CPinBOSS Study - Cerebral Palsy in the British Orthopaedic Surgery Surveillance Study

The Cerebral Palsy in the British Orthopaedic Surgery Surveillance Study (CPinBOSS), funded by Action Medical Research (AMR), has been running since July 2019. This agreement requests HES data from NHS Digital for case ascertainment purposes. HES data is required to ensure that no potential cases that should be included in the study are missed. AMR will not have access to NHS Digital's Hospital Episode Statistics (HES) data/collected data from the study or specify any aims or objectives of the study. All people living with cerebral palsy, aged between 5-16 years, that have been referred for Single Event Multi Level Surgery (SEMLS) by either the treating clinicians or gait lab teams will be entered onto a database, forming the service evaluation aspect of the trial. The aim of the study is to undertake a national surveillance of this surgical activity encompassing all NHS Trusts in England. All of these patients and data are to come from participating NHS Trusts, which have appropriate Research & Development (R&D) and Sponsor approvals to participate within CPinBOSS. University of Oxford holds the data collected and only members of the CPinBOSS study team (all substantive employees of Oxford University) will have access to this database. The database will not be shared with any other institutions.

SEMLS is a surgical intervention that involves a minimum of two surgical procedures (bony or soft tissue) undertaken at a minimum of two different levels (e.g. hip and knee or thigh and calf) with the objectives to improve walking function within cerebral palsy patients. SEMLS has evolved over the past 30 years to replace repeated episodes of limited surgery. There are major differences between the Trusts that perform SEMLS in terms of patient selection and the choice of the specific surgical interventions. The primary objective of CPinBOSS is to identify the total number of patients that are eligible for SEMLS across all Hospitals in England, i.e. the incidence of children with cerebral palsy who fulfil the criteria for this type of surgery and to look at the variation in the surgeons’ criteria in selecting children for surgery by analysing the children’s clinical characteristics.

SEMLS operation would typically take a whole operating day of a surgical team. Therefore, any given surgical team is unlikely to be undertaking more than 4-5 such operations per month at the very maximum. The data collected from the recruiting NHS Trusts thus far support this estimate. With this in mind, it is highly unlikely that overwhelmingly large numbers will be involved. With the small numbers involved per centre it is expected that case ascertainment will be straightforward through comparison of HES and local PI submitted data.

This research was recently prioritised as one of the top-10 research priorities in children’s orthopaedics in a Delphi consensus amongst the British Society for Children’s Orthopaedic Surgery (BSCOS) members. The Delphi technique is a well-established approach to answering a research question through the identification of a consensus view across subject experts. This research was also prioritised in the top three topics at a recent James Lind Alliance Priorities Setting Partnership on paediatric lower limb surgery. The CP Cohort study is similar to the successful British Orthopaedic Surgery Surveillance Study (BOSS study), previously delivered in association with the BSCOS, confirming the feasibility of utilising this network for patient recruitment. The study design of CPinBOSS, including case ascertainment through HES data, followed the example of the BOSS study, which has now been completed successfully.

To undertake this national surveillance study, University of Oxford reached out to all NHS Hospitals nationally and invited participation by all clinical teams undertaking this type of surgery for children with cerebral palsy. A total of 24 NHS Trusts confirmed they undertake such surgery and were invited to recruit for the study. CPinBOSS has 19 NHS Trusts open for recruitment at the time of application (with a further 5 Trusts in current setup) across England. The recruitment period is from July 2019 – March 2021 with 2 years follow-up period. HES data is requested to provide case ascertainment, i.e. ensure that no cases are missed. It is particularly important for the case ascertainment to be national in order to capture cases undertaken in Trusts that have not formally declared this activity and are not recruiting for the study. This will enable the CPinBOSS research team to confidently report that CPinBOSS has captured national activity data with reasonable accuracy. If there are centres undertaking small volume activity and are not engaged with the clinical community, this would constitute important information for a national surveillance / service evaluation.

