Improvement in the quality of health and care services depends on good-quality analytical support. We need to use data to identify areas of poor care, guide choices about priorities for care, improve efficiency and improve patient care. An organisation's analytical capability is their ability to analyse information and use it to make decisions. However, we know that in practice health and care systems are often not able to draw on high-quality analytical support. There is a shortage of people with the right skills and tools to do analysis, and collaborate with clinicians and managers on using their insights to improve care. This is exacerbated when the analysts we do have spend much of their time doing relatively low-value work – for example, compiling reports that aren't read. By investing in the analytical workforce, we will be able to unlock the full potential of data.

Better analysis is needed to support:

• clinical decision-making to help busy clinicians diagnose and manage disease
• innovation and change in the NHS, and to evaluate the success of new models of care and whether changes deliver the expected benefits
• effective board-level oversight of complex organisations and care systems
• better everyday management of the monitoring and improvement of the quality and efficiency of care
• senior decision-makers to respond better to national incentives and regulation
• the allocation of finite resources
• better understanding of how patients flow through the system
• new data and digital tools
• patients and the public in using information.

Advances in digital technologies have the potential to transform how care is delivered, but many of these benefits will not be fully realised by organisations without in-house analytical support. The current analyst workforce needs to develop its skill sets and be given leadership and support at senior levels in each organisation.

To get the most out of digital technologies, we need to recognise the importance of investing in the people who shape the information that is communicated and used. Though specialist academic, data-science roles are welcome, we also need people who can implement innovation. Where there has been investment in wider analytics (people, education, tools and techniques), there have been some favourable outcomes, as shown by the examples included in this report.

The challenge of developing analytical capability is not new. However, it is becoming ever more important as new data streams and tools emerge and opportunities for better analysis are missed.

We suggest action is needed at all levels of the system:

• National and local agencies should recognise the development of analytical capability as an important issue. They need to develop local strategies and new ways of working that support the development of the analytical teams in their organisations and identify key gaps.
• Investment in training and development tools to support better analysis is needed. A particular focus should be placed on developing in-house teams. Such development needs to apply not just to analysts but also to clinicians and managers and should encourage good links between the use of analysis and its application.
• The slow uptake of technologies such as open-source software, and techniques such as data-science and operational research, is of concern. These technologies and techniques can bring significant insight and value to health care services but are not being fully exploited at present.
• The sharing of skills and experience across organisations should be supported, to build a culture of 'build it once, share it to everyone'. The analytical community can be fragmented, with limited opportunity for teams to share learning or access specialist skills.
• As much emphasis should be placed on the application and translation of ideas as on their research and development. Investment in new digital technologies and in data-science methods needs accompanying investment in the local analytical workforce.
• The importance of the people who can make sense of data should be recognised, and investment made in the leadership of analytical roles, for instance by having a Chief Analytical Officer in each organisation.
• Most importantly, we need to set higher standards for the way information is used in delivering care and recognise the full potential of the datasets already held.

‘One of the greatest opportunities of the 21st century is the potential to safely harness the power of the technology revolution, which has transformed our society, to meet the challenges of improving health and providing better, safer, sustainable care for all.’National Information Board

The ability to exploit advances in digital technology in support of better health and better health care is a priority for health care services. The NHS Long Term Plan, published in January 2019, envisages a central role for technology, forecasting that technology will empower people, support health and care professionals to deliver better care, improve clinical efficiency and safety and improve population health overall. Alongside service innovations such as digital-first primary care and the NHS App, there are plans to create better digital infrastructure and build a digital-ready workforce that can make effective use of technologies as they are developed. This approach to digital infrastructure largely consists of the adoption of technology standards, to ensure that data are accessible and that it is possible for different systems to exchange information, and an expansion of the number of acute, mental health and ambulance trusts working with the NHS Global Digital Exemplars on information technology (IT) projects.

All these initiatives involve data, which the NHS has in abundance. A digital footprint is generated almost every time a person comes into contact with a service (Box 1). The volume of data will continue to grow as the NHS spends billions of pounds on its information systems. However, it is failing to make the most of the benefits that can flow from these systems, because there are not enough people with the right skills to use the information that is being collected.

Strategies aimed at exploiting the promise of data typically cover:

• innovation and investment in new technologies
• managing and accessing new data streams in new ways
• investing in the analytical workforce, which can use the data to provide actionable insight.

The third aspect tends to get the least attention – it is less glamorous and solutions are often challenging and long term. But it is an essential requirement if we are to exploit the advantages that new technologies offer.

