Choosing the right PACS can be a daunting prospect. The overarching strategic objectives described in our digital imaging strategy and future direction blog article is best described with a car industry analogy—and the strategic shift from petrol and diesel to electric.
Here we zoom in to look at the ‘must have’ latest features; the keyless entry, driver assist, and smartphone integration (to use the new car analogy again)—but in the world of imaging systems! This article explores key features you should insist on within your Picture Archiving and Communication System (PACS).
1. PACS Based Reporting (PBR)
The traditional method of radiology reporting has been to view the images in PACS, while simultaneously dictating the report into the Radiology Information System (RIS), using various integrated 3rd party voice recognition.
This has served in-hospital working very well with tight integration between RIS, PACS, and Voice Recognition (VR) systems—commonly called ‘desktop integration’. Ensuring any patient opened in RIS has their corresponding images displayed alongside in PACS.
This model was designed largely for in-house trusts reporting, where all 3 systems resided locally on high-speed networks. However, the evolution of reporting has moved the focus to ‘out of hospital’ (e.g. Community Diagnostic Centres and reporting hubs).
With hospital space being at a premium, having on-site reporters are not perceived as cost-effective or attractive from a job perspective (commuting extensively to site). The new norm of ‘working from home’ is an extremely possible and attractive proposition to reporters and can really assist any workforce strategy.
Subsequently, ‘off-prem’ reporting has become a focus mainly due to the prevalence of Community Diagnostic Centres (CDCs), the rapid increase of teleradiology reporting, and of course, COVID-19.
A modern alternative is now PACS-based reporting (PBR) where the report is transcribed directly into PACS (instead of RIS).
The Benefits Of PBR Are;
- Remote access, having 3 systems in 1 (PACS, RIS, and VR) hence less connectivity (bandwidth/latency) and synchronisation issues over a VPN over home broadband
- Structured interactive reporting where hyperlinks to relevant images are available to referring clinicians as far out as primary care (i.e. identifying pathology on the 2 relevant images out of 100)
- Ease of remote reporting, as only a PACS client with VR integration is required
- Reduced cost
- No synchronisation or replication issues
2. Mixed Economy Reporting And ‘Report Locking’
Perhaps contradictory to the previous section highlighting the benefits of moving to PBR, there are several situations where PBR may not be available or preferred. Examples include; during the transition to PBR, the obligation to report into a radiology element of an Electronic Patient Record (EPR) e.g. EPIC Radiant, or user preference.
In fact, a user may wish to report into RIS when on the hospital site working with a full three-monitor workstation set-up, but use PBR when at home when on a single or dual screen set-up (space saving). In shared (expensive) workstation environments of reporting hubs, a reporter that prefers PBR may be followed by a reporter that prefers RIS-based reporting.
The term ‘mixed economy reporting’ is used to describe the utopian solution to all scenarios described whereby users can choose what method they would like to use without obligation to only use one.
However, the critical element of mixed economy reporting is the ‘report lockout’ functionality. Critical being in terms of clinical safety.
In a mixed economy environment, there is the potential ability to report on the same examination in two different systems, by two different reports, at the same time. ‘User A’ could pick an examination of the PBR ‘to report’ list, while at the same time, ‘User B’ using an RIS-based system, could pick the same exam of the RIS ‘to report’ worklist. A similar issue arises with an ‘RIS arm’ of an EPR (e.g. Epic’s Radiant) and consideration for mixed economy vetting is necessary also.
The solution is ‘locking’ whereby as soon an exam is picked up for reporting in any system (by any method), it is immediately unavailable to report in any other system/methodology.
With PBR being a reactively new phenomenon, locking isn’t something that PACS and RIS companies have fully mastered—with scepticism over its proven ability.
Indeed, some providers are stipulating that mixed economies aren’t possible with their system (or often citing the other system’s inability) and subsequently stipulate either a RIS or PACS-based reporting environment, but not mixed.
Other important (and sometimes overlooked) considerations are having bi-directional integration between RIS and PACS and the ability to open each other system from within.
In any case, it’s apparent that it’s an important element that must be managed and standardised across regions.
3. Structured Reporting And Automation (Auto-Triggering)
Moving to PBR opens opportunities in terms of structured and interactive reporting. Traditionally radiology reporting is a collection of images and an associated plain text report. There may be a thousand images in a full body scan, however, the text report refers to the presence of abnormality in only (say) ten images. Hence, no specific connection between the abnormal images and the report. (BTW, this has been a major problem for imaging AI algorithms learning to read scans. While they can ingest the images and the reports, there was no way to identify the pathologies/abnormalities (referred to in the text report) for ‘machine learning’ to occur).