As the main objective of the study is to describe the incidence of the treatment across England , it is essential that University of Oxford do not miss any patients during the recruitment period. To ensure that University of Oxford have a full representative sample external sources will be used to check case ascertainment. Diagnostic CP codes will be searched within the Hospital Episode Statistics for England, the Patient Episode Database for Wales and the Scottish Morbidity Record and Cerebral Palsy Integrated Pathway Scotland (CPIPS) data (http://apcp.csp.org.uk/publications/cerebral-palsy-integratedpathway-scotland-cpips-dvd). The surgeon leads in hospitals with potential missed or duplicate cases will be contacted for confirmation. Only anonymised data will be reviewed.

HES data is requested for the recruitment period of the CPinBOSS study (July 2019 - Mar 2021) to ensure that University of Oxford have recruited all patients eligible for SEMLS. There are clear diagnostic (ICD) codes for diplegic cerebral palsy, these codes will be utilised to monitor hospital episodes throughout the study, which will be identified using the procedure codes provided.

The purpose of applying for HES data is to cross reference hospital admission for these patients with the information collected through the study. This will ensure that all cases of SEMLS are captured across England. In turn, this will contribute to the development of a comprehensive dataset which will reflect what is happening across the country with this patient population. Ensuring that University of Oxford capture all national activity in this clinical field and establishing the incidence of the condition is the central purpose of this study and cannot be achieved without the HES data. It is only with the cross referencing between data collected in the study with HES data that University of Oxford can ascertain complete case collection at national level.

University of Oxford are not seeking personal identifiable data, the request is for pseudonymised data. The data that is being sought will be utilised to identify NHS Trusts that have performed SEMLS over the specific recruitment period. The study team can then cross-reference with the recruitment statistics for CPinBOSS and identify if patients have not been included in the study. For example, the Oxford-based CPinBOSS Research group receive the HES data report indicating that in the first quarter of 2021 Hospital Trust X have undertaken 6 SEMLS operations. Trust X have provided the CPinBOSS research team data on 5 cases of SEMLS treated during the same period. The Study Coordinator of CPinBOSS contact the local PI at Trust X and inform them of the discrepancy in numbers, encouraging the PI to double check that no cases have been missed. No patient identifiable information is exchanged and only the total number of cases is discussed. There are then two potential outcomes: 1. The local PI at Trust X confirms that an extra case has been done but not included in the data provided to CPinBOSS. It is left with the local PI to encourage the local research team to upload the data of the additional case. The CPinBOSS team is no further involved with any chasing. No patient identifiable data are exchanged. 2. The local PI at Trust X confirms that NO extra case has been treated. The CPinBOSS team will then have no further involvement with this discrepancy. No patient identifiable data is exchanged.

Personal data will not be shared with the trusts in order to carry out this processing. Only counts will be shared.

All HES data will be held for 2 years post the end of the study (March 2025) so that research publications can be completed. It should be clarified here that a SEMLS operation would typically take a whole operating day of a surgical team. Therefore, any given surgical team is unlikely to be undertaking more than 4-5 such operations per month at the very maximum. The data collected from the recruiting NHS Trusts thus far support this estimate. With this in mind, it is highly unlikely that overwhelmingly large numbers will be involved. With the small numbers involved per centre it is expected that case ascertainment will be straightforward through comparison of HES and local PI submitted data.

University of Oxford and the CPinBOSS research team will conduct research following the General Data Protection Regulations (GDPR). Specifically, Article 6 1 (e) - Public Task and Article 9 2 (j) Archiving, Research and Statistics. Processing of the HES data is necessary for the research project to be conducted and will benefit the cerebral palsy population following dissemination of the peer reviewed published results, in line with Article 6 (e). In accordance with Article 9 (j) the research and statistical analysis is proportionate to the aim of the study and will respect the right to data protection and provide for suitable and specific measures to safeguard the fundamental rights and the interests of the data subject.

All data received from NHS Digital will be stored and managed by the University of Oxford (data processor and controller) and analysed by the trial statistician, a member of University of Oxford based at the Nuffield Department of Orthopaedics, Rheumatology and Musculoskeletal Sciences (NDORMS), on secure, password protected servers in a locked university building within NDORMS.