This report explores some of the ways in which good data analytics can support decision-makers. It also identifies some missed opportunities that flow from our limited ability to make sense of information relevant to health and health care. It is based on the Health Foundation’s experience of promoting innovative data analytics in health and social care. One of our projects in this area is the Advancing Applied Analytics programme, which is now supporting 23 teams to improve analytical capability in health and care services.

Better use of patient data can improve the quality and operational efficiency of health care in various ways (Table 1). However, as we discovered in previous work, health and care organisations often have problems accessing analytical skills when needed. For that work, we interviewed people across health and care systems in the UK, and many highlighted the problem of not having the right people to interpret the data and provide useful analysis to clinicians and managers (Box 2). This problem is not new or unique to health services. But it is a serious challenge if we want to make the most of the information collected and realise the benefits of investment in digital technology. Part of the answer is making sure that health care services have in place the systems designers, training and infrastructure necessary for new technology. It also means having people and teams who can help make sense of the growing mountains of data.

Table 1: Information and intelligence provided by data analysts – who it is for and how it is applied

Audience	Examples
General population	Identifying information sources that might be useful to patients. Testing the effects of different presentation styles aimed at patients and the general population. Looking at the effects of new information systems.
All users of health and care services	Designing and testing the effects of new approaches to sharing information with service users.
Clinical teams	Developing decision aids/tools that use data to help diagnose or manage diseases (eg risk scores and algorithms). Informing the design of improvement initiatives. Monitoring the quality of care delivered over time.
Service managers	Identifying service objectives and monitoring tools (performance indicators). Tracking new care initiatives and providing the information required to improve them. Devising and adapting mathematical modelling tools (eg to improve scheduling or patient flow). Synthesising and summarising the literature on the effectiveness of new interventions and service models. Supporting data collection and analysis of clinical audits.
Commissioners and planners	Assessing needs and priorities. Reviewing evidence of effectiveness and efficiency of service delivery. Assessing need and demand and forecasting for populations. Modelling capacity requirements and business planning. Agreeing evaluation frameworks and monitoring effects of service models. Monitoring the quality of services.
Those running organisations	Performance analysis. Assessment of the economic impacts of changes (eg new technology). Quality monitoring. Assessing probable effects of changes before they are made (eg closure of A&E departments). Forecasting demand for services (eg ahead of winter). Business and strategic planning.
System- and national-level decision-makers	Monitoring against strategic priorities. Developing and applying mathematical models to inform policy (eg vaccination or urgent care). Regulation of efficiency and quality. Resource allocation. Programme evaluation so that the NHS can learn from experience and improve.

This report has been written at a critical moment for the NHS workforce, with over 100,000 vacancies reported by trusts and problems attracting, retaining and motivating staff. Overcoming these wider problems is crucial to building analytical capability in the NHS. But it is equally important to recognise that analytical capability requires its own strategy, one that articulates clear roles for analytical teams in the health care system, as well as leadership models and approaches to supporting collaboration between analysts, clinicians and managers.

---

‘The failure to use information properly in health and care means people can experience unnecessary levels of preventable ill health. Those using services can suffer harm when it could be avoided, could live in greater pain and distress than they need to, and are less likely than expected to experience a full recovery. Every day, interactions with health and care services can waste people’s precious time. In addition, taxpayers do not get full value: the productivity benefits that come from effective use of new technology – doing more for less – are not widely realised.’ National Information Board

The consequences of a shortage of analytical expertise are not always obvious or immediately visible. The effects can be hidden in a range of suboptimal decisions and choices based on limited or inappropriate evidence. Some of the areas of health care in which good data analysis is critical are:

• clinical decision-making
• innovation and improvement in care
• board-level oversight of complex organisations and care systems
• everyday management
• responding to national initiatives and regulation
• resource allocation
• understanding patient flow
• supporting new data and new digital tools
• helping patients and the public to use information.

1. To support clinical decision-making

Support for clinical decision-making is one area where new technologies are changing quickly, as evidenced by the proliferation of tools and algorithms to help clinicians diagnose and manage disease. Development is rapid in both the public and private sectors,,, and in health care the digital future seems rich with possibility. As Robert Wachter noted, ‘Big-data techniques will guide the treatment of individual patients, as well as the best ways to organize our systems of care’.

