NHS Digital Data Release Register - reformatted
London School of Economics and Political Science (LSE)
Project 1 — DARS-NIC-354497-V2J9P
Opt outs honoured: No - data flow is not identifiable (Does not include the flow of confidential data)
Sensitive: Non Sensitive, and Sensitive
When: 2017/12 — 2020/02.
Legal basis: Health and Social Care Act 2012, Health and Social Care Act 2012 – s261(1) and s261(2)(b)(ii)
Categories: Anonymised - ICO code compliant
- Hospital Episode Statistics Admitted Patient Care
- HES:Civil Registration (Deaths) bridge
- Hospital Episode Statistics Accident and Emergency
- Patient Reported Outcome Measures (Linkable to HES)
- Hospital Episode Statistics Outpatients
- Civil Registration - Deaths
The London School of Economics (LSE) research team undertaking this application comprises researchers from LSE Health and Social Care and the LSE Research Laboratory. The research team will principally use HES and PROMs data to make three distinct contributions to health and social care within a single programme of research. The first contribution is to analyse the impact that various aspects of policy-development and reform of the NHS have had on patient outcomes, waiting times, and provider behaviour. The second contribution compares health care system performance in several countries with the aim of spreading best practice across different countries, with specific focus on lessons that can be learnt for health and social care policy within the NHS. The third contribution develops and tests a range of multi-dimensional indicators of health care quality and outcomes. All project outputs are published at an aggregate provider level, and the identification of individual patients or clinicians is not possible nor their behaviour identifiable. The purpose of receiving the data is to analyse the impact of on-going NHS reforms implemented between 2000 and the present day using pseudonymised non-sensitive patient-level data (further details of these reforms are provided below). These reforms were primarily associated with the introduction of two Acts of Parliament (the National Health Service Act 2006 and the Health and Social Care Act 2012) that changed both the organisational and payment structures of the NHS. The second Act of Parliament relies on on-going productivity gains to maintain efficient NHS output. The objective of the research programme is to assess the change in policy direction arising from these reforms, to better understand the impact of these reforms on patient outcomes and to improve the measurement of the impact of these reforms and to aid the efficiency with which they are implemented. Research deliverables are currently underway to examine the impact of the following reforms to the NHS and include the following: * Introduction of Payment by Results (2003-2006). * Implementation of waiting time targets for first outpatient appointment, elective surgery, and A&E attendances (2001-2005). * Introduction of the Quality and Outcomes Framework (2004). * Introduction of patient choice of hospital and GP surgery (2006-2008). * Introduction of independent care providers, both as part of the Independent Sector Treatment Centre programme and under the Any Qualified Provider programme (2003-2010). * Introduction of Walk in Centres for ambulatory care (2000-2010). * Introduction of the Alternative Provider of Medical Services GP contract (2004). * Expansion of primary care provision under the Equitable Access to Primary Care initiative (2007-2010). * Implementation of Equality and Human Rights Commission (EHRC) Memorandum of Understanding with the Care Quality Commission on equality and human rights in the context of healthcare (2011) * Introduction of Clinical Commissioning Groups in place of Primary Care Trusts (2013-2014). * Progress towards the Nicholson productivity challenge (2009-2015). * Introduction of new signals of hospital outcomes and performance, including the NHS Choices website (2006), Patient Reported Outcome Measures (2009), the NHS Staff Survey (2003), and various NHS Patient Satisfaction and Patient Experience surveys. Each of these research deliverables form part of the first contribution to health and social care outlined above, by analysing a specific policy introduced within the NHS over the last 15 years. Each of these deliverables also forms part of the third contribution to health and social care, in that each focuses on analysing the impact of a particular policy or reform on one or more multi-dimensional indicators of health care quality and outcomes. The research team has primarily progressed the second contribution to health and social care – a cross-country comparison of health care system performance – via a project entitled “Do Financial Incentives Trump Clinical Guidance?: The case of Hip Replacements in England and Scotland” that examines the impact of financial incentives on clinicians’ decisions in England and Scotland. Making further progress toward this second contribution to health and social care will be a major focus of the research programme over the next 2-3 years. Using patient-level data is necessary for risk adjustment at the patient level, thereby allowing control of confounding factors that will affect the analysis. Using patient-level data is also necessary to construct various inputs into the analysis -- for example, percentiles of distance from patient's GP surgery to hospital of admission, which are needed to construct indices that define hospital referral markets and potential areas of competition. All programme outputs produced thus far have been published at an aggregate provider level, where no individual patients or clinicians are identified or identifiable. The applicant intends to continue publishing project outputs on this basis, i.e. at the provider level and without identifying individual patients or clinicians.
