NHS Digital Data Release Register - reformatted

Telstra Health Uk Limited projects

Dr Foster Standard Extract Service Feed
Dr Foster direct feed
Summary Hospital-level Mortality Indicator (SHMI) data

236 data files in total were disseminated unsafely (information about files used safely is missing for TRE/"system access" projects).

Dr Foster Standard Extract Service Feed — DARS-NIC-68697-R6F1T

Type of data: information not disclosed for TRE projects

Opt outs honoured: Anonymised - ICO Code Compliant (Does not include the flow of confidential data)

Legal basis: Health and Social Care Act 2012 s261(1) and s261(2)(b)(ii), Health and Social Care Act 2012 - s261 - 'Other dissemination of information', Health and Social Care Act 2012 s261(2)(b)(ii)

Purposes: Yes (Supplier)

Sensitive: Sensitive, and Non-Sensitive

When:DSA runs 2019-09-01 — 2020-08-31

Access method: Ongoing

Data-controller type: TELSTRA HEALTH UK LIMITED

Sublicensing allowed: No

Datasets:

Civil Registration (Deaths) - Secondary Care Cut
HES:Civil Registration (Deaths) bridge
Hospital Episode Statistics Accident and Emergency
Hospital Episode Statistics Admitted Patient Care
Hospital Episode Statistics Critical Care
Hospital Episode Statistics Outpatients
Emergency Care Data Set (ECDS)
Civil Registrations of Death - Secondary Care Cut
Hospital Episode Statistics Accident and Emergency (HES A and E)
Hospital Episode Statistics Admitted Patient Care (HES APC)
Hospital Episode Statistics Critical Care (HES Critical Care)
Hospital Episode Statistics Outpatients (HES OP)

Objectives:

Data will be processed by Dr Foster under the GDPR lawful basis Legitimate Interest - Article 6(1)(f). Dr Foster has a legitimate interest in being able to provide tools and services that healthcare organisations will find useful and that will benefit the health and social care system. Without processing this data Dr Foster would not be able to deliver these tools and services, to the detriment of health professionals who use the products. Dr Foster's customers can be assured that the tools and services are based on evidence provided by data from a trusted source; there is no viable alternative as relying on public domain data would lead to gaps and not allow the same support for decision making. Using pseudonymised data allows Dr Foster to deliver evidence-based insight and analysis while minimising intrusion into a patient's privacy. If patients are uncomfortable with their data being used for purposes beyond their care then they may manage their patient data choices via the National Data Opt-Out.

Dr Foster Limited requires the pseudonymised HES datasets to help healthcare organisations achieve sustainable improvements in their performance, to gain insight and to inform decision making. This is delivered in three main strands:

Dr Foster tools and services to provide management information, analysis and clinical benchmarking through online products and services
Bespoke analytics to deliver customised projects to meet individual customer needs
Research for publication to provide thought leadership in the field of healthcare data analytics, with the aim of improving the planning, delivery and outcome of health and social care

The data being requested under this agreement flow through to Dr Foster via Imperial College London Dr Foster Unit (ICL DFU). Dr Foster Limited is a completely different legal entity from ICL DFU, Dr Foster Limited do not have any control over data held by ICL DFU.

Dr Foster funds ICL DFU and works closely with it on developing new methodologies to assist healthcare improvement. The development of new methodologies is achieved through the sharing of knowledge.

HES data is sent by NHS Digital to ICL DFU. ICL DFU further pseudonymise the data (described in more detail in the processing section of the agreement) and then transfer it securely to Dr Foster. Dr Foster cannot identify individuals in this data but the additional pseudonymisation process means that authorised Dr Foster NHS customers can take an additional service through ICL DFU to identify patients in Dr Foster tools. Dr Foster cannot identify individual patients in the data and does not have access to the re-identification service.

ICL DFU offer an independent service to Dr Fosters NHS customers which allows the customers to re-identify patients in Dr Foster tools. NHS customers with access to ICL DFUs re-identification service use it to support many processes critical to delivering high quality health care services, for example:
Mortality review
Case note review
Clinical coding review
Pathway analysis and design
Patient safety analysis
Clinician and specialty bench marking review

Dr Foster online products are used by:
NHS Provider Trusts Subscribed authorised users in customer organisations can view data that relate to their organisation at a record level. They cannot access record level HES data relating to other organisations.
Other NHS organisations Subscribed authorised users in customer organisations can view aggregated analysis which provides valuable insight but prevents any patients from being identified, in accordance with guidance provided by NHS Digital.
Care Quality Commission CQC can view aggregated analysis.

The Analytics team provides aggregate level and small number suppressed analysis and insight to a number of NHS customers including:
NHS Trusts
Clinical Commissioning Groups
Commissioning Support Units
Department of Health
NHS England
NHS Improvement
Care Quality Commission
Public Health England
National Institute for Health and Care Excellence

Dr Foster plan to also work with:
Non-NHS organisations providing services to benefit the NHS - It is proposed that these will be supplied with aggregate, small-number suppressed analyses.
Non-NHS organisations to benefit public health and social care - It is proposed that these will be supplied with aggregate, small-number suppressed analyses where benefits can be identified for the health and social care system. It is also proposed that algorithms or coefficients that have been derived through research on HES data may be provided directly to a customer for implementation on their own local data. Data provided in all outputs will be at an aggregate level and small number suppression will be implemented in line with HES analysis guidelines.

At present no analyses have been provided to non-NHS organisations. Any request for such analysis will be reviewed to determine if it benefits the health and social care system. Dr Foster will inform NHS Digital of analysis it provides to non-NHS organisations and will list this in any renewal or amendment to this agreement.

The Analytics team provides bespoke analytics and data science tailored to specific needs to identify clinical variation, efficiency savings, predict patient risk and improve patient outcomes.

They are a skilled team of experts in advanced healthcare analytics and data science including predicative analytics, machine learning techniques and advanced statistical methods. The team use this expertise to investigate issues and transform healthcare services.

Customers have access to Dr Fosters team of qualified data scientists, clinicians, statisticians, mathematicians, and economists. They supplement in-house analytical teams with Dr Fosters expert advice and guidance linking, modelling and visualising data and insight.

Dr Fosters aim is to help improve health and social care decision making and planning and, ultimately, to benefit patients.

Dr Foster use the data provided under this agreement to provide a management information function in the form of analysis and clinical benchmarking for healthcare organisations and to increase the power of predictive models for rare diseases, procedures and events. Dr Foster build standard casemix adjustment models for 259 diagnosis groups and 200 procedure groups which include some rarer conditions. Using all pseudonymised HES datasets means that Dr Foster have the most up to date information and can inform customers of potential issues around quality and in turn they can make better informed decisions for the improvement of healthcare and outcomes for patients.

At a high level Dr Foster analyses break down into the following:
Quality measures of healthcare services by providers/area/clinical interest/trend analysis
Variations in health outcomes
Health inequalities and needs analysis
Predictions
Performance data and changes in clinical practice
Management information
Efficiency Monitoring
Benchmarking
Contract Management and Variance Analysis
Activity Monitoring
National Target Performance
Pathway design, redesign and improvement.
Practice Performance Monitoring
Capacity and utilisation management
Cross checking of commissioning data
Systems to support and monitor the pattern of healthcare usage
Patient segmentation analysis
Overall data quality

NHS subscribers to the tools have access to analysis of HES data so that they can:
Track and trend performance, identify areas for efficiency savings and understand and influence demand and patient flow throughout the health and care system.
Investigate risk-adjusted quality, patient safety and clinical outcomes data including mortality, benchmark against other healthcare organisations and identify areas for improvement.

Dr Foster has provided objective insight and analysis since 1999.

Dr Foster requires all HES datasets in pseudonymised form to provide a wide array of relevant indicators to give end users as complete a picture of hospital performance as possible to allow health and social care organisations to effectively:
Monitor quality of services provided
Identify efficiency opportunities
Identify pathways where services can be improved for the benefit of patients

A data period of 15 years of historical data is essential to enable Dr Foster to:
Obtain longitudinal data on prior admissions for patients. Risk modelling will also require access to variables on prior admissions including previously recorded co-morbidities.
Create, update and maintain statistical risk models to enable the regular production of risk adjusted measures of mortality, quality and efficiency (including Hospital Standardised Mortality Ratio (HSMR) and CUSUM alerts as used by NHS organisations and regulators).

HSMR is the ratio of observed deaths that occurred following admission in a provider to a modelled expectation of deaths (multiplied by 100) on the basis of the average England death rates for 56 specific clinical groups given a selected set of patient characteristics for those treated there.

