ways, secondary care data provided by NHS Digital. The results of those interrogations, sometimes when combined with other data (eg workforce, size of population…etc), contribute to I5 Health does not, as a rule, seek to target children’s data or that of any other vulnerable group. However, there are occasions when analysis of such is specifically requested by the NHS (e.g. researching the variations in childhood asthma within the Brent CCG area) and i5 Health could not fulfil that sort of need without having the relevant historical data. In addition, i5 Health believe not having the data would limit the efficacity of their advice on Population Health Management generally. Providing less than complete data would detach the algorithms from reality, introduce bias and discrimination and therefore make them less effective when applied to the data of actual patients whose care is being decided on. reports requested by and for NHS decision makers. These include reports on: LEGAL BASIS FOR PROCESSING Article 6(1)(f) - 'processing is necessary for the purposes of the legitimate interests ...' and Article 9(2)(j) 'processing is necessary for archiving purposes in the public interest, scientific or historical research purposes or statistical purposes...' i5 Health have undertaken a Legitimate Interests Assessment and concluded that they can rely on legitimate interests for this processing. i5 Health’s legitimate interests are a necessary and lawful basis for processing and controlling SUS and HES data, as they enable i5 to assist NHS customers with the access and use of i5 analytics products and services, and the development of the same. i5 considers access to its analytics to be in the public interest as well as of critical value to end-user NHS customers, so they can better understand their patient pools and make better health and social care decisions. Legitimate interests also include for i5 Health, as a commercial for-profit company, to continue as a provider of leading analytics in the UK, and as well to support compliance with the clinical and non-clinical performance expectations of the NHS. The purpose of i5 solutions includes, though not limited to, utilisation by NHS customers to produce baselines for innumerable outcome metrics derived from key data combinations, to support better use of resources, staff, and services, ultimately resulting in better health and social care outcomes. It is important to note that access to SUS and HES data by i5 Health’s NHS customers is only ever to aggregated data, with numbers suppressed in line with ICO standards, and pseudonymisation always occurring prior to transfer of SUS and HES data to i5 Health. i5 Health processing activities are a necessary, targeted, and proportionate means of achieving the purposes of the legitimate interests outlined above, and these interests are not overridden by moral or ethical issues, and are balanced against any impact on individual rights. i5 Health uses the least intrusive means possible to achieve the analytics it provides to i5 Health customers, strictly within the UK only. The risks of any data breach are clearly understood by i5 Health, which is why i5 Health requires SUS and HES data to be pseudonymised before it is provided to the company and does not require access to the encryption key in order to process SUS and HES information. The solutions that i5 Health delivers are limited to the NHS customers like CCGs, STPs, CSUs, hospital trusts, NHS England, care quality commission registered providers, public health departments, and similar health care providers within the UK. The objective for processing is to provide support for commissioning activities, operational and financial analytics, comparators and indicators, data quality validation, and other critical insights as requested and directed by NHS customers on the basis of the pseudonymised SUS and HES data. i5 Health provides the services solely for the benefit of the NHS and its patients. i5 Health has developed Artificial Intelligence (AI) algorithms that interrogate, in various ways, secondary care data provided by NHS Digital. The results of those interrogations, when combined with other data (e.g. workforce, size of population…etc), contribute to reports requested by and for NHS decision makers. These include reports on: [5 paragraphs unchanged] - and reports on the patient care and financial benefits of specific activities eg e.g. Non-Medical Prescribing, digital initiatives, etc. There will never be any requirement or attempt by i5 Health to re-identify individuals.

