Congenital Heart Audit: Measuring Progress In Outcomes Nationally (CHAMPION) aand LAUNCHES QI: Linking AUdit and National datasets in Congenital HEart Services for Quality Improvement.

Text removed is struck through; text added is underlined. Unchanged paragraphs are summarised rather than repeated.

Objective for processing

Objective of this new 2022 amendment request University College London (UCL) require access to NHS England data for the purpose of the Congenital Heart Audit: Measuring Progress In Outcomes Nationally (CHAMPION) research project and LAUNCHES QI: Linking Audit and National datasets in Congenital Heart Services for Quality Improvement. DARS NIC 234297 initially only covered a study called “LAUNCHES QI: Linking Audit and National datasets in Congenital Heart Services for Quality Improvement”. In 2019 an amendment was successfully submitted to reuse the LAUNCHES dataset for a new study called CHAMPION “Congenital Heart Audit: Measuring Progress in Outcomes Nationally”. The CHAMPION study uses the same data extracted from LAUNCHES QI. University College London (UCL) request to add additional purpose and processing to the CHAMPION study to recalibrate the Partial Risk Adjustment in Surgery (PRAiS) risk model for mortality after paediatric cardiac surgery with new diagnostic and procedure codes and more recent National Congenital Heart Disease Audit (NCHDA) and Civil Registration (Deaths) data. UCL seek to submit an updated cohort of patients to NHS Digital to link to and receive pseudonymised Civil Registration Mortality life status data. UCL do not require any additional HES data under this agreement. LAUNCHES QI: Linking Audit and National datasets in Congenital Heart Services for Quality Improvement” aimed to indirectly improve services for congenital heart disease (CHD) by providing the first description of how CHD patients interact with the NHS acute sector and where variation in outcomes or service use exist. In 2019 an amendment was approved to reuse the LAUNCHES dataset for a new study called CHAMPION “Congenital Heart Audit: Measuring Progress in Outcomes Nationally”. UCL shall be applying to receive additional National Congenital Anomaly and Rare Disease Registration Service (NCARDRS) data (year 2018 of the data was previously received under data controller Public Health England; UCL intend to apply for year 2019 of the data) and to link both years of data to NCHDA data for the original purposes of CHAMPION. However, as the novation from Public Health England (PHE) to NHS Digital has only recently occurred UCL plan to submit this as a separate DARS amendment later because UCL need to prioritise the Civil Registration (Deaths) extract and the NCARDRS data will not be available until the end of 2022 / early 2023. Under previous iterations of this Agreement, UCL required pseudonymised data of life status, age at life status*, place of death and a subset of HES data for use in a study called “LAUNCHES QI: Linking AUdit and National datasets in Congenital HEart Services for Quality Improvement” (IRAS ID 246796). *Age at life status will be age at death for those who have died. For those who have no recorded date of death / have not died, the age at life status is the age at the date of Civil Registration (Deaths) extract. UCL are the sole data controller who also process data. [2 paragraphs unchanged] The activity is compliant with the principles of GDPR. Based on guidance, as this is for University research, the lawful basis for processing data is GDPR article 6(1)(e): ‘Processing is necessary for the performance of a task carried out in the public interest or in the exercise of official authority vested in the controller’. Also referred to as ‘Public Task’. As the research involves health data, which is included in the definition of special categories of personal data, it requires an additional condition for processing. Based on guidance, for health research this is GDPR article 9(2)(j), which details that processing is necessary for scientific and research purposes, subject to appropriate safeguards. The data are required for research purposes in the public interest – meeting the conditions in the DPA 2018 Schedule 1 Part 1 (4) - which GDPR Recital 52(2) determines is an appropriate derogation from the prohibition on processing special categories of personal data. The ways in which the processing of data will be of benefit to the public – thereby demonstrating that the processing is in the public interest. LAUNCHES QI is in the interest of the public as it aims to measure and improve services for those with CHD. The primary objective of the CHAMPION study is to indirectly improve services for congenital heart disease (CHD) by developing tools for routinely measuring CHD outcomes (including risk-adjusted survival in adults following intervention and long-term outcomes by diagnosis) that are considered meaningful and appropriate by stakeholders and can inform the delivery and commissioning of CHD services. LAUNCHES QI is a dataset analysis of five linked audit and national datasets which includes 143,862 patients with congenital heart disease that have been captured by the National Congenital Heart Disease Audit (NCHDA) since 2000, the core dataset defining the study CHD population. Included are all patients, with no age restrictions, who have had at least one intervention for congenital heart disease that has generated at least one entry in NCHDA. Patients undergoing cardiac intervention since April 2000 (when NCHDA began) are included in the study, but the requested records for these patients in HES is from 1997/98 (where available) until March 2017, since these patients may have interacted with NHS services prior to their cardiac intervention. NHS Digital Data was required up until the end of March 2018. The inclusion criteria into the study was anyone captured in the NCHDA audit since it began, up to March 2017 which was as far as was available at the time of application. The other datasets were then used to track the patient trajectories both before and since March 2017. Therefore, UCL requested data as far as was available at application to capture interactions of the patients with the health service. There is a future intention to request further data, but not as part of this Agreement specifically for LAUNCHES QI. Any request for further data will be subject to an updated application with NHS Digital. The secondary objectives include- CHD is a life-long condition, so obtaining 20 years of data (including 3 years prior to the NCHDA dataset) is crucial for establishing as complete trajectories of their service use as possible, which will enable the study to get a more detailed understanding of how patients interact with secondary care throughout their lives, how this varies across centres, and how it impacts on outcomes. National data is required to allow exploration of the different CHD services across the country. The linkage of the other 4 datasets to NCHDA at patient level will allow trajectories of care to be determined. 1. To define which cases of adults undergoing surgery or interventional cardiology should be included in the reporting and comparing of adult CHD outcomes; establish which outcomes are feasible and appropriate for routine monitoring and which factors influence outcomes; develop the analytical methods for taking account of differences in risk across case-mix. The study research dataset is generated by linking pseudonymised NCHDA data (English and Welsh centres) to the four additional sources: 2. To identify key diagnoses for which outcomes should be reported, establish which outcome metrics can be routinely measured for these diagnoses, and develop the analytical means to do so (including exploring non-diagnostic factors that can influence outcomes). 1) Paediatric intensive care data from the Paediatric Intensive Care Audit Network (PICANet) based at University of Leeds (English centres). 3. To capture the views of patients and parents/carers about which procedure- based outcomes for adults and which longer-term outcomes by diagnosis would be most meaningful and how these might inform their choice of provider or their understanding of the likely course of their conditions. 2) Adult intensive care audit data (Case Mix Programme, CMP) (English and Welsh centres). 4. To build prototype software tools that specify the automated mechanisms required to extract, generate, and visualise the outcomes defined in Objectives 1 and 2 from routinely collected data, including appropriate risk adjustment. 3) Hospital Episode Statistics (HES), (English centres). 5. To recalibrate the Partial Risk Adjustment in Surgery Model (PRAiS) risk model for mortality after paediatric cardiac surgery with new diagnostic and procedure codes and more recent data. 4) Death registrations (Civil Registration Deaths). (linked to identifiers from patients of English and Welsh centres). 6. To report on the risk adjustment of postoperative complications in children as well as adults. This objective was flagged as a priority across stakeholders. UCL received authorisation from HQIP to transfer personal identifiers from National Institute for Cardiovascular Outcomes Research (NICOR) (which curates the NCHDA dataset) to ICNARC, PICANet and NHS Digital. NICOR is located at Bart’s Health NHS Trust. UCL has also received authorisation from HQIP and ICNARC to receive the pseudonymised clinical information at University College London. UCL has Ethics (18/NS/0106) and CAG (18CAG0180) approval for the study to process the data and to link the datasets. The complications/morbidities following paediatric cardiac surgery have been audited by National Congenital Heart Disease Audit (NCHDA) since 2015/2016 and reported from 2019. However, interpretation has been limited as there is no risk adjustment and there is significant interest from stakeholders (incl. clinicians, patients and commissioners) to improve reporting. Two datasets were created: one containing the patient level, pseudonymised information of all hospital admissions, outpatient appointments and procedures for each CHD patient; the other a pseudonymised hospital admission level dataset covering all hospital stays for CHD patients. The final pseudonymised datasets are stored within UCL’s secure data safe haven. Patient identifiers were sent to NHS Digital from the NCHDA database. UCL requested that NHS Digital match the identifiers from the NCHDA dataset to HES and Civil Registration Deaths and extract the requested fields. NCHDA provided a record level LAUNCHES ID which should be transferred to each pseudonymised HES/Civil Registration Deaths record that matches. UCL additionally requested that NHS Digital generate a unique record-level Study ID that would be common between NHS Digital and UCL to facilitate data queries. The data was securely transferred with the record level study numbers to University College London. The LAUNCHES QI study was instigated and is led by the Principal Investigators based at the Clinical Operational Research Unit (CORU) at UCL. UCL have two members of the study team who have honorary contracts at UCL. Both are clinicians working at Great Ormond Street Hospital and their input as clinicians for CHD is valuable as they have experience with similar data and will be able to assist with analyses. This will increase the impact of the study. The sole funder, The Health Foundation, are involved in the study only to provide the award, as grant-funding. UCL will provide progress reports of the project through annual reports and award meetings. This refers to a funders role of how UCL are spending the funds and how the overall study is progressing, and has nothing to do with the data or analyses. The Health Foundation have no influence over the results. They are not permitted to access any record level NHS Digital data. To confirm, The Health Foundation has no involvement in determining the purpose and means of processing and does not carry out any data controllership activities. As in many studies now there are a number of collaborators providing an advisory role. Only UCL substantive employees and the two clinical researchers from Great Ormond Street Hospital with honorary contracts with UCL are working with the data. Clinical researchers will not have access to data until the signed contract/addendum letter is in place. The organisations involved as collaborators/advisors other than UCL are, Birmingham Children’s Hospital, Intensive Care National Audit & Research Centre, Leeds University, Royal Brompton NHS Foundation Trust, Great Ormond Street Hospital NHS Trust, Leeds Teaching Hospitals NHS Foundation Trust, University Hospital Southampton, and University Hospitals Bristol NHS Foundation Trust. These organisations will not have access to the data. The collaborators together with the UCL researchers make up the full study research team, with the collaborators providing insight into the aggregated results produced by the UCL researcher analysis. Only the UCL research team will be making decisions on, accessing and analysing the data and so only UCL will be the sole data controller. For clarity, no other organisations are determining the purpose or means of the processing. The CHAMPION study University College London (UCL) required the same data extracted for the LAUNCHES QI project (pseudonymised data of life status, age at life status, place of death and a subset of HES data) for use in a follow up study called “Congenital Heart Audit: Measuring Progress In Outcomes Nationally” (IRAS ID 266393). UCL are the sole data controller who also process data. UCL are asking for permission to use the LAUNCHES QI data extract for the purpose of a new project, “CHAMPION”. Amendment 2022: UCL are requesting a new extract of Civil Registration (Deaths) data (pseudonymised data of life status, age at life status, place of death) linked to an updated extract of the NCHDA identifiers provided for LAUNCHES, for the purposes both of the existing objectives and a new objective for CHAMPION as detailed below. CHAMPION aims to indirectly improve services for congenital heart disease (CHD) by developing tools for routinely measuring CHD outcomes (including risk-adjusted survival in adults following intervention and long-term outcomes by diagnosis) that are considered meaningful and appropriate by stakeholders and can inform the delivery and commissioning of CHD services. Amendment 2022: Improving NICOR’s main current reporting output (survival after paediatric surgeries using the PRAiS risk model) is within the overall aim but would be additional to the existing workstreams. The study will provide information on a) Which cases of adults undergoing surgery or interventional cardiology should be included in the reporting and comparing of adult CHD outcomes; establish