NHS Digital Data Release Register - reformatted
Private Healthcare Information Network (PHIN)
Project 1 — DARS-NIC-13906-G0F3F
Opt outs honoured: No - data flow is not identifiable (Does not include the flow of confidential data)
Sensitive: Sensitive, and Non Sensitive
When: 2016/04 (or before) — 2020/12.
Repeats: Ongoing, One-Off
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
- Patient Reported Outcome Measures (Linkable to HES)
PHIN has been appointed by the Competition and Markets Authority (CMA) to the role of ‘Information Organisation’ charged with implementing the remedies set out in the Private Healthcare Market Investigation Order 2014. The Order requires every operator of a private healthcare facility to supply PHIN with “information as regards every patient episode of all private patients treated at that facility, and data which is sufficiently detailed and complete to enable the information organisation to publish [a specified list] of performance measures by procedure at both hospital and consultant level.” [Private Healthcare Market Investigation Order 2014, Article 21.1] The Order required PHIN to “prepare and submit to the CMA for approval a five-year plan, which has been developed in conjunction with, and approved by, its members, setting out how it proposes to collect the information specified in this Order and the basis on which it may licence access to this information” [Article 24.1]. In response, PHIN produced its Strategic Plan 2015-2020 which outlines its proposals for implementing the remedies. CMA approved the plan. PHIN’s approved plan requires routine extracts of HES data from the HSCIC. This will include the necessary data on NHS-funded care delivered in independent hospitals and privately-funded episodes delivered in NHS hospitals. Receiving this data from HES, as opposed to directly from the private healthcare facilities, is advantageous both in minimising the duplication of data and data transmission routes – meaning that care providers will only have to transmit their information once (reducing both effort and the possibility of data transmission errors) – and crucially in ensuring consistency in data quality. Receiving HES data mitigates the risk of inconsistencies in the data resulting from PHIN and the HSCIC processing the raw data in different ways if both were to receive it directly from care providers. HES also contains important data that PHIN cannot replace by direct submission; this is principally the NHS-funded NHS-provided data comprising 85% of elective care activity that PHIN needs to provide the benchmarks against which to compare the private sector. That data is needed at hospital, consultant and procedure level. To report the whole of a consultants’ practice, it is necessary to consider the work that they do within the NHS as well as privately. The CMA’s Final Report states its expectation that data submitted by the private hospital operators to the information organisation (PHIN) should be “be fully comparable with that collected by the NHS to allow the information organization to report performance measures for the whole of consultants’ practices, both NHS and private, since this is the relevant basis on which to judge performance” [Article 11.486]. To achieve this PHIN must include in that comparison the 85% of elective episodes that are both NHS funded and provided. Hence data is needed for the NHS episodes. Although PHIN’s interest in consultants’ practice within the NHS is limited to those consultants that also have a private practice, PHIN need to receive all NHS episodes (non-emergency APC), because the HSCIC cannot determine from the data it holds which episodes belong to consultants who practice privately (where that private practice is conducted outside the NHS). This group of consultants is also a constantly changing population and as such it is not practical for PHIN to seek to amend the list of consultants for which it requires data, as gaps in a consultant’s data may be as a result of leave or switching from one private hospital to another. However, PHIN will not produce or publish indicators on consultants that do not have a private practice and such consultants will not be able to view their data via PHIN’s portal. Continuous inactivity will trigger the removal of a consultant’s indicator from the site. PHIN will establish the following rules to manage the circumstances where a consultant has either been inactive for a prolonged period of time or is no longer on the live list of GMC registered consultants. 1. Each month PHIN’s systems will automatically check the latest available GMC registration status of all consultants contained within PHIN’s database. Only data for “live” consultants (Registration Status is “Registered with a licence”) will be processed and published on the public web site. 2. As soon as there is 12 continuous months of no Admitted Patient Care data within the database for a particular consultant, they will automatically be removed from the public web site. This will apply to all their indicators as well as their public profile. 3. In such circumstances the consultant will be automatically notified by email of this event and PHIN's intention to remove them from the site. Unless PHIN receives a communication challenging this action, they will be removed. 