There is growing interest in the possibilities of ‘big data’ and artificial intelligence (AI). Although much of this is, for the time being, aspirational, it is clear that such developments will require a skilled workforce that can ensure that tools are implemented in the right way. As noted in the Topol review, development of the workforce is critical for advancements to be fully realised, as staff will need to understand the issues of data validity and accuracy. Indeed, it is possible that the future clinical teams will include data scientists and bio-informaticians.

Some new technologies are already in widespread use: for example, analytical tools such as predictive-risk algorithms that use historical information to make predictions about a future event. These might use patients’ electronic records to predict the probability that an individual might suffer an adverse event (such as requiring an emergency admission to hospital). These tools can help clinicians identify where to focus preventive services to avoid acute health problems occurring. Such tools cannot replace clinical judgement, but they might enhance it.

At their best, these predictive tools can operate within existing information systems and seamlessly provide input to a clinical decision. In many cases, they can function as standalone tools. But for the best results they will often require analytical support, which means that using them becomes something more than simply switching on a software module.

So, when applying predictive risk algorithms, several questions need to be considered:

• Are you able to extract the right data from operational systems?
• Can you analyse aggregate patterns across patients?
• Do you need to calibrate predictive models on local data?
• Will the model perform as expected, given the differences in the way information is collected and coded at a local level?
• What are the characteristics of high-risk patients and how can interventions be designed to improve the care they receive?

Other predictive tools are being tested as part of the Health Foundation’s Advancing Applied Analytics programme in areas such as general practice and mental health (Boxes 3 and 4).

2. To support innovation and improvement in care

The NHS is awash with innovations designed to deliver ‘better’ care, triggered by a desire to improve quality of care, the need for financial solvency, or both. Examples include integrated models of care, digital-first approaches to primary care, new algorithms to detect diseases, and the establishment of rapid diagnostic services to detect cancers. Despite the hunger for innovation, however, there is often no way to know whether these changes will actually improve care (eg around reducing emergency admissions).,,

While traditional evaluation can help understand what works, the process can also be too slow or too restrictive, requiring the process of care to remain unchanged until the study has been completed (or allowing only for small changes). What is needed is an approach to monitoring the effects of innovation in close to real time, so that teams (local or national) can ‘course correct’ along the way. An example of such an evaluation model is given in Box 5.

These evaluations can reveal higher-quality care, such as a recent evaluation from the Improvement Analytics Unit. It found that residents of care homes who received enhanced support experienced 23% fewer emergency admissions than expected. Even when an evaluation reveals that the intervention has not delivered the gains that were anticipated, it still produces valuable learning. In a complex environment, not every change will produce the intended effect and it is important to identify where the results were not as expected.

Implementing new approaches to the delivery of services can be challenging and take considerable time and effort from front-line teams. If these teams are going to have the best chance of improving patient care, they need better analytical support to help them understand the effects of their work to date and make improvements.

Questions that recur when supporting change are whether and how to provide help to clinical and managerial teams. It is often done using external consultants, but the downsides, apart from the initial expense, include the fact that the consultants’ skills and expertise are not transferred to the internal teams. The newly achieved solutions may not be sustainable. In 2016, the board of Taunton and Somerset NHS Foundation Trust took the bold step of making a big investment in developing permanent, internal improvement capability: creating the right culture, structure, tools and processes to enable and empower their workforce to improve from within. The decision was prompted by a concern that the use of external consultants had proved too costly and unsustainable. As a result, the Trust has successfully minimised spend on external consultants and now has a thriving team within the organisation to support improvement work. Box 6 outlines the model they have adopted.

3. To facilitate board-level oversight of complex organisations and care systems

‘IT and information is the ownership of the board and the senior management, if they’re not using it to run their organisation they’re in the wrong job. There is nothing more complicated than running the NHS and if we don’t do it as smart as any organisation in the world then we’re really betraying the taxpayer and we’re betraying our patients.’Matthew Swindells

Good-quality information and intelligence is critical for a board to be effective. Box 7 gives an example from a study looking at the quality of care. A study of NHS providers suggested that one of the key elements in achieving successful provider transformation is insight from data analysis that enables a fact-based understanding of problems, informed decision-making and performance-tracking.

The task of organising complex information and presenting it in ways that are meaningful and relevant to board-level decision-makers should be one of the fundamental roles of the analysts. However, in many cases, reporting at board level falls short of what is required and relies on long, unprocessed lists of tables. Sometimes, the problem is not a shortage of information but rather an 'overabundance of irrelevant information’.