which outcomes are feasible and appropriate for routine monitoring and which factors influence outcomes; develop the analytical methods for taking account of differences in risk across case-mix. b) Key diagnoses for which outcomes should be reported, establish which outcome metrics can be routinely measured for these diagnoses, and develop the analytical means to do so (including exploring non-diagnostic factors that can influence outcomes). The team will aim to inform the commissioning and delivery of CHD services: defining outcomes for adult CHD procedures, establishing longer-term outcomes by diagnosis, and building prototype software tools that specify the automated mechanisms required to extract, generate and visualise these outcomes. Amendment 2022: at the request of NICOR, the team will also additionally inform the commissioning and delivery of CHD services with the recalibration of the PRAiS risk model for 30-day mortality after paediatric cardiac surgery with new diagnostic and procedure codes and more recent NCHDA and Civil Registration (Deaths) data. The activity is compliant with the same principles of GDPR as LAUNCHES QI. Based on guidance, as this is for University research, the lawful basis for processing data is GDPR article 6(1)(e): ‘Processing is necessary for the performance of a task carried out in the public interest or in the exercise of official authority vested in the controller’. Also referred to as ‘Public Task’. As the research involves health data, which is included in the definition of special categories of personal data, it requires an additional condition for processing. Based on guidance, for health research this is GDPR article 9(2)(j), which details that processing is necessary for scientific and research purposes, subject to appropriate safeguards. The data are required for research purposes in the public interest – meeting the conditions in the DPA 2018 Schedule 1 Part 1 (4) - which GDPR Recital 52(2) determines is an appropriate derogation from the prohibition on processing special categories of personal data. The ways in which the processing of data will be of benefit to the public – thereby demonstrating that the processing is in the public interest CHAMPION is in the interest of the public as it aims to measure and improve services via the use of developed tools for those with CHD. [1 paragraph unchanged] Under a previous iteration of this data sharing agreement, UCL obtained permission to use the LAUNCHES QI dataset as the core dataset within CHAMPION. The LAUNCHES QI dataset was compiled as follows: The following NHS England Data will be accessed: All patients who have at least one entry in the National Congenital Heart Disease Audit (NCHDA) (data controller HQIP), up to March 2017 were included in the dataset. NICOR who oversee this audit, provided the personal identifiers NHS number, hospital number, date of birth and postcode for these patients to the data providers PICANet and ICNARC, and will provide these identifiers to NHS Digital. Patient name was also sent to PICANet and NHS Digital. Hospital Episode Statistics (HES)- The LAUNCHES QI sample is the entire cohort of patients (children and adults) who had at least one procedure related to congenital heart disease since 2000, and who appear on the NCHDA database up to March 2017 (n= 96,040), some with more than one record so a total of n= 143,862 records. The total LAUNCHES QI dataset sample size is subject to slight change following further data cleaning. • Outpatients (OP) To clarify: The LAUNCHES sample consists of all congenital heart disease procedures recorded in NCHDA between April 2000 and March 2017 from public hospitals (N=143,862 records). Together with NICOR and PICANet UCL worked on the identification of patients and revision of inconsistencies in dates and identifiers, which resulted in the current identification of n=96,041 patients (n=96,109 before quality assurance). The final number of patients may further change as UCL address any anomalies: as UCL link more datasets it may emerge those records thought to be for two patients turn out to be just one, and vice versa. For that agreement, the linkage to HES data was done at record level using the identifiers in the 143,862 NCHDA records from the LAUNCHES dataset. • Accident and Emergency (A and E) Each data provider identified in their respective datasets which records in the PICANET, ICNARC-CMP, HES and Mortality data pertained to those CHD individuals and returned to CORU at UCL the clinical data they hold on the matched individuals, pseudonymised with the record level unique ID (the LAUNCHES ID). • Admitted Patient Care (APC) These five national datasets have been linked at UCL to make up the LAUNCHES QI dataset. Civil Registrations of Deaths- Secondary Care Cut (CRD SCC) For CHAMPION, UCL also have permission from HQIP to access NACSA data and have requested NICOR match NACSA to NCHDA as they manage both audits. They will be able to flag the matching records by assigning the corresponding record's LAUNCHES ID, therefore the personal identifiers will not need to leave NICOR. UCL access National Congenital Anomaly and Rare Disease Registration Service (NCARDRS) data via NHS England. NCARDRS data will be disseminated to UCL for the purpose of the CHAMPION study under DARS-NIC-656875-B4G4P. NACSA records from 2015 to 2017 consist of 67,896 records and 58,508 patients. The number of patients this pertains to may be lower than this as more than one record may apply to one patient. Not all patients in NACSA may add to the total sample size, as some may already exist in the LAUNCHES QI dataset. 864 of NACSA patients linked to NCHDA (1,001 records). The pseudonymised linkage of the LAUNCHES QI dataset to NACSA at patient level will allow enhanced trajectories of care to be determined. The level of the data will be pseudonymised. Amendment 2022: UCL have requested the addition of NCHDA years 2017 to 2022 to increase not just the number of patients and records held but also to get more information since significant new fields were added in 2015. Based on an estimate, the number of patients will increase by approximately 50,000 to 146,040 patients in this linked dataset. The number of records pertaining to those patients is estimated to increase from 143,862 to approximately 192,500. However, the final number of patients and records will not be clear until final linkage and cleaning of the data at UCL. The Data will be minimised as follows: Under the previous iterations UCL received authorisation from HQIP to use the PICANet data, ICNARC to use the ICNARC CMP data, NICOR to use of the NCHDA data, and transfer of pseudonymised clinical information from the NACSA audit. UCL has Ethics (20/NS/0022) and CAG (20CAG0027) approval for the CHAMPION study to process the LAUNCHES QI data for the purposes (prior to this 2022 amendment) of CHAMPION and to link the NACSA dataset. • Limited to a study cohort identified by NHS England as meeting the following criteria; Patients (children and adults) who had at least one procedure related to congenital heart disease since 2000, and who appear on the NCHDA database up to March 2017 Amendment 2022: UCL has received ethics and CAG approval for this substantial amendment to: • Limited to data between 2003/04-2021/22* • Add an objective to the CHAMPION study (recalibrating the PRAiS risk model). *Data for LAUNCHES QI was minimised to data between 2003/04- 2017/18 for both HES and CRD SCC. Under a previous iteration of this DSA, UCL received CRD SCC records up to 2022 for the CHAMPION study. • To receive an updated pseudonymised extract of NCHDA data from March 2017 up to the most recent available date (2021 or 2022). UCL are the sole controller who also process data. • To transfer again the NCHDA identifiers from 2000 to NHS Digital (as was done for LAUNCHES), this time including from March 2017 to the most recent available date (2021 or 2022), and the LAUNCHES study ID to be linked to Civil Registration Mortality data. The lawful basis for processing personal data under the UK GDPR is: • For UCL to receive the Mortality data that linked to the identifiers up to the date of extract, pseudonymised with the LAUNCHES study ID (the pseudonymisation key assigned with the identifiers for the LAUNCHES study, to allow pseudonymisation and linkage of the datasets). Article 6(1)(e) - processing is necessary for the performance of a task carried out in the public interest or in the exercise of official authority vested in the controller; • To link the Civil Registration (Deaths) data and NCHDA data with the LAUNCHES study ID to use for the new objective. The lawful basis for processing special category data under the UK GDPR is: • To replace existing Mortality linked data in the CHAMPION dataset with the update, for the original purposes of CHAMPION. Article 9(2)(j) - processing is necessary for archiving purposes in the public interest, scientific or historical research purposes or statistical purposes in accordance with Article 89(1) based on Union or Member State law which shall be proportionate to the aim pursued, respect the essence of the right to data protection and provide for suitable and specific measures to safeguard the fundamental rights and the interests of the data subject. And subject to a separate NHS Digital amendment following this one: The processing is in the public interest because the research aims to measure and improve services via the use of developed tools for those with CHD. • To transfer the NCHDA identifiers from 2000 to the most recent available date to NHS Digital to link to years 2018 and 2019 of NCARDRS data. Amazon Web Services (AWS) is a processor acting under the instructions of UCL. AWS’ role is limited to secure back-up of data stored in UCL’s Data Safe Haven. • For UCL to receive the NCARDRS data for 2018 and 2019, with the linked LAUNCHES study ID if applicable. UCL uses offsite data centre services provided by VIRTUS data centre. VIRTUS does not have access to the data. UCL has submitted an amendment application to NICOR and their new data controller NHS England (Arden & GEM Commissioning Support Unit specifically) for the new NCHDA extract and linkage to Civil Registration (Deaths) data and NCARDRS for these purposes. Funding is provided by the National Institute for Health Research (NIHR). UCL has submitted an amendment application to PICANet and their data controller HQIP as the PICANet data within the CHAMPION dataset will be updated with the new Civil Registration (Deaths) extract. The funder(s) will have no ability to suppress or otherwise limit the publication of findings. The final pseudonymised datasets additionally containing NACSA data will be stored within UCL’s secure data safe haven. Organisations involved as collaborators/advisors for the CHAMPION study other than UCL are; UCL are requesting permission from NHS Digital to use the HES and mortality data for the new purposes of the CHAMPION study. Alder Hey Children’s NHS Foundation Trust, Barts Health NHS Trust, Newcastle University, Royal Brompton NHS Foundation Trust, Great Ormond Street Hospital NHS Trust, Leeds Teaching Hospitals NHS Foundation Trust, The Magdi Yacoub Institute, University Hospitals Bristol NHS Foundation Trust, Children’s Heart Federation (CHF), Little Hearts Matter (LHM), and The Somerville Foundation (TSF) Amendment 2022: UCL are requesting permission from NHS Digital to use the Civil Registration Death data for the new objective of the CHAMPION study (PRAiS risk model recalibration) as well as the existing objectives. No HES data is required under this (version 3) agreement. Organisations involved as collaborators/advisors for the LAUNCHES QI study other than UCL are; The CHAMPION study was instigated and is led by the Principal Investigators based at the Clinical Operational Research Unit (CORU) at UCL. UCL have two members of the study team who have honorary contracts at UCL. Both are clinicians working at Great Ormond Street Hospital and their input as clinicians for CHD is valuable as they have experience with similar data and will be able to assist with analyses. This will increase the impact of the study. The organisations involved as collaborators/advisors other than UCL are, Birmingham Children’s Hospital, Intensive Care National Audit & Research Centre, Leeds University, Royal Brompton NHS Foundation Trust, Great Ormond Street Hospital NHS Trust, Leeds Teaching Hospitals NHS Foundation Trust, University Hospital Southampton, and University Hospitals Bristol NHS Foundation Trust. The sole funder, The NIHR, are involved in the study only to provide the award, as grant-funding. They will also oversee progress of the project through annual reports and award meetings. The NIHR have no influence over the results. They are not permitted to access any record level NHS Digital data. These organisations will not have access to the data. UCL make all the final decisions as controller. As in many studies now there are a number of collaborators providing an advisory role. Only UCL substantive employees and clinical researchers on an honorary contract with UCL are working with the data. Clinical researchers will not have access to data until the signed contract/addendum is in place. The organisations involved as collaborators/advisors other than UCL are, Alder Hey Children’s NHS Foundation Trust, Barts Health NHS Trust, Newcastle University, Royal Brompton NHS Foundation Trust, Great Ormond Street Hospital NHS Trust, Leeds Teaching Hospitals NHS Foundation Trust, The Magdi Yacoub Institute, University Hospitals Bristol NHS Foundation Trust, Children’s Heart Federation, Little Hearts Matter, and The Somerville Foundation. These organisations will not have access to the data. UCL researchers make all the final decisions as data controllers. UCL have two members of the study team who have honorary contracts at UCL. Both are clinicians working at Great Ormond Street Hospital and their input as clinicians for CHD is valuable as they have experience with similar data and assist with analyses. The independent study advisory committee oversees the conduct of the project. Two Adult CHD patients provided feedback on the proposal and CHD patient charities the SF, CHF and LHM helped to design the online forums. These charities develop and moderate the forums, and representatives from each advise across all workstreams via the expert panel along with a patient involvement manager from the British Heart Foundation. The Somerville Foundation (TSF), Children’s Heart Federation (CHF) and Little Hearts Matter (LHM) will engage with potential participants for the online forums for i) adults and adolescents with CHD; ii) parents of children with less complex CHD conditions; iii) parents of children with the most complex CHD conditions respectively.

