NHS Digital Data Release Register - reformatted

LA-SER Europe Ltd

Project 1 — DARS-NIC-301834-K0S2Y

Opt outs honoured: No - data flow is not identifiable (Does not include the flow of confidential data)

Sensitive: Non Sensitive, and Sensitive

When: 2020/11 — 2020/11.

Repeats: One-Off

Legal basis: Health and Social Care Act 2012 – s261(1) and s261(2)(b)(ii)

Categories: Anonymised - ICO code compliant

Datasets:

  • Hospital Episode Statistics Outpatients
  • Hospital Episode Statistics Accident and Emergency
  • Hospital Episode Statistics Admitted Patient Care
  • Hospital Episode Statistics Critical Care
  • HES:Civil Registration (Deaths) bridge
  • Civil Registration - Deaths

Objectives:

Analytical Laser (trade name of LA-SER Europe Ltd, which is a Certara company) are requesting pseudonymised Hospital Episode Statistics (HES) datasets linked to mortality data for use in its study ‘’Clinical characteristics of adult haematological cancer patients with veno-occlusive disease, VOD, and their health resource utilisation and cost burden in England’’. Analytical Laser is the sole Data Controller who also process data. Veno-occlusive disease (VOD) is a rare but serious disorder of the liver. It has been suggested that chemo-radiotherapy is the primary cause of VOD and its occurrence may prevent patients from receiving a haematopoietic stem cell transplant (HSCT) or pose as one of the serious complications of HSCT. VOD can be more common among patients with haematological cancers than in the general population, although the reported incidence rates vary depending on the study conducted, and mainly reflects post-HSCT occurrence. VOD has a detrimental impact on patients' quality of life and significant economic burden to the health system. The disease can be potentially fatal, especially among patients who develop the severe form of the disease. VOD patients need intensive and expensive treatment, and occasionally organ support may be required. The use of HSCT as an effective treatment option in the care of haematological cancer have been growing over the last decade. The British Society of Blood and Marrow Transplantation has reported 3318 first HSCT in 2012 which increased by 23% to 4077 in 2017, an average 4% annual increase. There is currently a lack of data on haematological cancer patients, especially those with leukaemia, who developed VOD in England. This study will assist in obtaining valuable information in order to understand this patient group which has seldom been studied in England. It also falls within the bounds of recital 157 of the GDPR as the information generated from this study will be 'new knowledge of great value' to haematological cancer, which is a widespread disease (with a 5-year prevalence of 388.8 per 100,000 as reported by the UK's Haematological Malignancy Research Network). As VOD is a rare disease, a large data set is required to improve the likelihood of detecting its occurrence. The Hospital Episodes Statistics (HES) datasets being large, national data sources offer this unique opportunity which will otherwise be challenging using smaller and less representative cohorts, especially for the haematological cancer population in general and a subgroup of leukaemia patients where this disease is suspected to be more common. These HES datasets (Admitted care, Outpatient, Accident and Emergency, and Critical Care datasets), will also have more recent data on patients diagnosed with haematological cancers. Analytical Laser is a research consultancy organisation which proposed to undertake this study with an aim to improve the overall understanding of haematological cancer and VOD. Analytica Laser is the trade name of LA-SER Europe Ltd, which is a Certara company. LA-SER Europe Ltd also uses the term ‘Certara – Evidence & Access’ or in short ‘Certara EVA’ to describe itself. The data is being processed by Analytica Laser, LA-SER Europe Ltd, at its main office in London and at no other part of Certara. Analytica Laser, LA-SER Europe Ltd, is determining the purposes and how the data should be processed and as such is the data controller. All control for utilising the data from NHS Digital for the purpose of the study will remain within Analytica Laser. Therefore, all data controllership reside within Analytica Laser and no other companies of Certara. The funder making this study possible is Amgen, which include payment for the fees incurred by Analytica LASER and for all advisory inputs from experts (including an Haematological Cancer Consultant with extensive experience in veno-occlusive disease, see below), as well as for the results dissemination. Amgen is supporting this study with the general aims to continue understanding haematological cancer and its evolving treatment paradigm in the real-world setting. This type of arrangement or setup is not unique to this agreement and quite common in the medical research world. Amgen is not playing any role in the conduct of the study and will not receive any data, except for a report with aggregated data (with small number suppressed as per HES data analysis guidelines) which will also be made available to the public. Amgen will not have any influence over the design, outputs or dissemination of the outputs nor the ability to suppress any outputs derived form this study. Amgen is one of the world’s leading biotechnology companies. Amgen is a values-based company, deeply rooted in science and innovation to transform new ideas and discoveries into medicines for patients with serious illnesses. Analytica Laser developed the objectives of the study and initiated the protocol. The advice of a clinical expert in Haematology with extensive experience in VOD was sought in order to understand how VOD is diagnosed and treated in routine clinical practice. Any information provided was