NHS Digital Data Release Register - reformatted

NHS Midlands And Lancashire Commissioning Support Unit

Project 1 — DARS-NIC-05206-L1V6D

Opt outs honoured: N, Y

Sensitive: Non Sensitive, and Sensitive

When: 2016/12 — 2018/02.

Repeats: One-Off

Legal basis: Health and Social Care Act 2012, Section 42(4) of the Statistics and Registration Service Act (2007) as amended by section 287 of the Health and Social Care Act (2012)

Categories: Anonymised - ICO code compliant, Identifiable

Datasets:

  • Diagnostic Imaging Dataset
  • Hospital Episode Statistics Outpatients
  • Hospital Episode Statistics Accident and Emergency
  • Office for National Statistics Mortality Data
  • Patient Reported Outcome Measures (Linkable to HES)
  • Hospital Episode Statistics Admitted Patient Care
  • Mental Health and Learning Disabilities Data Set
  • Hospital Episode Statistics Critical Care
  • Bridge file: Hospital Episode Statistics to Diagnostic Imaging Dataset
  • Bridge file: Hospital Episode Statistics to Mental Health Minimum Data Set
  • Bespoke Monthly Extract : SUS PbR A&E
  • Bespoke Monthly Extract : SUS PbR APC Episodes
  • Bespoke Monthly Extract : SUS PbR OP
  • Bespoke Monthly Extract : SUS PbR APC Spells
  • Bespoke Extract : SUS PbR A&E
  • Bespoke Extract : SUS PbR APC Episodes
  • Bespoke Extract : SUS PbR APC Spells
  • Bespoke Extract : SUS PbR OP
  • Bridge file: Hospital Episode Statistics to Mortality Data from the Office of National Statistics
  • Patient Reported Outcome Measures

Benefits:

As illustrated within the examples above, the work provides customers (CCGs, Trusts, Local Authorities for the purposes of public health and social care, CQC, Public Health England, Department of Health), Clinical senates, Strategic clinical networks, NHS England, Monitor, Trust Development Agency (TDA)), and health charities) with understanding and insight that enables them to make the best decisions about the healthcare services they commission or provide. Improved decision making will have a direct effect on the quality of care and outcomes for patients. The CSU's work does not go outside the health and social care arena and outcomes will only be used by health and social care organisations. In relation to each of the services stated above, the specific benefits are :- a. The QIPP opportunity packs provide estimates of the potential scale of opportunity if best performance levels of comparator CCGs/trusts are achieved. These packs will help CCGs/trusts to prioritise their QIPP programmes to maximise potential benefits. The Quality, Innovation, Productivity and Prevention (QIPP) programme is a large-scale programme developed by the Department of Health to drive forward quality improvements in NHS care, at the same time as making up to £20 billion of efficiency savings by 2014/15. b. The PROMS project will help both patients and clinician make the best decisions about whether to undergo major surgery. It will help to minimise both the financial cost of carrying out operations that are unlikely to have a beneficial impact for the patient and avoid painful and debilitating surgery for some patients who are unlikely to gain significant benefit. c. The modelling of future mental health demand will provide commissioners with detailed projections about the potential future demand for services. Understanding this is key to the effective functioning of commissioner organisations if they are to make sound decisions about changes to future service requirements to meet expected demand or to commission new targeted preventative services to mitigate against projected demand growth. d. The impact of demographics project will provide CCGs with valuable insight into how healthcare need changes not only as we get older but also as we approach the end of life. Given that healthcare utilization increases significantly at this time a clearer understanding of the patterns of utilization at this stage of life will enable commissioners to provide the most appropriate and cost effective services for these patients. The CSU are not the end user of the outputs they produce however they regularly receive positive feedback from their customer base and currently receive repeat custom from around 75% of customers. Regarding the CSU's customer base for the above projects: As mentioned above, with the introduction of STPs, the CSU's customer base has rapidly become a collective local ‘health and care economy’, comprising of a number of different organisation types within the NHS. For this reason, all the parties involved in STPs are referred to as customers . This is because they are now all party but also because, depending on how the STP operates, in some cases different programmes of work within the STP are split between constituents and so the CSU might actually be directly commissioned by any of them on behalf of the collective. Local authorities are often participants and therefore customers of facilitated modelling exercises that the CSU run to assist a health economies to estimate the demand and supply of health and social care services. Walsall MBC employees for example are currently participating in an exercise of this type in the Black Country. See purpose C. Public Health England have recently added the CSU to a framework agreement to supply data analytics and impact assessments. Work via the framework agreement will be issued over the next few months. The CSU anticipate that specifications for these projects will be published by PHE in the public domain. NHS England have recently commissioned the CSU to model the potential shifts in activity and the implications for accessibility that might arise form a potential reconfiguration of PPCI, vascular surgery and hyper acute stroke services across the West Midlands. This is in support of STPs across the region and therefore those who are party to STPs. Along with NHS England, NHS Improvement (previously Monitor and the Trust Development Agency) have responsibility for assessing and signing off Sustainability and Transformation Plans( STPs). The CSU have supported the Black Country, Staffordshire and Hereford and Worcestershire STP footprints to develop their STPs. NHS England and NHS Improvement are the ultimate customers for this work. In 2014/15, Cancer Research UK (CRUK) commissioned the CSU to model the future demand for endoscopy services in the UK. The CSU are in discussions with CRUK to test the feasibility of estimating the impact of cancer awareness campaigns on the identification of non-cancer diseases (e.g. do lung cancer awareness campaigns increase the diagnosis of COPD) and the subsequent impact on the demand for health services. That national work is now also being used to support individual STPs in addressing the future challenges they face in terms of endoscopy capacity.