Processing activities

Data will not be handled by any additional third party organisations. Data will not be accessed outside the United Kingdom. No new data will be disseminated under this agreement. The below processing activities occurred under previous iterations of this DSA. LAUNCHES QI The LAUNCHES QI study dataset was generated by linking pseudonymised National Congenital Heart Disease Audit (NCHDA) data (English and Welsh centres) to the four sources: The completed data flows for all data processing under this agreement are as follows: 1) Paediatric intensive care data from the Paediatric Intensive Care Audit Network (PICANet) based at University of Leeds (English centres). 1) Data flows to NHS Digital 2) Adult intensive care audit data (Case Mix Programme, CMP) (English and Welsh centres). National Institute for Cardiovascular Outcomes Research (NICOR) securely transferred a file to the NHS Digital. This file contained patient identifiable information (NHS Number, name, postcode, date of birth, local hospital patient ID) for all patients in the National Congenital Heart Disease Audit (NCHDA), and unique study ID (NCHDA record-level LAUNCHES Study ID). 3) Hospital Episode Statistics (HES), (English centres). (NHS England) 2) NHS Digital identify common records between NCHDA data and HES and Civil Registration Deaths data, including the requested derived fields. NHS Digital generate a unique record-level Study ID for each identified record. 4) Death registrations (Civil Registration Deaths). (linked to identifiers from patients of English and Welsh centres) (NHS England) 3) Data flows from NHS Digital These national datasets have been linked at UCL to make up the LAUNCHES QI dataset i) NHS Digital return HES data for all individuals in the NCHDA cohort to University College London. The unique study ID (LAUNCHES QI record level study number and HES record-level study number) and requested derived fields are appended to the end of every episode record returned to University College London. The LAUNCHES sample consists of all congenital heart disease procedures recorded in NCHDA between April 2000 and March 2017 from public hospitals. ii) Civil Registration Deaths derived fields for all individuals in the NCHDA cohort returned to University College London. The unique study ID (LAUNCHES QI record level study number and HES record-level study number) are appended to the end of every record. UCL received authorisation from HQIP to transfer personal identifiers from National Institute for Cardiovascular Outcomes Research (NICOR) (which curates the NCHDA dataset) to ICNARC, PICANet and NHS Digital. NICOR was located at Bart’s Health NHS Trust, it is now located at NHS Arden and Greater East Midland Commissioning Support Unit (Arden & GEM). UCL has also received authorisation from HQIP and ICNARC to receive the pseudonymised clinical information at University College London. 4) The linkage strategy is that NICOR provided the personal identifiers NHS number, hospital number, date of birth and postcode to NHS Digital, ICNARC and PICANet from the NCHDA data. Patient name was also sent to PICANet and NHS Digital. Each identified in their respective datasets which records pertain to those CHD individuals and return to UCL the requested clinical and administrative data they hold on the matched individuals (UCL would not receive data for patients that do not match), pseudonymised with the LAUNCHES record level ID and local dataset record level IDs. UCL received from NICOR the pseudonymised clinical data of the NCHDA dataset and did not receive any personal identifiers. Linking (at UCL) was made possible via the pseudonymised unique study ID (LAUNCHES QI record level study number and HES record-level study number). Each data provider identified in their respective datasets which records in the PICANET, ICNARC-CMP, HES and ONS data pertained to those CHD individuals and returned to CORU at UCL the clinical data they hold on the matched individuals, pseudonymised with the record level unique ID (the LAUNCHES ID). UCL then linked HES and Civil Registration Deaths data received from NHS Digital to pseudonymised clinical data received from NICOR (NCHDA data), PICANet and ICNARC (CMP data), via the LAUNCHES QI unique record-level study number and the LAUNCHES Patient IDs (which will be contained within the NCHDA data set). UCL did not receive any personal identifiers from any of the data sources. UCL required the record level NHS Digital study identifiers in case of any queries about specific records with NHS Digital. For CHAMPION, UCL also have permission from HQIP to access The National Adult Cardiac Surgery Audit (NACSA) data and have requested NICOR match NACSA to NCHDA as they manage both audits. The data is stored and processed within the UCL Identifiable Data Handling Solution (IDHS) called the Data Safe Haven (DSH). The data is held within a secure environment where all statistical analyses will be undertaken. Access to this record level data will be limited to only the UCL LAUNCHES QI team" to allow for additional analysts at UCL, Access to this record level data will be limited to only members of the CHAMPION team, who are all substantive employees at UCL or clinical researchers on an honorary contract with UCL. Clinical researchers will not have access to data until the signed contract/addendum letter is in place. Staff accessing the UCL data safe haven attend training in its use and security procedures. Staff are also required to complete mandatory annual Information Governance and GDPR training. Each study working on the data safe haven has what is known as its own ‘share’, where the study specific data is kept. Access to this share is granted only by the Principle Investigators who request access for each user. Any team member leaving the study has their access revoked. NICOR flag matching records by assigning the corresponding record's LAUNCHES ID, therefore the personal identifiers will not need to leave NICOR. Re-identification is not permitted under this data sharing agreement. LAUNCHES QI- No linkage, other than that described within the agreement is permitted and no further data linkage will be undertaken. 1) Data flows to NHS England The data processing activities described applies to all processing under this Data Sharing Agreement. National Institute for Cardiovascular Outcomes Research (NICOR) securely transferred a file to the NHS England. This file contained patient identifiable information (NHS Number, name, postcode, date of birth, local hospital patient ID) for all patients in the National Congenital Heart Disease Audit (NCHDA), and unique study ID (NCHDA record-level LAUNCHES Study ID). As part of LAUNCHES analysis; admission, procedure, diagnosis, discharge, and provider information are required for all matched patients. UCL are also requesting derived fields, age at admission/appointment to four decimal places* and age at discharge to four decimal places. UCL are not requesting any dates of birth for this project to prevent identifiability of the data and so LAUNCHES will use ages at health service interaction to determine each patient’s treatment trajectory. In- and out-patient HES data is critical in identifying the treatment that each patient has received. In addition, HES A&E data are required to identify adverse outcomes such as unplanned emergency treatment following discharge from CHD surgery or through deterioration in a patient’s health. The requested Civil Registration Deaths derived fields of life status, age at life status to four decimal places and place of occurrence of death (home/hospice/hospital/care home/other communal establishment/elsewhere), are vital to complete the patient trajectories. 2) NHS England identified common records between NCHDA data and HES and Civil Registration Deaths data, including the requested derived fields. NHS England generated a unique record-level Study ID for each identified record. *Age to 4 decimal places is age presented as whole years with decimal places replacing month, day, and time units. This allows UCL to determine order of events that occur within 24 hours of each other (which can and does happen for children’s heart surgery). 3) Data flows from NHS England UCL need the 4 decimal places age to provide enough time resolution to distinguish the order of procedure for it will allow UCL to sequence operations on the same day for the same patient as long as they are more than one hour apart and a time of procedure is entered for each record. i) NHS England returned HES data for all individuals in the NCHDA cohort to UCL. The unique study ID (LAUNCHES QI record level study number and HES record-level study number) and requested derived fields were appended to the end of every episode record returned to University College London. There will be some records where there is no time stamp (only a date) where UCL will not be able to determine the sequence by age at operation alone, but UCL wish to use time information where it is available. ii) Civil Registration Deaths derived fields for all individuals in the NCHDA cohort were returned to University College London. The unique study ID (LAUNCHES QI record level study number and HES record-level study number) were appended to the end of every record. The following milestones determines the age to 4 decimal places; age at admission, age at appointment/procedure, age at discharge, age at life status. 4) The linkage strategy is that NICOR provided the personal identifiers NHS number, hospital number, date of birth and postcode to NHS England, ICNARC and PICANet from the NCHDA data. UCL receive month and year of birth and the further age updates depends on the number of interactions with the health service. The other information would be the clinical fields that were requested. Each data provider identified in their respective datasets which records pertain to those CHD individuals and return to UCL the requested clinical and administrative data they hold on the matched individuals (UCL did not receive data for patients that do not match), pseudonymised with the LAUNCHES record level ID and local dataset record level IDs. UCL recieved from NICOR the pseudonymised clinical data of the NCHDA dataset and did not receive any personal identifiers. Linking (at UCL) was made possible via the pseudonymised unique study ID (LAUNCHES QI record level study number and HES record-level study number). Methods of analysis for the study will include: CHAMPION Study 1. Data cleaning and descriptive analysis of individual datasets. NICOR provided the NACSA clinical and administrative data for the years requested which where possible were linked to and pseudonymised with the LAUNCHES record level ID and local dataset record level IDs. Any that did not have a LAUNCHES record level ID were pseudonymised with the local dataset record level ID only. UCL received pseudonymised clinical data of the NACSA dataset from NICOR and did not receive any personal identifiers. 2. Develop and update clinical coding maps. The Data will not be transferred to any other location. 3. Establish and examine variations in longitudinal patient trajectories. Data is stored on servers at UCL. 4. Identify candidate metrics to inform routine quality improvement and assurance. Amazon Web Services provides cloud hosting services to UCL and will store the data as contracted by UCL. The results of all analyses will be published in aggregate form, with small numbers suppressed in line with HES guidance. No identifiable data will be held by University College London, therefore no identifiable data will be released. Only UCL substantive employees and clinical researchers on an honorary contract with UCL are working with the data. The CHAMPION Study All personnel accessing the Data have been appropriately trained in data protection and confidentiality. The completed data flows are as follows (carried out under previous iterations of this Agreement): The Data will be accessed by authorised personnel via remote access. 1) Data flows to NHS Digital: The Controller(s) must confirm and provide evidence upon audit by NHS England that access via any remote device complies with the data security obligations within this DSA and the Data Sharing Framework Contract. There will be no further data flows to NHS Digital. For remote access: Amendment 2022: - Remote access will only be from secure locations situated within the territory of use (as further restricted elsewhere within the DSA if so done) stated within this DSA; As was done for LAUNCHES, National Institute for Cardiovascular Outcomes Research (NICOR) will securely transfer a file of identifiers to NHS Digital. This file will contain patient identifiable information (NHS Number, postcode, date of birth, sex) for all patients in the National Congenital Heart Disease Audit (NCHDA), and unique study ID (NCHDA record-level LAUNCHES Study ID). This will consist of the NCHDA identifiers from 2000 to 2021/2022 (most recent available year at the time of transfer). - Access controls granting users the minimum level of access required are in place; 2) Data flows from NHS Digital: - Remote access is only via secure connections (e.g., VPNs or secure protocols) to protect data; Data previously supplied under version 0 of this Agreement was incomplete and was resupplied under version 1. - Multifactor authentication (MFA) is required for remote access; Amendment 2022: - Device security, including up-to-date software and operating systems, antivirus software, and enabled firewalls are utilised for the remote access; NHS Digital will identify common records between NCHDA data and Civil Registration Deaths data, including the requested derived fields. NHS Digital will provide a unique patient-level Study ID for each identified record. - All remote access is undertaken within the scope of the organisation’s DSPT (or other security arrangements as per this DSA) and complies with the organisation’s remote access policy. 3) Data flows from NHS Digital The above applies in addition to any condition set out elsewhere within the DSA (e.g. who may carry out processing, and for what purpose). Amendment 2022: The data will not leave England or Wales at any time. Civil Registration Deaths derived fields for all individuals in the NCHDA cohort will be returned to University College London. The unique study ID (LAUNCHES QI record level study number and a patient level ID) will be appended to the end of every record. There will be no requirement and no attempt to reidentify individuals when using the Data. 4) The linkage strategy was that NICOR provided the NACSA clinical and administrative data for the years requested which where possible were linked to and pseudonymised with the LAUNCHES record level ID and local dataset record level IDs. Any that did not have a LAUNCHES record level ID were pseudonymised with the local dataset record level ID only. UCL received from NICOR the pseudonymised clinical data of the NACSA dataset and did not receive any personal identifiers. Amendment 2022: The linkage strategy is that NICOR will provide the personal identifiers NHS number, date of birth, sex and postcode to NHS Digital from the NCHDA data. NHS Digital will identify in their datasets which records pertain to those CHD individuals and return to UCL the requested clinical and administrative data they hold on the matched individuals (UCL will not receive data for patients that do not match), pseudonymised with the LAUNCHES record level ID and local dataset record level IDs. UCL will receive from NICOR the pseudonymised clinical data of the NCHDA dataset and will not receive any personal identifiers. Linking (at UCL) is made possible via the pseudonymised unique study ID (LAUNCHES QI record level study number patient level ID. Amendment 2022: UCL will then link Civil Registration Deaths data received from NHS Digital to the CHAMPION dataset, the pseudonymised clinical data received from NICOR (NCHDA data), PICANet and ICNARC (CMP data), via the LAUNCHES QI unique record-level study number and the LAUNCHES Patient IDs (which will be contained within the NCHDA data set). The Civil Registrations Deaths data will also be linked only to NICOR (NCHDA data) as a separate dataset to analyse for the new objective, to recalibrate the PRAiS algorithm. UCL will not be receiving any personal identifiers from any of the data sources. UCL will use the record level NHS Digital study identifiers in case of any queries about specific records with NHS Digital. For the LAUNCHES QI dataset securely stored currently at UCL, admission, procedure, diagnosis, discharge, and provider information are held for all matched patients. UCL also hold the derived fields, age at admission/appointment to four decimal places and age at discharge to four decimal places. UCL will not hold any dates of birth for CHAMPION QI to prevent identifiability of the data and so CHAMPION QI will use ages at health service interaction. In- and out-patient HES data is critical in identifying the treatment that each patient has received. In addition, HES A&E data was required to identify adverse outcomes such as unplanned emergency treatment following discharge from CHD surgery or through deterioration in a patient’s health. The Civil Registration Deaths derived fields of life status, age at life status to four decimal places and place of occurrence of death (home/hospice/hospital/care home/other communal establishment/elsewhere), are vital to complete the patient trajectories. Amendment 2022: The requested Civil Registration Deaths derived fields are the same as those already held at UCL: life status, age at life status to four decimal places and place of occurrence of death (home/hospice/hospital/care home/other communal establishment/elsewhere). These fields are vital to complete the patient trajectories with the update of the CHAMPION dataset, and also vital to complete the new objective or recalibrating the PRAiS risk model. UCL are not requesting any updates of HES data as part of this amendment. Methods of analysis for the study will include: 1. Data cleaning and descriptive analysis of the NACSA dataset. 2. Establish robust outcomes for adult CHD procedures 3. Establish longer-term outcomes by diagnosis Amendment 2022: Methods of analysis for the amendment will include: 1. Data cleaning and descriptive analysis of the linked NCHDA and Civil Registration Deaths data extract for recalibration of the PRAiS model UCL will: 2. Re-calibrate the PRAiS risk model; Update the software that implements PRAiS with the new calibration; Make the updated software accessible to all paediatric congenital heart centres in the UK and Republic of Ireland. 3. Update the Civil Registration Deaths data in the CHAMPION dataset for any further analyses of robust outcomes for adult CHD procedures and to establish longer-term outcomes by diagnosis. The results of all analyses will be published in aggregate form, with small numbers suppressed in line with HES guidance. No identifiable data will be held by University College London, therefore no identifiable data will be released. All organisations party to this agreement must comply with the Data Sharing Framework Contract requirements, including those regarding the use (and purposes of that use) by Personnel (as defined within the Data Sharing Framework Contract i.e.: employees, agents, PhD students and contractors of the Data Recipient who may have access to that data).