CUSUM is short for cumulative sum. The charts show the cumulative sum of the differences between expected outcomes and actual outcomes over a series of patients. The total difference is recalculated for each new patient and plotted on a chart cumulatively (i.e. where one patients difference ends the next one starts). They are used to detect small sustained increases in risk relating to quality outcomes, such as mortality, long length of stay and readmissions and usually for a single provider and diagnosis/procedure group. When a number of patients have a negative outcome one after the other, the CUSUM could hit a threshold, triggering an alert. This could indicate that the situation is out of control for that particular strata.

In simplistic terms for the example of mortality, every time a patient dies the graph goes up, every time a patient survives it goes down. The amount that it goes up or down depends on the risk, so if a low risk patient dies it will go up more than if a high risk patient dies.

The CUSUM technique is associated with false positive and false negative states. A false positive is when the CUSUM gives an alarm when in actual fact performance is at an acceptable level and the rate of the poor outcomes has not changed.

It is necessary for reasons of public interest in the area of public health, in particular to ensure high standards of quality and safety of health care. Although Dr Foster are not a public body they provide services to help public healthcare organisations monitor and improve their services. Processing is designed to benefit patients and society as a whole through better healthcare.

Dr Foster require the complete pseudonymised HES datasets including Civil registration mortality data to help healthcare organisations achieve sustainable improvements in their performance, to gain insight and to inform decision making.

Civil Registration (Deaths) - Secondary Care Cut data are requested to provide more timely and accurate analysis and insight for Dr Fosters customers. It is essential for performing survival analyses and so represents a very valuable source of data. It will improve the output of Dr Fosters products for the benefit of NHS customers and for the broader improvement of health and social care for the public.

These data are extremely critical because mortality information may be a surrogate metric for success of medical care. Therefore this dataset will enable identification of factors that drive successful treatment of a patient. Cause and date of death may also be used to identify trends in causes of death in particular groups of patients. Historical death data are necessary for identifying trends.

The capability to link the already held HES datasets with mortality data could provide valuable insights into how and why some patients with the same condition and at the same stage can have very different outcomes. In addition it would be used to:
Compare hospital mortality rates for in-hospital deaths with rates for all deaths to evaluate the effect of differential discharge policies
Calculate total post-operative mortality rates, e.g. when comparing operative techniques such as laparoscopy and open approaches
Assess potential quality of care issues by comparing the cause of death with the reason(s) for admission, e.g. for surgical patients who are discharged within 30 days of the procedure but who die at home and whether the death is related to their disease process or to complications of treatment
Develop and validate indicators of quality and safety of healthcare, particularly by consultant and hospital
Show variations in performance by unit and socio demographic stratum
Predict risk and adjust risk of indicators and variations and any other methodological aspects as they arise
Establish seasonal patterns of mortality
Supports organisations in delivering their Learning from Deaths agenda and timely mortality reviews
Help organisations improve quality of care and identify where they could do more to help patients and their families

30 day mortality (both in and out of hospital) is a well published and accepted standard for comparing post-operative and post-admission hospital mortality. Having the linked death data will allow us to provide this outcome, which will improve engagement with clinicians, and allow comparisons with other published analyses.

Dr Foster process the minimum data necessary to meet the purpose and build privacy into their designs, starting with the prior pseudonymisation, consisting solely of the replacement of HESID with FOSID, of the data. Dr Foster ISO 27001 certificate provides assurance over the security of the data

Dr Foster have no requirement to re-identify the individuals within the data they receive and will make no attempt to carry out any re-identification.

In the context of the above statement, while Dr Foster and NHS Digital recognise that the inclusion of record-level Date of Death linked to all four HES datasets will theoretically increase the risk of re-identification, for clarity it is pointed out that:
a) as stated above, Dr Foster will not re-identify; in addition, given the volumes of data, identification serves no purpose to Dr Foster anyway
b) raw Date of Death allows Dr Foster to carry out better analyses and provide a better service to their customers:
- using an alternative such as death 30/60/90 days from Discharge does not enable the calculation of median survival rates, nor does it enable Dr Foster to produce aggregated survival curves
- in some cases, Dr Foster calculates mortality rates with reference to particular Procedures not to Discharge; the above flag would not support this as day 0 would need to be set at different events in different cases for different purposes
c) record-level data is only shared by Dr Foster with customer Trusts only for those patients treated by that Trust
d) those Trusts already have a means of re-identifying those patients: the re-identification service provided by the Dr Foster Unit at Imperial College

Yielded Benefits:

Dr Foster measure benefits through customer feedback for their products and services: Clinicians and analysts use Healthcare Intelligence Portal (HIP) on a daily basis to analyse new patient safety alerts, high standardised mortality ratios and individual cases. Dr Foster investigate areas of concern in clinical coding and data analysis and look to improve future patient care through retrospective review and analysis. Northampton General Hospital NHS Trust (February 2015) Our data quality team use HIP to identify patients with missing or duplicate information and whether the problem is ongoing and needs a change in process to rectify it or whether its due to individual oversight. Northern Devon Healthcare NHS Trust (February 2015) We carried out a review of dermatological deaths and the data in HIP identified that the deaths were due to cellulitis. We reviewed treatment options based on this and identified where improvements could be made. Wrightington, Wigan and Leigh NHS Foundation Trust (February 2015) We use HIP to undertake many audits and reviews at any one time. For example, ten sets of case notes are reviewed each month by an emergency department consultant as part of our programme to improve performance on sepsis. The Royal Bournemouth and Christchurch Hospitals NHS Foundation Trust (February 2015) We find Dr Fosters combination of knowledgeable experts and powerful tools enormously helpful in our work to improve the quality of care we are providing to our patients. With Dr Fosters help weve made significant progress in understanding quality and its drivers, and identifying how we can make sustainable improvements in our hospitals. Dr Fosters insightful analysis, practical recommendations and ongoing support help us extract maximum value from our data, and their impact is far-reaching. Dr Jeremy Rushmer, Medical Director, North Cumbria University Hospitals NHS Trust (2016) Case study Lancashire Teaching Hospitals NHS Trust (January 2015) HIP is used on at least a weekly basis by the corporate and business intelligence teams and clinical staff to: inform and direct the Trusts mortality and morbidity review processes scrutinise care standards and their impact on patient outcomes provide analysis and reassurance to the board, governors and the public monitor trends in readmissions and complications and investigate if these were justified clinically Instigated several quality improvement initiatives including: Improved documentation of complexity in perinatal conditions that has: o reduced mortality ratios o increased income o engaged clinicians in a wider quality initiative introducing an enhanced model of care for potentially vulnerable babies Development of an improvement programme for patients suffering from chronic obstructive pulmonary disease across the whole care pathway in the local health economy Case study University Hospital of South Manchester NHS Foundation Trust (January 2015) The Trust specialises in cardiac surgery activity, for which it is a tertiary centre, and performs a high number of coronary artery bypass grafts (CABG) and heart valve replacements. Dr Fosters Practice and Provider Monitor enabled benchmarking of productivity and efficiency measures, giving the user the ability to compare mean-price-per-spell at both procedure and diagnosis HRG level. The Trust used the Dr Foster tool to look more deeply at other influences on the efficiency of the pathway compared with others and highlighted that the length-of-stay for these procedures was one of the longest of its peer group. This clearly has an impact on income as the amount earned per bed day is lower, and the capacity to put more patients through the system is reduced. From a patients point of view, this is also good news: a longer length-of-stay may increase risk. UHSM then used Practice and Provider Monitor to move through the specialties to highlight areas of variance and focus on where they could improve productivity and efficiency across the Trust.

Expected Benefits:

Expected benefits are:
Enabling NHS acute trusts to measure, compare and benchmark key quality indicator trends focusing on risk adjusted measures of mortality, readmissions and length of stay in hospital.
Providing evidence to instigate clinical audit and investigations related to quality of care, such as highlighting potential poor clinical coding or quality/efficiency concerns.
Validating other mortality indicators such as HSMR, CUSUM alerts and crude mortality.
Enabling NHS acute trusts and commissioners to use performance information to identify, quantify and act on opportunities to improve efficiency of health services.
Understanding areas of best practice amongst Dr Foster customers and facilitate interactions with other customers who are not performing as well to support quality and efficiency improvement.
Helping clinicians and managers by providing independent and authoritative analysis of the variations that exist in acute hospital care in a way that is meaningful for them and that is understandable to patients and the public.
Highlighting topics of interest to the health industry and wider public to enable discussion and improvement in healthcare provision.
Publication of articles around variations of healthcare within the NHS is in the public interest and supports the government agenda for transparency by promoting choice and accountability within the NHS.
Maintaining the focus of the organisations on improvement.
Raising public and professional awareness
Providing valuable insights into how and why some patients with the same condition and at the same stage can have very different outcomes
Supporting organisations in delivering their Learning from Deaths agenda and timely mortality reviews, for those patient 30 day discharge
Helping providers improve quality of care and identify where they could do more to help patients and their families

How these benefits will be measured
Benefits are ongoing as the outputs described above are used within NHS Trusts internal monthly reporting and quality processes. Dr Foster services allow performance of NHS Provider Trusts to be monitored and trended over time and therefore provide customers with the ability to measure changes in quality and performance particularly in instances where customers have been alerted and they have worked with them to understand the causes of worse than expected performance.

