INSIGHT: The Health Data Research Hub for Eye Health
Lead Research Organisation:
University Hospitals Birmingham NHS Foundation Trust
Department Name: UNLISTED
Abstract
"There is an increasing demand for ‘Big Data’ to enable innovation in health technologies and improve patient care. INSIGHT-DIH creates an NHS-academic-industry partnership to make anonymised datasets accessible and provide critical services that respond to user needs (patients/industry/other), including ‘real-world evidence’ and ‘smart’ clinical trials for new drugs or devices, and accelerating AI-algorithm development for the NHS and beyond.
INSIGHT-DIH is focused on eye disease and its application to wider health, including diabetes C19and dementia. INSIGHT-DIH turns routine eye imaging - currently >25million/yr (UK)- into an exceptional asset for innovation as a convenient, uniquely valuable health snapshot.
INSIGHT-DIH uses anonymised large-scale data and advanced analytics to bring new clinical insights: from detection, diagnosis and referral, to new treatments and personalised healthcare. INSIGHT-DIH unlocks this opportunity by pairing unique anonymised data assets to hubs of expertise where patients and clinicians can, with industry, more efficientlyexplore the safety and use of new AI systems. This will accelerate the pace with which AI is able to translate to real patient benefit for the NHS and accelerate innovation across all sectors."
INSIGHT-DIH is focused on eye disease and its application to wider health, including diabetes C19and dementia. INSIGHT-DIH turns routine eye imaging - currently >25million/yr (UK)- into an exceptional asset for innovation as a convenient, uniquely valuable health snapshot.
INSIGHT-DIH uses anonymised large-scale data and advanced analytics to bring new clinical insights: from detection, diagnosis and referral, to new treatments and personalised healthcare. INSIGHT-DIH unlocks this opportunity by pairing unique anonymised data assets to hubs of expertise where patients and clinicians can, with industry, more efficientlyexplore the safety and use of new AI systems. This will accelerate the pace with which AI is able to translate to real patient benefit for the NHS and accelerate innovation across all sectors."
Technical Summary
"T he INSIGHT Digital Innovation Hub (INSIGHT-DIH) focuses on Eye Disease and ‘Oculomics’ , a first-of-its-kind linkage of routinely collected, anonymised high-dimensional eye-imaging data from the major ocular diseases impacting blindness to wider health including diabetes, dementia, cardiovascular disease and inflammation/immunity . INSIGHT-DIH partners NHS (University Hospitals Birmingham NHSFT, Moorfields Eye Hospital NHSFT), charities (Action Against Age-related Macular Degeneration) and industry (Roche, Google), to enable innovation and transform patient care.
Eye/vision problems affect 1.3 billion people worldwide but is highly amenable to digital innovation. In particular, ocular imaging can be used to provide novel insight into both eye and systemic health (“oculomics”), from childhood to old age. The INSIGHT-DIH network gathers data from across the whole of the UK health sector - NHS clinics, imaging diagnostics, laboratories, primary care/optometry – which will be linked and anonymised within the DIH to build the first holistic data-record combining routine systemic and eye health data with unparalleled high-order retinal imaging centred on patient benefit.
This positions the INSIGHT-DIH as an ideal exemplar for making the NHS ‘AI-ready’ . Eyecare data is standardised, large-scale (10% of all NHS appointments; >25 million scans/yr), and with proven suitability for artificial intelligence (AI) health applications (e.g. automated retinal diagnosis).
INSIGHT-DIH will support innovation, making existing inaccessible datasets discoverable , and bringing scale and efficiency to dataset aggregation and curation of anonymised routinely collected data , utilising our existing UK network.
We will address the ‘implementation gap’, with NHS partners being accelerated in digital maturity to AI-ready work-flows.
Disease-focused image-linked datasets respond to NHS/industry needs and DIH priorities:
(1) diabetes and diabetic eye disease;
(2) age-related macular degeneration ;
(3) dementia;
(4) community cohort for AI-intervention development/evaluation to improve screening and access to care.
Datasets will be scaled towards national coverage.
Additional services include: rapid-response real world evidence (RWE); ‘smart clinical trial’ capability for pharmacological/device/AI-interventions; and ‘clinical/technical Faculties’ of domain experts.
