Share
Embed
Research in a Regional Radiology System
Yorkshire Academic Health Science Network and NIHR study day exploring the role of the Yorkshire Imaging Collaborative in imaging research.
Dr Daniel Fascia
December 11, 2019
Transcript
42 slides
-
The Yorkshire Imaging Collaborative Research in a Regional Radiology System
Dr Daniel Fascia Clinical Lead of The Yorkshire Imaging Collaborative @yorks_imaging yorkshireimaging.org -
Every patient in Yorkshire should be able to attend an
appointment and have full availability of their medical images and associated reports at the point of care delivery -
The Bigger Picture • Ultimate goal is a UK wide
radiology network • Regional collaborative networks are recommended and endorsed by the Royal College of Radiologists • and supported by NHS Improvement ◦ £8.4M awarded so far • Radiology workforce must redesign itself for success ◦ A workforce in crisis ◦ 29% retirement in next 5-years ◦ Insufficient trainees (skill + number match) ◦ 9 - 55% vacant consultant posts -
PHASE B Connect as regional network PHASE A 8 x
Trusts New PACS Co-procurement Deploy PACS Learn + Improve X 8 Transformation “Masterplan” <vendor(s)> Co-procurement Enact Transformations Regional Index Iteratively improve provider model -
Aire Hull Hg NLAG C&H Mid Yorks Leeds Brad Sharing
Solution Procurement Q3 2019 CLINICAL IMAGING INTEROPERABILITY Interoperating with Xero Viewer “Images and reports visible cross Collaborative” -
Q3 2019 CLINICAL IMAGING INTEROPERABILITY Leeds Hg Leeds Teaching Hospitals
CT Scan Report Qui ut fugiat dolore culpa velit adipisicing amet commodo. Aliquip voluptate sint irure ex minim ea irure nulla laboris nostrud deserunt velit ex reprehenderit. Non ut elit Lorem id sint id occaecat occaecat. Eiusmod excepteur Lorem duis ea labore proident excepteur velit nisi et voluptate. Laborum voluptate elit sint aliqua Lorem magna ea elit aliqua dolore. Harrogate Hospital CT Scan Report Qui ut fugiat dolore culpa velit adipisicing amet commodo. Aliquip voluptate sint irure ex minim ea irure nulla laboris nostrud deserunt velit ex reprehenderit. Non ut elit Lorem id sint id occaecat occaecat. Eiusmod excepteur Lorem duis ea labore proident excepteur velit nisi et voluptate. Laborum voluptate elit sint aliqua Lorem magna ea elit aliqua dolore. -
TOTAL INTEROPERABILITY IMAGING & REPORTS REGIONAL PATIENT INDEX Aire Hull
Hg NLAG C&H Mid Yorks Leeds Brad Q3 2020 TOTAL RADIOLOGY WORKFLOW INTEROPERABILITY -
Advanced Interoperation PATIENT JOURNEY GAME CHANGER GP Community Imaging Centre
Regional Hospital (Local to patient) Imaging Pathway MDTM Tertiary Hospital (Local to patient) Specialist Care Pathway Complete Imaging Integration without GP input Clinical Pathway -
FOR PATIENTS Same service everywhere Flexibility in appointments Find the
capacity Images and reports available -
FOR RESEARCH Regional Population Index Central taxonomy (labelling) Binary data
aggregation Metadata aggregation -
PROCESS & WORKFLOW COMMON PRACTICES SHARED CORE SERVICES BUSINESS INTELLIGENCE