One study looking at how boards work on improving the quality of care ranked organisations in terms of the maturity of their approach to quality improvement (QI). They found that organisations with high levels of QI maturity received reports in which the data were clear and readable, and in which different sources of data were discussed together (eg data on staffing levels considered alongside data on staff wellbeing and patient experience). By contrast, reports to boards with low levels of QI maturity were characterised by a large volume of data, which was often not clearly presented, reviewed in silos and not linked to improvement actions.

An experienced NHS manager observed that ‘the NHS gathers a massive amount of data but largely fails to use it intelligently. Energy is misplaced … most is spent downloading and gathering data, followed by preparing reports, analysing data, and ultimately using the data to make decisions’.

A recurring issue is how board-level reports handle statistical uncertainty. Most measure-ment contains some degree of uncertainty arising from chance variation and basic statistical methods; this is a widely accepted way to distinguish a systematic trend from an ambiguous one. Yet one study found that of a total of 1,488 charts found in board reports, only 6% acknowledged the role of chance. This presents a risk that boards react to changes in metrics that are the result of chance and do not reflect any real change to the underlying care processes – wasting time and resources. One approach to addressing these problems is to move from Red Amber Green ratings to statistical process control charts, as suggested by NHS Improvement in the initiative Making Data Count.

4. To improve everyday management

In terms of basic operational management, there are many opportunities for good analysis to make everyday tasks more efficient. New software tools allow better reporting and allow managers and clinicians to access information closer to where the decisions get made. Examples include Qlik, Tableau and Beautiful Information (Box 8).

Wrightington, Wigan and Leigh NHS Financial Trust has developed a suite of analytical apps that support the organisation and its provision of healthcare from ward to board. Their most renowned app supports their A&E department in monitoring demand (both current and predicted), wait times, decisions to admit and other aspects of patient flow. The app has become the ‘single version of the truth’ and supports both the department and the Trust executives in their decision-making. Since its introduction, the app has helped reduce the median length of stay by 30 minutes: improving discharge levels, reducing delays and minimising readmissions. Although the app may look simple, it contains complex algorithms that use things like weather data to determine how many people are likely to turn up at A&E in the hours and days ahead.

The Trust has developed other apps that support the organisation's referral-to-treatment times, theatre efficiency, budget management, outpatient and inpatient care, and monitoring of variations in care.

Mark Singleton, Associate Director of Information Management and Technology for the Trust, said ‘We are so lucky to have such as a fabulous Business Intelligence team that have developed a recipe for success when it comes to working with Clinical Services and producing ground-breaking apps that support the organisation in so many different ways but ultimately to ensure the organisation provides the best care for its patients.’

• A web-based app available on any smartphone, tablet or desktop platform.
• Provides aggregated, real-time data.
• A proactive management tool that highlights bed capacity and delays in the system.
• Available anytime and anywhere, in the hospital or off-site.
• Control over access to unlimited users.
• Developed by Beautiful Information, an NHS/private partnership.

Data analytics have also been beneficial to commissioners (Box 9).

5. To better respond to national initiatives and regulation

The local health and care agenda is often shaped by external demands from national government and arm’s-length bodies (Box 10). These will typically have a framework for accountability and performance assessment that is applied to local providers and commissioners of care. The tools used as the basis for these performance assessments frequently rely on complex analytical methods when defining performance targets or metrics, such as the Summary Hospital Mortality Index. Very often, local analytical teams are needed to interpret these national information measures and put them into local context.

For example, as part of the annual planning round, providers are obliged to generate demand forecasts for their key points of delivery (eg emergency admissions, outpatient referrals). For many organisations, these forecasts have been relatively naive in construction, ignoring core concepts like trends and seasonality. To bridge the gap and generate good-quality returns, NHS Improvement decided to develop univariate-time-series forecasting tools to help providers increase the level of sophistication and statistical rigour of their forecasts. The tools and outputs developed were delivered at scale using web-based interfaces and dedicated output files. The code driving the tools was developed on an open-source platform. It could then be shared with local analysts, who could re-use and amend it as necessary. This univariate-forecasting approach is now a standardised methodology for both NHS Improvement and NHS England.

It can be difficult for an individual organisation to ignore some of the approaches developed by national bodies, such as NHS England or the Care Quality Commission, when these are used in performance management or regulatory discussions. An organisation without sufficient analytical capability will be at a distinct disadvantage in such discussions.

6. To better allocate resources

Moving resource allocation from simply reinforcing historical funding patterns to a system that represents need has been a recurrent theme in health policy since the 1960s. It is an area in which local interests battle hard for their share of the pot, and one where good analytical support is essential to understanding the evidence and weigh different arguments.