CPinBOSS is funded by Action Medical Research (https://action.org.uk/).

This data application is trying to answer a specific question, how many cerebral palsy patients have SEML surgery across England, and the wider project including the whole of the UK. The data sought relates to specific lower limb surgeries that have been performed on patients with cerebral palsy aged between 5-16 years old with Gross Motor Function Classification System (GMFCS) levels 1-3. The GMFCS system is a 5-level classification that differentiates children with cerebral palsy based on the child's current gross motor abilities, limitations in gross motor function, and need for assistive technology and wheeled mobility.

The data applied for are specific products relating to type of surgery (multiple ICD codes), NHS Trust of where the operations are performed and the date of the operations. Due to SEMLS comprising of multiple lower limb surgeries and that there are no specific ICD codes for SEMLS, the application includes a large range of operations. The statistician and the Chief Investigator, both based at University of Oxford will be able to identify if the patients have received the specific type of surgery (SEMLS) based on this minimal dataset.

The data applied for is for record level data of Hospital Episode Statistics Admitted Patient Care, which includes the following:

• Date of admission – To identify when the potential patients who underwent multi-level surgery can be easily identified by the local hospital/Trust.

• Primary & secondary diagnosis codes – To allow the research team to confirm the patients meet the inclusion criteria for this study and they have diplegic Cerebral Palsy. The research team will then be able to contact the local site team to enquire whether a patient was missed during the recruitment phase of CPinBOSS.

• Date of operation and status - To identify when the missed SEMLS patients were admitted so local site clinicians can identify these patients.

• Total number of procedures per episode – To allow the research team to confirm the patients actually underwent Multi-Level Surgery. For the purpose of the CPinBOSS study the research team have defined SEMLS as the intervention that involves a minimum of two surgical procedures (bony or soft tissue) undertaken at a minimum of two different levels (e.g. hip and knee or thigh and calf). This field will facilitate whether patients meet this definition so should be included within the study.

• Duration of the episode and the type – To allow the research team to confirm the patients are indeed undergoing SEMLS

• What hospital/Trust performed the operation – To identify which Trust/hospital the operations were performed for the research team to contact and chase for details. This will also allow for a total number of operations performed per site.

• Patient data including Age – To ensure patients meet the inclusion criteria for CPinBOSS (5 – 16 years old). Patients that fall outside of this age range will not be required within the CPinBOSS study.

• Ethnicity & Sex– To ensure that the patients that undergo SEMLS are representative of the general population and there is not selection bias based on specific ethnicity or sex.

• Socio-economic including Index of Multiple Deprivation – To ensure the patients that undergo SEMLS are representative of the general population and that there is not selection bias based on socio-economic status. This can be used to identify whether only a certain population is operated on.

Demographic data are requested to ascertain that the sample is representative of the general population and that patient selection is not biased by demographics, in other words that surgery is not predetermined/biased based on a specific ethnicities or sex of patients. The data request has been minimised and is targeted to a specific population that has lower limb surgery. This population includes children, aged 5-16 years old, with cerebral palsy (GMFCS levels 1-3) that had lower limb surgery, within England, over the course of the recruitment period of CPinBOSS (July 2019 to March 2021).

The data application applies for a one-way data flow (i.e. from NHS Digital to University of Oxford). All data received from NHS Digital will be stored and managed by the University of Oxford (data processor and controller – data stored on University of Oxford servers only and these are backed up daily) ) and analysed by the trial statistician, a member of University of Oxford based at the Nuffield Department of Orthopaedics, Rheumatology and Musculoskeletal Sciences (NDORMS), on secure, password protected servers in a locked university building within NDORMS. All members of the CPinBOSS team have appropriate Good Clinical Practice (GCP) certificates and are appropriately trained in data protection and confidentiality. The Chief Investigator (CI) will oversee the whole project and the statistical analysis ready for dissemination. The data will be held for 2 years post end of the study for publication purposes.