UNIT LEARNING & DEVELOPMENT MESSENGER SERVICE Regional Dose Monitoring DIRECTORIES (searchable) Staff Machines Rooms Sites Trusts Organisations HUMAN RESOURCES Training Passport Cross site credentials INTERNAL STAFF BANK INSOURCED REPORTING MEDICAL PHYSICS CANCER Priority Group Scanning Protocols Reporting Standards Procedural Documents Patient Advice Paeds MSK Neuro H&N Gastro/Uro Trauma Nuclear Med Chest/Cardiac IR/Vascular SPECIAL INTEREST GROUPS RADIOLOGY SERVICE GROUP Remote Work ACCOUNTING CENTRAL IT FOR SHARING NETWORK PORTAL WEBSITE tScan tReport Standards Regionalised ON CALL PROCUREMENT GROUP 3rd party provider standards Remote Working Standards Priority standards Data Lifecycle Consent for research/AI REDUCING PROBLEMATIC VARIANCE -
PROCESS & WORKFLOW COMMON PRACTICES SHARED CORE SERVICES BUSINESS INTELLIGENCE
UNIT LEARNING & DEVELOPMENT MESSENGER SERVICE Regional Dose Monitoring DIRECTORIES (searchable) Staff Machines Rooms Sites Trusts Organisations HUMAN RESOURCES Training Passport Cross site credentials INTERNAL STAFF BANK INSOURCED REPORTING CANCER Priority Group Scanning Protocols Reporting Standards Procedural Documents Patient Advice Paeds MSK Neuro H&N Gastro/Uro Trauma Nuclear Med Chest/Cardiac IR/Vascular SPECIAL INTEREST GROUPS RADIOLOGY SERVICE GROUP Remote Work ACCOUNTING CENTRAL IT FOR SHARING NETWORK PORTAL WEBSITE Regionalised ON CALL PROCUREMENT GROUP 3rd party provider standards Remote Working Standards Priority standards Data Lifecycle Consent for research/AI REDUCING PROBLEMATIC VARIANCE MEDICAL PHYSICS tScan tReport Standards -
PROCESS & WORKFLOW COMMON PRACTICES SHARED CORE SERVICES BUSINESS INTELLIGENCE
UNIT LEARNING & DEVELOPMENT MESSENGER SERVICE Regional Dose Monitoring DIRECTORIES (searchable) Staff Machines Rooms Sites Trusts Organisations HUMAN RESOURCES Training Passport Cross site credentials INTERNAL STAFF BANK INSOURCED REPORTING CANCER Priority Group Scanning Protocols Reporting Standards Paeds MSK Neuro H&N Gastro/Uro Trauma Nuclear Med Chest/Cardiac IR/Vascular SPECIAL INTEREST GROUPS RADIOLOGY SERVICE GROUP Remote Work ACCOUNTING CENTRAL IT FOR SHARING NETWORK PORTAL WEBSITE PROCUREMENT GROUP Data Lifecycle Consent for research/AI REDUCING PROBLEMATIC VARIANCE MEDICAL PHYSICS tScan tReport Standards -
LIKE PILOTS RADIOLOGISTS NEED HELP - CLINICIAN DECISION SUPPORT -
MUNDANE TASKS (Lung Nodules) - SECOND READER (Breast) - INCREASING SCAN SIZE/VOLUME - SCAN RECYCLING (Vertebral #) - SMARTER WORKING (Just 60% of time direct clinical care) - PRIORITISATION (Abnormal > Urgent escalation) -
AI NEEDS: • Lots of data • Consent • Data
taxonomy • Incubation environment -
COMMON PRACTICES CANCER Priority Group Scanning Protocols Reporting Standards Paeds
MSK Neuro H&N Gastro/Uro Trauma Nuclear Med Chest/Cardiac IR/Vascular SPECIAL INTEREST GROUPS RADIOLOGY SERVICE GROUP Priority standards Data Lifecycle Consent for research/AI REDUCING PROBLEMATIC VARIANCE Lots of Data Well taxonomised Thanks to the work of our Special Interest Groups, data and practice standardisation are already underway -
ACADEMIA Prof Alejandro Frangi - Occult fractures - Cardiac Imaging
- Supply demand modelling -
PROJECT 1: Modelling Demand 2.6 Million Attendance Records 22 Years
Longitudinal Data Machine Learning trends FORWARD INSIGHT April 2025 CT SCANS 448 3 x Scanners Req CT DAY STAFF BAND 7 x 4 BAND 6 x 2 BAND 3 x 4 -
Model for the detection of meniscal tears on knee MR
• ImageNet transfer learning