NHS Digital Data Release Register - reformatted

The Nuffield Trust projects

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

🚩 The Nuffield Trust was sent multiple files from the same dataset, in the same month, both with optouts respected and with optouts ignored. The Nuffield Trust may not have compared the two files, but the identifiers are consistent between datasets, and outside of a good TRE NHS Digital can not know what recipients actually do.

Retrospective analysis of the impact of Royal Voluntary Service Home from Hospital scheme on NHS hospital use. — DARS-NIC-86623-P4F4D

Type of data: information not disclosed for TRE projects

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

Legal basis: Informed Patient consent to permit the receipt, processing and release of data by the HSCIC, Health and Social Care Act 2012 – s261(1) and s261(2)(b)(ii), Health and Social Care Act 2012 – s261(2)(b)(ii)

Purposes: No (Research)

Sensitive: Non Sensitive, and Non-Sensitive

When:DSA runs 2018-04-01 — 2021-03-31 2017.09 — 2017.11.

Access method: One-Off

Data-controller type: THE NUFFIELD TRUST FOR RESEARCH AND POLICY STUDIES IN HEALTH SERVICES

Sublicensing allowed: No

Datasets:

  1. Hospital Episode Statistics Admitted Patient Care
  2. Hospital Episode Statistics Accident and Emergency
  3. Hospital Episode Statistics Outpatients
  4. Hospital Episode Statistics Accident and Emergency (HES A and E)
  5. Hospital Episode Statistics Admitted Patient Care (HES APC)
  6. Hospital Episode Statistics Outpatients (HES OP)

Objectives:

The Nuffield Trust’s overarching purpose is to help provide objective research and analysis that boosts the quality of health policy and practice, and ultimately improves the health and health care of people in the UK. The Nuffield Trust are an independent research group overseen by a board of Trustees including a number of senior NHS clinicians, managers and academics.

This application relates to a piece of evaluation work the Nuffield Trust is undertaking funded Royal Voluntary Service (RVS) to evaluate their ‘Home from Hospital’ programme in Leicestershire.

In order to do this the Nuffield Trust propose using linked data to compare the outcomes for project beneficiaries versus a matched control group who did not receive the scheme’s services. The primary outcome measures will be hospital activity in the period after referral to the scheme.

The Nuffield Trust have a successful track record in such studies – most recently in the evaluation of the Cabinet Office and NHS England funded scheme to test the impact of volunteers on hospital utilisation. This work will add to the growing evidence base on the effectiveness of increasingly common programmes in use in the English NHS.

The Hospital from Home scheme commenced in October 2014 and runs into early 2017. The key output of the Nuffield Trusts work will be an independent assessment of the strength of evidence that RVS’s scheme has any favourable impact on older people’s use of hospital services.

The Nuffield Trust will not be sharing any data with any third party. All published or otherwise shared information will aggregated data with small numbers suppressed in line with the HES Analysis guidance.

The Royal Voluntary Service was founded in 1938 to support local communities. With over 35,000 volunteers, their aim is to help older people stay active, independent and able to continue to contribute to society.

Yielded Benefits:

This work has been delayed by approximately 6 months. At present the Nuffield Trust are actively working on the analysis but no outputs have yet been produced. The revised planned end date is late summer 2018.

Expected Benefits:

The Nuffield Trust evaluation will provide an independent assessment of the impact of RVS’s Home from Hospital Scheme. This scheme aims to use volunteers and voluntary sector staff to help support older people’s timely discharge from hospital wards. It is one of a number of similar schemes increasingly being commissioned by local authorities, CCGs and acute trusts in England.

The evidence the Nuffield Trust provide will be of benefit to these commissioners, both in England and beyond, as they consider funding these types of services over others on a longer-term basis. It may also provide new information for RVS and other charities about specific aspects of their scheme that appear to provide the greatest benefits.

The evaluation itself will measure the impact of the schemes on hospital usage: including emergency admissions, re-admissions, length of stay and A&E attendances over ~9 to 12 months post referral.

Evidence of the success or failure of the service will be beneficial to patients in the wider sense that if such schemes are shown to lead to improved care after admission to hospital, similar schemes will be more likely to receive funding from local commissioners elsewhere. The Nuffield Trust analyses may help these schemes identify specific subgroups of individuals who might be best targeted for care.

This study will be a useful addition to a growing set of Nuffield Trust evidence on the impact of the voluntary sector on use of the acute sector in England: evaluation of British Red Cross ‘Support at Home’ scheme (2012); evaluation of a Cabinet Office ‘social action’ fund (2016), evaluation of Age UK’s Integrated Care Programme (findings due 2017).

Outputs:

The prime analytical outcome of the Nuffield Trust data processing will be the identification of a group of people (within the HES data) who had extremely similar characteristics to the five projects’ service recipients, but who didn’t receive a service. For example if 1,500 people by early 2017 received a service from RVS, Nuffield will have identified this group in HES data, in addition to another 1,500 individuals to be used as our evaluation’s matched control group.

Differences between these two groups in terms of their subsequent use of hospital services will be analysed to test whether there is any evidence that the services have had any impact on admissions to hospital or lengths of stay, adjusting for remaining differences between the two groups.

An evaluation report will be submitted to RVS in the second half of 2017. Shortly afterward, the Nuffield Trust will publish the results in a brief Nuffield Trust branded research report, possibly as part of a summary document reviewing all Nuffield Trust recent analyses of the impact of voluntary sector lead schemes. This will be made freely available on the Nuffield Trust website. The Nuffield Trust may also submit the findings to quality peer reviewed journals.

Processing:

The role of NHS Digital:

NHS Digital will receive identifiable person level information from RVS.

The person level information received by NHS Digital will cover all people recruited to the programme between October 2014 and early 2017, where those individuals have consented to sharing data with NHS Digital for evaluation purposes (subsequently referred to as service recipients).

The transferred information will consist of only:

• NHS number (if available)
• Name
• date of birth
• address including postcode
• gender

in addition to a non identifiable client ID added by the services to denote each unique individual.

NHS Digital will receive no other information about any service recipient.

This data will be transferred to NHS Digital using the NHS Digital’s own secure transfer facilities, under NHS Digital advice.

NHS Digital’s Data Linkage Service will process the person identifiers. For each service recipient, they will find the relevant pseudonymised identifier, the HESID, in the form held by the Nuffield Trust.

The Data Linkage Service will produce a file intended for the Nuffield Trust. This file will contain the HESID of each service recipient, alongside other limited pseudonymised information:

• LSOA of residence,
• age (/year of birth)
• gender.

It will also include information about the matching technique, and the non identifiable client ID.

NHS Digital will finally transfer this file securely to the Nuffield Trust.

Role of NHS Digital summary: NHS Digital will receive personal identifiers of people recruited to the RVS programme and will provide the relevant pseudonymised HESIDs of these individuals to the Nuffield Trust.

The role of the Nuffield Trust

The Nuffield Trust will receive from NHS Digital the list of HESIDs of people recruited to the services between October 2014 and early 2017 (service recipients).

The Nuffield Trust will receive at the same time a dataset from RVS. This dataset will contain no identifiable information, and will only contain the non identifiable client ID as a person identifier.

Other data contained in these files will include non identifiable details of services received by the people recruited to these services, eg date of referral to the service provider, number of minutes of support received, dates of services provided, etc.

As a first processing step the Nuffield Trust will link the HESIDs (from NHS Digital) and the non identifiable service information (from the seven service providers) using the non identifiable client ID.

The Nuffield Trust will then link this data to pseudonymised HES data and monthly MMES.

For each service recipient the Nuffield Trust will link to up to two years of HES data for the period just prior to the referral to the service provider. The Nuffield Trust will also link to all subsequent HES activity captured using the latest monthly HES datasets.

Therefore the Nuffield Trust will need to link to HES/MMES APC, OP and A&E data from 2012/13 through to 2016/17 (M13).

Using this data the Nuffield Trust will build person level analysis datasets which will characterise each service recipient in terms of: demographic characteristics, history of hospital use, and morbidity characteristics.

The key aim of the Nuffield Trust processing is to find a matched group of people who very closely share these same characteristics, but who did not receive a service from RVS, and to use this group as a pseudo-control group.

To do this the Nuffield Trust will build analysis datasets for the wider group of people in the hospital in which RVS’s services were offered but where the individuals did not receive a service, or for very similar hospitals, as defined by predictive risk models.

From within this group the Nuffield Trust will use prognostic matching and risk modelling techniques to find a ‘closest match’ control cohort for each of the seven intervention cohorts.