Processing activities

NHS Digital will provide i5Health i5 Health with one drop of the latest record level pseudo/anonymised pseudonymised SUS PbR and HES data via the Secure Electronic File Transfer (SEFT) system. A Database Analyst (DBA) from i5Health i5 Health will load the record level pseudonymised data into a database. The database will be managed locally and secured by the DBA with user access control. an i5 Health secure Record-level data will only be accessed by individuals within the Analytics Team, who have the authorisation from the Operations Director (who is also Caldicott Guardian), to access the data for the purpose (s) described, all of whom are substantive employees of i5 Health. database. The database will be managed locally and secured by the DBA with user access control. Data will only be accessed at the named processing location as set out in this application. Record-level pseudonymised data will only be accessed by individuals within the Analytics Team, who have the authorisation from the Operations Director (who is also Caldicott Guardian), to access the data for the purpose (s) described within this DSA, all of whom are substantive employees of i5 Health. The additional SUS PbR Data being provided will be linked to SUS PbR data already held, across the datasets (e.g. SUS PbR Episode data with SUS PbR A&E data); from National Level to GP Practice Level. There will be no requirement nor attempt to re-identify individuals within the datasets. National data is required as i5 Health provide reports from local through Regional to National levels. Multiple years of data are required in order to produce time-series and predictive modelling, historic data is retained to enable this. The data cannot be minimised by applying filters to specific conditions of relevance as the full data is needed in order to produce the outputs as outlined within the application. Data will only be accessed at the named processing location as set out in this Agreement. Data will not be made available to any third parties except in the form of aggregated outputs with small numbers suppressed in line with the HES Analysis Guide which covers suppression rules for SUS. The additional SUS PbR and HES data being provided will be linked to SUS PbR data already held, across the data sets (e.g. SUS PbR Episode data with SUS PbR A&E data); from National Level to GP Practice Level. There will be no requirement nor attempt to re-identify individuals within the data sets. National data is required as i5 Health provide reports from local through Regional to National levels. Multiple years of data are required in order to produce time-series and predictive modelling, historic data is retained to enable this. The data cannot be minimised by applying filters to specific conditions of relevance as the full data is needed in order to produce the outputs as outlined within this Agreement. For example, a report containing aggregate data for cohorts from localities to large geographical areas will be produced for NHS England. HES data itself has many diagnosis fields. This is critical for the development of i5 Health’s algorithms. A key case in point is the current major project for NHS England at Skipton House that i5 Health is centring. This is Evidence Based Interventions (EBI) referenced in the Yielded Benefits Purpose section under #2.1) Case Finding: “ Develop criteria of adverse outcomes based on treatment received to support the separation of patient cohorts that should or should not have received treatment. Such definitions may include readmissions, complications, further treatment needs and death. These criteria may also be used to develop further NICE guidance”. The breadth of the information needed for the EBI project is important because of the number of potential and different pathways of analysis. I5 Health need the maximum number of fields to permit clustering methods based on common co-morbidities used to investigate if commonalities exist amongst patient groups. The inclusion of Voluntary Sector Organisations (VSO) as recipients in no way changes the processing activities set out in this paragraph, but will be limited to aggregate data with small numbers suppressed in line with the HES analysis guide which covers suppression rules for SUS. . SUS PbR data provides a very different type of information. It is, by definition, costed data and grouped relating to finance – needing HIG codes and national tariffs. SUS PbR is the basis of i5 Health’s costing advice when it comes to using the i5 Commissioning Opportunities (i5 COP) set of algorithms that provide the value for CCGs of alternative (principally non-hospital) treatments. Processing specific to the i5 Coronavirus Health Risk Calculator: The pseudonymised medical records of patients admitted to hospital with Covid-19 will be analysed to establish the characteristics in their history that might be additional to/different from those identified in the earlier analysis of Influenza and historical (ie non-Covid-19) Coronavirus patients. The proposed Work Packages are as follows: WP1 - Collect/categorise data (15 days) • Three years of medical records with over 200 fields including 21 diagnosis and 21 procedures and any diagnosis present prior to Covid-19 • Categorise into balanced cohorts (Low, Medium, High and Very High) for optimal separation between Sensitivity and Specificity WP2 - Principal Component Analysis (PCA) (10 days) • PCA carried out on diagnosis codes to ensure only conditions used that are pertinent for prediction WP3 - Data split (15 days) • Create separate data sets for AI training, monitoring of overfitting and performance testing • Balancing sets to contain patients with various outcomes and avoid over-representation WP4 - Coding and Training (25 days) • Short term -- risk level prediction; Medium/Long term -- Healthcare Planning information • Train