Expected output

[3 paragraphs unchanged] Amendment 2022: An additional output for the new objective will be the update of the software that implements PRAiS with the new calibration. This updated software will be made accessible to all paediatric congenital heart centres in the UK and Republic of Ireland. [3 paragraphs unchanged] The challenges UCL encountered with the process or of accessing and linking datasets was reported in our publication in BMJ Open [45 words unchanged] into the use of health data for research and analysis in research, and has led to a session at this year’s the HSRUK conference 2022 led by the LAUNCHES research team, the funder the Health Foundation and the Health Quality Improvement Partnership (HQIP). (HQIP), and was presented at the OR65 conference in September 2023. A paper on the technical methods of the data linkage titled ‘Linkage [10 words unchanged] care and mortality national datasets to enable research focused on quality improvement’ has been accepted by published in BMJ Open for publication. in May 2022 https://doi.org/10.1136%2Fbmjopen-2021-057343. The team reported on the transition of patients from child to adult [11 words unchanged] to adult services for patients with congenital heart disease in England?’. This has been submitted to BMJ Heart. was published in Heart in November 2022 https://doi.org/10.1136/heartjnl-2022-321085 Five further manuscripts are in analysis or draft stage. A paper titled ‘Hospital resource utilization in a national cohort of functionally single ventricle patients undergoing surgical treatment’ was published in JTCVS Open in June 2023 https://doi.org/10.1016%2Fj.xjon.2023.03.014 A paper titled ‘Arterial Switch for Transposition of the Great Arteries: Treatment Timing, Late Outcomes, and Risk Factors’ was published in JACC: Advances in July 2023 https://www.jacc.org/doi/full/10.1016/j.jacadv.2023.100407 Six further manuscripts are in draft stage. [3 paragraphs unchanged] The CHAMPION study team plan to write the ten peer reviewed publications to be completed within six months of the study end. The study end date is the 30th September 2023. 31st May 2024. Therefore, the team are aiming for the publications to be completed by the end of 2023 2024 and will present findings to key stakeholders (e.g. professional societies, national audit [51 words unchanged] communications are available. The final communication will coincide with the final publication. Amendment 2022: An additional output for the new objective will be the update of the software that implements PRAiS with the new calibration. This updated software will be made accessible to all paediatric congenital heart centres in the UK and Republic of Ireland. Amendment 2023: Additional outputs for the new objective may be the publication of a paper and the presentation of findings as per the original objectives. [1 paragraph unchanged] The team have also arranged the dissemination of findings through the charities [31 words unchanged] to its members. Where appropriate, results will be promoted as press releases (2020-2023). (2020-2024). UCL will also submit annual reports (most recently May 2022) 2023) to the NIHR (the project funders) and partner with them to draw [43 words unchanged] ICNARC, and NICOR, websites as well as on the UCL CORU website. Dissemination will occur throughout the course of the CHAMPION study (2020-2023). The target dates for production and dissemination of the outputs are over the course of the DSA 2023-2026 [3 paragraphs unchanged]

Expected measurable benefits

[15 paragraphs unchanged] The researchers hope to generate the high-quality evidence necessary to guide the [49 words unchanged] publications and presentations will be completed following conclusion of the study in 2022. 2023. The timescale of the full benefit through changes in policies and in procedures is impossible to determine. [4 paragraphs unchanged] Amendment 2023: The new objective involves the analysis of post operative complications in children, the importance of which was flagged as a priority across stakeholders. UCL hope to add interpretation to the audit reporting, and identify opportunities for improvement. [9 paragraphs unchanged] • 4. Recalibrated PRAiS risk model and updated software that implements PRAiS with [6 words unchanged] all paediatric congenital heart centres in the UK and Republic of Ireland. Amendment 2023: 5. Analysis of post operative complications in children relevant to patients, clinicians, and commissioners. [5 paragraphs unchanged] Amendment 2023: The importance of the new objective to analyse postoperative complications in children was flagged as a priority across stakeholders. The complications/morbidities following paediatric cardiac surgery have been audited by NCHDA since 2015/2016 and reported from 2019. However, interpretation has been limited as there is no risk adjustment and there is significant interest from stakeholders (incl. clinicians, patients and commissioners) to improve reporting. [3 paragraphs unchanged]

Benefits reported

Data provided for the LAUNCHES-QI study in September 2020, under Version 0 of this Data Sharing Agreement, was incomplete has been was re-supplied under Version 1 of this Agreement (version 1). The data previously disseminated has been securely destroyed. Data Sharing Agreement. The corrected extract was received in February 2022. No yielded benefits have been achieved to date (December 2022) as a result of the delayed data. Outputs are ongoing, with no yielded benefits to date (October 2023) benefits would not be expected until after the completion of the study.

Text removed is struck through; text added is underlined. Unchanged paragraphs are summarised rather than repeated.

Objective for processing

University College London (UCL) requires pseudonymised data of life status, age at life status*, place of death and a subset of HES data for use in a study called “LAUNCHES QI: Linking AUdit and National datasets in Congenital HEart Services for Quality Improvement” (IRAS ID 246796). *Age at life status will be age at death for those who have died. For those who have no recorded date of death / have not died, the age at life status is the age at the date of civil registration deaths extract. Objective of this new 2022 amendment request DARS NIC 234297 initially only covered a study called “LAUNCHES QI: Linking Audit and National datasets in Congenital Heart Services for Quality Improvement”. In 2019 an amendment was successfully submitted to reuse the LAUNCHES dataset for a new study called CHAMPION “Congenital Heart Audit: Measuring Progress in Outcomes Nationally”. University College London (UCL) request to add additional purpose and processing to the CHAMPION study to recalibrate the Partial Risk Adjustment in Surgery (PRAiS) risk model for mortality after paediatric cardiac surgery with new diagnostic and procedure codes and more recent National Congenital Heart Disease Audit (NCHDA) and Civil Registration (Deaths) data. UCL seek to submit an updated cohort of patients to NHS Digital to link to and receive pseudonymised Civil Registration Mortality life status data. UCL do not require any additional HES data under this agreement. UCL shall be applying to receive additional National Congenital Anomaly and Rare Disease Registration Service (NCARDRS) data (year 2018 of the data was previously received under data controller Public Health England; UCL intend to apply for year 2019 of the data) and to link both years of data to NCHDA data for the original purposes of CHAMPION. However, as the novation from Public Health England (PHE) to NHS Digital has only recently occurred UCL plan to submit this as a separate DARS amendment later because UCL need to prioritise the Civil Registration (Deaths) extract and the NCARDRS data will not be available until the end of 2022 / early 2023. Under previous iterations of this Agreement, UCL required pseudonymised data of life status, age at life status*, place of death and a subset of HES data for use in a study called “LAUNCHES QI: Linking AUdit and National datasets in Congenital HEart Services for Quality Improvement” (IRAS ID 246796). *Age at life status will be age at death for those who have died. For those who have no recorded date of death / have not died, the age at life status is the age at the date of Civil Registration (Deaths) extract. [1 paragraph unchanged] LAUNCHES QI aims to indirectly improve services for congenital heart disease (CHD) (CHD)* by providing the first description of how CHD patients interact with the [32 words unchanged] the current service offer the most potential for improvement programmes. The team will link linked for the first time NCHDA (National Congenital Heart Disease Audit), PICANet (paediatric intensive care audit), ICNARC CMP (adult (Intensive Care National Audit and Research Centre Case Mix Programme - audit of patient outcomes from adult, general critical care units (intensive care and combined intensive care audit), care/high dependency units)) , Life status and place of death, and HES (Hospital Episode Statistics) data. This will aimed to provide information on: a) the challenges in linking national data sets and [17 words unchanged] interactions CHD patients have with different NHS services over time. The team will aim aimed to improve services by: describing patient care trajectories through secondary and tertiary [11 words unchanged] and policy; and exploring variation across services to identify priorities for QI. Measuring, reporting and learning from outcomes should drive quality improvement (QI), but this is particularly challenging for lifelong conditions such as CHD where outcomes need to be interpreted in the context of changing treatment options, service provision and the natural evolution of disease. Given the complex care trajectories of such patients, rich datasets and careful multi-disciplinary analysis is required to identify meaningful variations and opportunities for targeted QI. The study will produce: the first comprehensive understanding of care received by a complex population from birth to adulthood; a basis for creating a step change in how quality in CHD services is measured and improved. *In the medical environment, CHD is commonly used to describe Coronary Heart Disease. However for the purpose of this Data Sharing Agreement, CHD refers to Congenital Heart Disease. The activity is compliant with the principles of GDPR. Based on guidance, [157 words unchanged] public – thereby demonstrating that the processing is in the public interest. LAUNCHES QI is in the interest of the public as it aims to measure and improve services for those with CHD. LAUNCHES QI is a dataset analysis of five linked audit and national [20 words unchanged] (NCHDA) since 2000, the core dataset defining the study CHD population. Included will be are all patients, with no age restrictions, who have had at least one [12 words unchanged] in NCHDA. Patients undergoing cardiac intervention since April 2000 (when NCHDA began) will be are included in the study, but the requested records for these patients in [14 words unchanged] interacted with NHS services prior to their cardiac intervention. NHS Digital Data is was required up until the end of March 2018. The inclusion criteria into [69 words unchanged] future intention to request further data, but not as part of this extension Agreement. Agreement specifically for LAUNCHES QI. Any request for further data will be subject to an updated application with NHS Digital. [1 paragraph unchanged] The study research dataset will be is generated by linking pseudonymised NCHDA data (English and Welsh centres) to the four additional sources: [3 paragraphs unchanged] 4) Death registrations (Civil Registration Deaths). (Will link (linked to identifiers from patients of English and Welsh centres). UCL has received authorisation from HQIP to transfer personal identifiers from NICOR National Institute for Cardiovascular Outcomes Research (NICOR) (which curates the NCHDA dataset) to ICNARC, PICANet and NHS Digital. NICOR [34 words unchanged] for the study to process the data and to link the datasets. Two datasets will be were created: one containing the patient level, pseudonymised information of all hospital admissions, [15 words unchanged] dataset covering all hospital stays for CHD patients. The final pseudonymised datasets will be are stored within UCL’s secure data safe haven. Patient identifiers will be were sent to NHS Digital from the NCHDA database. UCL are requesting requested that NHS Digital match the identifiers from the NCHDA dataset to HES and Civil Registration Deaths and extract the requested fields. NCHDA will provide provided a record level LAUNCHES ID which should be transferred to each pseudonymised HES/Civil Registration Deaths record that matches. UCL additionally request requested that NHS digital Digital generate a unique record-level Study ID that will would be common between NHS Digital and UCL to facilitate data queries. The data is to be was securely transferred with the record level study numbers to University College London. [1 paragraph unchanged] The sole funder, The Health Foundation, are involved in the study only to provide the award, as grant-funding. They UCL will also oversee provide progress reports of the project through annual reports and award meetings. This refers to a funders role of how UCL are spending the funds and how the overall study is progressing, and has nothing to do with the data or analyses. The Health Foundation have no influence over the results. They are not permitted to access any record level NHS Digital data. To confirm, The Health Foundation has no involvement in determining the purpose and means of processing and does not carry out any data controllership activities. [1 paragraph unchanged] The organisations involved as collaborators/advisors other than UCL are, Birmingham Children’s Hospital, [82 words unchanged] decisions on, accessing and analysing the data and so only UCL will act as be the sole data controller. For clarity, no other organisations are determining the purpose or means of the processing. [1 paragraph unchanged] University College London (UCL) requires required the same data extracted for the LAUNCHES QI project (pseudonymised data of [32 words unchanged] ID 266393). UCL are the sole data controller who also process data. [1 paragraph unchanged] Amendment 2022: UCL are requesting a new extract of Civil Registration (Deaths) data (pseudonymised data of life status, age at life status, place of death) linked to an updated extract of the NCHDA identifiers provided for LAUNCHES, for the purposes both of the existing objectives and a new objective for CHAMPION as detailed below. [1 paragraph unchanged] Amendment 2022: Improving NICOR’s main current reporting output (survival after paediatric surgeries using the PRAiS risk model) is within the overall aim but would be additional to the existing workstreams. [2 paragraphs unchanged] Measuring, reporting and learning from outcomes should drive quality improvement (QI), but this is particularly challenging for lifelong conditions such as CHD where outcomes need to be interpreted in the context of changing treatment options, service provision and the natural evolution of disease. Given the complex care trajectories of such patients, rich datasets and careful multi-disciplinary analysis is required to identify meaningful variations and opportunities for targeted QI. The study will produce: the first analyses of their kind involving a wide range of sentinel diagnoses some of which may have different end points; an appropriate approach to risk stratification rather than developing a full risk adjustment model. Amendment 2022: at the request of NICOR, the team will also additionally inform the commissioning and delivery of CHD services with the recalibration of the PRAiS risk model for 30-day mortality after paediatric cardiac surgery with new diagnostic and procedure codes and more recent NCHDA and Civil Registration (Deaths) data. The activity is compliant with the same principles of GDPR as LAUNCHES [160 words unchanged] the public – thereby demonstrating that the processing is in the public interest. interest CHAMPION is in the interest of the public as it aims to measure and improve services via the use of developed tools for those with CHD. In this study UCL are requesting permission to use the LAUNCHES QI dataset as the core dataset within CHAMPION. The LAUNCHES QI dataset was compiled as follows: Both studies aims to produce: the first analyses of their kind involving a wide range of sentinel diagnoses some of which may have different end points; an appropriate approach to risk stratification rather than developing a full risk adjustment model. Measuring, reporting and learning from outcomes should drive quality improvement (QI), but this is particularly challenging for lifelong conditions such as CHD where outcomes need to be interpreted in the context of changing treatment options, service provision and the natural evolution of disease. Given the complex care trajectories of such patients, rich datasets and careful multi-disciplinary analysis is required to identify meaningful variations and opportunities for targeted QI. Under a previous iteration of this data sharing agreement, UCL obtained permission to use the LAUNCHES QI dataset as the core dataset within CHAMPION. The LAUNCHES QI dataset was compiled as follows: [2 paragraphs unchanged] To clarify: The LAUNCHES sample consists of all congenital heart disease procedures [74 words unchanged] two patients turn out to be just one, and vice versa. For this that agreement, the linkage to HES data will be was done at record level using the identifiers in the 143,862 NCHDA records from the LAUNCHES dataset. Each data provider identified in their respective datasets which records in the PICANET, ICNARC-CMP, HES and ONS Mortality data pertained to those CHD individuals and returned to CORU at UCL [7 words unchanged] matched individuals, pseudonymised with the record level unique ID (the LAUNCHES ID). [2 paragraphs unchanged] NACSA records from 2015 to 2017 will consist of approximately 102,276 records. 67,896 records and 58,508 patients. The number of patients this pertains to may be lower than this [20 words unchanged] sample size, as some may already exist in the LAUNCHES QI dataset. The number 864 of these NACSA patients will not be evident until linkage. linked to NCHDA (1,001 records). The pseudonymised linkage of the LAUNCHES QI dataset to NACSA at patient level will allow enhanced trajectories of care to be determined. UCL has received authorisation from HQIP to use the PICANet data, ICNARC to use the ICNARC CMP data, NICOR to use of the NCHDA data, and transfer of pseudonymised clinical information from the NACSA audit. UCL has Ethics (20/NS/0022) and CAG (20CAG0027) approval for the CHAMPION study to process the LAUNCHES QI data for the new purposes of CHAMPION and to link the NACSA dataset. Amendment 2022: UCL have requested the addition of NCHDA years 2017 to 2022 to increase not just the number of patients and records held but also to get more information since significant new fields were added in 2015. Based on an estimate, the number of patients will increase by approximately 50,000 to 146,040 patients in this linked dataset. The number of records pertaining to those patients is estimated to increase from 143,862 to approximately 192,500. However, the final number of patients and records will not be clear until final linkage and cleaning of the data at UCL. Under the previous iterations UCL received authorisation from HQIP to use the PICANet data, ICNARC to use the ICNARC CMP data, NICOR to use of the NCHDA data, and transfer of pseudonymised clinical information from the NACSA audit. UCL has Ethics (20/NS/0022) and CAG (20CAG0027) approval for the CHAMPION study to process the LAUNCHES QI data for the purposes (prior to this 2022 amendment) of CHAMPION and to link the NACSA dataset. Amendment 2022: UCL has received ethics and CAG approval for this substantial amendment to: • Add an objective to the CHAMPION study (recalibrating the PRAiS risk model). • To receive an updated pseudonymised extract of NCHDA data from March 2017 up to the most recent available date (2021 or 2022). • To transfer again the NCHDA identifiers from 2000 to NHS Digital (as was done for LAUNCHES), this time including from March 2017 to the most recent available date (2021 or 2022), and the LAUNCHES study ID to be linked to Civil Registration Mortality data. • For UCL to receive the Mortality data that linked to the identifiers up to the date of extract, pseudonymised with the LAUNCHES study ID (the pseudonymisation key assigned with the identifiers for the LAUNCHES study, to allow pseudonymisation and linkage of the datasets). • To link the Civil Registration (Deaths) data and NCHDA data with the LAUNCHES study ID to use for the new objective. • To replace existing Mortality linked data in the CHAMPION dataset with the update, for the original purposes of CHAMPION. And subject to a separate NHS Digital amendment following this one: • To transfer the NCHDA identifiers from 2000 to the most recent available date to NHS Digital to link to years 2018 and 2019 of NCARDRS data. • For UCL to receive the NCARDRS data for 2018 and 2019, with the linked LAUNCHES study ID if applicable. UCL has submitted an amendment application to NICOR and their new data controller NHS England (Arden & GEM Commissioning Support Unit specifically) for the new NCHDA extract and linkage to Civil Registration (Deaths) data and NCARDRS for these purposes. UCL has submitted an amendment application to PICANet and their data controller HQIP as the PICANet data within the CHAMPION dataset will be updated with the new Civil Registration (Deaths) extract. [2 paragraphs unchanged] Amendment 2022: UCL are requesting permission from NHS Digital to use the Civil Registration Death data for the new objective of the CHAMPION study (PRAiS risk model recalibration) as well as the existing objectives. No HES data is required under this (version 3) agreement. [3 paragraphs unchanged]