However, with interactive reporting, reference text in the report can be hyperlinked to the exact area on the most relevant/key images (that may be 5-10 images out of said 1000). Embedding patient data within reports.
This is particularly useful for non-radiology clinicians viewing images who physically want to see the abnormality/pathology. Similarly, as imaging expands out into the community via regional care records, this functionality will be available to General Practitioners (GPs) and community practitioners.
Another relevance of structure reporting is the use of automation and report triggers. As a practical example, this could be a plain film chest X-ray report statement along the lines of ‘evidence of new lung pathology found’ which would automatically trigger an appointment request for a CT lung scan, with a follow-up appointment with a lung cancer specialist booked at the specialist cancer hospital. Previously this would all have to be completed manually and causes unnecessary delays.
Another example could be automated flagging of urgent findings back into hospital ordering and EPR systems and the subsequent triggering of urgent care pathways. Similarly with automation back to community referrers/GPs. The potential for automated triggers from interactive radiology reports is significant.
And as waiting time targets and Key Performance Indicators (KPIs) reduce for cancer treatment, this will be a vital tool in the cancer treatment pathways.
4. Multi-Disciplinary Team (MDT) Imaging On Demand
Currently, cross-ICS Multi-Disciplinary Team (MDT) meetings require copies of images to be physically sent between trusts which involves up to 2 weeks of prep work by MDT co-ordinators who need to request and receive images through the (antiquated) Image Exchange Portal (IEP) system.
The meetings themselves are often made up of highly specialist consultants and is a really expensive resource for any trust to have them run inefficiently. The fact that patient images are unavailable on-demand across the region is unacceptable and is a policy objective to eliminate.
5. EPR And Shared Care Record Integration
There is a need for clinical staff outside of imaging departments to request diagnostics and review results. The system of record will not be PACS or RIS but will typically be the hospital EPR system.
Critical alerts are most relevant in EPR being the system that clinicians interact with most. As such, there is a need to integrate RIS and PACS with EPR systems. Such integration will allow the image and study to be stored in the PACS/RIS environment and linked through to the EPR in patient context for others to access, following the principle of write once and read many times (WORM).
In addition to integrating with EPRs, as primary care, community, mental health, and ambulance services also have specific and bespoke systems, there is also a requirement to integrate with intermediary (regional) share care records. Results communication is also a vital consideration (including any updates/addendums).
6. Mobile Devices And Apps
Parallel to the move to cloud, a growing trend is using web-based applications. Originally PACS were built for ‘on-prem’ in-hospital working. And the functionality of web browsers was also limited.
However, web functionality has since greatly improved, and accessing PACS through web browsers (e.g. Google Chrome) is now viable. The concept is a zero-footprint-based client, with no additional software installation.
In practice, this means that any computer can be used to access PACS without any need to install a beefy PACS client (which often requires IT admin rights and time-consuming physical installs).
And even subsequent to web and cloud evolvement has been app development. Most PACS providers now have dedicated iOS and Android apps that any user can download and (with the right access credentials) can access PACS on mobile devices with several studies demonstrating image quality similar to PACS workstations.
And probably the most interesting aspect is the ability to use the device to capture images and store directly to the patient’s medical imaging dataset.
This has far-reaching implications and potential. It’s a very exciting development for the likes of medical photography/illustration departments who previously needed dedicated devices/cameras and physical connection/uploading to PACS—particularly in the community. Information governance is a general concern that apps help to address (for e.g. no images taken are direct can be saved to the device.) The potential is clear.
A similar solution is necessary for handover information between CDCs and acute trusts when the immediate transfer is required (e.g. severe contrast reaction, evolving compartment syndrome, etc.).
A potential future application could extend to mapping a patient journey through care (photos and videos) saved to PACS, which would be accessible via the patient portal (to both patients and next of kin with required permissions).
Therefore, it is a key recommendation to also embrace the latest app technology that enables further integration with community and secondary care.
7. Imaging AI
To briefly summarise, such is the exponential scale of imaging growth, that imaging AI is seen as potentially having the biggest impact. Subsequently, it is a national recommendation to establish regional oversight and ICS level imaging AI boards, and to appoint clinical and operation AI leads to focus on imaging AI applications with co-ordination of shared learning and deployment at scale in a timely manner – ‘from concept to clinic’.