various AI models and topologies to evaluate which combinations works best WP5 - Test beds (15 days) • Testing of revised calculator with several Integrated Care Systems nominated by CSUs • Address issues and develop any additional functionality The number, richness and extent of the data sets NHS Digital currently provide to i5Health are essential for the training of the organisations Neural Networks that are at the core of i5 Artificial Intelligence (AI) modules. The result is the highest level of reliability when reports are created for commissioning and NHS England decision makers. i5Health has reviewed the requirement for the amount of data being supplied under this Agreement and assessed the data requested as necessary for the purposes of this Agreement. Data will not be made available to any third parties except in the form of aggregated outputs with small numbers suppressed in line with the HES Analysis Guide which also covers suppression rules for SUS. For example, a report containing aggregate data for cohorts from localities to large geographical areas will be produced for NHS England. The inclusion of Voluntary Sector Organisations (VSO) as recipients in no way changes the processing activities set out in this paragraph, but will be limited to aggregate data with small numbers suppressed in line with the HES analysis guide which covers suppression rules for SUS. [2 paragraphs unchanged] All organisations party to this agreement must comply with the Data Sharing Framework Contract requirements, including those regarding the use (and purposes of that use) by “Personnel” (as defined within the Data Sharing Framework Contract ie: employees, agents and contractors of the Data Recipient who may have access to that data. There will never be any requirement or attempt by i5 Health to re-identify individuals. The number, richness and extent of the data sets NHS Digital currently provide to i5Health are essential for the training of the organisations Neural Networks that are at the core of i5 Artificial Intelligence (AI) modules. The result is the highest level of reliability when reports are created for commissioning and NHS England decision makers. i5Health has reviewed the requirement for the amount of data being supplied through this renewal and assessed the data requested as necessary for the purposes of their agreement. All organisations party to this agreement must comply with the Data Sharing Framework Contract requirements, including those regarding the use (and purposes of that use) by “Personnel” (as defined within the Data Sharing Framework Contract - i.e. employees, agents and contractors of the Data Recipient who may have access to that data.

Expected output

[2 paragraphs unchanged] The outputs i5 Health Limited have provided over the last 24 months are (according to each purpose); [1 paragraph unchanged] • Initially first national report for Health Education Board of NHS England on the economic value of Non-Medical Prescribing (NMP) called on three categories of NHS Digital information (latest HES data in respect of long term health conditions (LTC); (LTC)); Nurses currently in the workforce, and Nurses using FP10 Prescription forms). Thereafter continued to update value of NMP to the country. The input from i5 Health formed the basis of a decision by NHS England to increase considerably the funding of NMP. It is expected that, at the request of Health Education England (HEE), i5 Health will provide during the year further updates, based on the NHS data, of the workforce, patient care and financial benefits continuing to arise from Non-Medical Prescribing. [1 paragraph unchanged] • A number of commissioning support reports are made for CCGs within the footprint of Business Intelligence partners, Arden & GEM CSU; this has been further facilitated by the installation within the server of Arden & GEM CSU of some of the i5 Health algorithms that are trained on the NHS data provided under this agreement. The work carried out at the request of Arden & GEM CSU within the NHS Milton Keynes CCG area on Commissioning Opportunities is expected to expand to other CCG clients of the CSU through to the end of the DSA contract year. NHS England has commissioned in 2018 i5 Health to provide commissioning reports for 32 CCGs and 5 Sustainability and Transformation Plans (STP) Partnerships (STPs) in London as well as analysis of the effect on London over [7 words unchanged] of Digital technology. i5 Health continues to analyse the data received in order order, on request, to update and add to the commissioning reports and the dashboards provided to the CCGs and STPs and the reports on the value of Digital technology. [1 paragraph unchanged] • As part of NHS England’s Electronic Palliative Care Co-ordination System (EPaCCS) programme, i5 Health Limited carried out an evaluation of data in respect of End of Life [8 words unchanged] project – the outcomes being dependent on changes in trends of data. The studies have been carried out at the specific requests of the various CCGs within England and Wales. Purpose #2.2) #2.4) Case Finding • In many parts of the country, a number of LTC patients are [82 words unchanged] This NHS England initiative is continuing. The case –finding role of i5 Health was further called on in the context of Social Prescribing for all of London (at the specific request of the Board of NHS England). i5 is currently working with NEL CSU to integrate this activity into the CSU’s portal, NELIE. Purpose #2.3) Urgent Care In respect of Arden & GEM CSU, it is expected that the Case Finding work carried out for its clients, Milton Keynes CCG, will be extended during the next to more of the 70 CCGs within the CSUs ‘footprint’. It is also expected that work started in Q1 2020 for the CSU’s clients, South Lincolnshire CCG, on an Internet of Things (IoT) initiative will continue through into 