Once the Nuffield Trust have identified the matched control groups, the Nuffield Trust will analyse all subsequent hospital activity – following recruitment to the service - comparing the behaviour of the intervention group (the actual service recipients) to the matched control group.

The Nuffield Trust will focus especially on differences between the two groups in terms of emergency hospital admissions, readmissions, lengths of stay and A&E attendances.

Role of Nuffield Trust summary: The Nuffield trust will identify characteristics of all RVS service recipients using HES data from the period prior to recruitment to the service. The Nuffield Trust will find very closely matched control groups, from the wider population. The Nuffield Trust will finally test for differences in subsequent hospital activity between the service recipients and the matched control group.

Only substantive employees of the Nuffield Trust will have access to the data and only for the purposes described in this document.

Note about third parties: At no point will person level data be transferred from the Nuffield Trust to any other external organisation. All information published and/or passed to partner organisations will be aggregated with small numbers suppressed as required in HES guidance.

The Nuffield Trust shall ensure access to data disseminated by NHS Digital is strictly prohibited and must not be accessed by Wavex Technology.

The care of frail older people, and the role of the Comprehensive Geriatric Assessment (Refs: NIC-383324-D6B8T, previously HESR019) — DARS-NIC-383324-D6B8T

Type of data: information not disclosed for TRE projects

Opt outs honoured: Y, Identifiable, Anonymised - ICO Code Compliant (, )

Legal basis: Health and Social Care Act 2012, Approved researcher accreditation under section 39(4)(i) and 39(5) of the Statistical Registration Service Act 2007; Health and Social Care Act 2012 – s261(7), Health and Social Care Act 2012 – s261(1) and s261(2)(b)(ii), , Health and Social Care Act 2012 – s261(2)(b)(ii), Approved researcher accreditation under section 39(4)(i) and 39(5) of the Statistical Registration Service Act 2007 ; Health and Social Care Act 2012 – s261(7)

Purposes: No (Research)

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

When:DSA runs 2018-05-01 — 2021-04-30 2017.06 — 2017.08.

Access method: One-Off

Data-controller type: THE NUFFIELD TRUST FOR RESEARCH AND POLICY STUDIES IN HEALTH SERVICES

Sublicensing allowed: No

Datasets:

  1. Hospital Episode Statistics Admitted Patient Care
  2. Civil Registration (Deaths) - Secondary Care Cut
  3. Patient Reported Outcome Measures (Linkable to HES)
  4. Unmatched
  5. Civil Registrations of Death - Secondary Care Cut
  6. Hospital Episode Statistics Admitted Patient Care (HES APC)

Objectives:

The Nuffield Trust is an independent research group chaired by Andrew McKeon and overseen by a board of Trustees including a number of senior NHS clinicians, managers and academics. The Nuffield Trust undertake work for the public good and within a research governance framework.
In these studies Nuffield Trust are seeking to improve the NHS’s ability to identify and implement good practice in terms of efficient and effective health care for patients.
The projects are:
• Surveillance of outcomes and health service use for frail older people. This work is aiming to develop better indicators to identify good quality provision of care for older people. The focus of the first is to help the NHS spot effective care via innovative use of linked person-level datasets. The Nuffield Trust will try to identify areas of the country that are caring well for those aged over 65 with long term conditions and multiple co-morbidities. The project will develop new measures that make use of care pathways to provide a more sophisticated classification of patient types. The metrics will then be applied to the whole country to identify areas in England where the care pathways seem better than expected in terms of promoting better quality of care for older people. These areas will then be explored using qualitative methods to assess the nature of the successful service.

• An evaluation of the Comprehensive Geriatric Assessment (CGA) – Nuffield Trust were co-applicants with the University of Newcastle, and the work funded by the NIHR. The CGA is a multidisciplinary process designed to assess a frail older person’s medical conditions, mental health, functional capacity and social circumstances. However information is lacking on the types of patients that might benefit the most. The aim of this work is to describe existing models of care and to validate tools to deliver CGA on a hospital wide basis. No record level data is provided to the University of Newcastle.

Older people are the major users of acute hospitals, yet there is a growing perception that care for older people is sub-optimal (Patient’s Association report 2011, Health Service Ombudsman’s report 2011). Comprehensive geriatric assessment (CGA) is defined as ‘a multidimensional, interdisciplinary diagnostic process to determine the medical, psychological, and functional capabilities of a frail older person in order to develop a coordinated and integrated plan for treatment and long-term follow-up’. CGA improves outcomes for frail older people, including survival, cognition, quality of life and reduced length of stay, readmission rates, long term care use and costs.

CGA is the accepted gold standard method of caring for frail older people in hospital, documented in numerous randomised controlled trials and associated systematic reviews and meta-analyses. It is unclear, however, which types of patients benefit most and how CGA should best be targeted to achieve maximum impact.

In addition, individual patient characteristics and frailty markers that best predict improved outcomes from CGA is unknown. It is essential that factors identifying benefit can be derived from routine hospital data to facilitate service level evaluation of health outcomes and health and social care costs, but the feasibility of this is unknown. The number of people who might benefit from CGA is estimated to be between 15-50% of older inpatients, based on the prevalence of cognitive dysfunction as a marker for frailty.

This work will therefore explore clinical markers of frailty that are quick and simple to use, and can map to larger datasets from a health and social care perspective. This in turn will provide accurate data on the numbers involved in different hospital settings which are required to plan and resource appropriate models of service for the projected increasing admissions of older people to hospital.

Both studies will rely on analysis of older individuals who are likely to be frail and have multiple long term conditions. Information on mortality will be important for such a group, as death rates for specific subgroups are likely to be high.

PROMs data are requested so that Nuffield Trust would be able to carry out appropriate statistical analyses to help explore how well the results of the pre- or post-operative questionnaires correlate with proxy outcomes from HES. Although the PROMs data focus on patients awaiting one of the relevant procedures, Nuffield Trust expects to be able to make inferences about the general applicability of our results to the older population across England, as appropriate.

Analyses of linked HES, ONS and clinical datasets are part of a wider project to evaluate interventions to improve care of frail older people aged over 75 in hospitals. The analyses are focused on developing tools for identifying people who are frail in hospitals, understanding the implications of frailty for individuals and the health service, and quantifying the benefits of better care, namely “comprehensive geriatric assessment” which is an integrated management approach.

Date of death is required because survival is an important outcome for people over 75, it is common (up to half of frail older people die over a two year period), and it is a key indicator of the quality of care. It also has major implications for use of hospitals, which increases at end-of-life and is necessary to estimate the impacts of frailty on hospital use, as well as the potential effects of better care.

The Nuffield Trust wish to convert date of death into an anonymised form of the data, such as week of death, or mortality within specified time-frames from hospital admission to further aggregate the data.

Yielded Benefits:

The Nuffield Trust have contributed to a final NIHR report for the Hospital Wide (HoW) CGA study which will be submitted for peer review at the end of January, with publication expected in late 2018. This report will include a tool (excel based, reliant only on aggregated data with small numbers suppressed) for local NHS use - to determine how many people in an area might benefit from CGA. A paper on the development of a frailty risk score for older people in acute care was submitted to a peer reviewed journal in November 2017 and a decision is pending. Further papers for peer reviewed journals are planned including one focusing on long term outcomes of frailty in the community and one focused on hospital based populations. These are expected to be submitted to peer reviewed journal in Spring 2018.

