Where are we going anyway and what

Where are we going anyway? …. and what are the chances of getting there via the NPf. IT Dr Keith Foord Consultant Radiologist, East Sussex Hospitals, United Kingdom www. esht. nhs. uk keith. foord@esht. nhs. uk or secretary@pacsgroup. org. uk

Objective 1 for this group? Complete Integration of RIS and PACS or as near as possible, with some points from history

History 1970 -2001 • 1970 s – First RIS systems – To manage departmental workflows and store information • Late 1980 s/early 1990 s – First operational PACS – But did not link information in RIS with images • Mid 1990 -2001 - Image centric PACS with RIS interfaces • Incompatible communication protocols forced ‘Brokers’ • Image centric – PACS image DB has to be additionally populated with information INTRODUCED to the system • Some RIS functions have to be duplicated in PACS • Problems with correlation of RIS & PACS data - requires administrator intervention to correct

History 2001+ • RIS centric PACS – The RIS is prime and controls information flows, – – – including images Simplifies information management RIS becoming integrated – integrated Brokers or ‘Brokerless’ IHE integration profiling Provides DICOM Modality Worklist (MWL) directly to modalities Uses DICOM Modality Performed Procedure Step (MPPS) – if supported by both modality and RIS

Communication issues between IS databases, PACS and modalities 20/11/03 Keith D. Foord HL 71 HIS HL 7 i/f SPF or ‘Gateway’ Nov. 20 2003 Foord, Keith D. HL 72 RIS HL 72 DICOM HL 7/DICOM I/f = PACS Broker SPF Modality DICOM PACS DICOM

Unidirectional RIS/PACS Many RIS vendors have provided Uni-directional data to PACS via a PACS Broker. Data not sent back to RIS to update fields related to the exam. If RIS does not support DICOM MWL or modality does not support MWL Demographic data must be entered manually at modality – high risk of errors. Errors manually corrected at the Archive or QA station, Reducing productivity and delaying availability of images. If not corrected images ‘orphaned’ and not available.

Unidirectional RIS/PACS I/f without Modality DICOM MWL Archive RIS HL 7 PACS Broker DICOM minus MWL Modality QA station DICOM data, no MWL Manual input of data. Prone to error Non – MWL Modality Manual correction of data to match RIS data If not done up to 20% of studies are ‘orphaned’ Reporting Workstation

Unidirectional RIS/PACS I/f with Modality DICOM MWL Archive RIS HL 7 PACS Broker DICOM Data incl MWL Modality Reporting Workstation

Bi-directional RIS/PACS Data on start/finish exam, procedure changes, resource utilisation, number of images and series in study if sent back to RIS enhance QA, increase productivity and allow full integration into Integrated Clinical Systems. To do this both RIS and Modality must support not just MWL but also DICOM Modality Performed Procedure Step (MPPS)

Bi-directional RIS/PACS I/f with DICOM MWL and Modality Performed Procedure Step installed in both RIS and Modality Archive RIS MWL/MPPS HL 7 + DICOM PACS Broker DICOM + HL 7 DICOM MWL/MPPS Modality Reporting Workstation

Integrated RIS/PACS with DICOM MWL and Modality Performed Procedure Step installed in both RIS and Modality RIS/PACS Internal HL 7 - DICOM & DICOM – HL 7 transactions Demographics MWL MPPS Archive DICOM General Purpose Worklist (if provided allows choice Of WS independent of PACS Vendor) MWL/MPPS Modality Reporting Workstation

Voice RIS Broker PACS Internal Transactions Voice RIS PACS Internal Transactions PACS companies which have acquired RIS company products. Still basic brokering, but added internal HL 7/DICOM transactions. De-novo combined RIS-PACS products. Some internal interfacing plus Internal HL 7/DICOM transactions. Voice RIS Internal Transactions PACS Different vendors with all the HL 7/ DICOM transactions in RIS within a ‘PACS integration module’. Advantage – best of breed

Complete Integration of RIS into PACS: Dream or Reality? With an old non HL 7 RIS – forget it With an old HL 7 Brokered RIS – limited With a new HL 7(IHE) RIS - very nearly a reality with a PACS integration module - this allows freedom to choose best RIS and best (IHE) PACS With a same vendor combined RIS-PACS – internal HL 7/DICOM transactions …. But what about the modalities, DICOM MWL and MPPS? Don’t forget the need to integrate the HIS and Integrated Clinical Systems too!