The main outcome of the presentation at the Wennberg International Collaborative was simply to raise awareness about the research findings amongst policymakers and practitioners. Earlier work compiled by the research team (Cooper et al 2011) was cited by the then Prime Minister in support of the Health and Social Care Act 2012, which expanded choice and competition within the NHS. The research also helped to ensure that these reforms did not introduce price competition between health care providers, as had initially been proposed. The work on hospital quality and on choice and competition has been used by NHS Monitor (now NHS Improvement) in their measurement and analysis of hospital efficiency measurement, which forms one strand of their work to identify inefficient hospital trusts. The project lead has since worked with NHS Improvement to aid this strand of their work, and the work in this area also led to him becoming an adviser to the UK Competition and Markets Authority investigation into the Private Health Care Market which led to a range of measures being implemented in 2014 (see https://www.gov.uk/cma-cases/private-healthcare-market-investigation).
The applicant’s research programme has multiple outputs that are being, and will continue to be, continuously disseminated to policymakers and policy analysts, via the types of channels outlined in the previous sections. The aim of the research is to benefit English health and social care by contributing to a better understanding of the impacts of past and existing health and social care policies. In so doing, the applicant hopes to contribute to more informed policymaking in the future. Previous success in disseminating the applicant’s research to policymakers is attested to by the fact that, in 2012, Prime Minister Cameron referred to their research outputs in a speech in support of his reforms to the NHS. Further evidence of the success of these dissemination efforts in delivering benefits to health and social care is attested to by the attached letter of support (SD2) for the applicant’s research programme from the Prime Minister’s adviser for health and adult social care. Referring to the draft project output “Does Competition Improve Public Hospitals’ Efficiency? Evidence from a Quasi-Experiment in the English National Health Service”, as well as to past research outputs produced by the applicant, states that the applicant’s research “has informed the policy thinking at the highest levels of government and materially impacted policy formation for the better. His work serves as a prime example of how research can improve policy and make a positive impact that is felt outside of academia.” LSE's research on NHS Walk-In Centres, which had already been disseminated to policymakers in Monitor and Department of Health at the time of last application, has now percolated down to CCG level and is being used to inform commissioning decisions. See for example: • http://www.dorsetccg.nhs.uk/Downloads/aboutus/CCG%20Board/18%20March%202015/09.7%20x%20Appendix%201%20180315.pdf. • https://www.bristolccg.nhs.uk/media/medialibrary/2016/09/bccg_front_door_rapid_evidence_review2016-09-16.pdf.
The expected outputs consist of research reports and published papers, and discussions and presentations to UK health and social care policymakers, policy analysts, and clinicians. A list of outputs to date is provided below. The primary target audience for these outputs is the health policy community. This includes policymakers, as a key aim of the research is to investigate the impact of recent changes to health and social care policy, with an ultimate objective of influencing future policy formation. However, it also includes other policy analysts, such as (but not limited to) those at the Health Foundation, the King’s Fund and Nuffield Trust, who, while they may not be directly involved in the policy formation process, do have an important influence over the terms under which health policy is debated and therefore formulated. A secondary target audience is the medical community that has been responsible for implementing many of the policy reforms analysed, and whose decisions therefore determine the success of these reforms, and the nature of their impacts. The main outputs of these reports and papers will be estimates of the statistical relationship between different variables and the pursuit of establishing causal linkages between policy reform and health and social care outcomes. A few of these variables (e.g. whether a patient is discharged as dead) appear directly in HES, but most (e.g. length of hospital stay, patient severity aggregating across multiple diagnoses, hospital productivity, or intensity of competition to which a hospital is exposed) are constructed using multiple underlying HES variables. HES data will only feature directly in the outputs of these research projects in tables of summary statistics that report properties of key variables used in the analysis (such as minimum, maximum and average values). No data or datasets using HES data will be published or made available, either at the individual patient level or at an aggregate level, as part of this programme of work. The data will not be used for any commercial purpose. Several project outputs are already available in draft form, and are expected to be published in final form over the next 1-5 years. Draft project outputs already produced In the applicant’s field of work, the life cycle of a research project can be roughly summarised as follows: work in progress presentation (slides only); unpublished mimeo; Working Paper (which may also be a final output, or may be progressed to academic journal submission); submission to academic journal (if relevant); and publication of final report or journal article. Irrespective of the final publication location, the research team takes substantial effort to disseminate its outputs by presenting findings to policymakers and policy analysts. The following research projects are at work in progress stage (slides available on request): • Ted Pinchbeck: “Taking care of the budget? Clinical decisions and Patient Outcomes under recent NHS reforms”. • Tom O’Keeffe and Matthew Skellern: “Do altruistic hospitals and profit-maximising hospitals respond differently to competition?”