Processing activities

Data will not be handled by any additional third party organisations. Data will not be accessed outside England and Wales. the United Kingdom. The planned data flows for all data processing under this agreement are as follows: LAUNCHES QI The completed data flows for all data processing under this agreement are as follows: [1 paragraph unchanged] National Institute for Cardiovascular Outcomes Research (NICOR) will securely transfer transferred a file to the NHS Digital. This file will contain contained patient identifiable information (NHS Number, name, postcode, date of birth, local hospital [10 words unchanged] Disease Audit (NCHDA), and unique study ID (NCHDA record-level LAUNCHES Study ID). 2) NHS Digital will identify common records between NCHDA data and HES and Civil Registration Deaths data, including the requested derived fields. NHS Digital will generate a unique record-level Study ID for each identified record. [1 paragraph unchanged] i) NHS Digital will return HES data for all individuals in the NCHDA cohort to University [9 words unchanged] level study number and HES record-level study number) and requested derived fields will be are appended to the end of every episode record returned to University College London. ii) Civil Registration Deaths derived fields for all individuals in the NCHDA cohort will be returned to University College London. The unique study ID (LAUNCHES QI record level study number and HES record-level study number) will be are appended to the end of every record. 4) The linkage strategy is that NICOR will provide provided the personal identifiers NHS number, hospital number, date of birth and postcode to NHS Digital, ICNARC and PICANet from the NCHDA data. Patient name will was also be sent to PICANet and NHS Digital. Each will identify identified in their respective datasets which records pertain to those CHD individuals and return to UCL the requested clinical and administrative data they hold on the matched individuals (UCL will would not receive data for patients that do not match), pseudonymised with the LAUNCHES record level ID and local dataset record level IDs. UCL will receive received from NICOR the pseudonymised clinical data of the NCHDA dataset and will did not receive any personal identifiers. Linking (at UCL) is was made possible via the pseudonymised unique study ID (LAUNCHES QI record level study number and HES record-level study number). UCL will then link linked HES and Civil Registration Deaths data received from NHS Digital to pseudonymised [23 words unchanged] Patient IDs (which will be contained within the NCHDA data set). UCL will did not be receiving receive any personal identifiers from any of the data sources. UCL will use required the record level NHS Digital study identifiers in case of any queries about specific records with NHS Digital. The data is stored and processed within the UCL Identifiable Data Handling Solution (IDHS) called the Data Safe Haven (DSH). The data will be is held within a secure environment where all statistical analyses will be undertaken. Access to this record level data will be limited to only four members of the UCL LAUNCHES QI team, team" to allow for additional analysts at UCL, Access to this record level data will be limited to only members [102 words unchanged] each user. Any team member leaving the study has their access revoked. [16 paragraphs unchanged] The planned completed data flows are as follows: follows (carried out under previous iterations of this Agreement): [2 paragraphs unchanged] Amendment 2022: As was done for LAUNCHES, National Institute for Cardiovascular Outcomes Research (NICOR) will securely transfer a file of identifiers to NHS Digital. This file will contain patient identifiable information (NHS Number, postcode, date of birth, sex) for all patients in the National Congenital Heart Disease Audit (NCHDA), and unique study ID (NCHDA record-level LAUNCHES Study ID). This will consist of the NCHDA identifiers from 2000 to 2021/2022 (most recent available year at the time of transfer). [1 paragraph unchanged] Data previously supplied under version 0 of this Agreement was incomplete and will be was resupplied under this version (v1) 1. 3) The linkage strategy is that NICOR will provide the NACSA clinical and administrative data for the years requested which where possible will be linked to and pseudonymised with the LAUNCHES record level ID and local dataset record level IDs. Any that do not have a LAUNCHES record level ID will be pseudonymised with the local dataset record level ID only. UCL will receive from NICOR the pseudonymised clinical data of the NACSA dataset and will not receive any personal identifiers. Amendment 2022: For the LAUNCHES QI dataset securely stored currently at UCL, admission, procedure, diagnosis, discharge, and provider information are held for all matched patients. UCL also hold the derived fields, age at admission/appointment to four decimal places and age at discharge to four decimal places. UCL will not hold any dates of birth for CHAMPION QI to prevent identifiability of the data and so CHAMPION QI will use ages at health service interaction. In- and out-patient HES data is critical in identifying the treatment that each patient has received. In addition, HES A&E data are required to identify adverse outcomes such as unplanned emergency treatment following discharge from CHD surgery or through deterioration in a patient’s health. The requested Civil Registration Deaths derived fields of life status, age at life status to four decimal places and place of occurrence of death (home/hospice/hospital/care home/other communal establishment/elsewhere), are vital to complete the patient trajectories. NHS Digital will identify common records between NCHDA data and Civil Registration Deaths data, including the requested derived fields. NHS Digital will provide a unique patient-level Study ID for each identified record. 3) Data flows from NHS Digital Amendment 2022: Civil Registration Deaths derived fields for all individuals in the NCHDA cohort will be returned to University College London. The unique study ID (LAUNCHES QI record level study number and a patient level ID) will be appended to the end of every record. 4) The linkage strategy was that NICOR provided the NACSA clinical and administrative data for the years requested which where possible were linked to and pseudonymised with the LAUNCHES record level ID and local dataset record level IDs. Any that did not have a LAUNCHES record level ID were pseudonymised with the local dataset record level ID only. UCL received from NICOR the pseudonymised clinical data of the NACSA dataset and did not receive any personal identifiers. Amendment 2022: The linkage strategy is that NICOR will provide the personal identifiers NHS number, date of birth, sex and postcode to NHS Digital from the NCHDA data. NHS Digital will identify in their datasets which records pertain to those CHD individuals and return to UCL the requested clinical and administrative data they hold on the matched individuals (UCL will not receive data for patients that do not match), pseudonymised with the LAUNCHES record level ID and local dataset record level IDs. UCL will receive from NICOR the pseudonymised clinical data of the NCHDA dataset and will not receive any personal identifiers. Linking (at UCL) is made possible via the pseudonymised unique study ID (LAUNCHES QI record level study number patient level ID. Amendment 2022: UCL will then link Civil Registration Deaths data received from NHS Digital to the CHAMPION dataset, the pseudonymised clinical data received from NICOR (NCHDA data), PICANet and ICNARC (CMP data), via the LAUNCHES QI unique record-level study number and the LAUNCHES Patient IDs (which will be contained within the NCHDA data set). The Civil Registrations Deaths data will also be linked only to NICOR (NCHDA data) as a separate dataset to analyse for the new objective, to recalibrate the PRAiS algorithm. UCL will not be receiving any personal identifiers from any of the data sources. UCL will use the record level NHS Digital study identifiers in case of any queries about specific records with NHS Digital. For the LAUNCHES QI dataset securely stored currently at UCL, admission, procedure, diagnosis, discharge, and provider information are held for all matched patients. UCL also hold the derived fields, age at admission/appointment to four decimal places and age at discharge to four decimal places. UCL will not hold any dates of birth for CHAMPION QI to prevent identifiability of the data and so CHAMPION QI will use ages at health service interaction. In- and out-patient HES data is critical in identifying the treatment that each patient has received. In addition, HES A&E data was required to identify adverse outcomes such as unplanned emergency treatment following discharge from CHD surgery or through deterioration in a patient’s health. The Civil Registration Deaths derived fields of life status, age at life status to four decimal places and place of occurrence of death (home/hospice/hospital/care home/other communal establishment/elsewhere), are vital to complete the patient trajectories. Amendment 2022: The requested Civil Registration Deaths derived fields are the same as those already held at UCL: life status, age at life status to four decimal places and place of occurrence of death (home/hospice/hospital/care home/other communal establishment/elsewhere). These fields are vital to complete the patient trajectories with the update of the CHAMPION dataset, and also vital to complete the new objective or recalibrating the PRAiS risk model. UCL are not requesting any updates of HES data as part of this amendment. [4 paragraphs unchanged] Amendment 2022: Methods of analysis for the amendment will include: 1. Data cleaning and descriptive analysis of the linked NCHDA and Civil Registration Deaths data extract for recalibration of the PRAiS model UCL will: 2. Re-calibrate the PRAiS risk model; Update the software that implements PRAiS with the new calibration; Make the updated software accessible to all paediatric congenital heart centres in the UK and Republic of Ireland. 3. Update the Civil Registration Deaths data in the CHAMPION dataset for any further analyses of robust outcomes for adult CHD procedures and to establish longer-term outcomes by diagnosis. [1 paragraph unchanged] NHS Digital reminds all All organisations party to this agreement of the need to must comply with the Data Sharing Framework Contract requirements, including those regarding the use (and purposes of that use) by “Personnel” Personnel (as defined within the Data Sharing Framework Contract ie: i.e.: employees, agents agents, PhD students and contractors of the Data Recipient who may have access to that data).