Objective 2 for this Group The same complete Integration of RIS-PACS and non-Radiological Images

Example: UGI Tumour management Text and image data gathered at initial presentation and diagnosis plus local staging – followed by centre assessments History + added History Clinical examination Blood tests Endoscopy Histopathology CT/CXR/Ultrasound Abdo ECG/PET/Endoultrasound Spirometry/Cardiac NM Text Images Text General Condition of Patient Text Tumour Type TNM Images

PACS needs to store more than Radiology images ! Colposcopy Endoscopy Histopathology ECGs HIS Cytology RIS Blood films Broker PACS Internal Transactions EEGs Medical photographs EPR Voice Videos Opthalmology Dermatology

But…. we need the same standard of integration as with a modern Radiology RIS-PACS ‘X’IS/PACS Internal HL 7/DICOM/XML transactions DICOM images DICOM 2 o capture HIS HL 7 and / or XML data Non-DICOM images Archive EPR ‘X’IS Web server Viewing

Objective 3 for this Group Full Integration of RIS-PACS and non. Radiological Images into a comprehensive National Integrated Clinical Information System working with NPf. IT

Integrated National Multi-IS/PACS ASP Archive Wider NHS Local ICRS Wider NHS

RIS systems installed in UK by supplier Huge and long term International efforts have gone into protocol optimisation and framework standards with RIS and PACS to make them fully synergistic - DICOM HL 7 IHE These deep integration issues need to be matched by other Clinical IS systems – not just ‘top layer’ with XML but using From www. pacsgroup. org. uk data HL 7 and DICOM Old RIS systems have been a compromise and need replacing wholesale across the country to make PACS fully efficient, but must not be replaced ‘with just any’ new RIS

So. . What are the chances of getting there via the NPf. IT? cont. . §The best ‘buy’ PACS, from an LSP view, might not be the best clinical PACS. *LSPs appear to have 3 or 4 recognised suppliers each, so this is unlikely to be a problem §The best ‘buy’ PACS may not integrate well with the best ‘buy’ RIS or particularly an historic RIS! §Integrated RIS-PACS or a RIS with an Pacs Integration Module/DICOM MWL/DICOM MPPS may not be available from your LSP §Even if they are is your imaging equipment base up to it? §If the LSP has only one system per clinical speciality and these come from multiple sub-suppliers how will these fit ‘deeply’ with existing clinically satisfactory systems? *LSPs appear to have only ONE prime EPR supplier each, so this may be a problem with some hereditary systems. §What if clinicians on the ground don’t like what the LSP supplies – could there be clinical IT wastelands?

With thanks to Simon Daniell’s friend “Messages to NPf. IT………………” 1. 2. 3. 4. 5. 6. 7. 8. A good specification which must be achievable is paramount. This is usually acknowledged by the purchaser but they fail to recognise the responsibility this places on them. Where risk persists, you must have a work around solutions. This often means spending more money in the early phases on alternative solutions; each being dropped as their need diminishes. The prime contractor must identify the risks at the outset, but to declare the risks fully to the purchaser before contract award may reduce their chance of winning. The bigger, or more complex, the system the more important it is to manage the risk. Purchasers can relax too much when they force their supplier into fixed price contracts involving significant development. If the supplier gets into trouble it can rebound on the purchaser, especially with regard to timescale and even occasionally cost. If one major sub-contractor falls down there can be considerable cost impact on the other sub-contractors. The easy way to select a supplier of a development system is on cost, where he who underestimates most wins. He who has never implemented such a new system before is the more likely to underestimate. He who does not have ‘buy in’ from the end users advances at peril.