Once the data is processed and provided to the research team, the statistician will identify the sites that have performed SEMLS and the total number of patients operated on during the previous quarter. This will provide the research team with in-depth knowledge of total number of SEML surgeries conducted at different hospitals across the UK. The research team cross-check number of patients recruited by each site to the HES data to identify whether the CPinBOSS trial has captured all patients that have had SEMLS over the recruitment period. Without the HES data University of Oxford cannot ensure that a full national representation of the incidence of this condition is collected.

There will be no attempt made to reidentify the patients from the data supplied by NHS digital under this agreement.

If the figures are different between the numbers that the Trust has specified and the numbers coming from HES, it would highlight potential missingness to the designated individual Trust identifying patients for the study. Based on the minimal dataset provided by HES it is likely that the treating clinicians, would be able to identify if the case is truly missing, or if it is a coding error. If truly missing, then the CPinBOSS team will prompt the local Trust team to enter the case details. If a coding error, the Trust team would be asked to highlight this to the study team. At no time would the study team (other than the individual's treating clinicians) be able to identify the individual. This would allow for a more detailed dataset and able to identify all patients fulfilling the inclusion criteria at participating sites across the England. This will ensure that the study result offers a reliable estimate of eligible participants, and ensures that the dataset represents the true sample of participants - to ensure that the results are widely generalisable, without bias emerging through missingness.

Security Assurance from University of Oxford Information Governance Team:

• As described in the Data Security and Protection toolkit, a list of all systems for storing or processing NHS Digital Data are on the asset register which is monitored regularly by the Information Governance Manager and The Research Centre IT Manager. The University of Oxford statutes include regulations relating to the use of Information Technology systems.

• Appropriate information security controls are implemented to protect all IT facilities, technologies and services used to access, process and store University information. The IT security baseline consists of approximately 80 specific requirements covering the following domains: Access control, System Acquisition and Development, Change Management, Incident Management, Monitoring and Logging, Network Security, Operational Security and Vulnerability management.

• The data from NHS Digital will be processed and stored safely and securely as per the Data Security and Protection Toolkit. Oxford University has a comprehensive Information Security Policy and all members of NDORMs are expected to abide by the departmental Information Security Policies

The data from NHS Digital will not be stored on cloud solutions at any time.

The department has a comprehensive policy and procedure for reporting breaches

Data will be processed and stored within University of Oxford servers, which are backed up daily. No data is transferred off site at any time.

All outputs will be aggregated with small numbers suppressed in line with the HES analysis guide.

The primary output from this HES application is to double check and provide case ascertainment that University of Oxford have recruited all cerebral palsy patients (aged 5-16 years with GMFCS levels I - III) that have had SEML surgery over the recruitment period (Jul 2019 - Mar 2021). This will be analysed by the statistician and reported back to the trial management team. The level of data received from NHS Digital will be pseudonymised data which will reduce the risk of identifying patients.

The data will be written up into a high impact peer-reviewed research papers, such as the British Medical Journal or the Bone Joint Journal, whilst being presented at National and International conference following the completion of the study.

The aim will be to present the project at the British Society for Children's Orthopaedic Surgery and The European Paediatric Orthopaedic Society conferences. This will enable the research team to provide an update on the total number of SEML surgery performed over the recruitment period.

The data gained from CPinBOSS and HES will be utilised to inform and instruct policy, guidelines and frameworks on how patients with Cerebral Palsy are have their conditions managed across all Trusts within England. There is a current lack of coherence between Trusts and one of the major goals of this research is to inform policy, standardise treatment and care across the UK. This policy decision making will take place through the professional body BSCOS and its educational committee. BSCOS will then produce guidance/policy on best practice in this area by setting agreed standards for SEMLS treatment within the CP population that would emulate the practices of the centres that produce the best results. This study will also help to inform feasibility and design of future trials. However, should clinical trials not be feasible, this study will provide definitive prospective comparison of conventional vs minimally invasive SEMLS vs natural history that should have major impact on clinical practice.