INSIGHT-DIH is underpinned by public trust and co-development, with a further £1 million investment in public and patient involvement and engagement (PPIE) as part of this bid, and creation of a Data Trust (Open Data Institute; AAAMD). From first contact to project completion, we will ensure excellence in service, with a view to sustainable partnerships, value return to all parties and a transformation in patient care."
Eye/vision problems affect 1.3 billion people worldwide but is highly amenable to digital innovation. In particular, ocular imaging can be used to provide novel insight into both eye and systemic health (“oculomics”), from childhood to old age. The INSIGHT-DIH network gathers data from across the whole of the UK health sector - NHS clinics, imaging diagnostics, laboratories, primary care/optometry – which will be linked and anonymised within the DIH to build the first holistic data-record combining routine systemic and eye health data with unparalleled high-order retinal imaging centred on patient benefit.
This positions the INSIGHT-DIH as an ideal exemplar for making the NHS ‘AI-ready’ . Eyecare data is standardised, large-scale (10% of all NHS appointments; >25 million scans/yr), and with proven suitability for artificial intelligence (AI) health applications (e.g. automated retinal diagnosis).
INSIGHT-DIH will support innovation, making existing inaccessible datasets discoverable , and bringing scale and efficiency to dataset aggregation and curation of anonymised routinely collected data , utilising our existing UK network.
We will address the ‘implementation gap’, with NHS partners being accelerated in digital maturity to AI-ready work-flows.
Disease-focused image-linked datasets respond to NHS/industry needs and DIH priorities:
(1) diabetes and diabetic eye disease;
(2) age-related macular degeneration ;
(3) dementia;
(4) community cohort for AI-intervention development/evaluation to improve screening and access to care.
Datasets will be scaled towards national coverage.
Additional services include: rapid-response real world evidence (RWE); ‘smart clinical trial’ capability for pharmacological/device/AI-interventions; and ‘clinical/technical Faculties’ of domain experts.
INSIGHT-DIH is underpinned by public trust and co-development, with a further £1 million investment in public and patient involvement and engagement (PPIE) as part of this bid, and creation of a Data Trust (Open Data Institute; AAAMD). From first contact to project completion, we will ensure excellence in service, with a view to sustainable partnerships, value return to all parties and a transformation in patient care."
Publications
Zhou Y
(2023)
A foundation model for generalizable disease detection from retinal images.
in Nature
Yim J
(2020)
Predicting conversion to wet age-related macular degeneration using deep learning.
in Nature medicine
Wintergerst MWM
(2021)
Structural Endpoints and Outcome Measures in Uveitis.
in Ophthalmologica. Journal international d'ophtalmologie. International journal of ophthalmology. Zeitschrift fur Augenheilkunde
Wicks P
(2020)
Going on up to the SPIRIT in AI: will new reporting guidelines for clinical trials of AI interventions improve their rigour?
in BMC medicine
Wen D
(2022)
Characteristics of publicly available skin cancer image datasets: a systematic review
in The Lancet Digital Health
Wagner SK
(2023)
Association Between Retinal Features From Multimodal Imaging and Schizophrenia.
in JAMA psychiatry
Vasey B
(2022)
Reporting guideline for the early stage clinical evaluation of decision support systems driven by artificial intelligence: DECIDE-AI.
in BMJ (Clinical research ed.)
Vasey B
(2022)
Reporting guideline for the early-stage clinical evaluation of decision support systems driven by artificial intelligence: DECIDE-AI.
in Nature medicine
Unsworth H
(2022)
Building an evidence standards framework for artificial intelligence-enabled digital health technologies.
in The Lancet. Digital health
Tom E
(2020)
Protecting Data Privacy in the Age of AI-Enabled Ophthalmology.
in Translational vision science & technology
Thygesen JH
(2022)
COVID-19 trajectories among 57 million adults in England: a cohort study using electronic health records.
in The Lancet. Digital health
Thomas CN
(2022)
Emerging therapies and their delivery for treating age-related macular degeneration.
in British journal of pharmacology
Thaller M
(2021)
Negative impact of COVID-19 lockdown on papilloedema and idiopathic intracranial hypertension.
in Journal of neurology, neurosurgery, and psychiatry
Taylor-Phillips S
(2022)
UK National Screening Committee's approach to reviewing evidence on artificial intelligence in breast cancer screening.