When these will be achieved
As a majority of benefits are achieved on an ongoing basis, it is not possible to outline a specific target date for achievement of the benefits outlined as they are reliant on a range of factors outside of Dr Foster's control. However, whenever there are areas of particular concern about performance against key indicators, Dr Foster act immediately to alert relevant stakeholders and offer assistance in better understanding and addressing them.

Outputs:

Outputs are delivered through:

Dr Foster online tools and services including the Healthcare Intelligence Portal
Bespoke analytics
Research for publication

Specific outputs include benchmarked or standardised healthcare indicators and analysis such as mortality (Summary Hospital-level Mortality Indicator (SHMI)/HSMR), LOS (Length of Stay), admission trends, readmission rates, patient safety indicators, referral patterns, market share analysis etc.

Outputs will be used by customers to investigate clinical quality, performance and business development, specifically:
Assess and manage clinical quality and patient safety within NHS Organisations
Identify pathways where there is potential for improvement
Identify areas of best practice either within the Provider Trust or local/national health economies
Better understand how they compare to other Provider Trusts with similar case mixes
Identify improvements in operational efficiency
Understand patient outcomes
Identify and understand market activity
Monitor the impact of implemented changes
Identify variations in outcomes

The above outputs depend on processing of HES and Civil Registration data.

Specific outputs dependent on the processing of Civil Registration mortality data are:
Analysis of cause of death
Analysis of death following discharge, 7, 14, 30 days
Comparative analysis of cause of death and deaths following discharge
Development of outputs to further help users understand patient outcomes through analysis of survival rates
Analysis of variation in mortality across geographical boundaries
Support customers with out of hospital mortality queries
Additional level of insight for customers to investigate the care pathway for their patients

Subscribers to the tools have continual access which allows them to meet their own internal target dates.
Outputs of bespoke analytics projects are dependent on the nature of the project and can include Tabulations, Dashboards, Reports, Spreadsheets, Presentations or Articles. Outputs may be surfaced through a number of tools including Microsoft Office suite (Excel, Word and PowerPoint etc) or other tools (Tableau, Qlickview) depending on the requirements of the customer. In some instances, algorithms or coefficients that have been derived through research on HES data may be provided directly to a customer for implementation on their own local data. Data provided in all outputs are at an aggregate level and small number suppression is implemented in line with HES analysis guidelines. Bespoke analytics projects are conducted on an ad hoc basis and target dates for delivery of outputs are thus defined upon commencement of each project.

Examples of previous publications produced by Dr Foster include the Hospital Guide that published analysis of the variations in acute hospital care for the benefit of healthcare professionals, patients and the public.

Dr Foster are aware that publications, whether inside or outside the NHS, must adhere to strict guidelines in terms of disclosure, and ensure that any such publications are aggregated and comply with small number suppression in line with the HES Analysis Guide and other relevant legislation. Analyses for use in publications can be in the form of text, tables, or other data visualisations such as diagrams/graphs using aggregate data. Publications will also meet standards as defined in the Terms and Conditions of the Data Sharing Agreement.

All outputs will contain only data that is aggregated with small numbers suppressed in line with the HES Analysis Guide.

Processing:

The data requested under this agreement are initially sent to Imperial College London Dr Foster Unit (ICL DFU) under a separate agreement NIC-12828. ICL DFU pseudonymise the data and then transfer it securely to Dr Foster. Imperial College London Dr Foster Unit (ICL DFU) receives data monthly from NHS Digital under agreement NIC-12828. This includes de-identified HES data used for research by ICL DFU which is also processed with further pseudonymisation for secure transfer to Dr Foster. The HES data will be linked to the civil registration data before being passed onto Dr Foster.

ICL DFUs additional pseudonymisation of the de-identified HES data removes the HESID supplied by NHS Digital and replaces it with a new encrypted value that cannot be re-identified by Dr Foster. A unique identifier, known as the FOSID, is added by ICL DFU to each row of data.

The FOSID is shared with Dr Foster. The FOSID allows Dr Foster to process data without being able to identify patients while allowing authorised NHS customers to use it with ICL DFUs re-identification service. The re-identification service allows authorised users at care providers to further investigate patients under their care. The FOSID is used to extract the NHSNO or LOPATID so the care provider can review patients records. Dr Foster do not have access to this re-identification service and cannot see any re-identified data this data all remains with ICL DFU under NIC12828.

The FOSID allows authorised individuals within Provider Trusts to identify their own patients indicated in Dr Fosters healthcare performance tools at an episode level. If the data flowed directly from NHS Digital to Dr Foster in pseudonymised form it would not have these ICL DFU generated FOSIDs, which are ultimately used for authorised NHS customers to re-identify patients in Dr Foster tools without the requirement for Dr Foster to process identifiable HES. The re-identification service allows ICL DFU to supply NHS provider trusts with NHS Number and LOPATID using Dr Foster healthcare performance tools without passing these identifiers on to Dr Foster. No patient identifiers will ever be passed to Dr Foster or any other organisation except the NHS provider trust from where the data originated.

The processing of the pseudonymised HES data is within the scope of the ISO 27001 certified Information Security Management System (ISMS). Dr Foster processes are subject to internal and external audit. All staff receive initial information governance training followed by mandatory monthly modules.

The pseudonymised HES data are held securely on Dr Foster systems with access permissions only granted as necessary to specific roles.

Descriptions of processing through Dr Foster tools, analytics and research services are provided below.

Tools
HES is processed through products in the Dr Foster toolkit to provide:
Linkage into spells and superspells, which can often span across financial years
HRG, Tariff and other PBR related fields, using the HRG Grouper software
Various clinical groupings, including CCS Diagnoses, Ambulatory Care Sensitive (ACS) conditions and Procedure Groups
Quality outcomes, including mortality, emergency readmission within 28 days, Long Length of stay and patient safety indicators
Patient-level predicted risks for these outcomes, based on national Logistic Regression models which are executed using R statistical software and updated monthly
Various other national benchmarks, including Length of stay percentiles and Standardised Admission Ratio benchmarks
Numerous efficiency-based metrics, including average length of stay, day case rate and potential bed days saved
Prescribed Specialised Services (PSS) groups, using the PSS Grouper software.

NHS Provider Trusts have record level access to the pseudonymised data relating to their organisation through their subscription to our tools. Other NHS organisations can see aggregate level analysis through their subscriptions.

Bespoke analytics
An extract of processed data is used for conducting bespoke analytical services and research to support the NHS and other organisations for the benefit of health and social care. On a project-by-project basis, the team may conduct additional processing on the data to derive metrics or conduct statistical analyses that are not implemented during routine processing of data for the Dr Foster toolkit.

The analytics team work directly with NHS customers and propose to also work with non-NHS organisations that are in turn working with the NHS and using outputs of analyses for NHS benefit. The team also propose to offer services to non-NHS organisations for the broader analysis and benefit of health and social care. Non-NHS organisations will only ever be presented with aggregated, small-number suppressed data in line with guidance from NHS Digital.

The scope of analytics projects are by their nature bespoke and customised to local needs, however in all cases, the purpose and objectives of the work must demonstrate benefit to the NHS or Health and Social Care. For most projects, analyses are conducted to provide additional insight from the data that cannot be gained through use of Dr Foster toolkit. Examples of bespoke analytics projects conducted by Dr Foster are:
Supporting the NHS Improvement Getting It Right First Time (GIRFT) programme by delivering specialty specific data packs. The data packs developed by Dr Foster, include the derivation of bespoke indicators from HES data to measure the quality and efficiency of care in a particular specialty.
Research on HES data to develop an algorithm to help identify frail patients within hospitals. The developed algorithm was implemented locally at an NHS Trust.
Defining a new set of bespoke service lines for an NHS Trust to present meaningful performance indicators based on how they organise their services.