Expected output

[3 paragraphs unchanged] Amendment 2022: An additional output for the new objective will be the update of the software that implements PRAiS with the new calibration. This updated software will be made accessible to all paediatric congenital heart centres in the UK and Republic of Ireland. [1 paragraph unchanged] The team have also arranged the dissemination of findings through the Children’s [90 words unchanged] website. Dissemination will occur throughout the course of the LAUNCHES study (2019-2023). The Health Foundation publishes details of the study https://www.health.org.uk/research-projects/launches-qi-linking-audit-and-national-datasets-in-congenital-heart-services. [1 paragraph unchanged] The challenges we UCL encountered with the process or accessing and linking datasets was reported in [79 words unchanged] the funder the Health Foundation and the Health Quality Improvement Partnership (HQIP). [10 paragraphs unchanged] We UCL have published an academic manuscript titled ‘What was the impact of the [17 words unchanged] A qualitative study using online forums’, in BMJ Open in September 2021. [1 paragraph unchanged]

Expected measurable benefits

[6 paragraphs unchanged] Benefits It is hoped benefits will involve: [4 paragraphs unchanged] Identifying candidate metrics to inform commissioning and drive change for CHD: Many [67 words unchanged] and the relevant national audits, and endorsement from NHS England, clinical bodies (BCCA, SCTS, RCSEd) (British Congenital Cardiac Association (BCCA), Society for Cardiothoracic Surgery in Great Britain and Ireland (SCTS), Royal College of Surgeons of Edinburgh (RCSEd)) & HQIP. Healthcare Quality Improvement Partnership (HQIP) . [1 paragraph unchanged] Provide a template for other high-resource, high-impact NHS services: The approach for linking audit data to measure quality and variations will aim to be applicable to other lifelong conditions, particularly those where patients are likely to require periods of intensive care such as Cystic Fibrosis or Renal Disease. The team will aim to share & build on the generalisable learning e.g.: technical & governance solutions [17 words unchanged] & Quality Assurance (QA); approaches to tracking patient trajectories through multiple datasets. [1 paragraph unchanged] The researchers will hope to generate the high-quality evidence necessary to guide the quality improvement of CHD [63 words unchanged] benefit through changes in policies and in procedures is impossible to determine. [2 paragraphs unchanged] Amendment 2022: This new objective involves the analysis of the NCHDA data linked to Civil Registration Deaths data. It aims to involve the update of the PRAiS (Partial Risk Adjustment in Surgery) risk model which is currently used by all UK centres performing paediatric heart surgery and NICOR. The model works very well, but was last calibrated on data up to 2015. Since then, practice has progressed and more diagnostic and procedure codes have been added. UCL propose to recalibrate the PRAiS risk model using more up to date data to incorporate these changes and future proof it for ongoing use. [4 paragraphs unchanged] Benefits It is hoped benefits will involve: [3 paragraphs unchanged] Amendment 2022: • 4. Recalibrated PRAiS risk model and updated software that implements PRAiS with the new calibration, made accessible to all paediatric congenital heart centres in the UK and Republic of Ireland. [2 paragraphs unchanged] The researchers are expected to generate the high-quality evidence necessary to guide the quality improvement of CHD services and inform decisions about national policies. Scientific manuscripts will be written detailing the findings. UCL expect the publications to contribute to the evidence and expert opinion for the development and update of clinical guidelines in congenital heart disease in the future. The dissemination via publications and presentations will be completed following conclusion of the study in 2023. The timescale of the full benefit through changes in policies and in procedures is impossible to determine. Amendment 2022: The risk of death faced by children undergoing heart surgery depends amongst other things on the heart condition they have, the operation they have, their age and weight and whether they have other coexisting health problems (known as comorbidities). In previous work UCL used 10 years of national data to develop an equation, called the PRAiS (Partial Risk Adjustment in Surgery) model, and computer software that implements it, using these risk factors to estimate the chance of death within 30 days of surgery. PRAiS is now used nationally and by the national audit body (NICOR). The model was improved in 2015 for 30-day mortality following paediatric cardiac surgery by incorporating more detailed information on comorbid conditions. Recalibration is required again now as practice has progressed and more diagnostic and procedure codes have been added. UCL aim to use more recent national CHD data to incorporate these changes and future proof the model for ongoing use. This recalibration is expected to help all clinical teams involved in the management of patients with CHD to monitor and improve their services. Scientific manuscripts will be written detailing the findings. UCL expect the publications to contribute to the evidence and expert opinion for the development and update of clinical guidelines in congenital heart disease in the future. The dissemination via publications and presentations aims to be completed following conclusion of the study in 2023. The timescale of the full benefit through changes in policies and in procedures is impossible to determine. [2 paragraphs unchanged]

Benefits reported

[1 paragraph unchanged] The corrected extract was received in February 2022. No yielded benefits have been achieved to date (December 2022) as a result of the delayed data.

Text removed is struck through; text added is underlined. Unchanged paragraphs are summarised rather than repeated.

Objective for processing

[4 paragraphs unchanged] The activity is compliant with the principles of GDPR. Based on guidance, [87 words unchanged] processing is necessary for scientific and research purposes, subject to appropriate safeguards. The data are required for research purposes in the public interest – meeting the conditions in the DPA 2018 Schedule 1 Part 1 (4) - which GDPR Recital 52(2) determines is an appropriate derogation from the prohibition on processing special categories of personal data. The ways in which the processing of data will be of benefit to the public – thereby demonstrating that the processing is in the public interest. LAUNCHES QI is a dataset analysis of five linked audit and national [113 words unchanged] until the end of March 2018. The inclusion criteria into the study is was anyone captured in the NCHDA audit since it began, up to March 2017 which was as far as was available at the time of application. The other datasets are were then used to track the patient trajectories both before and since March 2017. Therefore, UCL are requesting requested data as far as was available at application to capture interactions of the patients with the health service. There is a future intention to request further data, but not as part of this extension Agreement. Any request for further data will be subject to an updated application with NHS Digital. [11 paragraphs unchanged] As in many studies now there are a number of collaborators providing an advisory role. Only UCL substantive employees and the two clinical researchers from Great Ormond Steeet Street Hospital with honorary contracts with UCL are working with the data. Clinical researchers will not have access to data until the signed contract/addendum letter is in place. The organisations involved as collaborators/advisors other than UCL are, Birmingham Children’s Hospital, [86 words unchanged] analysing the data and so only UCL will act as data controller. For clarity, no other organisations are determining the purpose or means of the processing. [7 paragraphs unchanged] The activity is compliant with the same principles of GDPR as LAUNCHES [91 words unchanged] processing is necessary for scientific and research purposes, subject to appropriate safeguards. The data are required for research purposes in the public interest – meeting the conditions in the DPA 2018 Schedule 1 Part 1 (4) - which GDPR Recital 52(2) determines is an appropriate derogation from the prohibition on processing special categories of personal data. The ways in which the processing of data will be of benefit to the public – thereby demonstrating that the processing is in the public interest. [14 paragraphs unchanged]

Processing activities

Data will not be handled by any additional third party organisations. Data will not be accessed outside the UK. England and Wales. [7 paragraphs unchanged] 4) The linkage strategy is that NICOR will provide the personal identifiers [92 words unchanged] data of the NCHDA dataset and will not receive any personal identifiers. Linking (at UCL) is made possible via the pseudonymised unique study ID (LAUNCHES QI record level study number and HES record-level study number). [31 paragraphs unchanged]

Expected output

[2 paragraphs unchanged] The LAUNCHES QI study team plan to write the ten peer reviewed [8 words unchanged] the study end. The study end date has now been extended to April 2022. October 2023. Therefore the team are aiming that the publications will be completed by Autumn 2022 Summer 2023 and will present findings to key stakeholders (e.g. professional societies, national audit [14 words unchanged] briefing documents. UCL will disseminate to the public through a project website (https://www.ucl.ac.uk/operationalresearch/domains/congenital_heart_disease/launches), (https://www.ucl.ac.uk/operationalresearch/domains/congenital_heart_disease/launches) , and via social media (Twitter @UCL_CORU) and blogs. Updates to the website [12 words unchanged] communications are available. The final communication will coincide with the final publication. [1 paragraph unchanged] The team have also arranged the dissemination of findings through the Children’s Heart Federation. Where appropriate, results will be promoted as press releases (2021-2022). (2022-2023). UCL will also submit have submitted annual reports (most recently in April 2022) to the Health Foundation (the project funders) and will partner with them to draw on their networks and skills in dissemination [15 words unchanged] as downloadable leaflets and case studies, research highlights, blogs and webinars. UCL will also publish has published details of LAUNCHES QI on the PICANet, ICNARC and NICOR websites as [5 words unchanged] CORU website. Dissemination will occur throughout the course of the LAUNCHES study (2019-2022). (2019-2023). Publications and conferences to date The challenges we encountered with the process or accessing and linking datasets was reported in our publication in BMJ Open in July 2021: ‘The road to hell is paved with good intentions: the experience of applying for national data for linkage and suggestions for improvement’ https://bmjopen.bmj.com/content/11/8/e047575.long. This paper was presented at the HSRUK conference 2021, has been widely cited, featured in the Government-commissioned Goldacre Review into the use of health data for research and analysis in research, and has led to a session at this year’s HSRUK conference led by the LAUNCHES research team, the funder the Health Foundation and the Health Quality Improvement Partnership (HQIP). A paper on the technical methods of the data linkage titled ‘Linkage of National Congenital Heart Disease Audit data to hospital, critical care and mortality national datasets to enable research focused on quality improvement’ has been accepted by BMJ Open for publication. The team reported on the transition of patients from child to adult services in a paper titled: ‘How effective is transfer from paediatric to adult services for patients with congenital heart disease in England?’. This has been submitted to BMJ Heart. Five further manuscripts are in analysis or draft stage. [3 paragraphs unchanged] The CHAMPION study team plan to write the ten peer reviewed publications to be completed within six months of the study end. The study end date of 31st December 2022 has received nine months of extension with NIHR and will result in a new study end date of is the 30th September 2023. Therefore, the team are aiming for the publications to [70 words unchanged] communications are available. The final communication will coincide with the final publication. [1 paragraph unchanged] The team have also arranged the dissemination of findings through the charities Children’s Heart Federation , Federation, Little Hearts Matter, The Somerville Foundation, and British Heart Foundation who will [16 words unchanged] to its members. Where appropriate, results will be promoted as press releases (2020-2023 ). (2020-2023). UCL will also submit annual reports (most recently May 2022) to the NIHR (the project funders) and partner with them to draw [54 words unchanged] website. Dissemination will occur throughout the course of the CHAMPION study (2020-2023). Publications to date We have published an academic manuscript titled ‘What was the impact of the first wave of COVID-19 on the delivery of care to children and adults with congenital heart disease? A qualitative study using online forums’, in BMJ Open in September 2021. This paper was part of the 'Eliciting views of patients and families’ objective of CHAMPION. Recommendations were made to improve patient experiences of care and potentially their outcomes, based on findings from our online discussion forums.