used in conjunction with knowledge gained from other studies, especially on how to identify VOD cases. The opinion from a clinician on how to identify cases of interest is common in research to ensure that a study has high internal validity. Neither the Haematological Cancer Consultant involved in the study nor Amgen are controlling how the data should be processed. Should there be any contribution from their parts, Analytica Laser decides whether to consider them for inclusion. The data will be held and processed by Analytica Laser only and no other parties mentioned. Therefore, the haematological cancer consultant involved in the study will only view analysed, aggregated data (including summary statistics) and will not see any pseudonymised individual-level data that Analytica Laser will receive directly from NHS Digital. Record level data access is restricted to substantive Analytica Laser employees who are responsible for this data. Prior to sharing any aggregated information with the haematology cancer consultant, these will be reviewed in line with the HES analysis guidelines, including suppressing data on small numbers. The developed algorithm will be an adaptation and possibly a modification, if required, of the US algorithm. For developing the algorithm, only aggregated data with small numbers suppressed will be shared with the Haematological Cancer Consultant who will then provide their opinion on whether it closely resembles how VOD cases are identified in real world practices. Any adjustment of the algorithm, for example to consider any potential outliers, will be carried out by epidemiologists with clinical background within Analytica Laser. The Haematological Cancer Consultant with experience in VOD will play further advisory role in helping Analytica Laser interpret the findings of the study within a clinical context. The Consultant Haematologist will also assist Analytica Laser to disseminate the findings to peers and patients within the UK. To note, the Consultant Haematologist is not representing their NHS institution for the advisory role in this study and is assisting Analytica Laser with advice as a self-employed expert. There is no plan to involve any other experts for developing the algorithm to identify VOD. In routine practice, VOD is usually misdiagnosed or overlooked. In a data source like HES, the use of VOD’s ICD 10 codes (K76.5 and K71.8) to identify all patients with the disease will therefore be insufficient because some patients who potentially had the disease would not have received these codes within their HES records. For these reasons other criteria will be required to identify the disease, such as the disease’s typical symptomatology. As part of this study, Analytica Laser is developing an algorithm to identify VOD cases, by adapting the Baltimore and Seattle and other clinical criteria for the diagnosis of VOD (Carreras et al., Biol Blood Marrow Transplant. 2011. (11):1713-20; McDonald et al. Hepatology. 1984;4(1):116-22; ; Jones et al., Transplantation. 1987. 44(6):778-83). This study therefore provides an opportunity to construct an algorithm (based on an adaptation and possibly a modification, if required, of the US algorithm) to identify VOD and to describe the characteristics of haematological cancer who develop the disease. Such algorithms were previously applied using MarketScan and Medicare data in the USA by Zhang et al (Clin Ther. 2018; 40(10):1711-1719.) but not assessed in relation to or applied on the HES national datasets in the UK. Analytica Laser will apply and adjust the algorithm accordingly to HES data and other parameters to be decided during the conduct of the analysis. The Consultant Cancer Haematologist will advise Analytica Laser on the clinical validity of each variant of the algorithm. Only analysed and aggregated data, (with small numbers suppressed in line with HES analysis guide), generated from the algorithm will be seen by the Consultant Haematologist. Analytica Laser will make the final decision on which version of the algorithm should be adopted. Under article 6(1)(f) of the GDPR, Analytica Laser would process the requested data for the purposes of legitimate interests pursued by the controller. Analytica Laser has therefore completed a Legitmate Interest Assessment (LIA) to highlight the legitimate interest proposed under this project. The purpose being that the processing of this data will provide information, in aggregated form, required to describe the characteristics and the healthcare utilisation and cost burden of VOD in haematological cancer patients. This will improve the general understanding of the disease and, in turn, will improve healthcare provision for patients who undergo HSCT and other cancer treatment. Analytica Laser is providing consultancy and analytics for this study in order to generate results which will be 'new knowledge of great value' for the study of haematological cancer, which is a widespread disease (with a 5-year prevalence of 388.8 per 100,000 as reported by the UK's Haematological Malignancy Research Network). The study is necessary as without this study, the above will not be possible and the impact of VOD on the haematological cancer patients and on the NHS will not be known. Also, there will a continued lack of data for the UK to assist decision making to improve outcomes, especially as VOD patients are anecdotally described as requiring intense level of care and have a poor prognosis. Although there is a need to understand the occurrence of VOD among HSCT recipients, it is equally important to gather findings on its occurrence among patients who received other treatments, such as chemotherapy or radiotherapy, or as a result of disease progression. Importantly, VOD is rare and difficult to diagnose. From a large sample of haematological cancer