Expected Benefits:

The benefits of the work are seen in terms of decisions made by healthcare commissioners and providers, when thinking about the types of services needed to deliver benefits to patients, as well as by policy makers.
• Surveillance of outcomes and health service use for frail older people
The frail older people work is aiming to develop better indicators to identify good quality provision of care for older people. The work consists of devising and testing new metrics - some of which are drawn from HES data - and then validating these in a number of ways:
- by identifying characteristic areas of good practice;
- by matching outcomes with PROMs;
- by engaging with a number of key experts in the field with whom the Nuffield Trust already have existing links, including via the CGA project that is also mentioned in this application.
In the long term the Nuffield Trust anticipate that the metrics they develop will be used by the care system to monitor and to promote good quality care for older people. Some of this will be integrated with the CGA project and benefit from the methods for dissemination outlined below.
Alongside the reports and peer-reviewed papers, key outputs will be tools that can be applied locally for looking at changes over time which will be supported by web-based materials. These tools will enable decision makers use their own data to monitor key outcome measures over time to identify, for example, whether new care initiatives are improving outcomes or quickly spot sudden deteriorations so that they can be acted upon in a timely fashion. The Nuffield Trust are known experts in this field: for example, the Trust currently have a request from NHS England for work in this area.
• Comprehensive Geriatric Assessment (CGA) evaluation.
The Nuffield Trust’s analyses and project deliverables will contribute to a much more detailed understanding of current models of inpatient care for older people in the UK. The Nuffield Trust will also improve the clarity of definition of the key elements of CGA and its use in hospital settings. Outputs will include a means of evaluating services for frail older people using HES data, a methodology for assessing the relationship between frailty markers and long-terms patient outcomes and costs, and a set of tools that will facilitate service redesign and long-term planning for commissioners (see below).
The Nuffield Trust will use a range of new and existing organisations to support the dissemination of finding from the project, for example Clinical Commissioning Groups (CCGs), NIHR Collaborations for Leadership in Applied Health Research and Care (CLAHRCs) (team members are involved with two CLAHRCs currently and have excellent networks with the remaining seven across England), and Academic Health Science Networks (AHSNs). AHSNs will play a pivotal role in promoting the uptake of innovation and best practice and the team will work with the AHSNs to promote adoption and spread of best practice recommendations and the use of resources to support change arising from the findings of the project.
The Nuffield Trust will produce a population based report in mid to late 2017. This will include a tool (Excel or web based, reliant only on aggregated data with small numbers suppressed) for local use – to determine how many people in an area might benefit from CGA. The tool will be freely available for use by any interested member of the public (not just commissioners), via download from the Nuffield Trust’s website. The tool will only consist of aggregated data derived from HES and will not use ONS data.

A further report about the relation of frailty measures to admissions will be produced in late 2017 early 2018.

The outputs of the above tool will be combined with a series of health system performance measures that relate specifically to the care of frail older people at area and provider level. These performance metrics, combined with the estimated number of CGA beneficiaries, will allow the Nuffield Trust to estimate the numbers, outcomes, resource use and costs for this group of service users. These estimates will then be used to produce the final “What if” interactive models to be provided to service providers and commissioners, allowing them to explore the scope for modifications to services to (for example) reduce costs or service utilisation, or evaluate the effectiveness of service interventions targeting the frail cohort. Specifically, the final tool would demonstrate the relationship between the scale of patient benefits, resource use and costs based on variable assumptions concerning: numbers receiving CGA, patient type and risk, and relative effectiveness.
In addition, which individual patient characteristics and frailty markers best predict improved outcomes from CGA are currently unknown. It is essential that factors identifying benefit can be derived from routine hospital data to facilitate service level evaluation of health outcomes and health and social care costs. The number of people who might benefit from CGA is estimated to be between 15-50% of older inpatients. Accurate data on the numbers involved in different hospital settings are required to plan and resource appropriate models of service for the projected increasing admissions of older people to hospital. These tools will permit commissioners to plan and resource accordingly.
In addition, by analysing differences between patients recorded as receiving CGA and those who are not, Nuffield Trust will also be able to evaluate the effectiveness of the CGA intervention, in terms of cost and patient outcomes, and also characterise groups of elderly patients who are most likely to benefit from CGA.

Outputs:

All outputs will be aggregate with small number suppressed in line with the HES Analysis Guide.

Anticipated dates of study reports are listed. All may also include presentational web material (for example slideshows and blog posts), in addition to presentations given in person at relevant research or policy conferences, etc.
• Surveillance of outcomes and health service use for frail older people
The Nuffield Trust anticipates producing a Nuffield Trust final research report to be available in mid to late 2017. At the same time, the Nuffield Trust also aims to produce papers for peer-reviewed journals such as BMJ Quality and Safety and Age and Ageing, as well as a paper targeted towards a more methodological journal. The Nuffield report will be placed in the public domain, and will contain data at no greater granularity than aggregate with small numbers suppressed.
• Comprehensive Geriatric Assessment (CGA) evaluation.
The Nuffield Trust will produce a population based report in mid to late 2017. This will include a tool (excel or web based, reliant only on aggregated data with small numbers suppressed) for local NHS use – to determine how many people in an area might benefit from CGA. A further report about the relation of frailty measures to admissions will be produced in late 2017 early 2018.

This HESIDs will enable the primary outputs of the research programme to be generated. Specifically Nuffield Trust will be able to determine the extent to which clinical frailty scales predict short, medium and longer term hospital use, survival, the frequency of emergency events and costs related to health activity.

Processing:

Only substantive employees of the Nuffield Trust will have access to the data and only for the purposes described in this document.

The ONS data will be processed in accordance with the standard ONS terms and conditions.

Data will be matched to the pseudonymised/non-sensitive HES data held by the Nuffield Trust under existing agreement NIC-384572-J7P6Y.

The NHS Digital technical team will provide the requested datasets using Nuffield Trust’s encryption to enable them to link the data with the HES data they already hold.

• Surveillance of outcomes and health service use for frail older people. Specific elements of the work will;
1. Develop a series of metrics (indicators), from routinely collected data that reflect the quality of care of older people.
2. Validate those metrics against patient reported health status and clinical databases.
3. Use the validated metrics as surveillance indicators to find regions of high quality care of older people
4. Confirm the surveillance findings by partnering with the relevant authority to understand the local context
5. If evidence of high quality care is established then an in-depth qualitative analysis will be performed to detail the local arrangements and assess the potential for transferability

• Comprehensive Geriatric Assessment (CGA). Specific elements of the work will;
1. Identify the characteristics of recipients of CGA in existing settings
2. Identify the characteristics of those who appear to benefit from CGA
3. Stratify local populations to test who might benefit from CGA more widely
4. Develop interactive tools for providers and commissioners to explore scope for service modifications
5. Assess the relationship between frailty markers (recorded and proxy) and long term care outcomes

Amendment (data linkage update) for CGA analysis:
NHS Digital will receive identifiable person level information from each one of three universities (Newcastle University, Southampton University and Nottingham University) who hold detailed clinical datasets relating to frail older people, containing clinical frailty markers and whether the patient received CGA.
The transferred information will consist of only: NHS number, full name, date of birth and address including postcode, in addition to a non identifiable linkage ID. This data is required in order to ensure a robust match at NHS Digital to ensure a high standard in data quality. NHS Digital will receive no other information about any service recipient.
This transfer of information to NHS Digital will occur just once. It is anticipated that there will be 2,000 individuals in total.
This data will be transferred to NHS Digital using NHS Digital’s own secure transfer facilities, under NHS Digital advice.
NHS Digital will process the person identifiers. For each service recipient, they will find the relevant pseudonymised identifier, the HESID, in the form held by the Nuffield Trust.
NHS Digital will produce a file for the Nuffield Trust. This file will contain the HESID of each service recipient, alongside other limited pseudonymised information: LSOA of residence, age and gender. It will also include information about the matching technique, and the non identifiable linkage ID.
NHS Digital will finally transfer this file securely to the Nuffield Trust.
The data received from NHS Digital will then be linked, via the non-identifiable linkage ID provided by the sites, to the de-identified clinical datasets held by the Nuffield Trust, and to HES and ONS mortality data via the HESID.

Additional note – third parties
The Nuffield Trust are not seeking permission for any third parties to access these data, even where these third parties are study partners. The use of this data will be limited to Nuffield Trust for the purposes outlined above only. Data published or provided to third parties will be limited to aggregated data, at area, organisational or cohort-level all with small numbers suppressed in line with HES analysis guide.
The Nuffield Trust will perform its analyses using statistical software including SAS, R and stata.
The Nuffield Trust shall ensure access to data disseminated by NHS Digital is strictly prohibited and must not be accessed by Wavex Technology.

Project 3 — DARS-NIC-18674-W8Q8K

Type of data: information not disclosed for TRE projects

Opt outs honoured: N ()

Legal basis: Informed Patient consent to permit the receipt, processing and release of data by the HSCIC

Purposes: ()

Sensitive: Non Sensitive

When:2017.03 — 2017.05.

Access method: One-Off

Data-controller type:

Sublicensing allowed:

Datasets:

  1. Hospital Episode Statistics Admitted Patient Care

Objectives:

The Nuffield Trust’s overarching purpose is to help provide objective research and analysis that boosts the quality of health policy and practice, and ultimately improves the health and health care of people in the UK. Nuffield Trust is an independent research group overseen by a board of Trustees including a number of senior NHS clinicians, managers and academics.