Benefits reported

Data provided for the LAUNCHES-QI study in September 2020, under Version 0 of this Data Sharing Agreement, was incomplete and will be has been re-supplied under this Agreement (version 1). The data previously disseminated will be has been securely destroyed. The corrected extract was received in February 2022. No yielded benfits benefits have been achieved to date as a result of the incomplete delayed data.

Unchanged: Expected measurable benefits.

Text removed is struck through; text added is underlined. Unchanged paragraphs are summarised rather than repeated.

Datasets: + HES-ID to MPS-ID HES Accident and Emergency; + HES-ID to MPS-ID HES Admitted Patient Care; + HES-ID to MPS-ID HES Outpatients

Objective for processing

University College London (UCL) requires pseudonymised data of life status, age at life status, status*, place of death and a subset of HES data for use in a new study called “LAUNCHES QI: Linking AUdit and National datasets in Congenital HEart Services for Quality Improvement” (IRAS ID 246796). UCL are *Age at life status will be age at death for those who have died. For those who have no recorded date of death / have not died, the sole data controller who also process data. age at life status is the age at the date of civil registration deaths extract. UCL are the sole data controller who also process data. [3 paragraphs unchanged] LAUNCHES QI is a dataset analysis of five linked audit and national datasets which will include up to approximately 144,000 includes 143,862 patients with congenital heart disease that have been captured by the National [79 words unchanged] patients may have interacted with NHS services prior to their cardiac intervention. CHD NHS Digital Data is a life-long condition, so obtaining 20 years required up until the end of data (including 3 years prior to March 2018. The inclusion criteria into the study is anyone captured in the NCHDA dataset) is crucial for establishing audit since it began, up to March 2017 which was as complete far as was available at the time of application. The other datasets are then used to track the patient trajectories of their service use both before and since March 2017. Therefore, UCL are requesting data as possible, which will enable the study far as available to get a more detailed understanding of how patients interact with secondary care throughout their lives, how this varies across centres, and how it impacts on outcomes. National data is required to allow exploration capture interactions of the different CHD services across patients with the country. The pseudonymised linkage of the other 4 datasets to NCHDA at patient level will allow trajectories of care to be determined. health service. CHD is a life-long condition, so obtaining 20 years of data (including 3 years prior to the NCHDA dataset) is crucial for establishing as complete trajectories of their service use as possible, which will enable the study to get a more detailed understanding of how patients interact with secondary care throughout their lives, how this varies across centres, and how it impacts on outcomes. National data is required to allow exploration of the different CHD services across the country. The linkage of the other 4 datasets to NCHDA at patient level will allow trajectories of care to be determined. [8 paragraphs unchanged] The LAUNCHES QI study was instigated and is led by the Principal Investigators based at the Clinical Operational Research Unit (CORU) at UCL. University College London UCL have two members of the study team who have honorary contracts at UCL. Both are clinicians working at Great Ormond Street Hospital and their input as clinicians for CHD is valuable as they have experience with similar data and will be able to assist with analyses. This will increase the only organisation to require access to impact of the record level data supplied from NHS Digital. study. [1 paragraph unchanged] As in many studies now there are a number of collaborators providing an advisory role. Only UCL substantive employees and the two clinical researchers from Great Ormond Steeet Hospital with honorary contracts with UCL are working with the data. The organisations involved as collaborators/advisors other than UCL are, Birmingham Children’s Hospital, Intensive Care National Audit & Research Centre, Leeds University, Royal Brompton NHS Foundation Trust, Great Ormond Street Hospital NHS Trust, Leeds Teaching Hospitals NHS Foundation Trust, University Hospital Southampton, and University Hospitals Bristol NHS Foundation Trust. These organisations Clinical researchers will not have access to data until the data. UCL researchers make all the final decisions as data controllers. signed contract/addendum letter is in place. The organisations involved as collaborators/advisors other than UCL are, Birmingham Children’s Hospital, Intensive Care National Audit & Research Centre, Leeds University, Royal Brompton NHS Foundation Trust, Great Ormond Street Hospital NHS Trust, Leeds Teaching Hospitals NHS Foundation Trust, University Hospital Southampton, and University Hospitals Bristol NHS Foundation Trust. These organisations will not have access to the data. The collaborators together with the UCL researchers make up the full study research team, with the collaborators providing insight into the aggregated results produced by the UCL researcher analysis. Only the UCL research team will be making decisions on, accessing and analysing the data and so only UCL will act as data controller. The CHAMPION study University College London (UCL) requires the same data extracted for the LAUNCHES QI project (pseudonymised data of life status, age at life status, place of death and a subset of HES data) for use in a follow up study called “Congenital Heart Audit: Measuring Progress In Outcomes Nationally” (IRAS ID 266393). UCL are the sole data controller who also process data. UCL are asking for permission to use the LAUNCHES QI data extract for the purpose of a new project, “CHAMPION”. CHAMPION aims to indirectly improve services for congenital heart disease (CHD) by developing tools for routinely measuring CHD outcomes (including risk-adjusted survival in adults following intervention and long-term outcomes by diagnosis) that are considered meaningful and appropriate by stakeholders and can inform the delivery and commissioning of CHD services. The study will provide information on a) Which cases of adults undergoing surgery or interventional cardiology should be included in the reporting and comparing of adult CHD outcomes; establish which outcomes are feasible and appropriate for routine monitoring and which factors influence outcomes; develop the analytical methods for taking account of differences in risk across case-mix. b) Key diagnoses for which outcomes should be reported, establish which outcome metrics can be routinely measured for these diagnoses, and develop the analytical means to do so (including exploring non-diagnostic factors that can influence outcomes). The team will aim to inform the commissioning and delivery of CHD services: defining outcomes for adult CHD procedures, establishing longer-term outcomes by diagnosis, and building prototype software tools that specify the automated mechanisms required to extract, generate and visualise these outcomes. Measuring, reporting and learning from outcomes should drive quality improvement (QI), but this is particularly challenging for lifelong conditions such as CHD where outcomes need to be interpreted in the context of changing treatment options, service provision and the natural evolution of disease. Given the complex care trajectories of such patients, rich datasets and careful multi-disciplinary analysis is required to identify meaningful variations and opportunities for targeted QI. The study will produce: the first analyses of their kind involving a wide range of sentinel diagnoses some of which may have different end points; an appropriate approach to risk stratification rather than developing a full risk adjustment model. The activity is compliant with the same principles of GDPR as LAUNCHES QI. Based on guidance, as this is for University research, the lawful basis for processing data is GDPR article 6(1)(e): ‘Processing is necessary for the performance of a task carried out in the public interest or in the exercise of official authority vested in the controller’. Also referred to as ‘Public Task’. As the research involves health data, which is included in the definition of special categories of personal data, it requires an additional condition for processing. Based on guidance, for health research this is GDPR article 9(2)(j), which details that processing is necessary for scientific and research purposes, subject to appropriate safeguards. In this study UCL are requesting permission to use the LAUNCHES QI dataset as the core dataset within CHAMPION. The LAUNCHES QI dataset was compiled as follows: All patients who have at least one entry in the National Congenital Heart Disease Audit (NCHDA) (data controller HQIP), up to March 2017 were included in the dataset. NICOR who oversee this audit, provided the personal identifiers NHS number, hospital number, date of birth and postcode for these patients to the data providers PICANet and ICNARC, and will provide these identifiers to NHS Digital. Patient name was also sent to PICANet and NHS Digital. The LAUNCHES QI sample is the entire cohort of patients (children and adults) who had at least one procedure related to congenital heart disease since 2000, and who appear on the NCHDA database up to March 2017 (n= 96,040), some with more than one record so a total of n= 143,862 records. The total LAUNCHES QI dataset sample size is subject to slight change following further data cleaning. To clarify: The LAUNCHES sample consists of all congenital heart disease procedures recorded in NCHDA between April 2000 and March 2017 from public hospitals (N=143,862 records). Together with NICOR and PICANet UCL worked on the identification of patients and revision of inconsistencies in dates and identifiers, which resulted in the current identification of n=96,041 patients (n=96,109 before quality assurance). The final number of patients may further change as UCL address any anomalies: as UCL link more datasets it may emerge those records thought to be for two patients turn out to be just one, and vice versa. For this agreement, the linkage to HES data will be done at record level using the identifiers in the 143,862 NCHDA records from the LAUNCHES dataset. Each data provider identified in their respective datasets which records in the PICANET, ICNARC-CMP, HES and ONS data pertained to those CHD individuals and returned to CORU at UCL the clinical data they hold on the matched individuals, pseudonymised with the record level unique ID (the LAUNCHES ID). These five national datasets have been linked at UCL to make up the LAUNCHES QI dataset. For CHAMPION, UCL also have permission from HQIP to access NACSA data and have requested NICOR match NACSA to NCHDA as they manage both audits. They will be able to flag the matching records by assigning the corresponding record's LAUNCHES ID, therefore the personal identifiers will not need to leave NICOR. NACSA records from 2015 to 2017 will consist of approximately 102,276 records. The number of patients this pertains to may be lower than this as more than one record may apply to one patient. Not all patients in NACSA may add to the total sample size, as some may already exist in the LAUNCHES QI dataset. The number of these patients will not be evident until linkage. The pseudonymised linkage of the LAUNCHES QI dataset to NACSA at patient level will allow enhanced trajectories of care to be determined. UCL has received authorisation from HQIP to use the PICANet data, ICNARC to use the ICNARC CMP data, NICOR to use of the NCHDA data, and transfer of pseudonymised clinical information from the NACSA audit. UCL has Ethics (20/NS/0022) and CAG (20CAG0027) approval for the CHAMPION study to process the LAUNCHES QI data for the new purposes of CHAMPION and to link the NACSA dataset. The final pseudonymised datasets additionally containing NACSA data will be stored within UCL’s secure data safe haven. UCL are requesting permission from NHS Digital to use the HES and mortality data for the new purposes of the CHAMPION study. The CHAMPION study was instigated and is led by the Principal Investigators based at the Clinical Operational Research Unit (CORU) at UCL. UCL have two members of the study team who have honorary contracts at UCL. Both are clinicians working at Great Ormond Street Hospital and their input as clinicians for CHD is valuable as they have experience with similar data and will be able to assist with analyses. This will increase the impact of the study. The sole funder, The NIHR, are involved in the study only to provide the award, as grant-funding. They will also oversee progress of the project through annual reports and award meetings. The NIHR have no influence over the results. They are not permitted to access any record level NHS Digital data. As in many studies now there are a number of collaborators providing an advisory role. Only UCL substantive employees and clinical researchers on an honorary contract with UCL are working with the data. Clinical researchers will not have access to data until the signed contract/addendum is in place. The organisations involved as collaborators/advisors other than UCL are, Alder Hey Children’s NHS Foundation Trust, Barts Health NHS Trust, Newcastle University, Royal Brompton NHS Foundation Trust, Great Ormond Street Hospital NHS Trust, Leeds Teaching Hospitals NHS Foundation Trust, The Magdi Yacoub Institute, University Hospitals Bristol NHS Foundation Trust, Children’s Heart Federation, Little Hearts Matter, and The Somerville Foundation. These organisations will not have access to the data. UCL researchers make all the final decisions as data controllers.