patients, it will be possible to identify adequate number of patients with VOD, as it is a rare disease. The processing of this data falls under article 9(2)(j) so that research can be conducted to generate much required evidence to assist clinicians, patients and the pharmaceutical industry to better understand the disease and develop treatment strategies to manage haematological cancer patients. In order to achieve these goals, demographic and clinical data as recorded on the HES datasets will be required and analysed as per the study objectives. No identifiable data will be required for this study. Only substantive staff (epidemiologists, statisticians and data analysts) with training in data protection at Analytica Laser in London will have access to the data for the purpose of conducting the study. Although pseudonymised data is being received, the risk of identifying any patients will be removed by Analytica Laser by suppressing any analysis/findings on small numbers of individuals to prevent their potential identification in line with HES analysis guide. It was estimated that 194,000 patients will be identified using the ICD codes for haematological cancers. This estimation was carried out using the reported epidemiology data by the National Cancer Registration and Analysis Service Cancer Analysis System and further calculations. About 320 patients (and possibly more) with VOD will be expected. This estimation was based on the aggregated data reported in the NHS Digital’s annual HES Episodes updates for cases seen in NHS England and coded under ICD code K76.5 and K71.8, the number cases expected with HSCT reported in the Clinical Commissioning Policy: Haematopoietic Stem Cell Transplantation and other literature. VOD is potentially an under-diagnosed condition. Therefore, using the algorithm, a sizeable number of patients will be identified based on symptomatology and treatments, which will be assessed with the assistance of the clinical expert(s). Amgen is supporting this study with the general aims to continue understanding haematological cancer and its evolving treatment paradigm in the real-world setting. Venous Occlusive Disease is a rare but clinically serious side effect associated with high intensity chemotherapy and/or following HSCT in patients suffering acute lymphoblastic leukemia. Recent advances in treatments for those patients have shown to present with increased incidence of VOD and there is limited evidence on the most up to date intervention to manage this. In addition to the impact of patients, the economic burden of VOD can be significant but again no recent data are available to help guide clinical decision making taking into consideration all relevant outcomes. Although Amgen is a commercial entity engaged in the developing treatment for improving health outcomes in general, this study is not aimed at investigating any specific treatment. It should be noted that this study is neither investigating the performance of any of Amgen's products nor will be used as a pivotal study for any marketing authorisation applications. There is no commercial element at the core of this study as the funds from Amgen and the expertise from Analytica Laser and the Consultant Haematologist are making this study possible. This type of arrangement or setup is not unique to this application and quite common in the medical research world. Evidence to improve healthcare need studies like these and input from key players, in the absence of any public funds on researching rare diseases such as VOD. Amgen is a commercial entity that engages in the manufacturing of therapies for serious diseases in different therapy areas including haematological cancers. Hepatic Veno-occlusive disease (VOD) is a potential life threatening complication of hematopoietic cell transplantation (HSCT) and it is rare in routine practice. VOD is often misdiagnosed or overlooked but can have a large impact on outcomes and costs by potentially threatening the success of an expensive procedure such as HSCT, and there is a need to improve the understanding of the disease. Amgen has in its portfolio, products for management of haematological malignancies and has an interest in sponsoring studies that allow for appropriate estimation of the cost of managing care for haematological malignancies including those of complications such as VOD. The evidence generated will contribute to the understanding of cost effectiveness and budget impact for hematological malignancies therapies and thus provide scientific evidence for Health Technology Assessments of these therapies by agencies such as NICE (National Institute for Health and Care Excellence). This study in particular is not aimed to investigate any specific treatment that Amgen and/or its competitors commercialise. The funder does not have any influence on the outcomes of the study and therefore will not suppress any of the findings. The data will be received from NHS Digital's by secure electronic file transfer onto a fully encrypted (AES 256 bit or equivalent) laptop which will then become a stand alone device which will be used for the analysis in the London office of Analytica Laser. A copy of this data will be backed up on an encrypted hard drive. All devices with data will be locked in a fireproof safe in the London office, which is in a security managed building with a receptionist at the entrance who checks all visitors. There are CCTV surveillance throughout the day in the building. Primary Objective of the study: 1. To describe the demographic and clinical characteristics of haematological cancer patients who developed VOD by applying an algorithm using ICD-10 codes, adapted Baltimore and Seattle and other clinical criteria for diagnosis of VOD among haematological cancers Secondary Objectives: 2. To describe the mortality