At a national level, the analysis can be complex and often involves expert teams advising on government strategies. As the National Audit Office noted, ‘Given the amount of money involved – equivalent to nearly £1,400 per person each year – the way in which the Department [of Health and Social Care] and NHS England allocate funding to local commissioners is a crucial part of the way the health system works. These decisions are complex, involving mathematical formulae and elements of judgement.’

There is some evidence that national policy of resource allocation had an impact on reducing inequalities between areas. Between 2001 and 2011, the proportion of resources allocated to deprived areas in England compared with that allocated to more affluent areas was associated with a reduction in absolute health inequalities from causes amenable to good health care.

Aside from the implementation of national allocation methods, there is also a need for better analysis to support ways to identify local priorities in allocating resources. As Geraldine Strathdee, Chair of the National Mental Health Intelligence Network, noted, without benchmarking data, NHS resources are allocated on the basis of historical patterns, guesswork or the ‘loudest voice’.

7. To understand patient flow

The past few years have seen a recognition of the importance of understanding the way patients flow through the care system. Often, the best way to achieve that is through the use of sophisticated methods such as modelling, yet their uptake has been patchy.

Good examples do exist: for example, work to understand demand for long-term care in Kent (Box 11). Others have used tools such as simulation and queueing theory to look at scheduling community mental-health assessments. In Sheffield, local teams have applied simulation modelling to evaluate the reconfiguration of stroke services in Sheffield and South Yorkshire and this work has been integral to decision-making.

The challenge in adopting these tools routinely has been linked to a lack of capacity in health services; too few staff members are felt to have the training or ability to use the models. This is especially relevant considering current concerns with managing urgent and emergency care and flows.,

The Health Foundation has been promoting local work to look at patient flow through the service. One recent review of a flow training programme in Wales noted the challenges of accessing the analytical time needed to support this work: ‘Another substantial constraint was that none of the local health boards had sufficient capacity and expertise in data collection and analysis to provide ongoing support to clinicians and make sure they had the right information in the right format for effective decision-making.’

8. To support new data and new digital tools

Another reason to support analytical capability is that it could open the way to new data-driven technologies, for example machine-learning algorithms and AI, that could help with the diagnosis and management of health conditions. Data are essential to the development of these technologies, and the NHS has some of the best health care data in the world. The NHS Long Term Plan envisages a role for private-sector companies, with an aim to ‘encourage a world leading health IT industry in England with a supportive environment for software developers and innovators’. Successful delivery of this innovation agenda is likely to depend on joint working between NHS teams and industry. NHS analytical teams have a lot to contribute in this area. They understand how NHS data are being collected and why, and they can act as a valuable bridge between NHS clinicians and managers and data scientists in industry.

To take advantage of new methods, analysts must have access to the right software tools, particularly open-source programming tools that allow analysts to learn from each other (eg R and Python). In the early 2000s, open-source software began to gain acceptance, even among the sceptics. Today, open-source software is practically embedded in large, commercial organisations such as Facebook, Google, Twitter and banking and blue-chip corporations. They are taking full advantage and seeing the benefits of its power and scale. Some, such as Facebook, are actively developing and sharing their software tools within the wider open-source community, and some of those tools have been embraced in health care (eg Prophet).

The NHS has been much slower in accepting and seeing the value of open-source software, although the Department of Health and Social Care has recently announced that the newly created NHSX will ensure that all source code is open by default. One of the main barriers to wider deployment of open-source tools has been the reluctance of IT staff to install open-source software on secure health care systems. One analytics manager confided: ‘Getting an open-source application installed on my NHS laptop was a lengthy and arduous process. IT professionals were particularly risk averse to deploy software on their systems especially given the highly sensitive information they contain. We battled with the reluctance because we saw the value add it would give us. Now the software is installed, we are starting to realise the value add, why it’s so popular, and the awesome things it can do – you can see the reason why all the big organisations have embraced it. We are doing some things very differently now, it has allowed us to work more on our methodologies rather than churn.’

Open-source tools such as 'R', a statistical programme that is gaining recognition in commercial and public-sector applications, present an opportunity. Open-source offerings include generalised tools such as Python (used as the platform for GCSE computing) as well as niche analytical disciplines such as JaamSim (used for discrete-event simulation). Open-source tools enable analysts to explore and sample an array of new analytical tools and techniques they can ultimately deploy in their organisations. The challenge for the future is to take advantage of these exciting analytical tools by building analytical capability. This includes the networks needed to allow collaboration, such as the NHS-R Community (Box 12).