Clinical AI
Clinical AI can be defined as specific AI algorithmic scan reading tools to identify pathologies/abnormalities and ultimately assist reporters. There is an abundance of AI algorithms (in radiology, digital pathology, and ophthalmology) on the market with many mini/test projects across regions including Brainomix (several silo projects), Veolity, Qure, Annalise, Khieron, and Aidence (lung health checks/screening) with plans to expand it wider.
Funding awards for projects include NHSX AILabs, SBRI, and NIHR. While this is greatly encouraging there isn’t currently a mechanism to share learning between sites, hence the potential of duplication for pilot/testing of the same software or application area at different trusts. ‘Deployment at scale’ through cloud is a major benefit here.
Operational AI
Operational AI is using AI to optimise reporting workflow. This is much less prevalent in the media and/or tools developed. However, (unbiased) industry insiders comment that while there is a lot of hype about clinical AI, operational AI could be key to optimising reporting workflow.
Examples include iRefer (up-to-date computer-aided imaging referral advice) and xWave (clinical decision support and vetting system).
For reporting, when a reporter logs on, the system should know their specialism and can allocate the most appropriate and urgent work from across the region. Providing the reporter with the most appropriate information at exactly the right time. The system will learn the operator’s work patterns (learning from their click patterns) and allocate work while preventing ‘cherry picking’.
8. Insourcing And Outsourcing — Realignment At Scale
The overuse and cost of outsourcing (teleradiology) is well documented nationally. The current estimates are around £200m annually on outsourced teleradiology services. As well as the imbalance with regard to insourcing, some Imaging Networks (IN) are in the pilot stage of developing an in-house, cross-site Radiology Reporting Collaborative (RRC).
The aim is to reduce the number of radiology images that are outsourced by utilising the skills of radiology reporters within the network. The use of dedicated VDIs (described later) pioneered by companies like Axon Diagnostics and Fujitsu.
In-sourcing refers to renumeration of current employee reporters to perform extra reporting on a cost-per-unit basis (similar to the teleradiology companies model). The potential benefit is reduction in cost by eliminating the teleradiology company margin.
Out-sourcing, or ‘teleradiology’ as it’s known as in the radiology world, is essentially a buyer’s marketplace for unit reporting services. Different hospitals negotiate different rates with the same companies. There are really only 4 main teleradiology companies in the UK with perhaps 2-3 peripheral/upcoming providers.
Each trust’s rate depends on the strength of the negotiator with resultant rate disparity across the region. Often there is an absence of shared learning or sharing of best rates across a region. This was largely a result of the competitive environment created by CCGs and independence of trusts.
With the development of imaging networks, a joint approach could standardise the rates in the region while in parallel have significant buying power to negotiate best rates (using the best negotiators) saving money in the process. The savings here could be substantial.
It appears that most expensive aspect of outsourcing is Out of Hours (OoH) reporting. This is where the big 4 companies actually make the most of their profit. The model is favoured by trusts as it reduces the need for reporters ‘on-call’ at night which can result in 2 days leave as compensation. Given current daily clinical workforce shortages, night time outsourcing is seen as a solution to having more staff available during daily hours and hence the expensive OoH work.
Models used by outsourcing companies include ‘follow the sun’ models where reports for OoHs in the UK are performed in locations with daytime hours (e.g. all of Wales OoH service is provided from Australia).
In parallel, traditional problems revolved around IG whereby data could not leave England. Private providers found ways around this however through cloud and streaming services. There is potential for INs to also look at alternative models.
A key objective for any IN or region is to pool its resources to establish an insourced reporting hub. These are ‘intense reporting farms’, dedicated to high throughput reporting productivity in a non-clinical setting. There could be several across regions, (linked virtually) combined with home reporting—designed and located around reporter convenience (i.e. could be in a facility close to major access routes).
Access to the latest high-speed network technology (e.g. VDIs), highest specification (shared/pooled) reporting workstations, and availability of cloud-deployed and workflow tools described elsewhere.