This application relates to a piece of evaluation work the Nuffield Trust is undertaking funded by Age UK, to evaluate Age UK’s Integrated Care Programme.

The role of the Nuffield Trust will be to provide an independent evaluation of Age UK’s Integrated Care Programme (ICP) across several areas in England (Portsmouth, N Tyneside, Blackburn, E Lancashire, Guildford, Sheffield, Kent). In order to do this we propose using linked data to compare the outcomes for project beneficiaries versus a matched control group who do not receive the schemes’ services. The primary outcome measures will be hospital activity in the period after referral to the schemes.

The Age UK programme commenced in Summer 2015, and will run until Autumn 2017. The key output of our work will be an independent assessment of the strength of evidence that Age UK ICP has any favourable impact on older people’s use of hospital services.

The Nuffield Trust will not be sharing any data with any third party. All published or otherwise shared information will be at highly aggregated levels with small number suppressed as required in HES guidance.

Expected Benefits:

The Nuffield Trust evaluation will provide an independent assessment of the impact of Age UK’s Integrated Care Scheme. This scheme aims to use volunteers and voluntary sector staff to help support older people’s formal and informal care choices. It is one of a number of similar schemes increasingly being commissioned by local authorities, CCGs and acute trusts in England.

The evidence the Nuffield Trust provide will be of benefit to these commissioners, both in England and beyond, as they consider funding these types of services over others on a longer-term basis. A local pilot study in Cornwall reported large falls in emergency admissions after referral to the Age UK ICP. The Nuffield Trust want to test the interpretation of this study and believe that this analysis (with larger numbers and adjustment for regression to the mean and other factors) will be able to provide a more definitive view on the nature of the scheme’s impact on hospital care activity.

The evaluation itself will measure the impact of the schemes on hospital usage: including emergency admissions, re-admissions, length of stay and A&E attendances over ~9 months post referral.
Evidence of the success or failure of the service will be beneficial to patients in the wider sense that if such schemes are shown to lead to improved care, similar schemes will be more likely to receive funding from local commissioners elsewhere. This analyses may help these schemes identify specific subgroups of individuals who might be best targeted for care.

This study will be a useful addition to a growing set of Nuffield Trust evidence on the impact of the voluntary sector on use of the acute sector in England: evaluation of British Red Cross ‘Support at Home’ scheme (2012); evaluation of a Cabinet Office ‘social action’ fund (findings to be published summer 2016), evaluation of Stroke Association’s Stroke Recovery Service (findings due 2017).

Outputs:

The prime analytical outcome of the Nuffield Trusts data processing will be the identification of a group of people (within the HES data) who had extremely similar characteristics to the seven projects’ service recipients, but who didn’t receive a service. For example if 3,000 people by Autumn 2016 received a service from Age UK, Nuffield will have identified this group in HES data, in addition to another 3,000 individuals to be used as the evaluation’s matched control group.

Differences between these two groups in terms of their subsequent use of hospital services will be analysed to test whether there is any evidence that the services have had any impact on admissions to hospital or lengths of stay, adjusting for remaining differences between the two groups.

A final study report will be submitted to Age UK in summer 2017. Shortly afterward, The Nuffield Trust will publish the results in a Nuffield Trust branded research report, alongside helpful information (for example web based presentations). These will be made freely available on the Nuffield Trust website. The Nuffield Trust may also submit our findings to quality peer reviewed journals.

Processing:

The role of NHS Digital:
NHS Digital will receive identifiable person level information from each one of the seven local Age UK organisations.

The person level information received by NHS Digital will cover all people recruited to the ICP service between Summer 2015 and Autumn 2016, where those individuals have consented to sharing data with NHS Digital for evaluation purposes (subsequently referred to as service recipients).

The transferred information will consist of only:
• NHS number (if available)
• full name
• date of birth
• address including postcode
• gender
in addition to a non identifiable client ID added by the services to denote each unique individual.

NHS Digital will receive no other information about any service recipient. The extent, therefore, of any risk of NHS Digital receiving this personal information is that NHS Digital will have knowledge that person A, for example, received a service from Age UK’s ICP, but they will not know what specific service was received, or when.

This data will be transferred to NHS Digital using NHS Digital's own secure transfer facilities.

NHS Digital will process the person identifiers. For each service recipient, they will find the relevant pseudonymised identifier, the HESID, in the form held by the Nuffield Trust.

NHS Digital will produce a file intended for the Nuffield Trust. This file will contain the HESID of each service recipient, alongside other limited pseudonymised information:
• LSOA of residence,
• age (/year of birth)
• gender,

It will also include information about the matching technique, and the non identifiable client ID.
NHS Digital will finally transfer this file securely to the Nuffield Trust.

Role of NHS Digital summary: NHS Digital will receive personal identifiers of people recruited to seven Age UK services, and will provide the relevant pseudonymised HESIDs of these individuals to the Nuffield Trust.

The role of the Nuffield Trust:
The Nuffield Trust will receive from NHS Digital the list of HESIDs of people recruited to the services between Summer 2015 and Autumn 2016 (service recipients).

The Nuffield Trust will receive at the same time a dataset from each of the seven local Age UK organisations. These datasets will contain no identifiable information, and will only contain the non identifiable client ID as a person identifier (note that consent is also received from all service recipients to share data with the Nuffield Trust).

Other data contained in these files will include non identifiable details of services received by the people recruited to these services, e.g. date of referral to the service provider, number of minutes of support received, dates of services provided, etc.

These data will be transferred securely to the Nuffield Trust using our own secure file transfer (SFTP) system.

As a first processing step Nuffield Trust will link the HESIDs (from NHS Digital) and the non identifiable service information (from the seven service providers) using the non identifiable client ID.

The Nuffield Trust will then link this data to pseudonymised HES data

For each service recipient the Nuffield Trust will link to up to three years of HES data for the period just prior to the referral to the service provider. The Nuffield Trust will also link to all subsequent HES activity captured using the latest quarterly HES datasets.

Using this data the Nuffield Trust will build person level analysis datasets which will characterise each service recipient in terms of: demographic characteristics, history of hospital use, and morbidity characteristics.

The key aim of the Nuffield Trust's processing is to find a matched group of people who very closely share these same characteristics, but who did not receive a service from the service providers, and to use this group as a pseudo-control group.

To do this the Nuffield Trust will build analysis datasets for the wider group of people in the areas in which these services were offered but where the individuals did not receive a service from the service provider, or for very similar areas.

From within this group the Nuffield Trust will use prognostic matching and risk modelling techniques to find a ‘closest match’ control cohort for each of the seven intervention cohorts.

Once the Nuffield Trust have identified the matched control groups, they will analyse all subsequent hospital activity – following recruitment to the service - comparing the behavior of the intervention group (the actual service recipients) to the matched control group.

The Nuffield Trust will focus especially on differences between the two groups in terms of emergency hospital admissions, readmissions, lengths of stay and A&E attendances.

Role of Nuffield Trust summary: The Nuffield Trust will identify characteristics of all Age UK ICP service recipients using HES data from the period prior to recruitment to the service. The Nuffield Trust will find very closely matched control groups, from the wider population. The Nuffield Trust will finally test for differences in subsequent hospital activity between the service recipients and the matched control group.

Note about third parties: At no point will person level data be transferred from the Nuffield Trust to any other external organisation. All information published and/or passed to partner organisations will be aggregated as required in HES guidance.

Project 4 — DARS-NIC-384572-J7P6Y

Type of data: information not disclosed for TRE projects

Opt outs honoured: No - data flow is not identifiable (Does not include the flow of confidential data)

Legal basis: Health and Social Care Act 2012, Health and Social Care Act 2012 – s261(1) and s261(2)(b)(ii)

Purposes: ()

Sensitive: Non Sensitive

When:2017.06 — 2017.02.

Access method: Ongoing

Data-controller type:

Sublicensing allowed:

Datasets:

  1. Hospital Episode Statistics Accident and Emergency
  2. Hospital Episode Statistics Admitted Patient Care
  3. Hospital Episode Statistics Outpatients

Objectives:

The Nuffield Trust is an independent research group overseen by a board of Trustees including a number of senior NHS clinicians, managers and academics. The Nuffield Trust undertakes work for the public good and within a research governance framework.

The purposes for receiving HES data falls into the following categories :-
1. Evaluations of the impact of innovations in health and social care on hospital utilisation

In an effort to improve the quality of health care and reduce the financial pressure on the NHS, efforts are being made to deliver more care in community settings, with the aim of preventing unnecessary and expensive admissions to hospital. The Nuffield Trust is developing methods to evaluate how well these interventions perform.