Prior to the final scientific write ups of the trial, results will be communicated through the CPinBOSS website, aimed at paediatric orthopaedic surgeons treating children with cerebral palsy as well as paediatricians and physiotherapists. Newsletters will be published during the recruitment period and can be found here: https://www.ndorms.ox.ac.uk/clinical-trials/current-trials-and-studies/cpinboss-study. The results will also be presented at British Society for Children’s Orthopaedic Surgery conference in March 2021, along with further national and international meetings. The CPinBOSS study team will liaise with the parent/patient organisations which have been engaged throughout the conception and design of this study (STEPS, Action Cerebral Palsy) to produce lay summaries and infographics for the wide dissemination of the results to this group of children and their families.

The total number of SEML surgeries that are performed on a yearly basis in the UK with children with cerebral palsy is unknown and there has been no large scale data collection on this. CPinBOSS is aiming to answer this primary question through a service evaluation whilst recruiting all NHS Trusts that perform SEMLS across the UK. The primary purpose of the study is to describe the incidence of the condition through identifying all the children treated nationally over a period of 18 months. This aim cannot be achieved with data collection alone as this would rely only on individual sites/hospitals entering data voluntarily. This is likely to lead to selective reporting that would affect the reliability and credibility of the study. It is hoped that cross referencing the data collected directly from the hospitals by the research team to those recorded in HES would ensure that University of Oxford do not miss cases and that the collected data truly represent the English activity over 18 months. This data sharing agreement will allow the research team to identify whether patients are being missed and not enrolled in the CPinBOSS study. This would facilitate chasing sites to report any unreported cases.

CPinBOSS is primarily a service evaluation with the primary outcome of identifying the total number of cerebral palsy patients that undergo Single-Event Multi-Level Surgery across the UK. The CPinBOSS research team have opened 19 recruiting NHS Trusts across England to report the total number of operations performed on cerebral palsy patients that meet the study inclusion criteria. Within the service evaluation there is a nested consented cohort where patients are approached to consent to be identified by the CPinBOSS research team. The patients that are recruited for the cohort are approached and consented directly by the research nurses and local PIs at the hospitals. The Oxford research team are not involved in the consent process. The patients are seen at either the gait lab or in clinic and then are approached by the local PI/research nurses and invited to consent to CPinBOSS cohort. This is different from the dataset that will be provided by the DARS team. There will be no transfer of personal identifiable data to/from the local hospitals or NHS Digital. It will just be the total number of operations performed over a specific period at that NHS Trust/hospital which will be discussed between the local site and the CPinBOSS research team.

CPinBOSS is not altering the care of the patients at the local NHS Trust as the only difference between the consented cohort and the service evaluation is that parents/guardians and children complete a consent/assent form, baseline patient reported outcome measures (PROMS), and follow-up PROMS at 1 and 2 years. It is hoped that this will allow the research team to identify whether the patients have improved their walking and mobility functions post-surgery and whether they are happy with the overall treatment and care they received. The patients that are identified via HES data will be included within the service evaluation and not be approached for the consented cohort. This is because the research team would not have received baseline PROMS prior to surgery.

The secondary objectives are to review the regional variation in total number of operations and to describe the clinical indication and surgeon decision making in the surgical management of these patients. It is hoped this will enable the research team to answer these objectively and provide recommendations for best practice, agreed standards and prioritizing future research.

It is hoped that publishing papers in high impact peer-reviewed journals will increase the evidence base around SEMLS within the CP population. There is little known about the management of these patients across England and it is hoped that this research project and the anticipated published results will help to increase knowledge and promote future research within this field.

CPinBOSS also aims to instruct policy and to produce agreed frameworks on the management of all CP patients across the UK that have been referred for SEMLS. Currently, there are no agreed standards in the care and management of these patients. This is why this HES application is so essential for the research team to achieve its goals in providing agreed policies that would significantly benefit the patients and to standardise practice within the UK. It is hoped this will result in a more agreed and suitable patient management pathway so that patients will receive the best possible treatment with the most up-to-date research supporting this.

Yielded Benefits is not a requirement for new applications.