4. If activity subsequently appears for the same consultant then they will automatically re-appear within the web site and an email will notify them of this fact. PHIN also requires linked HES/PROMS data to deliver “procedure-specific measures of improvement in health outcomes, as agreed by the information organisation and its members to be appropriate” (CMA Order Article 21.1(j)). The CMA’s Final Report states: 11.571 In order to facilitate the analysis and publication of meaningful performance statistics, we would expect the data provided by the private hospital operators to: c) contain diagnostic and procedure coding1096 for each episode in order to allow for risk-adjustment where appropriate—diagnostic coding should include full details of patient co-morbidities; In addition to the public access to the PHIN website that everybody will have, PHIN will grant licensed access to its information to consultants and to providers of private care. Secure, authenticated access will be granted to information specific and appropriate to the particular consultant: this will include detailed (but pseudonymised) views of the care for which they are directly responsible, alongside relevant totals, averages and benchmarks. Commercial confidentiality will be respected alongside patient confidentiality, and no party will have inappropriate access to details related to their peers and competitors. As such, no party will have access to the “database”. Users will be required to accept licence terms covering information governance, intellectual property and so on when they access the Portal. There will be no charge for this licensed access. This service is not explicitly mandated by the CMA Order but is described on pages 14 and 17 of the PHIN Strategic Plan 2015-2020. As stated in the Chairman’s foreword (p.3), one of PHIN’s aims is to “help private providers continuously improve their care and clinical outcomes”. PHIN will do this by enabling them “to see and understand performance measures in context including with peer group benchmarks” (p14). This is the type of service routinely provided within the NHS by the HSCIC (NHS Comparators), Dr Foster and many other means. However, there has never been any central collation of data in private healthcare and hence no information on comparative performance from which to learn. The lack of that information particularly disadvantages the standalone (usually charitable) hospitals, as the larger national providers can at least compare between their own hospitals. It has also meant that the CQC has a very limited view of private activity and quality, and the CQC has asked PHIN to help address that gap. Confidence in the data underpinning any analysis and performance measurement is essential. For that reason, it must be open to scrutiny and checking, such that providers and consultants can validate that the data is accurate and fix it where it is not. It is for that purpose that PHIN make the record-level data available for checking, only to the providers and consultants with whom the data originated as those responsible for the episode. They can then assure themselves that all of their data is present and accurately recorded.
2018 has seen significant milestones from the CMA Order met. Below are the key external milestones PHIN has delivered over the last 12 months, as PHIN build towards increased transparency of quality and safety. November 2017 - Consultant portal launched. Providing whole practice episode record data to consultants for the first time, with the ability to review and feedback on data inaccuracies for their private practice. June 2018 - Consultant measures review and sign-off. Consultants asked to review and approve data for their activity numbers and length of stay for both NHS and private activity. Consultants from across specialties have been actively working with PHIN and their hospitals to review and improve their private clinical practice data, and over 1,600 consultants have approved their first performance measures and are now searchable on PHIN’s website with their activity. While NHS HES attribution of activity has not thus far had sufficient accuracy to support publication for all consultants, the option to publish performance measures based on whole practice measures has been seen as a major benefit by many consultants who have been keen to promote the full breadth and volume of their activity. This allows PHIN to publish a complete picture of consultants activity to assist patient choice. February 2018 - Two additional hospital measures published. Patient Experience and Health Outcomes Participation both published on PHIN’s website. For patient satisfaction, over 66% of providers have provided sufficient PROMs or QPROMs data to publish a meaningful participation score. These enable patients to gain a fuller picture of the experience of care in different settings and lays the foundation for the comparison of outcomes within the private sector and between the private sector and the NHS. September 2018 - Consultant measures publication. First two measures published for 1,000 consultants with private practice, and volume transparency introduced for hospitals. PHIN has sufficient quality data to publish volume and length of stay for 326 hospitals. This covers over 90% of elective procedures in the private healthcare market. This enables patients to assess the experience that particular consultants and hospitals have in performing the procedures they may require.