Outputs:

QIPP opportunity packs (Objective A) – these packs contain a variety of comparative charts and tables plus bespoke ‘deep dive’ reports identifying the potential scale of opportunities available relating to areas of activity that may be amenable to common admission avoidance strategies. These packs are bespoke for an individual CCG/trust will be only be provided for CCGs/trusts who commission a pack. The CSU are continuing to produce these packs for a significant number of CCGs across the country. In 2015 the CSU provided around 30 CCGs with bespoke QIPP opportunity packs and have continued to further develop the product to offer comparisons with national nearest neighbours and bespoke comparator sets. In addition, the CSU has supported a number of CCGs to further explore issues highlighted by the packs by providing more in depth analysis using the data. The CSU anticipate a similar level of demand for this product in the coming months in order to support healthcare systems to develop their Sustainability and Transformation Plan (STP) plans. The packs are expected to be produced before the end of December 2016. PROMS decision support tool (Objective B) – the output of this project will be a web based tool that allows clinicians to input patient characteristics and receive an estimate of likely benefit of undergoing the procedure. The tool contains no patient data. There is currently no confirmed target date for the tool's completion, however it is still being developed and the CSU are keen to continue to refine and test the methodology with a view to piloting it with an interested partner CCG or other NHS organisation. Mental Health activity modelling (Objective C) -the output of this project will be a final written report containing charts, tables and written commentary. Aggregated summary data tables may also be provided and these will not be at patient level and small numbers will be suppressed. The report will be provided only to the project commissioner and will not be published or circulated. Mental Health intervention specific modelling -the output of this project will be a final written report containing charts, tables and written commentary. Aggregated summary data tables may also be provided and these will not be at patient level and small numbers will be suppressed. The report will be provided only to the project commissioner and will not be published or circulated. The CSU has carried out mental health modelling projects using the Mental Health dataset in Warwickshire and Shropshire which provided estimates of the future levels of demand for mental health services in the medium term. The CSU are due to commence work on new modelling projects for Walsall CCG and Sandwell and West Birmingham CCG imminently, with completion expected by end of February 2017. Impact of demography (Objective D) – the primary outputs of this work will be written reports containing appropriate charts, tables and written commentary. The report will be provided to the commissioning organisation and may be shared with other interested organisations with the permission of the commissioner. The secondary output will be in the form of an improved methodology to estimate the impact of population changes on healthcare utilisation. This methodology will be applied in future modelling projects but it would not require the re-use of the mortality data. The CSU have made progress in understanding how proximity to death may be applied to provide better estimates of future healthcare demand but a final methodology is not yet complete and this objective is still on-going. The CSU has carried out two projects looking at patterns of acute healthcare utilisation for patients in the 12 months prior to death. The first was a regional project commissioned by the West Midlands Strategic Clinical network which benchmarked patterns of acute healthcare utilisation for patients in the last 12 months of life. The second was a more in depth local analysis again using the ONS and HES data sets to further investigate seemingly high levels of utilisation in Dudley CCG and to provide the CCG with a more robust understanding of local provision of care to patients in their final year of life. Data presented within all outputs is aggregated with small numbers suppressed in line with the HES Analysis Guide.