An established team of Analysts conduct all bespoke analytics projects. All analysts undergo training on handling sensitive records and are highly conversant in national guidelines to protect patient confidentiality. All outputs produced by the team are at an aggregate level and small numbers are suppressed in line with the HES Analysis Guide Guidelines. (Where there is any doubt, the Dr Foster Head of Information Governance or SIRO will provide guidance and if required contact NHS Digital.) Outputs are typically provided in data tables to NHS customers. Dr Foster may also produce data visualisations (e.g. bar charts, box plots, funnel plots) based on the aggregate-level information from the data tables.

Research
Dr Foster conduct research on HES data to provide thought leadership in the field of healthcare data analytics and to develop and refine methodologies for evaluating, monitoring and improving performance within healthcare organisations. Dr Foster publish the outputs of research for the benefit of the NHS, health and social care, and the public. Such content may be published directly by Dr Foster as articles to journalistic/media entities, or within academic journals. Dr Foster may also collaborate with other research groups by providing aggregate analyses to support publications expected to benefit the NHS and health and social care.

No record level data will be transferred outside of England and Wales.

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 be no linkage of data supplied under this agreement to any other datasets.
Data will only be accessed and processed by substantive employees of Dr Foster.

Dr Foster direct feed — DARS-NIC-392201-S6C3W

Type of data: information not disclosed for TRE projects

Opt outs honoured: No - data flow is not identifiable, Anonymised - ICO Code Compliant, No (Does not include the flow of confidential data)

Legal basis: Health and Social Care Act 2012 - s261 - 'Other dissemination of information', , Health and Social Care Act 2012 s261(2)(a)

Purposes: Yes (Supplier)

Sensitive: Non Sensitive, and Sensitive, and Non-Sensitive

When:DSA runs 2020-10-13 — 2022-01-12 2020.11 — 2024.09.

Access method: Ongoing, One-Off

Data-controller type: TELSTRA HEALTH UK LIMITED

Sublicensing allowed: No

Datasets:

HES:Civil Registration (Deaths) bridge
Civil Registration - Deaths
Emergency Care Data Set (ECDS)
Hospital Episode Statistics Critical Care
Hospital Episode Statistics Outpatients
Hospital Episode Statistics Admitted Patient Care
Hospital Episode Statistics Accident and Emergency
Civil Registration (Deaths) - Secondary Care Cut
HES-ID to MPS-ID HES Accident and Emergency
HES-ID to MPS-ID HES Admitted Patient Care
HES-ID to MPS-ID HES Outpatients
Civil Registrations of Death - Secondary Care Cut
Hospital Episode Statistics Accident and Emergency (HES A and E)
Hospital Episode Statistics Admitted Patient Care (HES APC)
Hospital Episode Statistics Critical Care (HES Critical Care)
Hospital Episode Statistics Outpatients (HES OP)

Objectives:

Aim and purpose of this application:
Dr Foster has provided objective insight and analysis since 1999. Dr Foster’s aim is to help health and social care organisations make better and faster decisions with data and insight and, ultimately, to benefit patients. This is delivered in three main strands:
• Dr Foster tools and services – to provide management information, analysis and clinical benchmarking through online products and services
• Bespoke analytics – to deliver customised projects to meet individual customer needs
• Research for publication – to provide thought leadership in the field of healthcare data analytics, with the aim of improving the planning, delivery and outcome of health and social care
This application is to request the direct dissemination of data from NHS Digital to Dr Foster.
The requested data are:
• Hospital Episodes Statistics (HES)
• Civil Registration (Deaths) - Secondary Care Cut (CRD) plus bridge files
• Emergency Care Data Set (ECDS)
Dr Foster currently receives these via the Dr Foster Unit at Imperial College London (DFU) under agreement DARS-NIC-68697-R6F1T with NHS Digital. DFU first receive this data under agreement DARS-NIC-12828-M0K2D.

Dr Foster will end this process and replace it with the direct receipt of data from NHS Digital. A transition period will be required to ensure that Dr Foster can continue to provide its NHS customers with its services. This will include testing of outputs so that NHS customers can be assured that they are receiving the same quality of service. The end of the transition period will coincide with the expiry of agreement DARS-NIC-68697-R6F1T with NHS Digital, but Dr Foster will work to complete this sooner. Dr Foster requests that, during this transition period, it receives data via DFU and directly from NHS Digital. At the end of this transition period, data will be deleted securely so that Dr Foster is only processing HES, CRD and ECDS data provided under this agreement. Under this agreement Dr Foster will cease offering any services to trusts which would enable them to re-identify patients. Dr Foster would not have the ability to identify any patients

Dr Foster is requesting pseudonymised data only under this agreement.

Data controller
Dr Foster are the sole Data Controller who also process the data provided under this agreement.

General Data Protection Regulation legal bases
Dr Foster process the data under General Data Protection Regulation (GDPR) articles 6(1)(f) (legitimate interests) and 9(2)(j) (archiving in the public interest). Dr Foster determined the legal bases by undertaking a legitimate interests assessment and a data protection impact assessment. These documents are maintained and updated as necessary by Dr Foster.

Dr Foster has a legitimate interest in being able to provide tools and services that healthcare organisations will find useful and that will benefit the health and social care system. Without processing this data, Dr Foster would not be able to deliver these tools and services. Withdrawing these would be to the detriment of health professionals who use them. Dr Foster's customers can be assured that the tools and services are based on evidence provided by data from a trusted source. There is no viable alternative as relying on public domain data would lead to gaps and not allow the same support for decision making. Using pseudonymised data allows Dr Foster to deliver evidence-based insight and analysis while minimising intrusion into a patient's privacy.

Dr Foster has a legitimate interest in developing new ways to help its customers improve their services and utilise the data to its full potential. This will be used to drive improved patient care, which is in the broader interest of everyone using the health and social care services in England.

The processing of this data is also necessary for archiving purposes in the public interest, scientific or historical research purposes or statistical purposes in accordance with Article 89(1) of GDPR. The processing is proportionate to its aims, respects data protection rights and provides suitable and specific measures to protect the rights and interests of individuals. It is necessary for reasons of public interest in the area of public health, in particular to ensure high standards of quality and safety of health care. Although Dr Foster are not a public body, they provide services to help public healthcare organisations to monitor and improve their services. Processing is designed to benefit patients and society as a whole through better healthcare.

Purpose of the request
Dr Foster require the requested data to help healthcare organisations achieve sustainable improvements in their performance, to gain insight and to inform decision making.

The required data to meet the purpose are:
HES critical care
HES Admitted Patient Care
HES Outpatients
HES Accident and Emergency
Civil Registration (Deaths) - Secondary Care Cut
HES:Civil Registration (Deaths) bridge
Emergency Care Data Set (ECDS)

Dr Foster use the data provided under this agreement to provide a management information function in the form of analysis and clinical benchmarking for healthcare organisations and to increase the power of predictive models for rare diseases, procedures and events. Dr Foster build standard casemix adjustment models for 259 diagnosis groups and 200 procedure groups which include some rarer conditions. Casemix is a system that measures hospital performance, aiming to reward initiatives that increase efficiency in hospitals. It also serves as an information tool that allows policy makers to understand the nature and complexity of health care delivery. Using all the requested datasets means that Dr Foster have the most up to date information and can inform customers of potential issues around quality and in turn they can make better informed decisions for the improvement of healthcare and outcomes for patients.

At a high-level Dr Foster analyses break down into the following:
• Quality measures of healthcare services by providers/area/clinical interest/trend analysis
• Variations in health outcomes
• Health inequalities and needs analysis
• Predictions
• Performance data and changes in clinical practice
• Management information
• Efficiency Monitoring
• Benchmarking
• Contract Management and Variance Analysis
• Activity Monitoring
• National Target Performance
• Pathway design, redesign and improvement.
• Practice Performance Monitoring
• Capacity and utilisation management
• Cross checking of commissioning data
• Systems to support and monitor the pattern of healthcare usage
• Patient segmentation analysis
• Overall data quality

The data allow Dr Foster to provide a wide array of relevant indicators to give end users as complete a picture of hospital performance as possible to allow health and social care organisations to effectively:
• Monitor quality of services provided
• Identify efficiency opportunities
• Identify pathways where services can be improved for the benefit of patients

Civil Registration (Deaths) - Secondary Care Cut specific purposes
Civil Registration (Deaths) - Secondary Care Cut data are requested to provide more timely and accurate analysis and insight for Dr Foster’s customers. It is essential for performing survival analyses and so represents a very valuable source of data. It will improve the output of Dr Foster’s products for the benefit of NHS customers and for the broader improvement of health and social care for the public.