Out-of-Hours (OoH) reporting hubs have already been successfully deployed across a handful of imaging networks. Scaling up this model could have great potential nationally. Details of one such case study on the CAMRIN hub can be found online here;
‘The aim was to improve the quality of its round-the-clock radiology service through shared imaging and reporting, and the provision of a more streamlined workload for trainee residents. The hub is based at Broadgreen Hospital with a second site at Aintree, where national trauma requirements require the physical presence of a radiologist to provide instant opinion for trauma studies. Seven of the network’s thirteen trusts now have access to the hub.’
The potential benefits of reporting hubs are pooling resources, better access to specialist reporters for patients (and subsequent reduction in healthcare inequality), reduction in on-call frequency for reporters, and reduction in expensive OoH teleradiology. There’s also potential for completely realigning employment contracts of reporters whereby employing directly with ‘NHS regional reporting services’ with a subsequent ‘reporter passport’ to report for all sites across the region, as opposed to the traditional contracting to individual trusts model.
There is a caveat and school of thought that outsourcing companies can actually manage the service more cost-effectively than the NHS. This basically evolves from the fact that the NHS pay salaries whereas private providers pay per report with much less contractual and financial obligations.
In summary, there is a significant opportunity to completely realign both in-sourcing and out-sourcing of reporting across the regions as imaging moves from individual sites to wider imaging networks (multi-sites across regions).
There’s potential to pool resources, standardise and reduce costs and utilise different reporting models. National policy recommendations are to move to regional core central business management models (at scale) for both in and outsourcing radiology reporting.
9. Remote Access Via VDI/HTTPS
Current access to PACS and RIS remotely is largely through a Virtual Private Network (VPN) that connects back to the trust campus network. This is often shared, and slow for image reporting (not equivalent to in-hospital performance). Synchronisation between RIS, PACS, and Voice Recognition (VR) is also a major consideration, as is latency. In practice, to date, access via VPN in a RIS reporting environment has been challenging and limitational.
A more recent innovation in this space is access via Virtual Desktop Infrastructure (VDI), which improves performance with some options available for dedicated (i.e. not shared with other trusts services) PACS VDI solutions (at significant expense). All these are naturally routed back to trust given this is the natural home of PACS. Several sites within each ICS have implemented VDIs (individually).
A significant consequence of moving RIS and PACS to cloud is that trust campus networks can be completely removed from the equation and direct access to the cloud PACS via a secure internet gateway (HTTPS) is a more sensible and viable solution. It is similar to a VDI solution at a fraction of the cost (and can be deployed across sites at scale with cloud).
Access is also a single-entry solution, as opposed to per trust (depending currently on location and PACS IT/cloud config), and is useful for cross-site working and teleradiology providers. As a collective, many ICS/INs that are cloud already are now looking at HTTPS access as the future.
10. Virtual Support Tools
Various reporter decision support tools are available online. This often requires an integration (and subscription service). Teaching and training such as using qCheck and other similar tools such as RAIQC and Collective Minds.
11. Microsoft Teams, Skype, And Messenger/Chat Functionality
Research has shown that up to 30% of junior doctors time is spent chasing diagnostic results. In practical terms, this means if a junior doctor wants to speak with a reporter, they must walk to the radiology department and find an available and appropriate specialist reporter (often never in the same place due to hot reporting).
Through the use of integrated messaging, chat, and video functionality, reporters can be reached instantly online. Reporters can also set their online availability.
Need to discuss a case? Look online to see who is available and quickly send them a message or video call request direct to discuss (which can be set to pop up during a reporter’s reporting session). Special screen sharing and anatomical linking tools can also enable users to synchronise screens during the call also.
Over the past 36 months, it seems everyone in the NHS has been using Microsoft Teams for meetings to good effect. Screen sharing is a commonly used function. Integration of MS Teams and PACS applications is essentially MDT on-demand.
Setting up a MS Teams call between several sites/ICSs that use MS Teams with integrated PACS to discuss cases on demand is now possible (there are considerations required around image quality).
Hence, in summary, in much the same way as modern life has evolved to messenger, chat, and video calling functionality, diagnostic imaging can also benefit greatly from such tools.
About The Author
Pauric Greenan is an experienced PACS consultant that has been PACS lead on several high-profile PACS projects in both the UK and Ireland. Pauric is internationally available for PACS consulting. Contact him here for more details.
Did you know that we are able to fulfil a variety of your imaging diagnostic needs?
Choice Health is highly experienced in clinical trial and research imaging, PACS consultancy, telehealth, imaging AI and medical physics.
Would you like a free consultation to see if we’re a match? If so, do not hesitate to reach out via our contact page.