Processing:

The data will be stored on a secure server and accessed through a SQL server database by a small group of named analytical staff working within the Strategy Unit of the CSU. Those staff are based at the premises detailed in this application (Kingston House). The data in its raw form will not be loaded into any tool or provided as part of any product or output. All outputs will contain only data which is aggregated, with small numbers suppressed in line with the HES Analysis Guide. SUS PBR As detailed in the “Objectives section” (Objective A) accessing the national SUS PBR data will enable the CSU to offer the QIPP packs to all CCGs/trusts in England as well as allowing the CSU to improve the packs through the use of better comparative groups (i.e. nearest neighbours). In producing these packs the data required is extracted using SQL server and analysed using MS Excel to produce the charts and tables included within the packs. PROMS As detailed in the “Objectives” section (Objective B) the PROMS data will be used to develop a decision support tool, PROMS data will be extracted from SQL server and analysed using appropriate statistical analysis software (STATA or R) in order to establish the relationship between a range of patient characteristics (e.g. age, gender, co-morbidities) and the procedure outcomes based on PROM scores. The tool that will be developed will not contain any patient data. The tool that will be provided to the customer(s) will only contain a mathematical algorithm based on the established statistical relationships between patient characteristics and outcomes. Mental Health Minimum Dataset (MHMDS) The MHMDS will be used to model expected future activity levels and capacity requirements within CCGs after taking into account the impact of projected demographic changes and also the potential impact of mental health prevention strategies, admission avoidance strategies and length/intensity of treatment reduction strategies. Patient level data is required to enable the CSU to adjust and remove activity in line with expected changes. Using patient level data also allows the CSU minimise the impact of overestimating impacts as a result of double counting which is not possible with aggregate data. As outlined in the “objective” section the data will be used in two ways firstly it will be used to provide supporting benchmarking and historical trend analyses to support modelling parameter setting. For this aspect of the project data extracts will be produced using SQL server and downloaded into MS Excel to produce the charts and tables required. Secondly it will be used to create a model to estimate future activity levels after accounting for changes in demographics and the impact of changes to service provision. The model will be constructed using SQL server to process the data applying any modelling factors and parameters. Aggregate output files from SQL server will be downloaded and analysed in MS Excel in order to produce the required charts and tables for inclusion in reports. The dataset will also be used to develop prospective intervention specific models to estimate changes in mental health team activity levels and the scale of potential savings as a result of the introduction of specific strategies to reduce the need for mental health services. These strategies may include, for example, schemes to increase early diagnosis of mental health conditions. This will help the CCG to better understand the costs and benefits of proposed changes allowing them to make better decisions about the effective use of commissioning resources. As with the higher level modelling in order to develop specific intervention impact models requires the production of benchmarking and trend analyses to help the customer to make judgments on the likely scale of impact of specific interventions. These judgments are incorporated into the model so it is important that they are based as far as possible on the best available data available. These prospective models will be constructed within SQL server and aggregate outputs downloaded into MS Excel to produce required outputs. The reports and any accompanying data tables will contain only data which is aggregated, with small numbers suppressed in line with the HES Analysis Guide. ONS mortality data As detailed in the “Objectives” section (Objective D) the ONS mortality data combined with the national HES data will be used to understand how the nature and scale of healthcare utilisation changes as a result of changes in demographics. It will also allow the CSU to develop a new approach to estimating the impact of an ageing population on future healthcare demand. As with the other datasets the ONS data will be stored within a SQL server database and the data required for this analysis will be extracted and analysed within MS Excel or other appropriate statistical software packages such as STATA or R in order to establish the mathematical relationship between proximity to death and healthcare utilisation which can be used in future (and potentially some of the current modelling work outlined in this document). During these data transfers into appropriate analysis software packages the data will not leave the secure environment. Any other projects that may make use of this work (for example the NHS England Fit For the Future programme) would only utilise the methodology derived from this project and would not use the actual ONS data. The reports and any accompanying data tables will contain only data which is aggregated, with small numbers suppressed in line with the HES Analysis Guide. Across all of the above processing, processing will be only carried out by CSU staff with the appropriate governance and access. The data will not be used to link at record level to other datasets (other than where already provided in linked or bridging form by NHS Digital). The data may however be linked to organisational level data such as already exists within the public domain. For clarity, the DSCRO may not process the data for the CSU other than initially downloading the data and storing it on the servers accessible by the CSU, and hence is not listed as a data processor, furthermore, NHS England have confirmed they are happy for this DSA to reflect the request for access to data to sit under the contract which results in NHSE having responsibility for the receipt, use, storage and any dissemination of the data by the CSU.