The projects are:
• Evaluation of the Integrated Care ‘Pioneers’. These are models of care aimed at reducing the impact of boundaries between care providers. This work is in partnership with the DH Policy Innovation Research Unit based at the London School of Hygiene and Tropical Medicine.
• Evaluation of new models of primary care. This is an evaluation of selected ‘scaled up’ GP practices, to explore development of new general practice organisations.
• Evaluation of a local scheme to deliver improved care for complex cases – as part of the PM’s Challenge Fund in Barking, Havering and Redbridge (funded by PM’s Challenge Fund).

Additional projects may be added to this list but will be subject to an amendment to this agreement being approved by NHS Digital.

2. Research studies involving the surveillance of patterns in hospital admission and costs at area level in England, aimed at identifying areas where innovation in service delivery is taking place.

In these studies The Nuffield Trust are seeking to improve the NHS’s ability to identify and implement good practice in terms of efficient and effective health care for patients.

The projects are:
• A DH funded project to look at ways to identify people who suffered avoidable serious harm. This will test whether HES data can be used as a screening tool to identify cases for specific audit. This work will be undertaken in conjunction with the London School of Hygiene and Tropical Medicine and Imperial College London.
• Patterns of urgent care use related to the development of ambulatory emergency medicine, a new approach being adopted in hospitals around the country. The Nuffield Trust will look at the impact on patterns of acute hospital use and long terms outcomes for patients. Update May 2016: This project is now closed and we no longer require processing data for this use (apart from auditing purposes).
• An assessment of the impact of alcohol on hospital services. This study aims to evaluate whether alcohol is an increasing burden on acute hospital services, will attempt to identify geographical areas where hospital alcohol teams are working well and aims to share opportunities identified for improvement. UPDATED INFORMATION OCT 2016 : This project is now closed and no longer requires the processing data for this use (apart from auditing purposes).
Additional projects may be added to this list but will be subject to an amendment to this agreement being approved by NHS Digital.

3. Research studies relating to the efficiency of health services and level of competition in the English NHS.

The projects:
• Descriptions of differential patterns of hospital use (admissions, lengths of stay, child health indicators) by area and provider. This work aims to identify good practice in the efficient use of hospital resources.

4. Linkage of HES data to linked datasets provided by NHS Digital

A number of other projects require access to HES data, but also make use of a linked dataset such as HES-ONS. These are subject to separate applications to NHS Digital, but would not require a separate data release. Instead the applicant would make use of the data provided under this agreement. Such projects typically cover 10 years of HES data.

5. Linkage of HES data to HES ids provided by NHS Digital regarding a specific cohort.

Such uses would again be subject to separate data agreements, since they would have a different legal basis and potentially involve patient consent. They would be considered separately under different applications but would not require a separate data release.

6. Informing the public debate about hospital use

The Nuffield Trust regularly acts to improve the quality of public debate on use of hospital services by publishing responsive research, which helps focus the debate on evidence. Trigger for this work include a specific issue suddenly coming to national prominence, or an individual or organisation making an assertion which is easily tested using data already available. As an independent research organisation and registered charity, with independence from party politics overseen by the board of trustees, such interventions are carefully considered to ensure that an evidence-based statement may add value to the overall debate. They are not provided at the request of any individual organisation.

7. Evaluations of the impact of innovations in health and social care on hospital utilisation

In an effort to improve the quality of health care and reduce the financial pressure on the NHS, efforts are being made to deliver more care in community settings, with the aim of preventing unnecessary and expensive admissions to hospital. The Nuffield Trust is developing methods to evaluate how well these interventions perform.

The project involves an evaluation of Virtual Wards in Devon. This study aims to evaluate a multidisciplinary care management scheme which was delivered to individuals in their homes. This work is intended to follow up an earlier study funded by NIHR which was only able to capture the first hundred patients admitted to the Virtual Ward. Over subsequent years, six thousand individuals have been admitted to the scheme. These were chosen as they were judged to have a high risk of hospitalisation. Nuffield will test whether the post-Virtual Ward hospital admissions were low within this group, compared to a similar cohort of people from other parts of the country selected from HES. UPDATED INFORMATION OCT 2016: This project is now closed and no longer requires processing of the data for this use (apart from auditing purposes).

8. Research studies involving the surveillance of patterns in hospital admission and costs at area level in England, aimed at identifying areas where innovation in service delivery is taking place.

In such studies The Nuffield Trust are seeking to improve the NHS’s ability to identify and implement good practice in terms of efficient and effective health care for patients.
The project involves evaluating the implementation of Quality Standards in the London region, which were developed and introduced to reduce variability in and improve patient care. The study aims to:
- investigate variation in outcomes and adherence to standards across London
- determine if there is an association between degrees of standard implementation and outcomes
- evaluate innovative outcome measures to investigate the impact of standards on patient care and for monitoring standard adherence

UPDATED ADDITIONAL PROJECT information OCT 16

The following describes new projects the Nuffield Trust are requesting access to HES data for;

9. Research studies identifying models of medical generalism used in smaller hospitals and exploring their strengths and weaknesses in treating older and more complex patients from patient, professional and service perspectives.

The rising numbers of older and more complex patients is considered to be one of the most pressing problems facing the NHS. Although they receive the most resource-intensive care, their problems are less likely to be accurately diagnosed and have more adverse outcomes than other age groups. The emerging consensus is that current models of hospital care, which are heavily based around specialists delivering disease-specific care, serve these patients poorly, as it is often fragmented and poorly co-ordinated. A revival of medical generalism has been suggested to provide better and more cost-effective care. The reality, however, is that there is a paucity of evidence on which to base new models of medical generalism. Smaller hospitals provide an ideal environment in which to investigate models of medical generalist care, as their patient population is older and more vulnerable, while their size creates constraints on their income, capacity and staffing.

The overarching aim of this research, therefore, is to identify the models of medical generalism used in smaller hospitals and explore their strengths and weaknesses from patient, professional and service perspectives. More specifically, Nuffield Trust will be using HES data to create a classification of patients that might benefit from general medical care and, based on this classification, provide a descriptive analysis of the workloads of smaller hospitals.

10. Classifying readmissions and comparing readmission rates between the Netherlands and the UK based on national administrative data

The number of unplanned emergency readmissions to hospital have often been cited as a marker of quality of hospital care. Indeed, in England, readmissions have been used to influence hospital reimbursements. A number of studies, however, have shown that readmissions are complex and can be linked with a range of factors other than preventable or avoidable harms. If the quality of care at hospital level and individual patient characteristics are not the sole drivers of readmission then the additional factors must lie in the way different health systems manage patients. One way to consider the impact of systemic differences in health systems is to use international comparisons as a form of a natural experiment to see if patterns of readmission are similar or different. Though such comparisons cannot definitely identify the reasons behind differences, they can prompt useful questions on the effects of different health systems.

In this analysis Nuffield Trust want to test whether two different health systems demonstrate a fundamentally different pattern of hospital readmissions. Within any health system there are variations in readmissions rates between areas - the product of a host of patient and health system level factors influencing decisions and resource use. In order to understand the impacts of the broader health systems then we need to consider the overall distributions of readmissions and standardise – as far as possible – for differences at the patient level attributable to the underlying health problems.

The aim will be to analyse HES data covering admissions to NHS hospitals for selected years and calculate overall readmission rates. In parallel we will calculate the equivalent readmission metrics using the Dutch national data for the same time period. Nuffield Trust will then test for statistically significant differences in readmission rates between the two countries, and quantify to what extent any differences can be explained by patient-level factors (e.g. age, deprivation, severity) and the extent to which any variation could be explained by differences in the two health systems. Nuffield Trust will also attempt to distinguish between potentially preventable readmissions and other reasons for readmissions, using the administrative data, and produce a comparison between the two systems.

Expected Benefits:

Over the past five years The Nuffield Trust’s research studies, using NHS data, have been widely used to inform decision making and debate in health care. The Nuffield Trust publishes their reports on the Nuffield website and in peer reviewed journals where appropriate.

The Nuffield Trust list below recent reports of studies where they have made use of HES data;

There are many examples of The Nuffield Trust’s work being cited in parliamentary debates and select committees as well as used by national bodies including the Department of Health and NHS England, CQC and Monitor. Many of the projects have been funded by the Department of Health and NHS, and The Nuffield Trust work in partnership with NHS and other care organisations and with universities. The Nuffield Trust has also provided examples of their studies for NHS Digital to use as evidence to the health select committee.