rate among haematological cancer patients with VOD 3. To describe the resource utilisation and cost burden for haematological cancer patients who developed VOD In order to achieve these objectives, HES data for the whole of England on haematological cancer will be required. Being a rare disease and is not clustered within any specific regions, the chance of identifying VOD is improved with large, representative data source. This will also offer country-level estimates, with high external validity, and the large sample size will improve precision in the estimates. The patients will be selected based on the following criteria: • Adult (18 years old and over) patients with at least one ICD 10 code (C81.0 - 96.9; D45-47.9) for haematological cancer identified from the inpatient and outpatients datasets (irrespective of diagnosis date) between 01/07/2011 and 31/12/2018 AND using the following algorithm which will be optimised to HES data: • Developed VOD for the first time between 01/07/2011 and 31/12/2018 (index date) which will be defined (by adapting the Baltimore and Seattle Criteria for diagnosis of VOD) [Carreras et al., 2011; McDonald et al., 1984; Jones et al., 1987; Zhang et al., 2018] as: Having ICD-10 codes: K76.5 Hepatic veno-occlusive disease, K71.8 Toxic liver disease with other disorders of liver (includes among others VOD too) • OR, a prescription of defibrotide • OR, had a least 2 symptoms of VOD which include abnormal weight gain, ascites, hyperbilirubinemia, or hepatomegaly within 60 days post HSCT admission (adapting the Baltimore and Seattle Criteria for diagnosis of VOD)[Carreras et al., 2011; McDonald et al., 1984; Jones et al., 1987] • OR Multi-organ failure with liver failure within 60 days post HSCT admission AND • Had no evidence of chronic graft-versus-host disease (GVHD) any time before or 100 days post index date (for patients identified using HSCT and VOD symptoms or MOF with liver failure), OR no evidence of acute GVHD or total parenteral nutrition (TPN) cholestasis during 60 days before or 100 days post index date. In order to define VOD, data within the variables which contain diagnosis and procedural codes as well as the dates these were recorded will be required. The period of observation for this study will span from 1st January 2011 until the latest date available on the data sets. This timeframe was chosen in order to obtain data on contemporary practices, especially with regards to the use of latest antineoplastic treatment, including HSCT, and also to capture patients who are more recently diagnosed with VOD. More recent data will provide up-to-date findings which will be more useful for decision making. Each patient will have a least 6 months of baseline data for identifying their comorbid profile. The maximum follow up time for a patient can be 7.5 years (or up to the latest data release). The data which will be accessed by substantive staff of Analytica Laser based in London for executing this study only. The first step will include data management where the data will be checked for quality and completeness and the variables and its data points will be labelled accordingly The data points which are essential for this project are broadly provided below: • Demographic data including age and gender • Clinical data on primary and secondary diagnoses during inpatient and outpatient care, procedures and medication used as cancer treatment (including HCT) and other related conditions, hospital, A&E, outpatient and critical care episodes, health resource group, mortality data, and date of these events or episodes The datasets requested and rationale for their use are: Admitted care dataset • demographic details to describe the patients' age and gender • primary and secondary reasons for admission from which patients with haematological cancer and VOD will be identified. These will also provide comorbidity data for describing the patients profile • number and duration of hospital stays and health resource groups for health resource utilisation description and cost burden calculations • treatment (procedure) data for describing care received Outpatient dataset - contains the data points as described under admitted care above but applicable to outpatient visits only, which will be utilised for the same reasons above Accident and Emergency dataset - contains data on emergency care visits which will used to identify health resource utilisation and cost burden Critical care dataset - contains data on reasons for critical care which will be utilised in definition of VOD and to identify severity of certain admissions, health resource utilisation and a cost burden Mortality data - contains cause and date of death to be utilised to calculate mortality rate and also for estimating total follow up time and time at risk in rate calculations. There are no immediate or cost-effective alternative approach that can be employed to recruit as many patients with haematological cancer over a 10-year period and among whom the rare disease VOD can be identified. To do so will require a complex research infrastructure with a large team of researchers and data collectors across the country. Generally, large funds are unavailable or difficult to source to conduct such longitudinal studies for rare diseases. The HES data source provides a unique opportunity to obtain valuable evidence on VOD as it already holds the data which are required to meet the objectives of the study. Therefore, it use will enable the evidence to become available sooner to improve current knowledge on VOD and subsequently improve its detection (in the wider haematological cancer patients and not only among those who had HSCT) and its management to improve outcomes (with more accurate and earlier detection leading to more appropriate and timely treatment being provided).