. • Zack Cooper and Stuart Craig: “Home for the Holidays: Evidence on the Relationship Between Prospective Payment, Length of Stay, and Patient Outcomes”. The following research projects are at mimeo stage (papers available on request): • Matthew Skellern: “The hospital as a multi-product firm: Measuring the effect of hospital competition on quality using Patient-Reported Outcome Measures”. • Zack Cooper, Steve Gibbons and Matthew Skellern: “Independent Sector Treatment Centres in the English NHS: Effects on neighbouring NHS hospitals”. The following research projects are at Working Paper stage: • Ted Pinchbeck: “Walk This Way: Estimating Impacts of Walk in Centres at Hospital Emergency Departments in the English National Health Service”, SERC Discussion Paper 167, http://www.spatialeconomics.ac.uk/textonly/SERC/publications/download/sercdp0167.pdf. • Zack Cooper, Stephen Gibbons, Simon Jones and Alistair McGuire: “Does Competition Improve Public Hospitals’ Efficiency? Evidence from a Quasi-Experiment in the English National Health Service”, CEP Discussion Paper 1125, http://cep.lse.ac.uk/pubs/download/dp1125.pdf. The following papers are under review for academic journal publication (papers available on request) • Alistair McGuire and Irene Papanicolas: “Do Financial Incentives Trump Clinical Guidance?: The case of Hip Replacements in England and Scotland” (for slides see: http://www.slideshare.net/OHENews/do-financial-incentives-trump-clinical-guidance-apr15). • Alistair McGuire and Irene Papanicolas: “Measuring and Forecasting Hospital Quality”. Dissemination of draft project outputs This is a non-exhaustive list of the formal and informal methods by which the applicant has disseminated their draft project outputs to policymakers, policy analysts and clinicians. The attached ESRC Outstanding Impacts Application Form by Professor Zack Cooper, a founding member of the applicant’s research team, outlines some of the applicant’s efforts to disseminate draft project outputs to policymakers through to 2013. Since 2013, draft outputs of the applicant’s research projects have been disseminated directly to policymakers in the following ways: • Seminar presentation to Department of Health, July 2015. • Seminar presentation to Office of Health Economics, May 2015 (attended by representatives from Monitor and Department of Health). • Presentation of research to Department of Health group, March 2015. • Meetings to report draft project outputs to representatives of Monitor and Department of Health. In addition, the draft project outputs have been presented to the following fora that have been attended by policymakers and policy analysts: • Health Economics Study Group (January 2015). • Royal Economics Society Meeting (March 2015). • LSE Spatial Economics Research Centre Conference (March 2015). • LSE STICERD Work in Progress Seminar (October 2013) (attended by representatives of Royal College of Surgeons). • Informal workshop involving representatives from King’s Fund, and former Prime Ministerial advisors and heads of regulatory bodies (June 2015). • Joint LSE-Dartmouth College workshop on Medical Practice Variations (September 2014). Finally, the research outputs have been reported in media sources widely read by health policymakers and thought leaders, including (this is a very incomplete list) the Health Services Journal, The Guardian, The New Statesman, The Daily Telegraph, and The Financial Times. For further details, see the attached ESRC Outstanding Impacts Application Form by Professor Zack Cooper. (SD1) Update October 2016: New Work in Progress • Alistair McGuire and Victoria Serra-Sastre (2016), “The relationship between new technologies and workforce in English hospitals”, October. New mimeos (available on request): • Tom O’Keeffe and Matthew Skellern (2016), “Do altruistic hospitals and profit-maximising hospitals respond differently to competition?”, April. • Tommaso Gabrieli, Mireia Jofre-Bonet, Alistair McGuire, and Matthew Skellern (2016), “Patients’ choice and hospital quality competition: Unintended impacts of the signals”, October. • Jose-Luis Fernandez, Alistair McGuire, and Maria Raikou (2016), “Coordinating Hospital Discharges: Bed Blocking in England”, October. New Working Papers: • Ted Pinchbeck (2016), “Taking Care of the Budget? Practice-level Outcomes during Commissioning Reforms in England”, SERC Discussion Paper 192, February, http://www.spatialeconomics.ac.uk/textonly/SERC/publications/download/sercdp0192.pdf. • Zack Cooper, Stephen Gibbons and Matthew Skellern (2016), “Does Competition from Private Surgical Centres Improve Public Hospitals’ Performance? Evidence from the English National Health Service” CEP Discussion Paper 1434, June, http://cep.lse.ac.uk/pubs/download/dp1434.pdf. New academic journal publications: • Irene Papanicolas and Alistair McGuire (2016), “Measuring and forecasting quality in English hospitals”, Journal of the Royal Statistical Society: Series A, May, ISSN 0964-1998. • Irene Papanicolas and Alistair McGuire (2015) “Do financial incentives trump clinical guidance? Hip replacement in England and Scotland”, Journal of Health Economics 44, pp.25-36, ISSN 0167-6296. Since the last application, LSE have disseminated their research to individual contacts in NHS Improvement, Department of Health, the Competition and Markets Authority, and the Health Foundation. In 2015, LSE presented their research on the tension between financial incentives and clinical guidance to a Department of Health Seminar. In September, LSE presented their research on the impacts of hospital competition to the Wennberg International Collaborative on unwarranted variations in health care utilisation and outcomes. This high-level forum included representatives from Monitor/ NHS Improvement, NHS England, and NHS Scotland, as well as senior representatives from other health care systems around the world. In November 2016, LSE will be presenting their research on the impacts of hospital competition to a seminar at the Competition and Markets Authority. Our findings have a direct bearing on the CMA’s decision-making process concerning hospital mergers. In November 2016, LSE will also be presenting their research on the impact of new medical technologies within the NHS to the Health Foundation. This will be a stepping stone to disseminating this research to widely to policymakers and the broader health policy community.