The benefits of The Nuffield Trust’s work are seen in terms of decisions made by healthcare commissioners and providers, when thinking about the types of services needed to deliver benefits to patients, as well as by policy makers.

The following provides benefits for each of the projects;

1. Evaluations of the impact of innovations in health and social care on hospital utilisation.

The study of Integrated Care Models will inform health care providers, commissioners and policymakers of the impacts of new forms of integrated care emerging under the banner of national ‘Pioneer projects’. This is a piece of applied research funded by DH and with a very wide audience. Throughout the study, the Nuffield Trust will be engaging with the Pioneer sites themselves and with others interested in developing new models of care. This aligns with national policy goals to provide a better care experiences through integration of services.

Over the past few years a range of new forms of GP organisations have emerged. These are seen as one solution to systemic problems facing primary care; however there is little empirical analysis of these new organisational forms. The Nuffield Trust’s study of new models of primary care will be looking at the impact of networks of practices in terms of the way they have changed patient care and service delivery. Such analysis will be critical to future NHS planning on how to organise primary care services. The results will be of key interest all those involved in general practice, primary care and the commissioning of out of hospital services.

UPDATE OCT 2016: As part of realising the expected benefits the Nuffield Trust has presented preliminary results to members of the New Models of Primary Care Network at quarterly meetings.
Improving access to primary has been a matter of concern at the highest levels of government. In 2013 the PM created a Challenge Fund to look at new models of care. As evaluation partners of Barking, Havering and Redbridge’s successful bid to adapt services in the local area, The Nuffield Trust’s work will give provide a judgement about how successful these schemes might be. This will directly impact on local commissioners, and have implications across the country.

2. Research studies involving the surveillance of patterns in hospital admission and costs at area level in England, aimed at identifying areas where innovation in service delivery is taking place.

In all such studies The Nuffield Trust are seeking to improve the NHS’s ability to identify and implement good practice in terms of efficient and effective health care for patients. The Nuffield Trust’s work on serious avoidable harms has the possibility of contributing to safer care and better methods for future monitoring systems.

Ambulatory emergency care (AEC) provides a model of care for patients who have urgent care needs, but do not necessarily warrant an acute hospital admission. Though a number of local evaluative studies have been undertaken there are no systematic analyses across a range of organisations providing AEC. Through analysing the impacts of these schemes we can identify which types are most successful in delivering better care for patients. These can serve as models of success for areas wishing to develop their own services.

UPDATE OCT 2016: As part of realising the expected benefits the Nuffield Trust has presented findings at British Association of Ambulatory Emergency Care Annual National Conference 2015. This project is now closed andno longer requires processing data for this use (apart from auditing purposes).

The Nuffield Trust’s study of alcohol aims to understand the burden on hospital services and to identify areas where preventative action is working well. This work will consider effectiveness of different models of service provision which will meet a specific recommendation for future work made by Public Health England following their survey of hospital alcohol teams. The findings will be of interest to Health and Wellbeing Boards and to local commissioners of alcohol services and aims to share learning of how the burden of alcohol to hospitals can be reduced.

UPDATE OCT 2016: As part of realising the expected benefits the Nuffield Trust has –
(1) Produced a paper "Alcohol-specific activity in hospitals in England" published on 22nd Dec 2015, along with blog "The sobering burden of alcohol on the NHS".
(2) Submitted an academic paper to the BMC Public Health "The impact of alcohol care teams on emergency secondary care use following a diagnosis of alcoholic liver disease - a national difference-in-difference study." on 12/1/2016 .
(3) Public Health England Conference abstract poster presentation "Hospital use before and after first recorded diagnosis of alcohol related liver disease in England: Opportunities for early intervention to reduce harm".
This project is now closed and no longer requires processing data for this use (apart from auditing purposes).

3. Research studies relating to the efficiency of health services and level of competition in the English NHS.

These pieces of work aim to identify good practice in the efficient use of hospital resources. The Nuffield Trust’s future modelling will be used (as their past modelling currently is) to inform public debate about health service provision and central planning assumptions about future needs.
UPDATE OCT 2016: As part of realising the expected benefits the Nuffield Trust has, through the QualityWatch (http://www.qualitywatch.org.uk/) programme, produced a series of reports, carried out several activities and had the following impact. This includes –
1. Hospital admissions from care homes (29th January 2015). Subsequent to publication this was referenced in the evidence for NICE guidance: Transition between inpatient hospital settings and community or care home settings for adults with social care needs (November 2015)
2. Mental ill health and hospital use (14th October 2015). Nuffield Trust has presented the findings at the Public Health England Conference (September 2015) and at the International Conference of Integrated care (May 2016).
Nuffield Trust has also had several contacts with local healthcare providers for guidance on how to apply the methods to carry out the same analysis locally.
Whilst it is hard to draw causality, Nuffield have also been made aware of NHS England undertaking work to drive improvements in the quality of physical health care provided by mental health providers to service users with severe and enduring mental ill health and will be developing a national clinical audit to underpin this under the National Mental Health CQUIN scheme for 2016/17.

4. Linkage of HES data to linked datasets provided by HSCIC
Individual benefits will be covered within separate applications

5. Linkage of HES data to HES IDs provided by HSCIC regarding a specific cohort
Individual benefits will be covered within separate applications

6. Informing the public debate about hospital use
The Nuffield Trust’s responsive analyses will improve the quality of public debate on hospital use by broadening the available evidence base, focusing the debate on evidence rather than assertion and potentially preventing poor policy decisions. This work is widely reported in the media and Nuffield regularly meet with senior policy-makers and leaders in the NHS to discuss thiswork.

7. Evaluation of Virtual Wards in Devon
The evaluation of Virtual Wards fits within a suite of work the Nuffield are undertaking on community based alternatives to hospital care. An earlier study of ours (funded by NIHR) included a very early assessment of Devon’s virtual wards [1], but the number of people recruited was too small to make any robust assessment of whether the scheme had reduced future unplanned admissions. In the following years 6,000 primarily older people have been provided with care in a home based virtual ward. This cohort size should give us the power to detect significant differences, even where these are relatively small. The overall aim is to provide evidence to the health service on methods of care which help people to stay independent of hospital for longer.
[1] Lewis GH, Georghiou T, Steventon A, Vaithianathan R, Chitnis X, Billings J, et al. Analysis of virtual wards: a multidisciplinary form of case management that integrates social and health care. Final report. NIHR Service Delivery and Organisation Programme; 2013.
UPDATE OCT 2016: This project is now closed and no longer requires processing data for this use (apart from auditing purposes)

8. Evaluating the implementation of Quality Standards in the London region
This work aims to assess the extent of the patient benefits delivered by the introduction of new quality standards and to evaluate a new set of outcome indicators that can be used to more accurately measure their benefits. Applying these standards has a potential direct impact on the quality of over a million care episodes a year – over 500,000 patients. Moreover the London standards will be incorporated into the national Keogh standards. As a consequence this work on standards and outcomes will be of relevance to every hospital in the NHS with prospects of being incorporated into national audit tools. Nuffield will be presenting the outcomes in peer-reviewed reports and using Nuffield's extensive network of contacts to target these towards key decision makers. This will be facilitated by the Trust’s communications team who are very experienced with this type of activity. There is also direct interest from NHS London who have expressed an interest in funding further pieces of work in this area.

UPDATED ADDITIONAL PROJECTS OCT 2016 – new projects the Nuffield Trust are requesting access to HES data for.

9. Research studies identifying and evaluating models of medical generalism

Three separate panels of experts, the Independent Commission for the Royal College of General Practitioners and the Health Foundation, The Royal College of Physicians of London’s (RCPL) Future Hospital Commission (FHC) and the General Medical Council’s Shape of Training review, have all recommended a revival in general medicine to better provide high-quality, cost-effective care. However, general medicine cannot be rapidly reintroduced, nor hospitals, let alone smaller ones, reconfigure services unless there is a clear understanding of patient need and how different models of current medical care meet these.

With NHS England’s Viable Smaller Hospitals workstream of the New Models of Care programme already underway, there is an urgent need for clear and comprehensive evidence to guide future policy and service reconfiguration.