The UK government, in its Life Sciences Industrial Strategy, sees health-service data as being of value to those developing new digital tools. For example, the Open Prescribing project uses existing data streams on prescribing collected from GP practices by NHS Digital. The raw data files are huge, with more than 700 million rows, so a team at Oxford have put together some analytical tools that are freely available to GPs, managers and the public.

Over the coming years, NHS organisations may wish to provide private companies with access to NHS data, for example to help with the development of new algorithms or drugs. The benefit for the NHS might include seconded data scientists working alongside NHS teams. In that situation, we'd recommend an emphasis on skills transfer, so that the NHS builds its capability to conduct analysis in a sustainable way.

9. To help patients and the public use information

Investing in data analytics enables new information flows to and from patients and the general public. This information can help people make better-informed decisions about their care, as well as contribute more effectively to the development of their local services. There are also opportunities to engage the public in decisions about how data are used.

Table 2 lists some of the ways in which these information flows operate. Although many revolve around the use of new technological tools, it is important to recognise that there is still a significant analytical role in reporting, presenting and understanding the data.

---

People make decisions under constraints. These might be knowledge constraints or constraints imposed by analytical ability. Decision-making is therefore based on heuristics: experience-based techniques for problem-solving, or ‘knowing by trying’. The ‘recognition heuristic’ is when people make a decision based on only one piece of information, recognition – the knowledge that many others have chosen the same option.

Analytical input can be considered too slow, misguided or irrelevant to the problem at hand. The divide between analyst and senior manager can be further widened by:

• The challenge of choosing the right analytical approach to fit the managerial problem. This is partly an issue of whether analysts have effective communication skills. It is also a question of whether analytical teams have the skills to unpick problems and questions from senior managers in ways that match the analysis they can do. This is an area in which external management consultants often excel.
• Expectations of what constitutes good (enough) analysis. If senior managers have little experience in data analysis, they may not be able to recognise the value that data analysis can bring and the difference that it can make.

One way to bridge the divide is to invest in people who span professional boundaries. In our previous report, Understanding analytical capability in health care: Do we have more data than insight?, many interviewees described the importance of analytical leadership, in particular people who understand the possibilities that good analysis can engender (and how analytical teams work), yet who can also engage with managers at the highest levels to influence and shape demand for analysis. Organisations with a well-developed analytical workforce also tend to have strong leaders who are influential within the organisation, whether these are chief analysts or in Chief Information Officer roles, and in some cases they may be strong clinical professionals. The implication is that enhancing the profile of good-quality analytics within an organisation must involve recognising the current generation of leaders as well as investing in the next generation. A key skill is to spot opportunities for analysis that other senior staff don’t see, and manage expectations around requests for analysis.

The right presence at the organisation’s decision-making level can help shift the relationship between decision-makers and their in-house analytical teams. The direction of change is from a relationship where the analysts unquestioningly provide whatever they have been asked to provide, to one where the requests received will address the problems and begin to be anticipated, and to an extent shaped, by the analytical team (Box 13).