Objectives:

To support contractual and strategic benchmarking across Midlands and Lancashire, for programmes such as planning commissioning and productivity, service quality and performance improvement, and activity and outcomes monitoring for local populations. The CSU needs: • The provision of analytically based intelligence for a range of CCGs for benchmarking of similar health economies or populations in England, not just in the CSU’s area. • To provide in depth analysis of all aspects of a specific service areas and allow comparisons with other CCG areas or health economies known to have better outcomes or new/different pathways. • To support large scale transformation projects that may impact several commissioners (CCGs) • Descriptive analyses of healthcare needs, demands or supply including comparisons between providers, commissioners and geographical areas, analysis over time and of the characteristics of patients and the services they receive. • Retrospective analyses exploring the reasons for observed changes in healthcare provision and health outcomes • Prospective modelling of the impact of planned or proposed changes in healthcare services on healthcare activity, travel times and resource use • Quantitative evaluations and monitoring estimating the impact of service redesign of improvement initiatives on healthcare and outcomes • To develop tools and information packs to support patients, clinicians, commissioners and providers to make informed decisions about healthcare service provision, organisation and strategy The specific services and products that will utilise the data are the following :- A. QIPP opportunity packs which provide a summary of performance, cost and activity levels for individual CCGs/trusts compared to other local CCGs/trusts. The packs include aggregate analysis in relation to QIPP priorities covering Inpatient, Outpatient and A&E but are subject to change in line with the QIPP programme. These packs were originally produced for those CCGs within the CSU's core geography (Birmingham and the Black County). However the CSU have now been requested to provide packs for a wider range of CCGs and trusts including all Staffordshire, Lancashire, Herefordshire, Worcestershire, Shropshire and Telford and Wrekin. The CSU have also had requests from as far afield as Cornwall. The value of these packs (as demonstrated by the willingness to pay) in supporting CCGs/trusts to assist with their statutory duty to commission/provide high quality and best value services for their populations is clearly proven and as such the CSU will be offering the packs to all CCGs trusts in England. In addition to the wider provision of packs the CSU's existing customers have also requested that the packs be enhanced to offer comparisons against national nearest neighbour comparators or bespoke comparators (for example Birmingham combined CCGs compared with other large cities). Customers for the packs also can request ‘deep dive’ analyses to explore identified opportunities in greater detail B. Development of decision support tools for clinicians to help them make better decisions when deciding whether a patient is suitable for Hip or knee replacement procedures. The development of the tools requires sophisticated statistical analysis to establish the relationship between a range of patient characteristics and procedure outcomes (as measured by PROMs data). The statistical relationships will be used within the tools whereby it will allow a clinician to input patient characteristics and provide an estimate of the likely benefit of the procedure for the patient. This additional information can help both the patient and their clinician make the best informed decision about whether to proceed with the operation. In order to ensure that that relationship is as robust as possible and to maximise the predictive power of the tool (which is vital given that the tool will be used to support important decisions about patient care) a full national dataset is required. In order to further validate the relationship and establish its robustness over time (which will be important for clinician and patient confidence in the tool) the CSU will be carrying out the analysis on all data years. The development of these tools will establish a prototype for the development of other similar products for other procedures where data is available through the PROMS dataset such as Varicose vein surgery etc. However for the purposes of this request the CSU are requesting only PROMS data relating to hip and knee procedures. A number of Local CCGs with programmes aimed at improving orthopaedic services (across all of Staffordshire for example) have confirmed that they plan to put this tool into practice on an initial pilot basis as soon as it is available. The CSU have also been approached by a number of other CCGs who have indicated that they would also be interested in applying the tool once its efficacy has been established. C. Projects on behalf of CCGs and Strategic Clinical Networks (part of NHS England) to model expected future Mental Health activity levels and capacity requirements within a CCG after taking into account the impact of projected demographic changes and also the potential impact of mental health prevention strategies, admission avoidance strategies and length/intensity of treatment reduction strategies. An integral part of this work is to elicit modelling parameters from clinicians and commissioning