Expected Benefits:

The study is being conducted to benefit all those who are involved in the care of haematological cancer and VOD, and mostly patients who will benefit from its improved detection or improved care with better knowledge that this study will generate. Analytica Laser will not send any data with identifiers to NHS Digital for linkages. Instead, a set of selection criteria based on International Classification of Disease (ICD) codes will be provided to NHS Digital in order to identify haematological cancer cases and a further set of such codes will be required to define VOD cases in this group of patients. Analytica Laser does not require any identifiers, such as names, NHS numbers or addresses, for this study. Instead, each study subject will be provided an encrypted code (encrypted HESID) by NHS Digital. As the cohort will be formed by NHS Digital, Analytica Laser will only receive a pseudonymised version of this cohort. As this study does not require any identifiable data, the risk of intrusion or personal detriment is low, whilst all appropriate steps would have been taken to minimise the risk of patients being re-identified from the study. The purpose of this study is to describe the clinical characteristics of haematological cancer patients who developed VOD. This study also seeks to estimate the healthcare resources utilisation (HRU) and disease cost burden to healthcare providers for caring for haematological cancer patients with VOD in England. A subgroup analysis will be carried for leukaemia patients, who are known to be more at risk of VOD. The evidence that this study seeks to generate for the UK is currently unavailable in the public domain. Therefore, this study will provide useful background information on VOD in haematological cancer to anyone with an interest in this disease area. Therefore, the overall legitimate interest in processing this data would be to make available findings that will improve general understanding of haematological cancer and VOD. The evidence that this study seeks to generate is currently unavailable in the public domain, and thus this study will provide vital background information to anyone with in interest in the care of patients with haematological cancer, including clinicians, drug developers and regulators (e.g. European Medicine Agency), Payers (e.g. NHS organisations) or Health Technology Assessors (e.g NICE), and health care organisation (for planning resource allocation). For clinicians, this information will increase their awareness of VOD among haematological cancer patients as a whole, and not only limited to those who received HSCT. Generally, monitoring for VOD may be carried out more closely for recipients of HSCT who are reported at increased risk of the disease. As there is limited or no data for its occurrence among non-HSCT patients in the UK, it is challenging to understand if the disease is under-ascertained in this group of patients. By describing who developed VOD, especially by the type of cancer treatment, this will provide valuable information to clinicians and others of its occurrence among HSCT and non-HSCT recipients. The published results will increase clinician's awareness on being more cautious in detecting this disease in the general haematology cancer population, and not specifically among those who had HSCT. In turn, it will also encourage more active monitoring for VOD symptoms irrespective of HSCT status. Healthcare practitioners such as oncologists, hepatologists, haematologists, HCT surgeons and will be benefit from this studythe findings because it will reduce bias to diagnose the disease, and thus will increase the likelihood of correctly diagnosing VOD in all patients affected by the condition. At national level for the health and social care system this evidence can be utilised to inform policy makers on the impact of VOD on the haematological cancer population and to allocate appropriate resources for its prevention as well as planning its effective management. This will lead to improvement in health resource utilization and reduction of overall costs due to early detection and management of the condition. Public health bodies (e.g. NHS and NICE) can utilise the scientific findings generated by this study will include also other stakeholders involved in the management of haematological cancer patients and VOD outside England. For the pharmaceutical sector, estimates of the disease burden and description of the characteristics of patients with VOD are important. In order to develop any treatments in haematological cancer in general, it is important to understand the disease area well and what other conditions or diseases may develop as a result cancer progression or through receiving standard of care. Haematology cancer patients will benefit as this study will increase general awareness about VOD and its impact on them and the healthcare system in England. This will lead to more active monitoring and improving its identification, which in turn will prompt for more immediate and appropriate treatment, thus improving patients’ outcomes. These outcomes will be measured in terms of disease progression, mortality and health resource utilisation. Studies in the futures will need to be conducted to estimate the impact of the findings on improving awareness for more accurate and early detection of VOD which can be measured in terms of reduced progression to its severe form, decrease in mortality rate and reduction of health resource utilisation. Currently, VOD may be misdiagnosed for other complications which lead to the wrong treatment being provided and therefore worsening outcomes, including disease progressing to an advance stage and death. The benefits include: 1. General awareness among the clinicians to detect the disease and provide the right treatment to the patients, thus improving their outcomes and preventing death. 2. Highlight the occurrence of VOD, especially among non-HSCT patients, and therefore reducing a bias to diagnose this condition in HSCT patients most, thus allowing patients to be identified for the correct treatment and reducing their mortality rate. 3. Resources will be promptly directed to the treatment of VOD and thus improving the patients outcomes A Consultant Haematologist, has echoed the above benefits to the patients and to their haematogical care in general. Under article 6(1)(f) of the GDPR, there is a legitimate interest to conduct this study to provide information that will improve our current understanding of the disease. This also falls within the bounds of recital 157 of the GDPR as the information generated from this study will be 'new knowledge of great value' to benefit patients with haematological cancer and VOD as a whole.

