Does anyone have a specific retention policy for externally imported imaging?
We have a very large volume annually and it is taking up valuable storage space.
Obviously the original acquiring institution would retain their original data and anything we acquired on site would be kept, but is there a relevant requirement to retain something externally imaged and imported?
Perfectly valid request - we have a number of customers who specifically identify imaging from external sources and choose a retention period of anything from 30 days to 1 year and then purge. The obvious downside is that if you do have imaging that is useful long term, you need a process for identifying it and moving it to another âorganisation codeâ or similar so that it no longer meets the criteria for deletion.
In theory, we keep any imported imaging that hasnât been fixed up for six months. Though we havenât been rigorously enforcing this and only just starting to clear the backlog that has built up. Our supplier says a large number of unverified studies could cause perfromance issues.
If your PACS allows it, you could be really clever and set up some rules that also take into account when the study was last viewed - for example (if source = IEP and insertion date = older than 2 years and last viewed = greater than 1 year - delete) that way you could potentially save space but also keep stuff that is being reviewed yearly as part of your follow up imaging. Otherwise you could save on storage but spend more on re-requesting imaging constantly.
Two weeks is the period listed in the various PACS textbooks for storing external images.
Sounds short, why two weeks� We store them for a year!
As there is no duty (or real established process) for the sending sites to notify destination sites of any corrections / changes / updates (e.g. new recons which may reveal new information or splits due to patient admin), 2 weeks was considered the âsafeâ length of retention for external medical records without re-verifying that the copy is still the most complete and current iteration of the record.
Two weeks was also the period in the original IG approval agreements for bbRad and IEP to be setup overseen by the NHS CfH (then HSCIC) IG team.
From a legal perspective, a fortnight guarantees compliance with all the various requirements around proportionate data retention and allows for timely discussions if the patient is to be physically transferred. It also stops âclutterâ building up in PACS databases, which prefer things to be booked in and matched up.
To negate the question completely, use XDS to access the most current records, or have a disclaimer appear on external imaging stating something like âthis imported imaging from another institution is over 2 weeks old. Be sure to check it remains the most complete and current record of the patients external treatmentâ.
Itâs entirely possible to store everything external for 6 months, 1 year, forever etc. and we all know what happens in practice is that external imaging gets held for seemingly random periods of time, but technically two weeks remains the âsafeâ period. Anything routinely over this should really be justified in a SOP or policy.
Having a âdirtyâ archive for imports is a common way to manage these imports, or using Image Lifecycle Management to purge all imports from station name âTempImportâ (or whatever your site uses).
Back to your original question Mike - no, there is no real requirement to retain external images and imported studies unless you reprocess or add to them on-site (e.g. using Terarecon on data from a site which doesnât have it). BUT, for all imaging used in a decision making process clinicians must at least retain a record that images x, y, z were consulted and used in the clinical decision making process (e.g. by a RIS entry or note by the clinician in the patients notes). If the external imaging was only used to decide if a patient should be physically transferred, no real record is needed at the destination end (as it would be recorded at the sending end).
In fact the burden falls more on sites keeping external images for longer than the two weeks that they are then continually checking they are the most complete and currentâŚ
A can of worms really, use XDS 
Hope this helps!
Al.
1 Like
We retain all of our imported exams. Each import is given a corresponding entry on our RIS system and then matched before they are transferred to PACS. This ensures that both the imports and locally acquired exams are accessible on PACS or RIS under one patient folder. We find that this is important, particularly when the hospital number is the common method of searching for studies, either directly or via the EPR - PACS API integration.
Having all exams under one folder makes it easier for the Radiologists to prepare and present their MDTâs and also for the Clinicianâs who are working in clinics.
The only exams which are deleted off Pacs are the unwanted TEST/QA images. These are captured in an automated worklist (following boolean logic rules) before they are manually deleted. I would never feel comfortable allowing an ILM script to perform the actual deletion task. There is still a need for a bit human quality assurance!
Very interesting thread (particularly in light of the emergence of imaging networks).
Why is it that different NHS Trusts have different retention policies? Maybe a tertiary referral centre might have a reason, but otherwise?
Copying data gives rise to duplicates. Which one is now the âmost up to dateâ record? What if additional series are added at the receiving site?
Duplicates (data orphans) are clinically unsafe.
Interestingly â in this thread I see a ground-up set of rules emerging. It would be great if:
i) We could all agree
ii) For it to become a national set of policies
iii) Getting our suppliers to establish automated behaviours in their RIS/PACS to enact these policies
iv) I am sure there will ne a data/sematic normalisation (helped by national policy) to aid the functionality we ask of our suppliers
Interesting topic and one that always provides difference in opinion.