This study has the potential to have a major impact at the national and international levels, as the debate so far around medical generalism has operated largely at an abstract level. This research will provide much needed evidence to ground the debate in the empirical and experiential realities of underlying patient need. It will provide a theoretically informed evidence base from which to take the debate forward. This will not only have relevance for unscheduled adult medical care, but could also be applied to other clinical areas, such as surgery and paediatrics. It also has relevance to larger hospitals, as well as informing much broader, international debates around matching medical workforce to growing patient need.
This research should influence decision making around:
• Ways of working in hospitals
• The education of doctors at undergraduate and graduate level
• The number and types of doctors required in the UK
• Continuing professional development for physicians
• Legislative and contractual arrangements for doctors
• The future of smaller hospitals and their role in the wider healthcare system

10. Classifying readmissions and comparing readmission rates between the Netherlands and the UK based on national administrative data

Despite the prominence of readmission rates in discourse around health service organisation, there are few empirical studies that compare the rates of readmission across different health systems.

This analysis will provide a thorough comparison between readmission rates in England and the Netherlands, and also allow a deeper understanding of which readmissions are truly preventable and those that occur for other reasons. Readmissions due to complications are a burden for patients and account for high healthcare costs: this research will provide insight into the types of readmissions that should really be included in hospital-level indicators to get a valid representation of quality improvement potential.

By way of further evidence as to the use made of HES, a non-exhaustive list of recent publication using HES data is provided below. All such articles are in the public domain, and many relate directly to current health practice or topics of interest, and have been commissioned by the NHS or the Department of Health.
Nuffield Trust Research Group - Recent publications that have used HES (October 2014)
• Holly Dorning and Martin Bardsley. Focus On Allied Health Professionals. Nuffield Trust September 2014.
• Theo Georghiou and Martin Bardsley. Exploring the costs of end of life care. Nuffield Trust September 2014.
• Ian Blunt, Martin Bardsley & Giovanni FM Strippoli. Predialysis hospital use and late referrals in incident dialysis patients in England: a retrospective cohort study. Nephrology Dialysis Transplantation. Nephrol. Dial. Transplant. (2014) doi: 10.1093/ndt/gfu284
• Steventon A, Bardsley M, Doll H, Tuckey E, Newman S. Effect of telehealth on glycaemic control: analysis of patients with type 2 diabetes in the Whole Systems Demonstrator cluster randomised trial. Submitted for publication. BMC Health Services Research. June 2014
• Steventon A, Bardsley M and Mays N. Effect of a telephonic alert system (Healthy Outlook) for patients with chronic obstructive pulmonary disease: cohort study with matched controls. Journal of Public Health Advance Access published July 10, 2014. pp. 1–9 doi:10.1093/pubmed/fdu042
• Blunt I, Bardsley M, Grove A, et al. Classifying emergency 30-day readmissions in England using routine hospital data 2004–2010: what is the scope for reduction? Emerg Med J Published Online First 26/3/14 doi:10.1136/emermed-2013-202531
• Lewis, G. H., Vaithianathan, R., Wright, L., Brice, M. R., Lovell, P., Rankin, S., & Bardsley, M. (2013, November 4). Integrating care for high-risk patients in England using the virtual ward model: lessons in the process of care integration from three case sites. International Journal of Integrated Care. Retrieved from http://www.ijic.org/index.php/ijic/article/view/URN%3ANBN%3ANL%3AUI%3A10-1-114754/2197
• Lewis GH, Georghiou T, Steventon A, Vaithianathan R, Chitnis X, Billings J, et al. Analysis of virtual wards: a multidisciplinary form of case management that integrates social and health care. Final report. NIHR Service Delivery and Organisation Programme; 2013.
• Steventon A, Tunkel S, Blunt I and Bardsley M. Effect of telephone health coaching (Birmingham OwnHealth) on hospital use and associated costs: cohort study with matched controls. BMJ. 2013 Aug 6;347:f4585. doi: 10.1136/bmj.f4585.
• Bardsley M, Doll H and Steventon A. Impact of telehealth on general practice contacts: findings from the whole systems demonstrator cluster randomised trial BMC Health Services Research 2013, 13:395 doi:10.1186/1472-6963-13-395
• Roberts A, Blunt I, Bardsley M. Focus On: Distance from home for emergency care. QualityWatch Report May 2014. Nuffield Trust/Health Foundation
• Georghiou T, Cooke M, & Bardsley M. How Representative Are Patients Who Access the Marie Curie Nursing Service of the Population of People Who Die Each Year in England? BMJ Supportive & Palliative Care, 3(1), 134–134. doi:10.1136/bmjspcare-2013-000453b.27
• Natasha Curry, Matthew Harris, Laura Gunn, Yannis Pappas, Ian Blunt, Michael Soljak, Nikolaos Mastellos, Holly Holder, Judith Smith, Azeem Majeed, Agnieszka Ignatowicz, Felix Greaves, Athina Belsi, Nicola Costin-Davis, Jessica D Jones Nielsen, Geva Greenfield, Elizabeth Cecil, Susan Patterson, Josip Car, Martin Bardsley Integrated care pilot in north west London: a mixed methods evaluation. Int J Integr Care 2013; Jul–Sep, URN:NBN:NL:UI:10-1-114735
• Chitnis, X. , Georghiou, T., Steventon, A., & Bardsley, M. J. (2013). Effect of a home-based end-of-life nursing service on hospital use at the end of life and place of death: a study using administrative data and matched controls. BMJ Supportive & Palliative Care, 1–9. doi:10.1136/bmjspcare-2012-000424
• Bardsley M, Blunt I, Davies S, Dixon J. Is secondary preventive care improving? Observational study of 10 year trends in emergency admissions for conditions amenable to ambulatory care. BMJ Open 2013; :e002007. doi:10.1136/bmjopen-2012-002007
• Davies A, Chitnis X, Bardsley M. Hospital activity and cost incurred due to unregistered patients in England: considerations for current and new commissioners. J Public Health first published online December 19, 2012 doi:10.1093/pubmed/fds098
• Clarke A, Blunt I, Bardsley M. Analysis of emergency 30 day readmissions in England using routine hospital data 2004-20010.Is there scope for reduction. Presented the Society for Social Medicine Annual Scientific Meeting. Journal of Epidemiology and Community Health. September 2012 Supplement. Doi:10.1136/jech-2012-201753.117
• Billings J, Blunt I, Steventon A, Georghiou T, Lewis G, Bardsley M. Development of a predictive model to identify inpatients at risk of readmission within 30 days of discharge (PARR-30). BMJ Open. 2012;00:e001667. doi:10.1136/bmjopen-2012-001667.
• Roland M, Lewis R, Steventon A, Adams J, Bardsley M, Brereton L, Chitnis X, Staetsky L, Tunkel S, Ling T. Case management for at-risk elderly patients in the English Integrated Care Pilots: observational study of staff and patient experience and secondary care utilisation. International Journal of Integrated Care – Volume 12, 24 July – URN:NBN:NL:UI:10-1-113731 / ijic2012-130 – http://www.ijic.org/
• Steventon A, Bardsley M, Billings J, Dixon J, Doll H, Hirani S, Cartwright M, Rixon L, Knapp M, Henderson C, Rogers A, Fitzpatrick R, Hendy J, Newman S. Effect of telehealth on use of secondary care and mortality: findings from the Whole System Demonstrator cluster randomised trial. BMJ 2012;344:e3874.

Outputs:

Anticipated dates of study reports are listed. All may also include presentational web material (for example slideshows and blog posts), in addition to presentations given in person at relevant research or policy conferences, etc.
1. Evaluations of the impact of innovations in health and social care on hospital utilisation
• - Integrated Care Models (Pioneers).Interim report in 2016, with annual reports from 2016 to 2020 (final report in 2020).
- New models of primary care
• Interim report, September 2015. Final Nuffield Trust research reports will be available Autumn 2016. Update April 2016: Nuffield Trust publication due in June 2016 and planning a subsequent separate academic research paper
- PM’s challenge fund in Barking, Havering and Redbridge
• Interim report to local areas, June 2015. Final Nuffield Trust research report, peer review articles - summer 2016.

2. Research studies involving the surveillance of patterns in hospital admission and costs at area level in England, aimed at identifying areas where innovation in service delivery is taking place.
- Avoidable harm project
• Reports co published with LSHTM. Final report to NIHR available January 2017. Also peer review articles at this time.
- Ambulatory emergency care project
• Nuffield Trust preliminary report (evaluation of pilot sites) – spring 2015 and if successful we will agree a more comprehensive longer term evaluation proposal. UPDATED OCT 2016: This project is now closed andno longer require processing data for this use (apart from auditing purposes).
- Assessment of the burden of alcohol on hospital services
• Nuffield Trust report completed by summer 2015, with peer reviewed papers to be submitted also in summer 2015. UPDATED OCT 2016: This project is now closed and we no longer require processing data for this use (apart from auditing purposes).