PHIN’s over-arching mission is two-fold: to enable patients to be able to make better informed choices about their healthcare providers and, through the provision of comparative information, to help private providers continuously improve their care and clinical outcomes. Whilst a small proportion (around 5%) of the 10 million or so patients encountering the UK independent hospital sector annually come from overseas, usually in a handful of central London hospitals, the overwhelming majority of patients are also NHS patients for most of their care (GP, maternity, A&E, end of life, emergency and most elective etc.), simply opting to take some elective care privately. Crucially, neither HSCIC nor NHS Choices has access to private episode data from independent hospitals (Private HES or “PHES” data), nor a mandate nor funding that would enable them to collect that data to form a full view of the private hospitals from which NHS funded care may also be being commissioned and delivered. Consequently, for example, the CQC has found that the data required to inform proper regulation is not routinely available for independent hospitals as it is for NHS providers. The CMA Order enables PHIN to licence this PHES data to interested external third parties to support information gaps such as these. Consequently, the Health and Social Care system has no means of properly understanding private healthcare including, for example, determining the extent to which patient deaths or complications following treatment in the private sector places a burden on the NHS when they result in emergency admissions into NHS hospitals. Similarly it is blind to the extent to which private patients require an emergency transfer of care to the NHS. PHIN aims to fill those gaps and address those deficiencies, by the methods described above, for the benefit of patients. Furthermore, as Patient Choice frequently includes NHS funded treatment in a private hospital, the PHIN website will be the only source of information for these patients which describes the totality of care provided by these hospitals, being the combination of their private and NHS funded activity. This is a much more comprehensive indication of their performance than the partial information available from the NHS Choices site which is only based on the NHS funded element of their workload. PHIN’s use of the data requested in this application will therefore facilitate new understanding and inform quality improvements within the private healthcare sector, along with facilitating improved regulation, commissioning and policy making, leading to improvements in the quality and management of care that will benefit UK citizens and tax payers generally.
PHIN will calculate and publish on its public facing website (www.phin.org.uk) the following indicators derived from a combination of HES and PHES data: • volumes of procedures undertaken (by hospital and by consultant); and • average lengths of stay for each procedure (by hospital and by consultant). • Procedure-specific measures of improvement in health outcomes, as agreed by the information organisation and its members to be appropriate Indicators will be presented as iconic, graphical and numerical visualisations, similar to NHS Choices and other public health websites, with the specific calculated values for the selected hospital or consultant presented within a statistically robust and comparative context which will include a sector average. Each indicator will be accompanied by interpretive and methodological information. Each indicator will also include explanatory information and descriptive information for each hospital and consultant. Information for the Public Members of the public will access the performance indicators at www.phin.org.uk. Indicators containing the requested data will be presented as iconic, graphical and numerical visualisations, with calculated values for each hospital and consultant presented within a statistically robust and comparative context which will include one or both of an independent sector and a NHS sector average and conforming to rules on small number suppression. Where appropriate, indicators will be risk adjusted using methodologies approved by relevant clinical and/or academic bodies. The CMA Order requires PHIN to subject these and all its methodologies to external, independent scrutiny (see CMA Order Article 24.5). Each indicator will be accompanied by interpretive and methodological information and each hospital and consultant will be accompanied by descriptive information drawn from other data sources but independent of and not linked to the data requested under this application. Data Quality and Data Validation by Private Healthcare Facilities For the sole and specific purpose of data quality and data validation, each of the three types of performance indicator prescribed in the CMA Order together with the underpinning pseudonymised record level data will be accessible to each private healthcare facility and to authorised individuals from PHIN. Essentially this will entail the hospital confirming that the numerator and denominator values are correct for each of their procedures that is going to appear on the