2021. In 2019, discussions started with NEL CSU, after the success of the Haringey CCG Atrial Fibrillation exercise, to integrate the Case Finding algorithms into the CSU’s server or at least partner with the CSU by providing an API link. The Coronavirus crisis has accelerated these considerations and NEL CSU is now benefitting from accessing the Risk Stratification service provided by the i5 Health Coronavirus Health Risk Calculator - up to 22 million uses through the DSA year (see explanation in 5 d iii) below). I5 Health has received funds from Innovate UK to commence the Calculator Work Packages on 15th July 2020. The 90 FTE days allocated would give an end date of the middle of October 2020 but i5 Health will endeavour to accelerate the process as time is of the essence. • Halton CCG needed to map out the Urgent Care pressure across all the supporting Hospitals. Linked with this exercise, they commissioned i5 Health, using the Commissioning Opportunity module (COP), to investigate the patient urgent care journey. (i5 Health Limited applied the same skills for the benefit of other North West CCGs including South Sefton CCG and Southport and Formby CCG and for Arden & GEM CSU). Purpose #2.3) Readmissions • Halton CCG asked for the assistance of i5 Health in analysing significant Readmissions issues. The analysis, based on the COP algorithms, got right to the heart of the problem and identified a significant number of patients that, under normal circumstances, should not have undergone readmission. This activity carried out by i5 has now formed part of the offering through all CSUs within the Elis group. Purpose #2.3) Outpatient Procedures • On behalf of Halton CCG, More recently, i5 Health performed analysis of into what has been happening in respect of Outpatient episodes and then developed some solutions to excessive use in Cardiology, Mouth/Head/Neck & Ears, Orthopaedic Non-Trauma and Urology. Purpose #2.3) ACS – Respiratory and Ear, Nose, Throat (ENT) • i5 Health established what might, currently, be the best opportunity for Halton CCG to reduce acute care activity and cost - dealing with Respiratory and Ear Nose Throat conditions. Besides analysing historic and current situation, i5 Health examined six case studies to establish, using a Population Health Management approach, what might be optimum strategies for to pursue (the product of this work is now being leveraged into the Case Finding activity). [1 paragraph unchanged]

Expected measurable benefits

[1 paragraph unchanged] The immediate benefits of the NMP analysis carried out by i5 Health were the increase in investment of £500,000 by the NHS in nurse prescribing training. More nurses who are able to prescribe will accelerate patient treatment, freeing up doctors to spend more hands on time with patients who need to be seen by a doctor. Activities carried out with regards to NMP analysis: - Global research of NMP experience - National research through personal interviews and group sessions - Data analysis and modelling to establish: - Local Impact - National Impact - Potential cost savings - Activity and cost outcomes [3 paragraphs unchanged] The financial benefit for the healthcare system varies from case to case. [12 words unchanged] economy two years ago resulted in a 10% reduction in Non-Elective admissions (NEL) thus saving hundreds of thousands of pounds annually. (NEL). [6 paragraphs unchanged] The Coronavirus Calculator tool is intended to benefit: • Individuals - to adjust their lifestyle and minimise the risk of infection. • Clinicians and Hospital Management for prioritising care, improving bed management and ensuring right facilities in ICUs • Clinicians for telephone assessment or remote consultations • Public Health authorities and governments, for planning for disease control, levels of quarantine and targeted shielding There are two particular points about this exercise that relate to Long Term Conditions: - The Calculator risk assessments can lead to persons, clinicians and government authorities taking more informed decisions that could reduce the effect of the condition Covid-19 itself (eg the screening exercise carried out for London) - The focus given by the NHS to Covid-19 during this pandemic has a negative effect on the treatment of Long Term Conditions in general; this might be because of delays in screening or of treatment itself. The Calculator can contribute to authorities carrying out more selective screening of populations which means more optimal use of resources and better availability of resources for other Long Term Conditions. Identification of patients that have (or risk having) an Long Term Condition (LTC) but do not appear on the GP's risk register can lead to better management of their health requirements Evidence Based Interventions (EBI) work which will continue into 2021: Consequent on delivery of Preliminary EBI Evidence Pack, working with GIRFT Clinical Coding Team, Expert Working Groups, Clinical Classifiers , Clinical Leads and Price Assessors to enrich the material in Preliminary EBI Evidence Pack. Continue working on answers to following two key questions : Question 1: 'Can administrative data help to select inappropriate procedures for consideration for the EBI programme?' Essentially this is about establishing, from longitudinal patient records, what criteria is present that presents an unacceptable outcome for a patient undergoing a specific procedure. Question 2: 'Can administrative data help us to clarify the criteria for intervention to contribute to our guideline development?' Develop criteria of adverse outcomes based on treatment received to support the separation of patient cohorts that should or should not have received treatment. Such definitions may include readmissions, complications, further treatment needs and death. These criteria may also be used to develop further NICE guidance.