Outputs:

The outputs will be generated using statistical software. All outputs will be in the form of aggregated data, with results on small number of patients being suppressed to prevent their likely identification in line with HES guidelines. A literature review conducted by Coppell et al (Biol Blood Marrow Transplant. 2010; 16(2): 157) found an incidence rate of VOD of 13.7% among 24,920 HSCT recipients. Some of the studies reviewed by these authors reported an incidence rate of up to 40%. Although regarded as a rare disease, HSCT recipients are particularly at risk. Morbidity and mortality are high among VOD sufferers. Zhang et al (Clin Ther. 2018; 40(10):1711) estimated that the mortality rate for patients with B-cell acute lymphoblastic leukaemia who developed severe VOD was 37.0%. Coppell et al (2010) reported an estimated mortality rate of 84.3% among severe VOD sufferers. The utilisation of health resources and cost burden is also high among VOD sufferers despite being rare. A study conducted in the US estimated the mean length of stay for patients who developed severe and non-severe VOD was 49.2 and 28.5 days per patient, respectively, and their overall medical cost (inpatient, outpatient and medical) within 100 days of post diagnosis was $86,953 and $22,047, respectively (Zhang et al., 2018). Sven et al (Bone Marrow Transplant. 2012 May;47(5):706) found an increased costs after allogeneic HSCT for patients who developed VOD, especially during the first treatment it increased significantly compared to the other patients in their study cohort. Generally, patients who develop severe VOD also require organ support in intensive care settings. Quality of life is also a major factor for patients following HSCT and this worsens rapidly among those who develop VOD. However, there is also a risk of developing VOD among patients who received chemotherapy and radiotherapy. There is currently a tendency to assess patients for its risk of developing and identification among HSCT recipients. However, this study is also seeking to describe the occurrence of VOD among patients who received chemotherapy and radiotherapy. If the disease can be identified among these other treatment categories, more awareness and advice (via various routes of scientific dissemination) will be provided to the healthcare practitioners for higher vigilance of its detection, which can also bring measures to prevent the disease becoming more severe. There is no estimate available on how many haematological cancer patients developed VOD in England using a large, representative cohort, especially among those who were treated with HSCT, chemotherapy or radiotherapy. VOD patients’ resource utilisation (including intensive care) and their associated cost have not been described for England. Without this evidence it will not be possible to support current assumptions on its occurrence, burden and cost for VOD care. This study is seeking to address this gap in evidence. The findings from the study will also provide a benchmark against which future assessment of its occurrence (especially in its severe form), resource utilisation and cost can be made. This study will be reported to all contributors by end of 2020 and aimed for submission for publication in a peer reviewed journal by Autumn 2020. A high impact journal (including the British Journal of Haematology, European Journal of Haematology, or Lancet Haematology) which will be accessible by clinicians based in the UK and abroad will be chosen. An abstract will also be submitted in Autumn 2020 for the American Society of Haematology Conference to be held Dec 2020 and/or also in Dec 2020/Jan 2021 for the British Haematological Society (BSH) Conference which will be held in March/April 2021 to increase exposure to the finding to clinicians and patients' group in the UK and worldwide. Both these conferences, especially BSH, are attended by clinicians (including haematologists), researchers, charities and patients groups, and will provide maximum exposure of the findings to those involved in caring for patients with haematological cancer and VOD in England. Abstracts presented at these conferences can be accessed for free by members of the public through their websites. To increase further awareness of its publication and findings, a press release will be made. Charities on Haematological Cancer, such as Leukaemia Care in the UK, will be informed of the finding and how the publication can be accessed. A clinical expert in haematology care, will provide his invaluable assistance in disseminating the findings at scientific conferences and to his peers via direct contact and scientific paper which Analytica Laser will aim to publish. The funder, Amgen, will also receive a report with the aggregated findings of this study. All publications and communications of the study finding will be in aggregated form with small numbers suppressed in line with the HES Analysis Guide. Therefore, no individual level data will be presented in any publication. Results to be reported will be in the form of numbers, percentages or proportions, means, median, interquartile range, overall range and other statistics such relative risk.