in The Lancet. Digital health
Taylor M
(2021)
Raising the Bar for Randomized Trials Involving Artificial Intelligence: The SPIRIT-Artificial Intelligence and CONSORT-Artificial Intelligence Guidelines.
in The Journal of investigative dermatology
Tan L
(2023)
Associations of antidiabetic drugs with diabetic retinopathy in people with type 2 diabetes: an umbrella review and meta-analysis.
in Frontiers in endocrinology
Subramanian A
(2022)
Symptoms and risk factors for long COVID in non-hospitalized adults.
in Nature medicine
Subramanian A
(2022)
Angiotensin-converting enzyme inhibitors and risk of age-related macular degeneration in individuals with hypertension.
in British journal of clinical pharmacology
Sounderajah V
(2021)
Developing a reporting guideline for artificial intelligence-centred diagnostic test accuracy studies: the STARD-AI protocol.
in BMJ open
Sounderajah V
(2021)
A quality assessment tool for artificial intelligence-centered diagnostic test accuracy studies: QUADAS-AI.
in Nature medicine
Sounderajah V
(2020)
Developing specific reporting guidelines for diagnostic accuracy studies assessing AI interventions: The STARD-AI Steering Group.
in Nature medicine
Solebo AL
(2022)
Development of a Nationally Agreed Core Clinical Dataset for Childhood Onset Uveitis.
in Frontiers in pediatrics
Shelmerdine SC
(2021)
Review of study reporting guidelines for clinical studies using artificial intelligence in healthcare.
in BMJ health & care informatics
Schwartz R
(2020)
Objective Evaluation of Proliferative Diabetic Retinopathy Using OCT.
in Ophthalmology. Retina
Schwartz R
(2020)
Objective Evaluation of Proliferative Diabetic Retinopathy Using OCT.
in Ophthalmology. Retina
Sachdeva G
(2023)
AI in Clinical Medicine - A Practical Guide for Healthcare Professionals
Rivera SC
(2020)
Guidelines for clinical trial protocols for interventions involving artificial intelligence: the SPIRIT-AI Extension.
in BMJ (Clinical research ed.)
Reekie IR
(2021)
The Cellular Composition of the Uveal Immune Environment.
in Frontiers in medicine
Pichi F
(2023)
Consensus-based recommendations for optical coherence tomography angiography reporting in uveitis.
in The British journal of ophthalmology
Pendleton SC
(2021)
Development and application of the ocular immune-mediated inflammatory diseases ontology enhanced with synonyms from online patient support forum conversation.
in Computers in biology and medicine
Panch T
(2020)
"Yes, but will it work for my patients?" Driving clinically relevant research with benchmark datasets.
in NPJ digital medicine
Ometto G
(2020)
Merging Information From Infrared and Autofluorescence Fundus Images for Monitoring of Chorioretinal Atrophic Lesions.
in Translational vision science & technology
Nath S
(2022)
Reinforcement learning in ophthalmology: potential applications and challenges to implementation.
in The Lancet. Digital health
Moraes G
(2021)
Quantitative Analysis of OCT for Neovascular Age-Related Macular Degeneration Using Deep Learning.
in Ophthalmology
Mollan SP
(2023)
Predicting the immediate impact of national lockdown on neovascular age-related macular degeneration and associated visual morbidity: an INSIGHT Health Data Research Hub for Eye Health report.
in The British journal of ophthalmology
Mateen B
(2020)
Improving the quality of machine learning in health applications and clinical research
in Nature Machine Intelligence
Martindale APL
(2024)
Concordance of randomised controlled trials for artificial intelligence interventions with the CONSORT-AI reporting guidelines.
in Nature communications
Liu Xiaoxuan
(2019)
Extension of the CONSORT and SPIRIT statements
in LANCET
Liu X
(2019)
A comparison of deep learning performance against health-care professionals in detecting diseases from medical imaging: a systematic review and meta-analysis.
in The Lancet. Digital health
Liu X
(2019)
Extension of the CONSORT and SPIRIT statements.
in Lancet (London, England)
Liu X
(2020)
Reporting guidelines for clinical trial reports for interventions involving artificial intelligence: the CONSORT-AI extension.
in Nature medicine
Liu X
(2022)
OCT Assisted Quantification of Vitreous Inflammation in Uveitis
in Translational Vision Science & Technology
Liu X
(2020)
Reporting guidelines for clinical trial reports for interventions involving artificial intelligence: the CONSORT-AI Extension.
in BMJ (Clinical research ed.)