Benefits reported

- For all of London, there is now an on-line process that allows each of the 32 CCGs and 5 STPs to establish the value of introducing of Social Prescribing within their areas. - Likewise for each area of London, the value of introducing 48 Digital initiatives has been made available on-line. Current estimates indicate a circa £800m annual value will accrue in five years time and over £3 billion in ten years time - CCGs that have received the i5 Commissioning Opportunity (COP) reports have been able to introduce the findings into their planning processes - Patients that, with the cooperation of NHS Digital, were able to be recognised as 'missing' from the LTC registers or, if not treated, would suffer from an LTC continue to be contacted by their GPs for screening purposes Benefits achieved; Purpose #1) Non-Medical Prescribing (NMP) • With increasing pressure on the availability of doctors in both primary and secondary care, there is a growing case for greater use of NMP i.e. prescribing by a non-doctor (e.g. nurse, pharmacist, etc..). That case is reinforced by the cost/benefit identified and the better levels of care demonstrated. Each year NHS Digital provides i5 with fresh data, i5 are able to update their estimates of the benefits of NMP. Those are communicated to NHS England (Education), NMP organisers and NMP practitioners during the year either in meetings or conferences. Such communication has now extended to Scotland. As noted in benefits, the increased funding of the education of NHS clinical practitioners has been a significant consequence of the i5 analysis. More specifically, i5 expect to provide further input of this nature in Q3 2018. Purpose #2.1) End of Life • The HES based studies so far are showing a disturbing picture particularly, though not exclusively, in respect of the frail and elderly in their last year of life. The concerns are around the high levels of admissions to hospitals and the distress to patients this causes. Further work needs to be done for some CCGs on identifying, with business cases, the alternative strategies that can answer the above concerns. June 2017 – Reports, at the request of NHS England, for all CCGs in London on consequences of investment in Digital initiatives – specifically in the context of End-of-Life plans [1 paragraph unchanged] • Identification of patients that have (or risk having) an Long Term Condition (LTC) but do not appear on Coronavirus the GPs risk register can lead to better management of their health requirements. i5 Health has created for the NHS and for individuals a Calculator that establishes the Health Risk of Coronavirus (accessible at https://coronavirusrisk.org). It uses past medical history of people to predict their risk levels if infected by the virus 2019-nCoV leading to Covid-19. It categorises people into low, medium, high and very high risk in order to facilitate shielding and protection strategies for escalation and de-escalation, to support hospitals in their management – particularly of ICUs – and to provide information for decision making by clinicians and individuals. i5 are right in the midst of an exercise with NEL CSU and North London CCGs, which started in Q4 2017, North East London (NEL) CSU has selected the medical records of circa 4,100,000 Londoners and successfully processed them through the i5 Health system in an exercise to identify those most at risk of Coronavirus. As data starts to come in about the new strain Covid-19, i5 Health will upgrade the Calculator accordingly and work with NEL CSU and Arden & GEM CSU to test the upgraded system. on the application of i5 case finding algorithms to address Atrial Fibrillation. As already noted, i5 NN NEL CSU has used the i5 Coronavirus Health Risk Calculator to provide the following: algorithms are updated and refined each time data is received from NHS Digital. • Mapping Covid 19 – the calculator enabled NEL CSU to create heat map of South West London to the establish the districts of greatest risks. Purpose #2.3) Urgent Care • SHIELD – the calculator supported NEL CSU in the SHIELD List Segmentation exercise as requested by the government. • Commissioning Opportunity (COP) is all about matching patient groups against successful healthcare • Specific Patient Targeting – GP’s who were provided re-identified lists by NHS Digital of own patients with their risk scores. initiatives and forecasting the effect of local implementation on patient care and budgets. By enabling NEL CSU to support the local health authorities in targeting efforts on specific districts at greater risk and deciding on increased shielding activities, with the potential to drill down to individual patients, it will allow the GPs to make decisions about their patients using the risk score calculated using the i5 Coronavirus Health Risk Calculator. Over the period, Arden & GEM CSU have been increasingly working closely with i5 to create a system Diagnosis Stratification whereby the reporting capabilities of i5 modules can be enriched for the provision of reports for the - i5 