Processing:

The processing activities will consist of: 1. Analytica Laser will provide the relevant information to NHS Digital (consisting of selection criteria utilising ICD 10 codes and procedural codes) on how to identify the population of interest, and no other information such as person identifiers will be required for this purpose and in the data requested. 2. For data minimisation the data will only be limited to the fields required for the study which will be obtained • from the inpatient, outpatient, A&E, Critical Care and mortality datasets for • on adults (18 years and above) with a diagnosis code for haematological cancer patients will be identified • using psuedonymised data on demographic of the patients, diagnoses and procedures which are relevant in haematological cancer and VOD, and the time that these events occurred will be used to describe the basic characteristics of the cohort and to be used to construct an algorithm for VOD identification will be required. The algorithm (which will be an adaptation and possibly a modification, if required, of the US algorithm) will be constructed using the information as set out below: Adult (18 years old and over) patients with at least one ICD 10 code (C81.0 - 96.9; D45-47.9) for haematological cancer identified from the inpatient and outpatients datasets (irrespective of diagnosis date) between 01/07/2011 and 31/12/2019 AND using the following algorithm which will be optimised to HES data: - Developed VOD for the first time between 01/07/2011 and 31/12/2019 (index date) which will be defined (by adapting the Baltimore and Seattle Criteria for diagnosis of VOD) [Carreras et al., 2011; McDonald et al., 1984; Jones et al., 1987; Zhang et al., 2018] as: - Having ICD-10 codes: K76.5 Hepatic veno-occlusive disease, K71.8 Toxic liver disease with other disorders of liver (includes among others VOD too) - OR, a prescription of defibrotide - OR, had a least 2 symptoms of VOD which include abnormal weight gain, ascites, hyperbilirubinemia, or hepatomegaly within 60 days post HSCT admission (adapting the Baltimore and Seattle Criteria for diagnosis of VOD)[Carreras et al., 2011; McDonald et al., 1984; Jones et al., 1987] - OR Multi-organ failure with liver failure within 60 days post HSCT admission AND - Had no evidence of chronic graft-versus-host disease (GVHD) any time before or 100 days post index date (for patients identified using HSCT and VOD symptoms or MOF with liver failure), OR no evidence of acute GVHD or total parenteral nutrition (TPN) cholestasis during 60 days before or 100 days post index date. • The date range provided for the study period can be updated to the last date that data is available from NHS Digital as the study was designed in the previous year based on the data which was available at that time. The study will benefit with including more recent data for the following reasons: 1. VOD is a rare disease and extending the overall sample size will improve the probability of detecting more VOD cases, 2. Increase follow up time to describe treatment received, health resource utilisation, cost and mortality rate, and 3. To provide more up-to-date evidence based on latest practices in England. • The data that NHS Digital will provide Analytica Laser will include records without any identifiers on admissions (inpatients and critical care) or on outpatients’ appointments for patients with haematological cancer or VOD (or the conditions that describes VOD) as identified with the ICD 10 codes and procedural codes as provided above. Any records of haematological cancer patients within the time specified (or to the latest date available) and their relevant data fields on admission or outpatient visit date (month and year) would be required by Analytica Laser to construct the algorithm described above and describe their characteristics and treatment received. The data fields will include reasons for admissions/outpatient appointments in the form of ICD 10 codes and procedures conducted. Their approximate age will be determined from the variable with part date of birth (month and year) and their comorbidity indices from the ICD 10 codes for their associated conditions that they have sought care for. The procedural codes will also assist determining what treatment they have received as surgery, chemotherapy and radiotherapy for their cancer and HSCT. • The developed algorithm will be an adaptation and possibly a modification, if required, of the US algorithm. The algorithm to identify VOD will be constructed and adapted to data for England. It will be validated utilising aggregated data for its accuracy to identify VOD. Clinical inputs from an expert clinician will be sought to refine and select the most specific algorithm; several potential algorithms may be created and the best one adopted, as decided by Analytica Laser. The algorithm is a paper trail and as part of the publication of the study results, it will be made available for other researchers and clinicians. • As the clinically validated algorithm cannot be applied by NHS Digital to test its applicability and fine tuned, Analytica Laser will carry this out until a final algorithm is obtained. Once finalised, the algorithm will be applied to their overall datasets to define the VOD and non-VOD cohorts. • Once these sub-cohorts’ cases are identified, their fields relevant to resources utilised and cost for inpatient, outpatients, A&E and critical care and mortality will be analysed to meet the objectives of the study. 