These data are extremely critical because mortality information may be a surrogate metric for success of medical care. Therefore, this dataset will enable identification of factors that drive successful treatment of a patient. Cause and date of death may also be used to identify trends in causes of death in particular groups of patients. Historical death data are necessary for identifying trends.

Linking HES datasets with mortality data could provide valuable insights into how and why some patients with the same condition and at the same stage can have very different outcomes. In addition, it would be used to:
• Compare hospital mortality rates for in-hospital deaths with rates for all deaths to evaluate the effect of differential discharge policies
• Calculate total post-operative mortality rates, e.g. when comparing operative techniques such as laparoscopy and open approaches
• Assess potential quality of care issues by comparing the cause of death with the reason(s) for admission, e.g. for surgical patients who are discharged within 30 days of the procedure but who die at home and whether the death is related to their disease process or to complications of treatment
• Develop and validate indicators of quality and safety of healthcare, particularly by consultant and hospital
• Show variations in performance by unit and socio demographic stratum
• Predict risk and adjust risk of indicators and variations and any other methodological aspects as they arise
• Establish seasonal patterns of mortality
• Support organisations in delivering their Learning from Deaths agenda and timely mortality reviews
• Help organisations improve quality of care and identify where they could do more to help patients and their families

30 day mortality (both in and out of hospital) is a well published and accepted standard for comparing post-operative and post-admission hospital mortality. Linking Civil Registration (Deaths) - Secondary Care Cut data with HES data allows Dr Foster to provide this outcome, which will improve engagement with clinicians, and allow comparisons with other published analyses.

Dr Foster process the minimum data necessary to meet the purpose and build privacy into their designs. Dr Foster have no requirement to re-identify the individuals within the data they receive and will make no attempt to carry out any re-identification.

In the context of the above statement, while Dr Foster and NHS Digital recognise that the inclusion of record-level Date of Death linked to all four HES datasets will theoretically increase the risk of re-identification, for clarity it is pointed out that:
a) as stated above, Dr Foster will not re-identify; in addition, given the volumes of data, identification serves no purpose to Dr Foster anyway
b) raw Date of Death allows Dr Foster to carry out better analyses and provide a better service to their customers:
- using an alternative such as “death 30/60/90 days from Discharge” does not enable the calculation of median survival rates, nor does it enable Dr Foster to produce aggregated survival curves
- in some cases, Dr Foster calculates mortality rates with reference to particular Procedures not to Discharge; the above flag would not support this as “day 0” would need to be set at different events in different cases for different purposes
c) record-level data is only shared by Dr Foster with customer Trusts only for those patients treated by that Trust

Emergency Care Data Set (ECDS) specific purposes
Dr Foster are requesting Emergency Care Data Set (ECDS), in parallel to HES A&E, so that it can:
• continue to deliver products and services that provide insight into and analysis of emergency care provision;
• develop enhancements to its services that utilise the greater detail provided by ECDS,
• transition its products from A&E data to reduce any interruption for NHS healthcare professionals that rely on these tools and services;
• quality assure its processing of ECDS.

Historical ECDS data is necessary for more useful trend analysis, research and quality assurance. The additional fields within ECDS will help Dr Foster to:
• improve understanding of the complexity of attending patients and the causes of rising demand;
• capture diagnostic data for richer information on the diagnosis with which patients are presenting to emergency departments;
• enhance the understanding of the value of emergency departments;
• enhance understanding of need, activity and outcomes;
• better understand patient pathways such as type 5 emergency admissions (same day emergency care), which are currently not coded within HES.

Mental health fields within ECDS will be used by Dr Foster:
• To help better understand the cohort of patients seen in the Emergency Department with mental health conditions who are both formally and informally detained under the Mental Health Act.
• To help emergency care departments understand how the above patients use their services and what affect this may have on departments.

Number of years requested
Dr Foster has reviewed its requirements to ensure that only the minimum data necessary is requested. It has refined its processing to require a reduced minimum of 11 years of historic data to deliver results to customers. (This is a reduction from the 15 years required under DARS-NIC-68697-R6F1T).
A data period of 11 years of historical data is essential to enable Dr Foster to:
• Obtain longitudinal data on prior admissions for patients. Risk modelling will also require access to variables on prior admissions including previously recorded co-morbidities.
• Create, update and maintain statistical risk models to enable the regular production of risk adjusted measures of mortality, quality and efficiency (including Hospital Standardised Mortality Ratio (HSMR) and Cumulative Sum Control Chart (CUSUM) alerts as used by NHS organisations and regulators).

Users of Dr Foster products and services
NHS subscribers to Dr Foster tools have access to analysis of the data so that they can:
• Track and trend performance, identify areas for efficiency savings and understand and influence demand and patient flow throughout the health and care system.
• Investigate risk-adjusted quality, patient safety and clinical outcomes data including mortality, benchmark against other healthcare organisations and identify areas for improvement.

Dr Foster online products are used by:
• NHS Provider Trusts– Subscribed authorised users in customer organisations can view data that relate to their organisation at a record level. They cannot access record level HES data relating to other organisations.
• Other NHS organisations – Subscribed authorised users in customer organisations can view aggregated analysis which provides valuable insight but prevents any patients from being identified, in accordance with guidance provided by NHS Digital.
• Care Quality Commission – CQC can view aggregated analysis.

The Analytics team provides aggregate level and small number suppressed analysis (in line with the HES Analysis Guide) and insight to a number of NHS customers including:
• NHS Trusts
• Clinical Commissioning Groups
• Commissioning Support Units
• Department of Health
• NHS England
• NHS Improvement
• Care Quality Commission
• Public Health England
• National Institute for Health and Care Excellence
The Analytics team also provides analysis with small numbers included as part of the Getting It Right First Time (GIRFT) programme. This is a national programme that identifies best practice and where changes can be made to improve care and patient outcomes. This is described more under ‘processing activities’.

Dr Foster also work with:
• Non-NHS organisations providing services to benefit the NHS – these are only supplied with aggregate, small-number suppressed analyses in line with the HES analysis guide.
• Non-NHS organisations to benefit public health and social care - these are only supplied with aggregate, small-number suppressed analyses where benefits can be identified for the health and social care system. It is also proposed that algorithms or coefficients that have been derived through research on HES data may be provided directly to a customer for implementation on their own local data. Data provided in all outputs will be at an aggregate level and small number suppression will be implemented in line with HES analysis guidelines.

Any request for such analysis is reviewed to determine if it benefits the health and social care system. Dr Foster will inform NHS Digital of analysis it provides to non-NHS organisations and will list this in any renewal or amendment to this agreement.

Dr Foster provided analysis on COVID-19 to the British Red Cross in May 2020. This was frailty analysis by Lower Super Output Area (LSOA) which included the following:

• Proportion of frail patients with mobility problems
• Proportion of frail patients with mobility problems and a fracture.

These were percentages only and included no small numbers.

Customers have access to Dr Foster’s team of qualified data scientists, clinicians, statisticians, mathematicians, and economists. They supplement in-house analytical teams with Dr Foster’s expert advice and guidance linking, modelling and visualising data and insight.

The Analytics team provides bespoke analytics and data science tailored to specific needs to identify clinical variation, efficiency savings, predict patient risk and improve patient outcomes.
They are a skilled team of experts in advanced healthcare analytics and data science including predicative analytics, machine learning techniques and advanced statistical methods. The team use this expertise to investigate issues and transform healthcare services.