Health have finalised an exercise with NEL CSU and NHS Haringey CCG which started in Quarter 4 of 2017 and led to the successful application of i5 Health case finding algorithms to the medical records of the population of the NHS Haringey CCG area in respect of Atrial Fibrillation. The results reported in Quarter 2 of 2019 confirmed the power of the tool. CSU’s customers. As reported to the Board of NHS Digital on 5th June 2018 - to that end, we have - i5 Health co-operated, in periods Quarter 2 to Quarter 4 of 2020 with Arden & GEM CSU and NHS Milton Keynes CCG on the application of the Case Finding Tool to the population of the NHS Milton Keynes CCG area in respect of: Hypertension, Atrial Fibrillation, Heart Failure, Diabetes Mellitus, Cancer, Asthma, COPD, Coronary Heart Disease, Palliative Care, Dementia, Depression (18+), Obesity (16+), Epilepsy (18+), Chronic Kidney Disease (16+), Stroke and TIA, Rheumatoid Arthritis (16+), Peripheral Arterial Disease, Osteoporosis (50+). commenced work with Arden & GEM on integrating i5 algorithms into the CSU’s server, GEMIMA (latest Social Prescribing meeting in w/o 11th June 2018). Work of this nature, with constant reference to the data received by i5 As noted last year, for all of London there is now an on-line process that allows each of the 32 CCGs and 5 STPs to establish the value of introducing of Social Prescribing within their areas. CCGs accessing the site have been able to introduce the findings into their planning processes during the year 2019/20. Patients that, with the cooperation of NHS Digital, were able to be recognised as 'missing' from the LTC registers or, if not treated, would suffer from an LTC continue to be contacted by their GPs for screening purposes, i5 Health advised at several Social Prescribing events during the year. Building on the above work, i5 Health carried out, in Quarter 3 of 2019, preliminary analysis in respect of the population within the NHS West Kent CCG area on Leg Ulcers and Diabetes and developed a proposal to be actioned post the NHS Kent and Medway CCGs merger by mid-2020 to the entire region. Health from NHS Digital, will progress throughout the year 2018/19. A proposal was made in Quarter 2 of 2019 to Public Health England to apply the Social Prescribing algorithms to the rest of the South of England which proposal is still awaiting budget approval in 2020. June 2017 – Report, at request of NHS England, for all CCGs in London on consequences of investment in Evidence Based Interventions (EBI). Digital initiatives – specifically in the context of Urgent Care In Quarter 1 of 2019, the Department of Data and Analytics at NHS England and NHS Improvement commissioned i5 Health to use its knowledge and understanding in Case Finding combined with other of its analytical skills to centre a study on Evidence Based Interventions (EBI). Purpose #2.3) Readmissions EBI work done during 2019 comprised: • Issues relating to specific surgeons were highlighted. i5 Health went beyond problem identification and a) Discovery Phase evaluation and made detailed and well considered recommendations – not just for the CCG but also its - Team approach: top down (Clinical) and bottom up (Data Analysis) with support from Leads experts providers and colleagues across primary and community care. - A focus on CCG/Provider areas for numbers of elective interventions; establishing clusters of OPCS Classification of June 2017 – Report, at request of NHS England, for all CCGs in London on consequences of investment in - Interventions and Procedures to check variability; increasing the criteria to include patient history and outcomes. Digital initiatives – specifically in the context of Readmissions b) Review Phase Purpose #2.3) Outpatients – 17 Original Procedures revisited by i5 Health and adjusted in the light of findings from Discovery Phase. • The i5 Health solutions included alternatives for 6,000 procedures currently costing over £1.5m. The c) 45 Additional Procedures Phase - Research and Identification on: solutions are being implemented and will result in significant cost savings. - Using i5 Health's AI tools to find and address 45 additional procedures of limited value. June 2017 – Report, at request of NHS England, for all CCGs in London on consequences of investment in - Prioritising them using various criteria including the most challenging in terms of costs, clinicians’ concerns, available evidence. Digital initiatives – specifically in the context of Outpatients - Reviewing data requirements for coding of clinical criteria for the 45 procedures. Purpose #2.3) ACS - Respiratory and ENT - Production of Preliminary EBI Evidence Pack for these procedures. • One of the many positive outcomes of the exercise has been the identification of over 200 patients clinically diagnosed with COPD in secondary care that are not on the GPs risk register and which are likely to be unmanaged. June 2017 – Report, at request of NHS England, for all CCGs in London on consequences of investment in Digital initiatives – specifically in the context of COPD