3. For this study, NHS Digital will provide a unique pseudonymised code (unique identifier) for each person on their datasets so that Analytica Laser can utilise this to link this person's records across these datasets. 4. These psuedonymised datasets will be received from NHS Digital on a secure, stand-alone encrypted device for processing and only accessed by designated staff (statisticians, data analysts and epidemiologists who are trained in data protection) at Analytica Laser based in London. 5. Access is limited to substantive employees of Analytica Laser who will be physically located in the London office and have received training in data protection and confidentiality. 6. No access to the data is possible by the funder (Amgen) of this study or the Consultant Haematologist (Clinical Advisor). 7. The designated staff at Analytica Laser will initially assess the data received from NHS Digital for quality and completeness and will liaise with the data release team for any issues. 8. Once the above process is completed, the analysis will be conducted to achieve the study objectives 9. Results that will be generated will be in aggregated form, with results on small number of patients being suppressed to prevent their potential identification, in line with HES analysis guide. 10. The results generated will be initially reported to clinicians with an interest in the field of haematology and VOD care and the funder of the study, Amgen, in aggregated form (no individual level data will be reported). 11. Once reviews from scientific peers and all contributors are obtained, these will be addressed and then reported back with an aim to prepare for abstract presentation at conferences and subsequently into a journal article publication 12. Abstracts will be presented to the conferences without any individual level data or identifiers 13. A draft manuscript without any individual level data or identifiers will be sent to a journal for their review and publication 14. Once accepted for publication the study findings will be disseminated into the public domain without any individual level data or identifiers. 15. Anyone in the public domain will be able to access the article through a journal which provide open access and relevant charities and patient group will be made aware of the findings so that they can reach patients who will benefit from this information. This is a retrospective observational study that involves secondary use of pseudonymised data collected during the course of routine care. No identifiers are required and there will be no backward tracing of any individuals, even if there is any remote possibility of doing so. Training is in place to educate substantive staff on how to safeguard data. In addition to suppressing results on small number of individuals as per HES analysis guide. All statistical analyses will be conducted using a statistical software. The analyses will be focused on describing the characteristics of patients with overall haematological cancer as well as the resource utilisation and cost burden associated with VOD. The same analyses will be performed also for the subgroup of leukaemia patients. The feasibility of identifying VOD among haematological cancer patients in the HES datasets will be assessed by describing number of patients identified using each criteria of the algorithm separately and by providing an overall number of VOD patients applying an algorithm in its entirety. The outcomes, demographic and clinical characteristics will be summarised using descriptive statistics for both primary and secondary objectives. Categorical variables will be summarised by number of observations and percentages (%). Continuous variables will be summarised using: mean, with standard deviation; minimum, first quartile, median, third quartile, and maximum. The rate of each study outcome will be calculated overall and separately for patients with haematological cancers. Frequency of variables reflecting resource use following development of VOD , such as hospitalisation, both elective and emergency, length of stay, outpatient visits at hospitals, treatment received, associated cost and death rate, will be estimated. Analytica Laser sought advice from a Consultant Haematologist with extensive knowledge in haematological cancer and VOD in order to understand how it is diagnosed and treated in routine clinical practice including complications that may be experienced through disease progression and treatments. The Consultant Haematologist will also provide clinical interpretations of the derived algorithms and the decision to employ the best algorithm will carried by solely Analytica Laser based on the clinical advice, information from the published medical literature and the data analysed so far for that purpose. As an advisor, the Consultant Haematologist will be assisting Analytica Laser interpret and disseminate the results of the study in a manner that is relevant to clinicians and patients, by paper or electronic-based communications, at scientific meetings, during interactions with relevant physicians and with patients groups throughout the UK. To note, the consultant haematologist did not play a role in the original conceptualisation of this study and in the definition of the data analytical methods. NHS Digital reminds all organisations party to this agreement of the need to 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 ie: employees, agents and contractors of the Data Recipient who may have access to that data).’