Yielded Benefits:

COVID-19 heatmaps (April 2020) Dr Foster’s awareness raising publications have provided professionals and the public with insight into the progression of COVID-19. Their interactive dashboard, first released in April 2020, uses heatmaps to show the spread of the disease. It also shows historical perspectives within the past 10 years by using respiratory and frailty data from HES. Frailty analysis for British Red Cross (May 2020) Dr Foster’s frailty analysis has been used by the British Red Cross create a COVID-19 vulnerability index for the UK, mapping clinical vulnerability, economic vulnerability, social vulnerability and other health and wellbeing needs. This is helping the British Red Cross focus help on the most vulnerable people whose needs aren’t being met. High Intensity User (HIU) report (January 2019) Dr Foster continues to raise public and professional awareness. Its High Intensity User (HIU) report of January 2019 uncovered important characteristics of HIU patients and patterns in their attendances of A&E. It showed that the vast majority of HIUs are living in the most deprived areas of England, suggesting that the most vulnerable members of society may be more prone to high intensity use. Smoking, drugs and alcohol all appear to play an important role in frequent A&E use, in relation to the most common reasons that HIUs are admitted to hospital. Dr Foster also measure benefits through customer feedback for their products and services. Case study - North Cumbria University Hospitals NHS Trust (2016) “We find Dr Foster’s combination of knowledgeable experts and powerful tools enormously helpful in our work to improve the quality of care we are providing to our patients. With Dr Foster’s help we’ve made significant progress in understanding quality and its drivers, and identifying how we can make sustainable improvements in our hospitals. Dr Foster’s insightful analysis, practical recommendations and ongoing support help us extract maximum value from our data, and their impact is far-reaching.” Case study – Lancashire Teaching Hospitals NHS Trust (January 2015) Dr Foster’s HIP is used on at least a weekly basis by the corporate and business intelligence teams and clinical staff to: • inform and direct the Trust’s mortality and morbidity review processes • scrutinise care standards and their impact on patient outcomes • provide analysis and reassurance to the board, governors and the public • monitor trends in readmissions and complications and investigate if these were justified clinically Instigated several quality improvement initiatives including: • Improved documentation of complexity in perinatal conditions that has: o reduced mortality ratios o increased income o engaged clinicians in a wider quality initiative introducing an enhanced model of care for potentially vulnerable babies • Development of an improvement programme for patients suffering from chronic obstructive pulmonary disease across the whole care pathway in the local health economy Case study – University Hospital of South Manchester NHS Foundation Trust (January 2015) The Trust specialises in cardiac surgery activity, for which it is a tertiary centre, and performs a high number of coronary artery bypass grafts (CABG) and heart valve replacements. Dr Foster’s Practice and Provider Monitor enabled benchmarking of productivity and efficiency measures, giving the user the ability to compare mean-price-per-spell at both procedure and diagnosis HRG level. The Trust used the Dr Foster tool to look more deeply at other influences on the efficiency of the pathway compared with others and highlighted that the length-of-stay for these procedures was one of the longest of its peer group. This clearly has an impact on income as the amount earned per bed day is lower, and the capacity to put more patients through the system is reduced. From a patient’s point of view, this is also good news: a longer length-of-stay may increase risk. UHSM then used Practice and Provider Monitor to move through the specialties to highlight areas of variance and focus on where they could improve productivity and efficiency across the Trust. Case study - Imperial College Healthcare NHS Trust (October 2018) Imperial College Healthcare NHS Trust has been a longstanding customer of Dr Foster with a dedicated Business Insight Manager based at the trust who has been delivering bespoke analytic support and expertise in clinical benchmarking. This dedicated, expert resource supports mortality monitoring, market share analysis and efficiency indicator benchmarking. Dr Foster data is now integrated into strategic planning and service redesign and has been used to explore growth opportunities for services previously provided by other trusts. As part of North West London’s Shaping a Healthier Future, Dr Foster’s analytic support has helped the trust in service redesign work, for example in integrating services previously provided by Ealing Hospital. The Deputy Chief Information Officer, Imperial College Healthcare NHS Trust stated that, “National benchmarking is possible, but you have to do a lot of work with the data yourselves. With Dr Foster tools the data is easy to access and we can make sure we are keeping pace with other high-performing organisations. Having a Dr Foster analyst on site has been very successful. […] Dr Foster understands what our objectives are and is able to carry out complex analysis on our behalf. It is a fast track way of getting good benchmarking. For example, our performance framework has over 100 different metrics. The Dr Foster tools are useful looking across the Sustainability and Transformation Plan (STP) area to understand what is happening with indicators such as length of stay.” Northampton General Hospital NHS Trust (February 2015) “Clinicians and analysts use Healthcare Intelligence Portal (HIP) on a daily basis to analyse new patient safety alerts, high standardised mortality ratios and individual cases. Dr Foster investigate areas of concern in clinical coding and data analysis and look to improve future patient care through retrospective review and analysis.” Northern Devon Healthcare NHS Trust (February 2015) “Our data quality team use HIP to identify patients with missing or duplicate information and whether the problem is ongoing and needs a change in process to rectify it or whether it’s due to individual oversight.” Wrightington, Wigan and Leigh NHS Foundation Trust (February 2015) “We carried out a review of dermatological deaths and the data in HIP identified that the deaths were due to cellulitis. We reviewed treatment options based on this and identified where improvements could be made.” The Royal Bournemouth and Christchurch Hospitals NHS Foundation Trust (February 2015) “We use HIP to undertake many audits and reviews at any one time. For example, ten sets of case notes are reviewed each month by an emergency department consultant as part of our programme to improve performance on sepsis.”

Expected Benefits:

Dr Foster has a legitimate interest in being able to provide tools and services that healthcare organisations will find useful and that will benefit the health and social care system. It has a legitimate interest in assuring customers that the tools and services are based on evidence provided by data from a trusted source, and that they can be confident in realising the benefits this brings. Expected benefits are:
• Enabling NHS acute trusts to measure, compare and benchmark key quality indicator trends focusing on risk adjusted measures of mortality, readmissions and length of stay in hospital.
• Providing evidence to instigate clinical audit and investigations related to quality of care, such as highlighting potential poor clinical coding or quality/efficiency concerns.
• Validating other mortality indicators such as HSMR, CUSUM alerts and crude mortality.
• Enabling NHS acute trusts and commissioners to use performance information to identify, quantify and act on opportunities to improve efficiency of health services.
• Understanding areas of best practice amongst Dr Foster customers and facilitate interactions with other customers who are not performing as well to support quality and efficiency improvement.
• Helping clinicians and managers by providing independent and authoritative analysis of the variations that exist in acute hospital care in a way that is meaningful for them and that is understandable to patients and the public.
• Highlighting topics of interest to the health industry and wider public to enable discussion and improvement in healthcare provision.
• Maintaining the focus of the organisations on improvement.
• Raising public and professional awareness
• Providing valuable insights into how and why some patients with the same condition and at the same stage can have very different outcomes
• Supporting organisations in delivering their Learning from Deaths agenda and timely mortality reviews
• Helping providers improve quality of care and identify where they could do more to help patients and their families
• Helping providers address pressures on the emergency care services by identifying opportunities to relieve these pressures
• Helping providers improve patient outcomes and experiences

Dr Foster has a legitimate interest in developing new ways to help its customers improve their services and utilise the data to its full potential. This will be used to drive improved patient care, which is in the broader interest of everyone using the health and social care services in England.

Dr Foster’s publications and research articles around variations of healthcare within the NHS is in the public interest and supports the government agenda for transparency by promoting choice and accountability within the NHS.

How these benefits will be measured
Benefits are ongoing as the outputs described above are used within NHS Trusts’ internal monthly reporting and quality processes. Dr Foster services allow performance of NHS Provider Trusts to be monitored and trended over time and therefore provide customers with the ability to measure changes in quality and performance particularly in instances where customers have been alerted and they have worked with them to understand the causes of worse than expected performance.

When these will be achieved
These benefits are achieved continually and are reliant on a range of factors outside of Dr Foster's control. However, whenever there are areas of concern about performance against key indicators, Dr Foster act immediately to alert relevant stakeholders to help in better understanding and addressing them.

Outputs:

Dr Foster provide commercial tools, insight and analysis to healthcare professionals. They also provide freely available online resources for healthcare professionals and the public (using anonymised outputs) to help improve professional and public understanding of health. For example, Dr Foster analysts have developed an indicator to help local public health teams monitor any increases in Covid-19 infections to help prevent lockdown situations. The interactive graph allows public health teams to monitor any increases in cases of Covid-19 by recent daily case rate compared to a prior eight-week daily case rate. It provides an early indication of a potential spike, allowing local authorities to act early and implement restrictions. This is available on the Dr Foster website: https://drfoster.com/2020/08/12/dr-foster-covid-19-hotspot-indicator/. Dr Foster also carried out a detailed analysis of accident and emergency (A&E) attendances nationally with the aim of uncovering common characteristics of High Intensity Users (HIUs) - people who attended 10 or more times in a 12-month period - and patterns in HIU attendances to provide valuable insight and a better understanding of the reasons they attend with such high frequency. This was published on the Dr Foster website and reported on in the media. Press coverage of Dr Foster’s freely available analysis helps disseminate knowledge and understanding further.

The majority of Dr Foster outputs are delivered through:

• Dr Foster online tools and services including the Healthcare Intelligence Portal
• Bespoke analytics
• Research for publication. Specific outputs include benchmarked or standardised healthcare indicators and analysis such as mortality (Summary Hospital-level Mortality Indicator (SHMI)/HSMR), LOS (Length of Stay), admission trends, readmission rates, patient safety indicators, referral patterns, market share analysis etc.