1. National Information Board. Personalised Health and Care 2020: Using Data and Technology to Transform Outcomes for Patients and Citizens. A Framework for Action. National Information Board, 2014 (www.digitalhealth.net/includes/images/news0254/PDF/0172_NHS_England_NIB_Report_WITH_ADDITIONAL_MATERIAL_S8.pdf).
2. NHS England. Digital Transformation [webpage]. NHS England (www.england.nhs.uk/digitaltechnology).
3. NHS England. The NHS Long Term Plan. Department of Health and Social Care, 2019.
4. Huesch MD, Mosher TJ. Using it or losing it? The case for data scientists inside health care. New England Journal of Medicine. 4 May 2017 (https://catalyst.nejm.org/case-data-scientists-inside-health-care).
5. NHS Confederation. NHS Statistics, Facts and Figures [webpage]. NHS Confederation. 14 July 2017 (www.nhsconfed.org/resources/key-statistics-on-the-nhs).
6. NHS England. Our 2015/16 Annual Report: Health and High Quality Care for All, Now and For Future Generations. TSO (The Stationery Office), 2016.
7. Bardsley M. Understanding Analytical Capability in Health Care: Do We Have More Data Than Insight? Health Foundation; 2016.
8. Farr M. Down and out in Manchester and Liverpool [webpage]. Beautiful Information. 14 June 2018 (http://beautifulinformation.org/news/developing-an-information-strategy).
9. Buchan J, Charlesworth A, Gershlick B, Seccombe I. A Critical Moment: NHS Staffing Trends, Retention and Attrition. Health Foundation, 2019.
10. Administrative Data Taskforce. Improving Access for Research and Policy. The UK Administrative Data Research Network, 2012 (https://esrc.ukri.org/research/our-research/administrative-data-research-network/administrative-data-taskforce-adt).
11. Beresford M. Demystifying Data: The Data Revolution and What it Means for Local Government. New Local Government Network, 2015 (www.nlgn.org.uk/public/wp-content/uploads/DEMYSTIFYING-DATA2.pdf).
12. Nesta. Analytic Britain: Securing the Right Skills for the Data-driven Economy. Nesta, 2015.
13. Wachter R. The Digital Doctor: Hope, Hype and Harm at the Dawn of Medicine’s Computer Age. McGraw Hill, 2015.
14. Health Education England. The Topol Review: Preparing the Healthcare Workforce to Deliver the Digital Future. Health Education England, 2019.
15. Lewis G, Curry N, Bardsley M. Choosing a Predictive Risk Model: A Guide for Commissioners in England. The Nuffield Trust, 2011.
16. NHS Improvement. Single Oversight Framework for NHS Providers [webpage]. NHS Improvement (https://improvement.nhs.uk/resources/single-oversight-framework/).
17. Bardsley M, Steventon A, Smith J, Dixon J. Evaluating Integrated and Community-Based Care: How Do We Know What Works? Nuffield Trust, 2013.
18. Purdy S, Paranjothy P, Huntley A, Thoma R, Mann M et al. Interventions to Reduce Unplanned Hospital Admission: A Series of Systematic Reviews. University of Bristol, 2012.
19. Lloyd T. Why Before-and-After Analyses Can Give Misleading Results [webpage]. Health Foundation. 28 September 2018 (www.health.org.uk/newsletter-feature/why-before-and-after-analyses-can-give-misleading-results).
20. Lloyd T, Wolters A, Steventon A. The Impact of Providing Enhanced Support for Care Home Residents in Rushcliffe. Health Foundation, 2017.
21. Health Foundation. Improvement Analytics Unit [webpage]. Health Foundation, 2019 (www.health.org.uk/funding-and-partnerships/our-partnerships/improvement-analytics-unit).
22. Gibbons A. Transforming How We Deliver Improvement Projects at Taunton and Somerset NHS Foundation Trust [webpage]. The Academy of Fabulous Stuff, 2018 (https://fabnhsstuff.net/fab-stuff/transforming-how-we-deliver-improvement-projects-at-taunton-and-somerset-nhs-foundation-trust).
23. Swindells M. The Importance of Digital Health in the Implementation of the NHS Five Year Forward View [video transcript]. The King’s Fund, 6 July 2016 (www.kingsfund.org.uk/audio-video/matthew-swindells-digital-health-nhs-five-year-forward-view).
24. McKinsey Hospital Institute. Transformational Change in NHS Providers. Health Foundation, 2015.
25. Ackoff RL. Management misinformation systems. Management Sciences. 1967; 14: B147–56.
26. Jones L, Pomeroy L, Robert G, Burnett S, Anderson JE et al. How do hospital boards govern for quality improvement? A mixed methods study of 15 organisations in England. BMJ Quality and Safety. 2017; 26: 978–86.
27. Schmidtke KA, Poots AJ, Carpio J, Vlaev I, Kandala N-B et al. Considering chance in quality and safety performance measures: an analysis of performance reports by boards in English NHS trusts. BMJ Quality and Safety. 2017; 26: 61–9.
28. NHS Improvement. Making Data Count. NHS Improvement; 2018.