In certain medical workflows, especially those involving specialised imaging disciplines such as neuroradiology or paediatric radiology, images may be subject to secondary review by specialists. These images, along with these new reports, are stored in the RIS and PACS in their original form as they were initially received.
If any additional images are submittedâwhether they are updates or intended to complement the existing studyâthe original reporting staff member should review these new images and modify their report if necessary.
Moreover, if the new report and received images influence patient management decisions and later face legal scrutiny, the images and reports as they were at the time of reporting must be preserved and could be used as evidence in such proceedings.
This is a great topic!
As regards imports, it seems to be a reasonable requirement that the external imaging becomes part of the âpermanentâ PACS record at the receiving facility in circumstances where the imaging was used to influence the management of the patient.
What I have recommended previously is that these studies are imported onto PACS and that a RIS entry is created for the study, back-dated to the original imaging date. A scanned document or other entry on the RIS will have a disclaimer that this is external imaging received on (the received date) and the study may not be a complete representation of what was acquired or reported at the sending hospital.
If they sent a report with the imaging, that can be another series on the PACS record of that imaging.
Letâs say the receiving hospital wants one of their experts to report the imaging. This is not done on the entry on RIS where the imported entry was made. It is done on another entry âreview of external imagingâ under the date that this review actually took place. The imported imaging and the review of this imaging will have two different accession numbers.
Why all this rigamarole?
Because it retains the chronological integrity of the patientâs imaging history at the receiving facility. You have a record of when the imaging was received and what imaging was sent. If the imaging is reported or referenced (such as in a subsequent MDT) then that activity is also correctly chronologically documented relative to this imported imaging and subsequent imaging at the receiving hospital.
Consider this scenario:
Letâs say Alex sends me a CT of a patient and unbeknownst to him, there was one series on the CT scanner not sent to PACS or some other fault happened where it isnât in the correct folder on his PACS at the sending hospital.
My guys get the CT, get it included on our PACS with an associated RIS entry and this CT is then discussed at an MDT. The patient then has treatment based on that MDT.
How does it look medico-legally if, when we find out that there was a missing CT series, we request it from Alex and get that included retrospectively in the imported CT on the receiving PACS?
If a third party ever investigates the sequence of events relating to this patient, they are going to ask why the MDT outcome was the way it was when this extra series clearly indicated a pathology which the MDT seems to have ignored!
As messy as it sounds, it may be that the correct way to handle this is to create yet another imported images folder (same date as the original scan date, but with a different accession number on the receiving PACS and RIS) and place only the extra series in that folder. If the new series provokes changes to the patientâs management then this can be reflected in a subsequent report on imported images, but the crucial thing is that this report will have a date after the MDT and it will have itâs own accession number.
It might be an idea to include relevant accession numbers in the MDT so that anybody can subsequently see which imaging was discussed. It is to protect the receiving team from the radiological equivalent of âcooked books.â
Also, there are discrepancies between how different sites handle the chronological placement of imported images on PACS. At one site I found out that they would routinely import external imaging onto their PACS and create the RIS and PACS entries with the date of receipt of the images, not the date of the examination!
When those studies were then viewed on PACS, the oldest imaging appeared as the most recent in the Display Protocol. It canât be good for sequentially evaluating whether a lesion has increased in size or not.
1 Like
Agree 100%, particularly the scenario of creating a RIS entry without a back dated entry. That could be misleading for anyone looking at the patientâs imaging history and create an unnecessary clinical risk
All very good points and interesting discussion!
The âtwo weeksâ is quite old as it dates from the CfH era (and the whole âmaximum 4 patients on an unencrypted CD in the postâ times).
Looks like thereâs a couple of things to work on with this:
- Processes for sending sites to notify destination sites that a change to the imaging or record has occurred (similar to the Associated Press system for âkillingâ or correcting errors, or adding further / later information) - a recall or alert process.
- The desired retention period will likely vary per site, but for those retaining for longer routinely, XDS may be more appropriate.
- Disposal in compliance with patients rights under GDPR and the revision of the Code of Practice for Medical Records Retention document due soon.
- It would be useful for PACS manufacturers to offer a âCaution: External Images - Accurate Only At Time of Importâ flag or banner option for display on the image window overlays for non locally acquired imaging.
We could draw up new formal guidance on this, but I suspect it may be tricky to reach consensus (which is why there is nothing more recentâŚ
Al.
1 Like