Processing activities

LAUNCHES QI has the necessary research ethics and section 251 approvals. A favourable opinion has been obtained from the North of Scotland Research Ethics Committee, reference number 18/NS/0106. Section 251 support has been received to ensure that the accessing, linking and processing of the datasets is in line with the common law duty of confidence (Ref: 18CAG0180). Data will not be handled by any additional third party organisations. Data will not be accessed outside the UK. The planned data flows for all data processing under this agreement are as follows: [8 paragraphs unchanged] The data is stored and processed within the UCL Identifiable Data Handling [30 words unchanged] will be limited to only four members of the LAUNCHES QI team, Access to this record level data will be limited to only members of the CHAMPION team, who are all substantive employees at UCL or clinical researchers on an honorary contract with UCL. Clinical researchers will not have access to data until the signed contract/addendum letter is in place. Staff accessing the UCL data safe haven attend training in its use [54 words unchanged] each user. Any team member leaving the study has their access revoked. [1 paragraph unchanged] Any linkage that could identify an individual is not permitted under this agreement. [1 paragraph unchanged] As part of LAUNCHES analysis; admission, procedure, diagnosis, discharge, and provider information are required for all matched patients. UCL are also requesting derived fields, age at admission/appointment to four decimal places and age at discharge to four decimal places. UCL are not requesting any dates of birth for this project to prevent identifiability of the data and so LAUNCHES will use ages at health service interaction to determine each patient’s treatment trajectory. In- and out-patient HES data is critical in identifying the treatment that each patient has received. In addition, HES A&E data are required to identify adverse outcomes such as unplanned emergency treatment following discharge from CHD surgery or through deterioration in a patient’s health. The requested Civil Registration Deaths derived fields of life status, age at life status to four decimal places and place of occurrence of death (home/hospice/hospital/care home/other communal establishment/elsewhere), are vital to complete the patient trajectories. The data processing activities described applies to all processing under this Data Sharing Agreement. As part of LAUNCHES analysis; admission, procedure, diagnosis, discharge, and provider information are required for all matched patients. UCL are also requesting derived fields, age at admission/appointment to four decimal places* and age at discharge to four decimal places. UCL are not requesting any dates of birth for this project to prevent identifiability of the data and so LAUNCHES will use ages at health service interaction to determine each patient’s treatment trajectory. In- and out-patient HES data is critical in identifying the treatment that each patient has received. In addition, HES A&E data are required to identify adverse outcomes such as unplanned emergency treatment following discharge from CHD surgery or through deterioration in a patient’s health. The requested Civil Registration Deaths derived fields of life status, age at life status to four decimal places and place of occurrence of death (home/hospice/hospital/care home/other communal establishment/elsewhere), are vital to complete the patient trajectories. *Age to 4 decimal places is age presented as whole years with decimal places replacing month, day, and time units. This allows UCL to determine order of events that occur within 24 hours of each other (which can and does happen for children’s heart surgery). UCL need the 4 decimal places age to provide enough time resolution to distinguish the order of procedure for it will allow UCL to sequence operations on the same day for the same patient as long as they are more than one hour apart and a time of procedure is entered for each record. There will be some records where there is no time stamp (only a date) where UCL will not be able to determine the sequence by age at operation alone, but UCL wish to use time information where it is available. The following milestones determines the age to 4 decimal places; age at admission, age at appointment/procedure, age at discharge, age at life status. UCL receive month and year of birth and the further age updates depends on the number of interactions with the health service. The other information would be the clinical fields that were requested. [6 paragraphs unchanged] The CHAMPION Study The planned data flows are as follows: 1) Data flows to NHS Digital: There will be no further data flows to NHS Digital. 2) Data flows from NHS Digital: Data previously supplied under version 0 of this Agreement was incomplete and will be resupplied under this version (v1) 3) The linkage strategy is that NICOR will provide the NACSA clinical and administrative data for the years requested which where possible will be linked to and pseudonymised with the LAUNCHES record level ID and local dataset record level IDs. Any that do not have a LAUNCHES record level ID will be pseudonymised with the local dataset record level ID only. UCL will receive from NICOR the pseudonymised clinical data of the NACSA dataset and will not receive any personal identifiers. For the LAUNCHES QI dataset securely stored currently at UCL, admission, procedure, diagnosis, discharge, and provider information are held for all matched patients. UCL also hold the derived fields, age at admission/appointment to four decimal places and age at discharge to four decimal places. UCL will not hold any dates of birth for CHAMPION QI to prevent identifiability of the data and so CHAMPION QI will use ages at health service interaction. In- and out-patient HES data is critical in identifying the treatment that each patient has received. In addition, HES A&E data are required to identify adverse outcomes such as unplanned emergency treatment following discharge from CHD surgery or through deterioration in a patient’s health. The requested Civil Registration Deaths derived fields of life status, age at life status to four decimal places and place of occurrence of death (home/hospice/hospital/care home/other communal establishment/elsewhere), are vital to complete the patient trajectories. Methods of analysis for the study will include: 1. Data cleaning and descriptive analysis of the NACSA dataset. 2. Establish robust outcomes for adult CHD procedures 3. Establish longer-term outcomes by diagnosis The results of all analyses will be published in aggregate form, with small numbers suppressed in line with HES guidance. No identifiable data will be held by University College London, therefore no identifiable data will be released. [1 paragraph unchanged]

Expected output

The results of the LAUNCHES QI study will be disseminated actively and extensively. The research team has strong [44 words unchanged] Foundation, Children’s Heart Federation, The British Heart Foundation and Little Hearts Matter. [1 paragraph unchanged] The LAUNCHES QI study team plan to write the ten peer reviewed publications to be completed within six months of the study end. The study end date is end of February 2021. has now been extended to April 2022. Therefore the team are aiming that the publications will be completed by summer 2021 Autumn 2022 and will present findings to key stakeholders (e.g. professional societies, national audit [14 words unchanged] briefing documents. UCL will disseminate to the public through a project website (https://www.ucl.ac.uk/operational-research/domains/congenital_heart_disease/launches), (https://www.ucl.ac.uk/operationalresearch/domains/congenital_heart_disease/launches), and via social media (Twitter @UCL_CORU) and blogs. Updates to the website [12 words unchanged] communications are available. The final communication will coincide with the final publication. [1 paragraph unchanged] The team have also arranged the dissemination of findings through the Children’s Heart Federation. Where appropriate, results will be promoted as press releases (2019-2021). (2021-2022). UCL will also submit reports to the Health Foundation (the project funders) [60 words unchanged] CORU website. Dissemination will occur throughout the course of the LAUNCHES study (2019-2021). (2019-2022). The CHAMPION study The results of the study will be disseminated actively and extensively. The research team has strong links with the Congenital Heart Services Clinical Reference Group, NHS England and clinical bodies including the British Congenital Cardiac Association, the Society for Cardiothoracic Surgery and the Royal College of Surgeons of Edinburgh. UCL also have strong links with CHD charities to capture their perspectives through online forums as patient and family views are crucial to measuring outcomes that matter and make sense. Charities include the Somerville Foundation, Children’s Heart Federation, The British Heart Foundation and Little Hearts Matter. Outputs will involve approximately ten publications in peer-reviewed Medical and Scientific Journals, oral and written presentations at national and international conferences. Target journals for the papers are Circulation, Heart, The Annals of Thoracic Surgery, and Archives of Disease in Childhood. The final outputs will only contain aggregate results with small number suppression, in line with the HES Analysis Guidelines. The CHAMPION study team plan to write the ten peer reviewed publications to be completed within six months of the study end. The study end date of 31st December 2022 has received nine months of extension with NIHR and will result in a new study end date of 30th September 2023. Therefore, the team are aiming for the publications to be completed by the end of 2023 and will present findings to key stakeholders (e.g. professional societies, national audit bodies, the Care Quality Commission, HQIP, commissioners and local hospitals) through meetings and short briefing documents. UCL will disseminate to the public through a project website (https://www.ucl.ac.uk/clinical-operational-research-unit/champion), and via social media (Twitter @UCL_CORU) and blogs. Updates to the website will be ongoing throughout the study and as and when publications and communications are available. The final communication will coincide with the final publication. UCL will ensure that lay summaries are provided (reviewed in collaboration with patients and parents on their Advisory Committee). The patients and parents on the advisory committee attend annual advisory group meetings to receive updates and to provide feedback on any aspect of the study. The team have also arranged the dissemination of findings through the charities Children’s Heart Federation , Little Hearts Matter, The Somerville Foundation, and British Heart Foundation who will send newsletters to members and news items on its website highlighting the work of the study to its members. Where appropriate, results will be promoted as press releases (2020-2023 ). UCL will also submit reports to the NIHR (the project funders) and partner with them to draw on their networks and skills in dissemination and spread to make an impact more widely, which may include generating accessible resources such as downloadable leaflets and case studies, research highlights, blogs and webinars. UCL will also publish details of CHAMPION on the PICANet, ICNARC, and NICOR, websites as well as on the UCL CORU website. Dissemination will occur throughout the course of the CHAMPION study (2020-2023).

Expected measurable benefits

The LAUNCHES QI study addresses important areas related to services in congenital heart disease (CHD) [8 words unchanged] information. In response to national recognition of the urgent need, the study will is expected to provide information regarding better system measures for identifying opportunities for quality improvement, trajectories of service use, and variation in CHD service provision. [2 paragraphs unchanged] 2. Little is known about patient trajectories through the NHS for lifelong [9 words unchanged] outcomes and sequence of contacts with the NHS over time). The study will is expected to generate important understanding about service use and variation, and using this to identify key areas to target improvements in health care quality. 3. Many specialised CHD services are commissioned directly by NHS England. To [11 words unchanged] necessarily uses fairly crude measures available in the current system. The study will is expected to directly address a stated need for better measures of service quality by NHS England by developing new robust service metrics. 4. Other lifelong conditions, particularly those where patients are likely to require periods of intensive care such as Cystic Fibrosis or Renal Disease, will are also expected to benefit from their approach for linking audit data to measure quality and variations. Benefits will include: involve: [3 paragraphs unchanged] This research will is hoped to benefit the provision of health care by: Identifying candidate metrics to inform commissioning and drive change for CHD: Many [20 words unchanged] necessarily uses fairly crude measures available in the current system. The study will is expected to directly address a stated need for better measures of service quality by [25 words unchanged] and endorsement from NHS England, clinical bodies (BCCA, SCTS, RCSEd) & HQIP. Generate important new knowledge about lifelong service use & variation for CHD, [29 words unchanged] contacts with the NHS over time). Leveraging the national audits, the study will is expected to establish patient trajectories for the CHD population for the first time, generating [9 words unchanged] this to identify key areas to target improvements in health care quality. [1 paragraph unchanged] With approximately Approximately 200,000 people are currently living with CHD in the UK and services are expensive, high profile and have enormous impact on patients’ and their family’s lives, dissemination is in the public interest. lives. With the possibility of applying findings to other lifelong conditions the magnitude of the impact is potentially even higher. The researchers will generate the high-quality evidence necessary to guide the quality [48 words unchanged] publications and presentations will be completed following conclusion of the study in 2021. 2022. The timescale of the full benefit through changes in policies and in procedures is impossible to determine. The CHAMPION study: The CHAMPION study addresses important areas related to services in congenital heart disease (CHD) about which there is currently little or no information. In response to national recognition of the urgent need, the study is expected to provide information regarding better adult outcome measures for identifying opportunities for quality improvement, trajectories of adult CHD patients, and longer- term outcomes for all CHD patients. It is hoped that this research will enable stakeholders including commissioners to assure the quality of services for people with CHD. By defining the scope of adult CHD practice, developing methods and prototype tools to risk-stratify adult outcomes and measuring longer-term paediatric and adult outcomes by diagnosis, UCL work is expected to enable commissioners to benchmark the outcomes achieved by other centres and clinical teams to target service improvements. The main areas of uncertainty addressed by the CHAMPION study are: 1. There is little known regarding trajectories and outcomes for adult CHD patients, following operations. The percentage of adults surviving certain procedures are publicly reported, but there is no overview by hospital because there is no way to take into account the complexity of each operation. 2. Little is known about what happens in the longer-term for a person with CHD. From both a patient and an NHS perspective, it is more relevant to understand what happens to people over the longer-term (and possibly many operations) rather than just one month after surgery. Benefits will involve: 1. Rigorous definition of adult CHD procedures, survival (and other) outcomes and risk-adjustment methods for national reporting of centre outcomes. 2. Objectively defined sentinel diagnoses and long-term outcome measures relevant to patients, clinicians and commissioners. 3. Prototype software for routine identification of sentinel diagnoses, calculation of long-term outcomes and generation of risk-adjusted ACHD outcomes and visualisations. It is hoped that this additional research will benefit the provision of health care by: This research is expected to generate a powerful set of tools to help all clinical teams involved in the management of patients with CHD to monitor and improve their services. It is also expected that this UCL research will enable patients to explore the quality of their services and better understand their conditions. Patients and families want greater transparency and a broader range of information than currently available. By developing new, more relevant outcome measures, informed by and co-created with patients and families, UCL hope to address these concerns and improve the information available for all. The researchers are expected to generate the high-quality evidence necessary to guide the quality improvement of CHD services and inform decisions about national policies. Scientific manuscripts will be written detailing the findings. UCL expect the publications to contribute to the evidence and expert opinion for the development and update of clinical guidelines in congenital heart disease in the future. The dissemination via publications and presentations will be completed following conclusion of the study in 2023. The timescale of the full benefit through changes in policies and in procedures is impossible to determine. There is evidence that better monitoring of the quality of services leads to improvement in services. CHAMPION (and LAUNCHES QI) both directly address the better monitoring of services for congenital heart services. In the main description of the study, UCL explain that the main objective is to develop meaningful outcomes for audit, and how they can be objectively and fairly measured and reported because the current reported outcome of 30-day survival is no longer appropriate on its own. The benefit is expected to come from national audit adopting these measures for use in their annual public reports and local hospitals adopting a subset as part of monthly local quality improvement – a high priority for national audit identified by NHS England after consultation with stakeholders in 2015 (and hence why Dept of Health commissioned this research).

Benefits reported

Yielded Benefits is not a requirement for new applications. Data provided for the LAUNCHES-QI study in September 2020, under Version 0 of this Data Sharing Agreement, was incomplete and will be re-supplied under this Agreement (version 1). The data previously disseminated will be securely destroyed. No yielded benfits have been achieved to date as a result of the incomplete data.