Data provided by HSCIC is stored on a dedicated secure data server housed within the LSE that is only accessible to researchers authorised to use the data. Within the Secure Server, the raw HES data is stored on a SQL Server, providing a second layer of security. Researchers extract only the HES records they need using an ODBC connection that securely imports the required data into Stata. Statistical analysis conducted in Stata is also restricted to take place on the Secure Server -- the data remains on the Secure Server at all times. Individual project data are held in working files on the Secure Server. All data held on the Secure Server is encrypted; see the attached System Level Security Policy for further details. Once final tables of results (e.g. regression tables, summary statistics) are produced, there is a monitored and highly restricted facility allowing researchers to remove such outputs from the Secure Server, to allow reproduction within reports and other deliverables. As noted above, these final outputs contain aggregate provider level data only, do not identify (or allow the identification of) any individual patients or clinicians, and comply with the HES Analysis Guidelines on suppression of small numbers. All data users are required to sign a data use agreement forbidding the removal of patient-level data from the Secure Server. All printing functionality on the Secure Server is disabled. HES and PROMs data have been merged together at the individual patient level using the epikey field provided for this purpose by HSCIC. A small number of other publically available data sources (e.g. North West England unemployment rates) have been merged to the HES/PROMs data, in order to allow researchers to control for demographic or socio-economic characteristics of health care providers or geographical areas at a given point in time. These data sources include postcodes and latitudes/longitudes of health care providers to help define referral markets and areas of potential competition; area deprivation indices to aid in the risk-adjustment of outcomes; and hospital-level data such as annual admissions and NHS Staff Survey results. These data sources are only ever merged on the basis of provider-level fields (e.g. trust code, site code, region of England, or MSOA) and date fields (year, financial year, quarter or month). While it is not feasible to provide an exhaustive list of data sources that will be merged into HES/PROMs or of HES/PROMs fields that will be used for merging – on the grounds that research is fundamentally a discovery process and it may become desirable, in the future, to incorporate new data sources, merged on the basis of hitherto unused (for merging) HES/PROMs fields -- as the research progresses, merging of data will only be undertaken at the provider level and therefore will not compromise the anonymity of patients or clinicians. Any additional data sources used are always (with a single exception, noted in the next paragraph) fully anonymised, publicly available data that do not contain any individual-level information, but report average characteristics of large-scale geographical areas or health care providers at a given point in time. As such, the highly aggregated data that is merged into HES/PROMs cannot be used for patient identification, and cannot increase the risk of patient identification beyond the level of risk that is inherent to the pseudonymised patient level HES data itself. In addition to the small number of fully anonymised, publicly available data sources that the applicant will merge to HES/PROMs, the applicant intends to merge one additional data source to HES/PROMs that is fully anonymised but is not publicly available – namely the World Management Survey or WMS (http://worldmanagementsurvey.org/) which was conducted for English NHS hospitals in 2006 and 2009. The WMS data consists of survey responses by individual hospital managers concerning hospital management practices. There are between zero and two survey responses per hospital trust, with one observation per survey response. This data source will be merged to HES/PROMs using the trust code field. The WMS data is fully anonymised in that it does not contain any personally identifiable information about the hospital managers that completed the survey, other than the trust code. It is not, however, publicly available, in the sense that a research application must be submitted and approved the WMS Oversight Committee in order to obtain the version of the data that contains trust codes. No record-level data will be shared outside of the organisations named in the agreement.