Outputs will be used by customers to investigate clinical quality, performance and business development, specifically:
• Assess and manage clinical quality and patient safety within NHS Organisations
• Identify pathways where there is potential for improvement
• Identify areas of best practice either within the Provider Trust or local/national health economies
• Better understand how they compare to other Provider Trusts with similar case mixes
• Identify improvements in operational efficiency
• Understand patient outcomes
• Identify and understand market activity
• Monitor the impact of implemented changes
• Identify variations in outcomes

The above outputs depend on processing of all the requested data.

Civil Registration mortality data specific outputs
Specific outputs dependent on the processing of Civil Registration mortality data are:
• Analysis of cause of death
• Analysis of death following discharge, 7, 14, 30 days
• Comparative analysis of cause of death and deaths following discharge
• Development of outputs to further help users understand patient outcomes through analysis of survival rates
• Analysis of variation in mortality across geographical boundaries
• Support customers with out of hospital mortality queries
• Additional level of insight for customers to investigate the care pathway for their patients

Timeframe for outputs
Subscribers to the tools have continual access which allows them to meet their own internal target dates.

Outputs of bespoke analytics projects are dependent on the nature of the project and can include tabulations, dashboards, reports, spreadsheets, presentations or articles. Outputs may be surfaced through tools including Microsoft Office suite (Excel, Word and PowerPoint etc) or other tools (Tableau, QlikView) depending on the requirements of the customer. In some instances, algorithms or coefficients that have been derived through research on the data may be provided directly to a customer for implementation on their own local data. Data provided in all outputs are at an aggregate level and small number suppression is implemented in line with HES analysis guidelines. Bespoke analytics projects are conducted on an ad hoc basis and target dates for delivery of outputs are thus defined upon commencement of each project.

Publications
Examples of previous publications produced by Dr Foster include the Hospital Guide that published analysis of the variations in acute hospital care for the benefit of healthcare professionals, patients and the public and insight articles published on the Dr Foster website.

In January 2020, Dr Foster undertook statistical analyses of abdominal aortic aneurisms and trans-catheter aortic valve implementations and found interesting correlations between surgeon annual volume and mortality. Following on from this, the Dr Foster team carried out an analysis that examined how the number of annual knee replacement procedures performed within a trust influences the rate of readmission. Insights from this are published at https://drfoster.com/2020/01/30/detailed-analysis-of-knee-replacement-annual-volume-reveals-its-significant-effect-on-readmission-rates/. Other Insights reports and briefings are available at https://drfoster.com/insights/.

Dr Foster provide an interactive dashboard on its website to provide information to help manage and predict the risk of COVID-19 for England. The respiratory and frailty data are from HES. The dashboard was initially published in April 2020 and is updated regularly at https://drfoster.com/2020/04/06/uk-covid-19-progression-dashboard.
Dr Foster are aware that publications, whether inside or outside the NHS, must adhere to strict guidelines in terms of disclosure, and ensure that any such publications are aggregated and comply with small number suppression in line with the HES Analysis Guide and other relevant legislation. Analyses for use in publications can be in the form of text, tables, or other data visualisations such as diagrams/graphs using aggregate data. Publications will also meet standards as defined in the Terms and Conditions of the Data Sharing Agreement.

All outputs will contain only data that is aggregated with small numbers suppressed in line with the HES Analysis Guide.

Processing:

Under this agreement, pseudonymised HES, ECDS and Civil Registration Deaths datasets will be transferred by NHS Digital to Dr Foster via Secure Electronic File Transfer (SEFT).

Dr Foster’s ISO 27001 certified Information Security Management System (ISMS)
The processing of the pseudonymised HES data is within the scope of the ISO 27001 certified Information Security Management System (ISMS). Dr Foster processes are subject to internal and external audit. All staff receive initial information governance training followed by regular mandatory modules spread across the year.

The pseudonymised data are held securely on Dr Foster systems with access permissions only granted as necessary to specific roles.

Dr Foster tools
The data are processed through products in the Dr Foster toolkit to provide:
• Linkage into spells and superspells, which can often span across financial years
• Healthcare Resource Group (HRG), Tariff and other Payment By Results (PBR) related fields, using the HRG Grouper software
• Various clinical groupings, including Clinical Classification Software (CCS) Diagnoses, Ambulatory Care Sensitive (ACS) conditions and Procedure Groups
• Quality outcomes, including mortality, emergency readmission within 28 days, Long Length of stay and patient safety indicators
• Patient-level predicted risks for these outcomes, based on national Logistic Regression models which are executed using R statistical software and updated monthly
• Various other national benchmarks, including Length of stay percentiles and Standardised Admission Ratio benchmarks
• Numerous efficiency-based metrics, including average length of stay, day case rate and potential bed days saved
• Prescribed Specialised Services (PSS) groups, using the PSS Grouper software.

NHS Provider Trusts have record level access to the pseudonymised data relating to their organisation through their subscription to our tools. Other NHS organisations can see aggregate level analysis through their subscriptions.

Bespoke analytics
An extract of processed data is used for conducting bespoke analytical services and research to support the NHS and other organisations for the benefit of health and social care. On a project-by-project basis, the team may conduct additional processing on the data to derive metrics or conduct statistical analyses that are not implemented during routine processing of data for the Dr Foster toolkit.

The analytics team work directly with NHS customers and also work with non-NHS organisations that are in turn working with the NHS and using outputs of analyses for NHS benefit. The team also offer services to non-NHS organisations for the broader analysis and benefit of health and social care. Non-NHS organisations will only ever be presented with aggregated, small-number suppressed data in line with guidance from NHS Digital.

The scope of analytics projects is by nature bespoke and customised to local needs, however in all cases, the purpose and objectives of the work must demonstrate benefit to the NHS or Health and Social Care. For most projects, analyses are conducted to provide additional insight from the data that cannot be gained through use of Dr Foster toolkit. Examples of bespoke analytics projects conducted by Dr Foster are:
• Supporting the NHS Improvement Getting It Right First Time (GIRFT) programme by delivering specialty specific data packs. The data packs developed by Dr Foster, include the derivation of bespoke indicators from HES data to measure the quality and efficiency of care in a particular specialty.
• Research on HES data to develop an algorithm to help identify frail patients within hospitals. The developed algorithm was implemented locally at an NHS Trust.
• Defining a new set of bespoke service lines for an NHS Trust to present meaningful performance indicators based on how they organise their services.

An established team of Analysts conduct all bespoke analytics projects. All analysts undergo training on handling sensitive records and are highly conversant in national guidelines to protect patient confidentiality. Outputs are typically provided in data tables to NHS customers. Dr Foster may also produce data visualisations (e.g. bar charts, box plots, funnel plots) based on the aggregate-level information from the data tables.
Outputs produced by the team are at an aggregate level and small numbers are suppressed in line with the HES Analysis Guide Guidelines. However, some analysis with small numbers included may be provided where necessary to the GIRFT programme overseen by NHS Improvement.

Getting It Right First Time (GIRFT) which relates only to hospital care.
“Getting It Right First Time (GIRFT) is a national programme designed to improve medical care within the NHS by reducing unwarranted variations. By tackling variations in the way services are delivered across the NHS, and by sharing best practice between trusts, GIRFT identifies changes that will help improve care and patient outcomes, as well as delivering efficiencies such as the reduction of unnecessary procedures and cost savings.” https://gettingitrightfirsttime.co.uk
GIRFT is overseen by NHS Improvement, which is now cooperating with NHS England as a joint enterprise, with the two organisations being referred to as ‘NHS England and Improvement’. To make sure that they comply with data protection obligations NHSI and NHSE have entered into a Joint Controller and Information Sharing Framework Agreement, details of which can be found at https://www.england.nhs.uk/nhse-nhsi-privacy-notice/joint/joint-controller-agreement/. As a result of this NHS Improvement will be working more closely with NHS England and it is their intention that the outputs of data analysis completed using data provided under this agreement will also be shared with NHS England to facilitate service improvement work with Trusts.
• All output of bespoke analysis is suppressed as per NHS Digital disclosure rules other than work undertaken for the NHS England/Improvement (NHSE/NHSI) GIRFT programme, where the Data Controller is required to provide unsuppressed (low count) data in these outputs to enable national clinical leads, who are NHS employees, to explore and understand low volume activity in their conversations with providers as part of the picture of the service that is being considered. These outputs are produced by the Data Controller and shared securely with the GIRFT programme.
All users working on the GIRFT program (and any organisation receiving a report as outlined below), where unsuppressed small numbers data may be visible, are informed of the terms of use which state that they must not seek to re-identify any individual from that data. Data with small numbers included will only be shared within the programme where it is deemed absolutely necessary, and only after a local risk assessment has taken place and where sufficient controls are in place to manage any risks. Such considerations include that the geography of the data makes any risk of re-identification remote, limited patient demographics are included, data is presented across a whole quarter/year or where any risk of re-identification is not possible without unreasonable effort.
The test that will be applied by GIRFT is that 'the requirements for the need to share the unsuppressed data outweigh any risks posed and that all mitigated actions would be taken'.
Dr Foster users only have access to the data necessary for them to carry out their tasks, are reminded of their responsibilities for confidentiality and data protection, and receive regular training. Within the GIRFT programme access is managed and only made available to users subject to approval and risk assessment.
Reports containing unsuppressed data may only be shared by NHSE/I with :
a. individual organisations to whom the data relates, and/or
b. with other network members (with the agreement of the individual organisation to whom the data relates), and/or
c. within NHSE/I in order to achieve the benefits outlined within the GIRFT programme (whilst NHSE and NHSI have an active data sharing agreement with NHS Digital covering the source data).