3. Research studies relating to the efficiency of health services and level of competition in the English NHS.
- Descriptions of differential patterns of hospital use by area and provider.
• Analysis of length of stay led to workshop in September 2014 (with Monitor), research report planned for winter 2014. Update May 2016: This project is now closed and we no longer require processing data for this use (apart from auditing purposes).

Child health study, to report 2016/17.

4. Linkage of HES data to linked datasets provided by HSCIC
Individual outputs will be covered within separate applications

5. Linkage of HES data to HES ids provided by HSCIC regarding a specific cohort
Individual outputs will be covered within separate applications

6. Informing the public debate about hospital use
The outputs of this work will be in the form of blogs, briefings, and/or presentation slideshows, posted on the Nuffield Trust website and made freely available to all. Due the responsive nature of the work, The Nuffield Trust are not able to provide prospective dates for these outputs

7. Evaluation of Virtual Wards in Devon
Nuffield trust report to be published by end of 2015. UPDATE OCT 2016: This project is now closed and no longer requires processing data for this use (apart from auditing purposes).

8. Evaluating the implementation of Quality Standards in the London region
• Nuffield Trust report to be published by mid-2016
The report will require quantitative analysis, the methods for which will be piloted in a Masters dissertation to be completed by the end of 2015. This work, which is overseen by a senior analyst, will lead to a peer reviewed publication to be submitted also in early 2016. The Nuffield Trust network will also be used to promote the findings amongst senior decision makers in the NHS and the Department of Health.

UPDATED ADDITIONAL PROJECTS OCT 2016 – new projects the Nuffield Trust are requesting access to HES data for.

9. Research studies identifying and evaluating models of medical generalism
Outputs in late 2017 will include a final report, an executive summary, and summary results for a lay audience. All will be made publically available on the Nuffield Trust website. Other planned mechanisms for dissemination include: the packaging and provision of on-going feedback to participating hospitals; workshops with user groups; face to face engagement with policy makers at national level; explicit knowledge transfer and exchange initiatives, such as working with networks such as the NHS Confederation. All data published/disseminated will be aggregated and no small numbers are anticipated but if they arise then they will be suppressed in line with HES analysis guidelines.
At the same time abstracts will be submitted to key conferences, such as Future Hospital Programme of the RCPL, Quality and Safety in Health Care Forum, the NHS Confederation Conference, as well as NIHR events.

10. Classifying readmissions and comparing readmission rates between the Netherlands and the UK based on national administrative data
Summary report published on Nuffield Trust website in winter 2016 and with an aim to publish in a reputable, peer reviewed journal by early 2017. The Nuffield Trust network will also be used to promote the findings amongst senior decision makers in the NHS and the Department of Health.
All data published will be aggregated and no small numbers are anticipated but if they arise then they will be suppressed in line with HES analysis guidelines.

Processing:

The data requested and already disseminated will be accessed and processed by substantive employees of The Nuffield Trust and only for the purposes described in the application.

Whilst the nature of detailed analysis in relation to each project varies, the broad context of processing is consistent. In summary :-
- The data is downloaded from NHS Digital and imported into SAS. The server is held on-site, and access is restricted to named individuals according to The Nuffield Trust’s security policy.
- The data is held within separate folders within the server.
- Remote access to the database is permitted, but only through Citrix via secure token (so processing is still carried out on site), and with local printing and downloading disabled.
- Only staff who have signed a confidentiality agreement and have received IG training are permitted access.
- All access to individual files is recorded, and a sample audited to investigate the existence of any adverse incidents, and ensure that appropriate access has been maintained.
- Once held in SAS, the researcher will view the data and select a specific cohort for each individual study. Commonly a process will initially take place to define the particular cohort of interest in terms of e.g. individual diagnostic codes or procedure codes. The researchers will use routinely available filter definitions where possible, but may amend these based on the nature of each study’s group of interest. Depending on the research a similar control group may be established.
- The individual researcher then analyses the data, before applying the relevant disclosure controls to any output. Software used will be SAS, R and stata; typically this will involve analysis on several outcome measures, risk adjustment and the construction of control groups.
- No record level data would be linked to this dataset, but it may be combined with publically available demographic or geographic data, for example in relation to local Trust performance
- Outputs are thus produced which consist of aggregate data (or indicator/statistical data) only.

As an example, for the assessment of the burden of alcohol on hospital services, The Nuffield Trust will look at national trends of alcohol related A&E attendances and inpatient admissions for alcohol related liver disease (ARLD) over the most recent decade. The Nuffield Trust will also identify a cohort of patients diagnosed with ARLD for the first time, and examine their prior and post diagnosis patterns of hospital use, in the context of a comparator group. Analyses will be undertaken at local authority level and will take into account provision of acute trust alcohol services.
In all such work, The Nuffield Trust analyse patterns of hospital activity by area, by year, by condition or by provider, developing comparative analyses and standardising for a range of episode level, or patient level variables – such as age, the presence of a long terms condition, prior patterns of use. The analyses commonly follow the health and care of a well-defined cohort of individuals over a lengthy period of time (for example the alcohol study will follow for ten years, the child health study – part of the differential patterns of health – will evaluate a cohort throughout childhood). Such analyses require complex processing for fair comparisons and to capture activity for whole populations – something that only nationally collated data can provide.

The datasets necessary for each of the studies are listed below. Where the start date is given as being earlier, OP data will be used from 2003/04 and AE data from 2007/08. *Study will require further years of data beyond those being requested in this application. :
- Integrated Care Models (Pioneers).
• APC, OP, AE 2004/05 to 2019/20*
- New models of primary care
• APC, OP, AE 2003/04 to 2015/16
- PM’s challenge fund in Barking, Havering and Redbridge
• APC, OP, AE 2004/05 to 2015/16
- Avoidable harms project
• APC, OP, AE 2003/04 to 2016/17*
- Ambulatory emergency care project. (Update May 2016: This project is now closed and we no longer require processing data for this use (apart from auditing purposes)).
• APC, OP, AE 2001/02 to 2015/16
- Assessment of the burden of alcohol on hospital services (Update May 2016: This project is now closed and we no longer require processing data for this use (apart from auditing purposes)).
• APC, OP, AE 2001/02 to 2013/14
- Descriptions of differential patterns of hospital use by area and provider.
• APC, OP, AE 1997/98 to 2015/16

For the study of Virtual Wards in Devon, the Nuffield Trust will use HES data from a small set of areas chosen as being similar to Devon (these areas are Somerset, Cornwall, Shropshire and Herefordshire). From these areas, the researchers will select a pseudo control group of individuals who shared characteristics with individuals who were admitted to the Virtual Ward scheme in Devon. These characteristics include age, sex, prior use of hospital services, and diagnostic history. The characteristics and hospital utilisation of the individuals admitted to Virtual Wards will be derived using a locally sourced pseudonymised data set (SUS). We will then compare future unplanned and other admissions to test for differences between the Devon Virtual Wards cohort and the pseudo control group.

For the evaluation of the implementation of Quality Standards in the London region The Nuffield Trust will look at variation in outcomes by London hospital/trust by carrying out cross section analysis at different time points, analysing the changes in extent of variation over time, investigating variations in trends and looking at these in different patient groups.
The datasets necessary for the new study is listed below. Where the start date is given as being earlier, OP data will be used from 2003/04 and AE data from 2007/08.
- Evaluation of Virtual Wards in Devon. (UPDATED OCT 2016: This project is now closed and no longer require processing data for this use (apart from auditing purposes).
• APC, OP, AE 2006/07 to 2014/15
- Evaluating the implementation of Quality Standards in the London region
• APC, OP, AE 2005/06 to 2015/16

UPDATED ADDITIONAL PROJECT information OCT 2016 – new projects the Nuffield Trust are requesting access to HES data for.

- Research studies identifying and evaluating models of medical generalism
• APC, OP, AE 2010/11 to 2015/16

- Classifying readmissions and comparing readmission rates between the Netherlands and the UK based on national administrative data
• APC 2013/14 to 2015/16

Additional note – third parties
The Nuffield Trust will not provide access to for any third parties to access record level data, even where these third parties are study partners. The use of this data will be limited to Nuffield Trust for the purpose outlined above only. Data published or provided to third parties will be limited to aggregated data, at area, organisational or cohort-level all subject to small number suppression in line with the HES Analysis Guide.