Liu X
(2022)
The medical algorithmic audit.
in The Lancet. Digital health
Description | Appointed to UK Government's Regulatory Horizons Council - Independent advice on areas of regulatory reform which will support innovation |
Geographic Reach | National |
Policy Influence Type | Contribution to a national consultation/review |
Impact | Advise government on priorities for regulatory reform that supports innovation whilst protecting citizens; first reports include medical devices and nuclear fusion. |
URL | https://www.gov.uk/government/groups/regulatory-horizons-council-rhc |
Guideline Title | CONSORT-AI |
Description | CONSORT-AI - International Guidelines for Reporting of Clinical Trials for AI Interventions |
Geographic Reach | Multiple continents/international |
Policy Influence Type | Citation in clinical guidelines |
Impact | Improving design and reporting of AI interventions, for safer more effective more inclusive AI interventions. |
URL | https://www.nature.com/articles/s41591-020-1034-x |
Description | Contribution to national CASMI report on Health Data, the NHS and the UK |
Geographic Reach | National |
Policy Influence Type | Citation in other policy documents |
Impact | Realising patient and NHS benefits from health and care data - from policy to practice Informing the discussion and increasing trust |
URL | https://acmedsci.ac.uk/file-download/73707502 |
Guideline Title | SPIRIT-AI (Equator guidelines) |
Description | SPIRIT-AI _ International Guidelines for Reporting of Clinical Trial Protocols of AI Interventions |
Geographic Reach | Multiple continents/international |
Policy Influence Type | Citation in clinical guidelines |
Impact | Guidelines to inform the design and reporting of the design of clinical trials in artificial intelligence systems. |
URL | https://www.nature.com/articles/s41591-020-1037-7 |
Description | Artificial Intelligence for Missing Data Imputation in Electronic Medical Records |
Amount | £10,312 (GBP) |
Funding ID | NE/T013982/1 |
Organisation | Natural Environment Research Council |
Sector | Public |
Country | United Kingdom |
Start | 04/2020 |
End | 03/2021 |
Title | Birmingham, Solihull and Black Country Diabetic Retinopathy Screening Set |
Description | Routinely collected data of latest regional diabetic screening service in the UK |
Type Of Material | Database/Collection of data |
Year Produced | 2020 |
Provided To Others? | Yes |
Impact | Supports discovery, real world evidence and development/validation of AI systems. |
URL | https://web.www.healthdatagateway.org/dataset/36886b21-12ff-45e7-82bc-fb5308c12450 |
Title | INSIGHT Databases |
Description | 10 Databases |
Type Of Material | Database/Collection of data |
Year Produced | 2022 |
Provided To Others? | Yes |
Impact | Available for researchers |
URL | https://www.insight.hdrhub.org/datasets |
Title | Moorfields AMD Dataset Release 002 |
Description | World's largest AMD linked structure-function dataset with longitudinal data of over ten years. |
Type Of Material | Database/Collection of data |
Year Produced | 2021 |
Provided To Others? | Yes |
Impact | A number of active studies, including impact of anti-VEGF therapies. |
URL | https://web.www.healthdatagateway.org/dataset/bf392537-82b4-4d71-ace0-e7bea3b167fb |
Title | Moorfields Eye Image Dataset Release 001 |
Description | World's largest single centre ophthalmic imaging dataset |
Type Of Material | Database/Collection of data |
Year Produced | 2020 |
Provided To Others? | Yes |
Impact | Multiple ongoing projects for discovery and real world evidence, including impact of diabetes on AMD. |
URL | https://web.www.healthdatagateway.org/dataset/6bfe44ef-9532-4986-b8a2-9c2cda4a89cc |
Title | UHB Eye Image Dataset Release 001 |
Description | Exceptional eye imaging dataset from a diverse urban population in Birmingham |
Type Of Material | Database/Collection of data |
Year Produced | 2020 |
Provided To Others? | Yes |
Impact | Can support real world evidence, biological discovery and validation of AI diagnostics. |
URL | https://web.www.healthdatagateway.org/dataset/21fb56d1-726f-40e3-a998-e4b34ba3e46f |
Description | International ophthalmic data standards |