stakeholders relating to expected impacts on activity levels as a result of planned changes or interventions. In order to do this the CSU produce a range of supporting analyses to help them to understand current activity levels, trends in activity and also how they compare with others. Provision of this supporting data is key to helping stakeholders to make considered and robust estimates based on a clear understanding of past progress and performance against other relevant comparators. In order to provide this comparative benchmarking the CSU require full national datasets covering multiple years. As the CSU are requesting the full set of historical data, they felt it important to clarify their rationale for doing so. In terms of the number of years of data requested, the CSU's professional experience has shown that providing longer term trends (in excess of 5 years) is often important, given the level of variation that exists, in order to evidence general trends. Being able to show local trends in the context of national trends is also essential for sophisticated interpretation. Shorter time series can often be misleading in this respect and as such could result in incorrect assumptions about future levels of demand. D. Projects on behalf of CCGs to understand how the nature and scale of healthcare utilisation changes as a result of changes in demographics. One specific aim of this work (for which ONS mortality data is required) is to investigate how patient need, demand and service utilisation changes towards the end of a person’s life. In addition it will also allow the CSU to develop a new approach to estimating the likely impact of an ageing population on future healthcare demand. The new approach will take into account not only the future size and age structure of a population but also changes in the proportion of the population who are estimated to be in their final months of life. It is also worth noting that NHS England have expressed interest in the CSU's development of this method of forecasting future demand as part of their national Fit for the Future programme (FFF). The project requires national level datasets in order that the analysis is as statistically robust as possible. It will also allow the CSU to establish the extent to which utilisation prior to death varies across the country. Benchmarking analysis (including historical trend analysis) will be carried out in order to provide estimates of the potential scale of opportunity for reducing acute healthcare activity (or developing alternatives to acute provision) for those patients at the end of life. Benchmarking and trend analysis will also enable the identification of those Trusts or CCGs who may be more advanced in end of life care provision. As part of this project the CSU will also be considering how patterns of utilisation at the end of life have changed over time (advances in medical technology and new treatments will certainly have had an impact on levels of service utilisation particularly for older people). Long term trends in excess of five years will be important in order to identify and have confidence in historical trends and applying these trends to future estimates. E. Other specific projects are: 1 Describing changes in acute utilisation over the long term provides insights that are lost when focusing on the most recent past. Striking reductions, for example, in casemix-adjusted length of stay following an emergency acute hospital admission or the frequency of admissions to psychiatric inpatient units only really become apparent when viewed over a long time frame. These longer term perspectives demonstrate the enormous positive changes that have been achieved in the past and can motivate and guide health economies seeking improvements in areas that seem equally intractable. To delete older data would eliminate the potential for these insights. The CSU have deployed this kind of longitudinal analysis (going back to pre 2000) recently in support of several Sustainability and Transformation Plans (STPs - compromising of CCGs, trusts, Foundation Trusts, Local Authorities and other key local partners) as they seek to address the requirements placed upon them nationally. 2 When explaining historical acute hospital utilisation rates, or forecasting future rates, the longer the time series, the more robust (on average) the explanation or forecast. Whilst for time series models, it might be argued the diminishing returns result from adding very old data points, this is not necessarily the case for causal models. 3 The CSU are frequently asked to model the potential implications of new models of care. These ‘new’ models are more commonly reinventions or adaptations of earlier models. The ‘NHS Five Year Forward View’ describes a number of new care models which move away from a purchaser-provider split in favour of lead-provider arrangements. To many these proposed models mirror or approximate arrangements that existed in the NHS prior to the development of primary care trusts. If analysed and interpreted appropriately, data relating to these earlier periods can provide useful insights into the unintended consequences of ‘new’ care models and the CSU are being asked to do this to support STPs and national Vanguards in meeting the national requirements placed upon them. Data will only be used for the purposes outlined above, and any requirement to change the purpose will be subject to a separate request to NHS Digital.