public web site. Time series analyses of the data will also help reveal unexpected patterns that may point to missing data. The portal within which this process will take place will include functionality for queries against the data to be automatically directed to the relevant (authorised) individual from the hospital site or group in question. Such queries will provide specific feedback on the highlighted issue and workflow will track their subsequent resolution and outcome. If necessary, data will be corrected at source and refreshes passed through to PHIN as part of the routine data submission process. This data will be made available through an online reporting tool via a secure portal requiring a validated username and password. User access will be granted in line with agreed data sharing protocols, where appropriate this will be approved by the local Caldicott Guardian. User’s login credentials will restrict the data to which each user has access, which means that users from specific hospitals will only be able to see record level data originating from their hospital. All record level data will be pseudonymised and contain no patient identifiable data and all data will at all times remain solely on PHIN’s servers – it will not be possible for users to move this data to another location. This data has no commercial value in that it relates solely to the care which the hospital itself provided. Data Quality and Data Validation by Consultants with NHS and Private Practice For the sole and specific purpose of data quality and data validation, each of the three types of performance indicator prescribed in the CMA Order together with the underpinning pseudonymised record level data will be accessible to each consultant and to authorised individuals from PHIN. The process whereby this data validation takes place is the same as that described above for hospitals. This data will be made available through an online reporting tool via a secure portal requiring a validated username and password. User’s credentials will restrict the data to which each has access, which means that a specific consultant will only be able to access his or her indicators and associated record level data. Furthermore, this validation process, whereby PHIN will require consultants to actively opt-in to having their activity published as performance indicators (by means of an electronic sign-off), may have the beneficial effect to the NHS of having consultants checking their HES data for errors for the first time. All record level data will be pseudonymised and contain no patient identifiable data and all data will at all times remain solely on PHIN’s servers – it will not be possible for users to move this data to another location. This data has no commercial value in that it relates solely to activity for which they were identified within the data as the responsible clinician. Timeline for Publication The CMA Report and associated Order requires that its indicators will be published from April 2017 onwards and that they must be based on at least 12 months of data. Thus PHIN has between now and January 2016 to develop and evaluate its indicator methodologies, from which point onwards all data flows must be identified, established and operational.
Periods of Data Requested At a minimum, PHIN will need data from 2013/14 onwards. This will enable it to begin analysis now on a complete 24 months of “official” HES data and ensure alignment and testing of case mix adjustment methods, their validation by relevant expert bodies and socialising the outputs with all relevant stakeholders - in particular the organisations and individuals covered by the order (approximately 500 private hospital operators and 12,500 consultants). The date by which PHIN must publish information for patients is 30 April 2017 and will be based on data for the calendar year 2016. Frequency of Data Requested HES data will be required on a monthly frequency. Although the CMA states only a quarterly data extract from the private hospital operators, PHIN’s view, noting that the NHS data processing to produce HES takes at least three months, is that in order for the data to describe the most current view of hospital and consultant activity it requires monthly updates. This frequency also coincides with the extract schedule for data received from the independent hospitals. The indicators on the website will be updated on a monthly basis so it would be optimum to synchronise the frequencies of data collection and reduce delays to data availability as much as possible. Furthermore, and to address the “provisional” nature of this data, PHIN is also requesting Annual Refreshes each year. Periodic indicators will cover a rolling 12-month period and will be recalculated to reflect the final data. Data Storage and Processing PHIN will not perform any record level linkage between the requested data and any other patient data it currently holds or may hold or have access to in the future. The data will only be used to fulfil the CMA requirement to provide information for the general public, and data will not be provided to private healthcare providers that could then be used for sales or marketing purposes. PHIN