29. Dixon-Woods M, Baker R, Charles K, Dawson J, Jerzembek G et al. Culture and behaviour in the English National Health Service: Overview of lessons from a large multimethod study. BMJ Quality and Safety. 2014; 23: 106–115.
30. Digital Health. How Tableau Has Transformed How Our NHS Foundation Trust Uses Data Within 6 Months [webinar]. Digital Health, 2017 (www.digitalhealth.net/event/how-tableau-has-transformed-how-our-nhs-foundation-trust-uses-data-within-6-months-2).
31. Beautiful Information. Performance [webpage]. Beautiful Information, 2018 (http://beautifulinformation.org/solutions/performance).
32. BJC Staff. New series on insights from the Bradford Healthy Hearts project. British Journal of Cardiology. 2017; 24: 130.
33. Burnett S, Mendel P, Nunes F, Wiig S, van den Bovenkamp H et al. Using institutional theory to analyse hospital responses to external demands for finance and quality in five European countries. Journal of Health Services and Research Policy. 2016; 21: 109–117.
34. Lilford R, Mohammed MA, Spiegelhalter D, Thomson R. Use and misuse of process and outcome data in managing performance of acute medical care: avoiding institutional stigma. Lancet. 2004; 363: 1147–54.
35. Hogan H, Zipfel R, Neuburger J, Hutchings A, Darzi A et al. Avoidability of hospital deaths and association with hospital-wide mortality ratios: retrospective case record review and regression analysis. British Medical Journal. 2015; 351: h3239.
36. Spiegelhalter D, Sherlaw-Johnson, Bardsley M, Blunt I, Woods C et al. Statistical methods for healthcare regulation: rating, screening and surveillance. Journal of the Royal Statistical Society Series A. 2012; 175: 1–47.
37. NHS Digital. Summary Hospital-level Mortality Indicator (SHMI) [webpage]. NHS Digital (https://digital.nhs.uk/data-and-information/publications/ci-hub/summary-hospital-level-mortality-indicator-shmi).
38. Gorsky M, Millward G. Resource allocation for equity in the British National Health Service, 1948–89: An Advocacy Coalition Analysis of the RAWP. Journal of Health Politics, Policy and Law. 2018; 43: 69–108.
39. National Audit Office. Department of Health and NHS England, Funding Healthcare: Making Allocations to Local Areas. Report by the Comptroller and Auditor General. National Audit Office, 2014.
40. Barr B, Bambra C, Whitehead M. The impact of NHS resource allocation policy on health inequalities in England 2001-11: Longitudinal ecological study. British Medical Journal. 2014; 348: g3231.
41. Noveck BS. Another NHS crisis looms – an inability to analyse data. The Guardian, 8 February 2017.
42. Fillingham D, Jones B, Pereira J. The Challenge and Potential of Whole System Flow: Improving Flow across Whole Health and Care Systems. Health Foundation, 2016.
43. MASHnet. Application of Simulation and Queueing Theory to Scheduling Community Mental Health Assessment (Simulation Study) [webpage]. MASHnet, 2019 (https://mashnet.info/casestudy/application-of-simulation-and-queuing-theory-to-scheduling-community-mental-health-assessment).
44. Pitt M, Monks T, Crowe S, Vasilaki C. Systems modelling and simulation in health service design, delivery and decision making. BMJ Quality and Safety. 2016; 25: 38–45.
45. Brailsford SC, Lattimer VA, Tarnaras P, Turnbull JC. Emergency and on-demand health care: modelling a large complex system. Journal of the Operational Research Society. 2004; 55: 34–42.
46. NHS Improvement. Good Practice Guide: Focus on Improving Patient Flow. NHS Improvement, 2017.
47. NHS Improving Quality. Population-Level Commissioning for the Future. NHS Improving Quality, 2013.
48. Crisp H. Delivering a National Approach to Patient Flow in Wales: Learning from the 1000 Lives Improvement Patient Flow Programme. Health Foundation, 2017 (www.health.org.uk/sites/health/files/PatientFlowWales.pdf).
49. Facebook Open Source. Prophet: Forecasting at Scale. Facebook, 2018 (https://facebook.github.io/prophet).
50. Department for Health and Social Care. NHSX: New Joint Organisation for Digital, Data and Technology. Department for Health and Social Care, 2019.
51. Allred M-L. What Does the Life Sciences Industrial Strategy Mean for the NHS [webpage]? Arden & GEM, 2018 (https://www.ardengemcsu.nhs.uk/newsinsight/what-does-life-sciences-industrial-strategy-mean-nhs).
52. OpenPrescribing. Explore England’s Prescribing Data [webpage]. OpenPrescribing, 2019 (https://openprescribing.net).
53. Castle-Clarke S. The NHS at 70: What Will New Technology Mean for the NHS and its Patients? Health Foundation, Institute for Fiscal Studies, The King’s Fund, Nuffield Trust, 2018.
54. Williams I, Brown H. Factors Influencing Decisions of Value in Health Care: A Review of the Literature. University of Birmingham; 2014 (www.nhsconfed.org/-/media/Confederation/Files/Publications/Documents/DOV_HSMC_Final_report_July_281.pdf).