Research
Dr Foster’s research provides thought leadership in the field of healthcare data analytics and develops and refines methodologies for evaluating, monitoring, and improving performance within healthcare organisations. Dr Foster publish the outputs of research for the benefit of the NHS, health and social care, and the public. Publication may be through its website, news or other media organisations or through academic journals. Dr Foster may also collaborate with other research groups by providing aggregate analyses to support publications expected to benefit the NHS and health and social care.

NHS Digital reminds all organisations party to this agreement of the need to 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 be no linkage of data supplied under this agreement to any other datasets.

Data will only be accessed and processed by substantive employees of Dr Foster.

Datasets requested under this agreement are on a rolling basis and be deleted on a rolling basis so that only 11 years of data are held.

Summary Hospital-level Mortality Indicator (SHMI) data — DARS-NIC-368020-R5L2K

Type of data: information not disclosed for TRE projects

Opt outs honoured: No - data flow is not identifiable, Anonymised - ICO Code Compliant, No (Does not include the flow of confidential data)

Legal basis: Section 42(4) of the Statistics and Registration Service Act (2007) as amended by section 287 of the Health and Social Care Act (2012), Health and Social Care Act 2012 – s261(1) and s261(2)(b)(ii), Health and Social Care Act 2012 s261(1) and s261(2)(b)(ii), Health and Social Care Act 2012 - s261 - 'Other dissemination of information', Health and Social Care Act 2012 s261(2)(b)(ii), Health and Social Care Act 2012 s261(2)(a)

Purposes: Yes (Supplier)

Sensitive: Sensitive, and Non Sensitive, and Non-Sensitive

When:DSA runs 2019-07-01 — 2020-06-30 2017.06 — 2024.09.

Access method: Ongoing

Data-controller type: TELSTRA HEALTH UK LIMITED

Sublicensing allowed: No

Datasets:

Summary Hospital-level Mortality Indicator (SHMI) data split by trust and diagnosis group
Hospital Episode Statistics Admitted Patient Care
Office for National Statistics Mortality Data
Summary Hospital-level Mortality Indicator
Summary Hospital-level Mortality Indicator (SHMI)

Objectives:

To produce/analyse statistics using births/deaths data solely to help the NHS perform its duties.

Yielded Benefits:

With the SHMI data provided, Dr Foster have been able to help their customers: - uncover and investigate some of the potential root causes of differences between the various mortality indicators and to investigate variations against peers. - Understand any variation between SHMI and HSMR at a summary level and what drives that variation. - Gain the insight they need to embed SHMI within their mortality management programme, alongside other mortality indicators, such as HSMR and Deaths after Surgery. - investigate and understand the impact of the inclusion of post-discharge mortality data (only available through the SHMI mortality indicator). - The ability to drill down and investigate by SHMI supergroup, CCS group or user-defined basket of diagnoses.

Expected Benefits:

Dr Foster provides all of its NHS customers free of charge with a tool to allow them to monitor and improve the quality and safety of care they provide by comparing the two leading mortality indicators in England: SHMI and HSMR. It enables their customers to perform a root cause analysis of their SHMI, in line with the requirements of the NHS Operating Framework.

“All hospital trusts, regardless of whether they are outliers, need to examine, understand and explain their SHMI and identify and act where performance is falling short. Should a trust be an outlier on any mortality measure it should scrutinise the underlying data to understand the reason and take appropriate action.”
The Operating Framework for the NHS in England 2012/13

As identified in the commissioning letter, with SHMI data provided, Dr Foster have been able to help their customers:
- uncover and investigate some of the potential root causes of differences between the various mortality indicators and to investigate variations against peers.
- Understand any variation between SHMI and HSMR at a summary level and what drives that variation.
- Gain the insight they need to embed SHMI within their mortality management programme, alongside other mortality indicators, such as HSMR and Deaths after Surgery.
- investigate and understand the impact of the inclusion of post-discharge mortality data (only available through the SHMI mortality indicator).
- The ability to drill down and investigate by SHMI supergroup, CCS group or user-defined basket of diagnoses.

Expected measurable benefits include:
• Enable customers to measure, compare and benchmark mortality and alerting those who have higher than expected mortality levels to encourage efforts to investigate and address these. Dr Foster's independent position is beneficial as it supports customer focus on information and data as opposed to anecdotal evidence.
• Identify mortality trends across hospitals.
• Instigate clinical audit and inform investigations related to quality of care, such as highlighting poor clinical coding or quality/efficiency concerns.
• Validate other mortality indicators – such as HSMR and crude mortality.
• Understand and quickly visualise SHMI & HSMR indicators side by side.

How will these be measured:
• By their nature, Dr Foster analytical tools allow the performance of customers to be monitored and trended over time. It can indicate changes to quality and efficiency performance particularly in instances where trusts have been alerted and Dr Foster has worked with them to understand the causes of worse than expected performance.

When will these be achieved:
• It is not possible to outline a specific target date for achievement of the benefits outlined above as they are reliant on a range of factors outside of Dr Foster immediate control. However, whenever there are areas of particular concern about performance against key indicators, Dr Foster acts immediately to make their customers aware and offer assistance in better understanding and addressing them.
• In addition benefits are ongoing as these outputs are used within NHS Trusts internal monthly reporting and quality processes.

Outputs:

The Dr Foster Dashboard Tool Online application is available to NHS Acute Trusts, only which compares the two leading mortality indicators in England – the SHMI and the Dr Foster Hospital Standard Mortality Ratio (HSMR). All data is aggregated with small numbers suppressed in line with the HES analysis guide and there are no links to any identifiers. It enables users to uncover and investigate some of the potential root causes of differences between these indicators and to investigate variations against peers.

Key outputs:
• Overview of SHMI and Hospital Standard Mortality Ratio (HSMR) - summary charts, trends and breakdowns.
• Graphical Dashboard comparing mortality measures side-by-side.
• Analyse national position, regional comparisons and custom peer groups.
• The ability to drill down and investigate by SHMI supergroup, Clinical Classification System (CCS) group or user-defined basket of diagnoses.

Typical end users
• Chief Executives
• Medical Directors
• NHS Managers
• Information Analysts
• Clinicians
• Nurses

Processing:

Landing

On landing the SHMI dataset will be recorded on the Dr Foster Data Asset Register (DAR) and allocated a unique Asset Tag, in addition a Date of Destruction will be recorded along with Acknowledgements required in the publication of these data.

Processing

NOTE: Data will flow to Dr Foster Ltd only. Data will be accessed by the named users within this agreement, i.e. substantive employees of Dr Foster Limited.

Once logged in the Data Asset Register, it is handed over to a named individual who will load these data onto a secure central processing server located at Dorset Rise, a ‘SHMI’ Extract, Transform & Load process (documented) will then be run to transform record level data and then appended into a aggregated SQL database (aggregated at Provider & Diagnosis group level). Once processed the data will then be quality checked and upon completion published to the live client facing Dr Foster Dashboard Tool.

Publication

SHMI data, which has been available to Dr Foster since 2011, will only be made available to NHS Trusts via the Dr Foster Dashboard Tool.

Note: This tool is provided Free of Charge to all NHS Trusts

Destruction

Raw SHMI data will be Blancco (CESG approved) file shredded with certificated evidence when Date of Destruction is applicable (identified on Dr Foster’s Data Asset Register via a monthly process).

Telecity is listed as a storage address for recovery purposes only. Telecity do not have access to the Dr Foster Ltd. Server or Server passwords and will not be accessing/backing up the data provided within this agreement.