stores and processes data in accordance with its Information Governance Policy, which is aligned to the NHS IG Toolkit and HES licence terms in the form of a System Level Security Policy. All record level data requested under this application will be stored on PHIN’s own servers hosted in Claranet’s ISO27001-accredited data centre in Bristol. Record-level data is held only electronically within this data centre. Access to record-level data is strictly limited to nominated PHIN-authorised persons required to process or check it. All database administration and Extract, Transform and Load (ETL) processing is under the control of named PHIN-authorised staff who can only access this data via a dedicated internet connection into PHIN’s offices. All computers accessing the data centre are located in PHIN’s offices, are password protected, encrypted and owned and administered by PHIN. PHIN uses SQL Server 2012 to store data and Tableau to analyse and report on it. In order to ensure consistency in methodology with other publicly available indicators (e.g. NHS Choices), PHIN agree to work with the HSCIC’s assurance processes to ensure their methodology is suitable peer-reviewed. Indicators based on aggregate data will be published on PHIN’s public facing website (www.phin.org.uk) through a series of reports. PHIN has a staging server that is only accessible internally where record level data is loaded and aggregate data processed. It has a SQL server port open and an ETL process that moves the output aggregate tables onto an output database server, which resides on internal Virtual Local Area Networks within PHIN’s network. The third layer is the web server that runs the website. This is accessed (controlled by firewall) through specific ports over the internet (different depending on whether the site requires an SSL certificate and is HTTP or HTTPS). The website can then communicate (through another firewall) with the output database server through the appropriate SQL port. Combining HES and PHES Data For the purposes of whole-hospital reporting, where an independent hospital is also providing NHS-funded care, HES data and the data describing the treatment of private patients in an independent hospital (PHES) data will be combined under the hospital’s site code. No patient or record-level data linkage will occur between these two datasets. For the purposes of consultant whole-practice reporting where a consultant provides care in both independent and NHS hospitals (where they are identified within the data as being the responsible consultant), HES and PHES data will be combined under the consultant’s GMC code. No patient or record-level data linkage will occur between these two datasets. For the avoidance of doubt, the PHES dataset that will used for the two purposes listed above will be pseudonymised in terms of patient identifiable data. Performance indicators produced as a result of combining the English PHES and HES data will only be compared to similar results for hospitals in Wales, Scotland and Northern Ireland. HES data will not be combined with its equivalent data from any of these Nations. Data Destruction PHIN will permanently destroy all record level HES data submitted as part of the monthly data flows as soon as it receives the associated Annual Refresh data. PHIN will hold a maximum of five years of finalised, annual data at any time, destroying older data on a rolling basis. Risk Adjustment and Standardisation Subject to the availability and quality of certain relevant data (e.g. diagnoses, patient age etc.), indicators will be adjusted with respect to these variables and calculated using appropriate statistical methodologies in order to support comparative analysis and presentation. Wherever possible such adjustments will be guided by NHS best practice in order to enable comparisons within and across healthcare sectors. The CMA Order requires PHIN to subject these and all its methodologies to external, independent scrutiny (see CMA Order Article 24.5). Calculation of Performance Indicators The HES data will be used to generate the indicators prescribed in the CMA Order, based on pseudonymised aggregated or combined PHES and HES data, conforming to rules on small number suppression and available at hospital, consultant and procedure level. The information will be published as indicators on a publically accessible website (www.phin.org.uk): • Volumes of procedures • Length of stay Similarly, procedure specific NHS-funded PROMS data will be used to generate the indicators prescribed in the CMA Order, based on pseudonymised aggregated or combined private and NHS-funded PROMS data, conforming to rules on small number suppression and available at hospital, consultant and procedure level. The information will be published as indicators on a publically accessible website (www.phin.org.uk): • Procedure-specific measures of improvement in health outcomes, as agreed by the information organisation and its members to be appropriate All outputs will be aggregated and will consist of respectively, relative values, median values and “scores”. Small number suppression will be applied where required in line with the HES Analysis Guide.