Organisation | American Academy of Ophthalmology |
Country | United States |
Sector | Charity/Non Profit |
PI Contribution | Consensus for international standards for ophthalmic imaging data - co-leadership with INSIGHT and the AAO representatives |
Collaborator Contribution | Consensus for international standards for ophthalmic imaging data - co-leadership with INSIGHT and the AAO representatives |
Impact | Published - AAO statement Pending - RCOphth statement In planning - data collection study, and a grand challenge |
Start Year | 2020 |
Description | International ophthalmic data standards |
Organisation | Royal College of Ophthalmologists |
Country | United Kingdom |
Sector | Academic/University |
PI Contribution | Consensus for international standards for ophthalmic imaging data - co-leadership with INSIGHT and the AAO representatives |
Collaborator Contribution | Consensus for international standards for ophthalmic imaging data - co-leadership with INSIGHT and the AAO representatives |
Impact | Published - AAO statement Pending - RCOphth statement In planning - data collection study, and a grand challenge |
Start Year | 2020 |
Description | AI Ethics Launch - NHSX Panel |
Form Of Engagement Activity | A talk or presentation |
Part Of Official Scheme? | No |
Geographic Reach | International |
Primary Audience | Professional Practitioners |
Results and Impact | Talk on Health Data Poverty, and impact of routinely collected data to over 300 attendees followed by a panel discussion |
Year(s) Of Engagement Activity | 2021 |
Description | Big Picture Podcast - INSIGHT and Health Data Poverty |
Form Of Engagement Activity | A broadcast e.g. TV/radio/film/podcast (other than news/press) |
Part Of Official Scheme? | No |
Geographic Reach | International |
Primary Audience | Public/other audiences |
Results and Impact | Regular podcast featuring leaders in medicine. We discussed AI, health data research and the role of INSIGHT in addressing health data poverty. |
Year(s) Of Engagement Activity | 2020 |
URL | https://podcasts.apple.com/gb/podcast/big-picture-medicine/id1500446262 |
Description | Bloomsbury Festival |
Form Of Engagement Activity | Participation in an activity, workshop or similar |
Part Of Official Scheme? | No |
Geographic Reach | National |
Primary Audience | Public/other audiences |
Results and Impact | Bloomsbury Festival 2020 - online due to pandemic; 21200 website visitors; 32000 audience of which 3500 were for digital events, 15000 for radio |
Year(s) Of Engagement Activity | 2020 |
Description | INSIGHT website |
Form Of Engagement Activity | Engagement focused website, blog or social media channel |
Part Of Official Scheme? | No |
Geographic Reach | International |
Primary Audience | Public/other audiences |
Results and Impact | Development of INSIGHT website |
Year(s) Of Engagement Activity | 2020,2021 |
URL | https://www.insight.hdrhub.org |
Description | Insights from INSIGHT |
Form Of Engagement Activity | A broadcast e.g. TV/radio/film/podcast (other than news/press) |
Part Of Official Scheme? | No |
Geographic Reach | National |
Primary Audience | Media (as a channel to the public) |
Results and Impact | Series of podcasts - 'Insights from INSIGHT' |
Year(s) Of Engagement Activity | 2020 |
Description | NIHR BRC Patient and public group |
Form Of Engagement Activity | A talk or presentation |
Part Of Official Scheme? | No |
Geographic Reach | National |
Primary Audience | Public/other audiences |
Results and Impact | Presented to broad groups of public and patients on 8 different occasions through the NIHR BRC through 2020. |
Year(s) Of Engagement Activity | 2020 |
Description | Westminster Health Forum - Regulation including Data |
Form Of Engagement Activity | A formal working group, expert panel or dialogue |
Part Of Official Scheme? | No |
Geographic Reach | National |
Primary Audience | Policymakers/politicians |
Results and Impact | Talk on 'Navigating the changing research landscape' Panel as part of Trust and Transparency |
Year(s) Of Engagement Activity | 2021 |
URL | https://www.westminsterforumprojects.co.uk/conferences/westminster-health-forum |