Partial mammography design history post (#363)

Co-authored-by: Ed Horsford <mail@edwardhorsford.com>

Author: dannychadburn

app/manage-breast-screening/2026/02/improving-the-partial-mammogram-process/additional-details.png

@@ -0,0 +1,162 @@
+---
+title: Improving the partial mammogram process
+description: We've been looking into what happens when a mammographer does not take all the images they wanted to during a breast screening appointment.
+date: 2026-02-02
+author: Danny Chadburn
+opengraphImage:
+  src: /manage-breast-screening/2026/01/improving-the-partial-mammogram-process/additional-details.png
+  alt: Options for mammographers to record additional details related to the mammogram
+tags:
+  - beta
+  - prototype
+  - breast screening
+---
+
+We've been looking into what happens when a mammographer does not get to take all the images they wanted to during a breast screening appointment.
+
+Sometimes this is known as a 'partial mammography', but in some circumstances it is not.
+
+This has to do with how much of the breast has been scanned rather than (as we wrongly assumed) the completeness of the appointment.
+
+## To partial, or not to partial
+
+If an appointment stops before any images have been taken, the 'Attended not screened' status can be applied (read our [Exit, pause and resume an appointment](/manage-breast-screening/2025/12/exit-pause-resume-appointment/) design history).
+
+But as soon as one mammogram has been taken, a whole host of scenarios open up that our service needs to cater for.
+
+The National Breast Screening System (NBSS) gives users a checkbox and a reason dropdown when they need to record partial mammography.
+
+![Screenshot of NBSS with the partial mammography option highlighted](nbss-partial.png)
+
+[Guidance on partial or incomplete screening mammography](https://www.gov.uk/government/publications/breast-screening-guidance-on-partial-or-incomplete-screening-mammography/breast-screening-guidance-on-partial-or-incomplete-screening-mammography) explains when this should and should not be used.
+
+A partial mammography indicates that enough images have been taken and the client is not coming back for more. This is reflected in the screening results letter sent to the participant, which explains that an assessment has been made based on a 'partial' scan of their breasts.
+
+Partial mammography **should** be recorded when the participant:
+
+- withdrew their consent (after at least one mammogram is taken)
+- couldn't cooperate (usually due to learning or language difficulties)
+- could not be positioned correctly (due to being in a wheelchair, or for other mobility issues)
+- had an implanted medical device that affected the ability to scan breast tissue
+
+A partial mammography **should not** be recorded when too few images were taken and another appointment could be arranged to take more. So if one of the scenarios above does occur but the participant is willing and able to come back to try again, then it is not a partial mammography.
+
+Grasping this concept was crucial for us as a design team when trying to understand the current setup.
+
+Partial mammography also does not apply when:
+
+- there has been a technical failure with X-ray equipment that prematurely ended the appointment
+- there's a limited amount of tissue to be scanned (for example, if the participant has had a unilateral mastectomy)
+- all breast tissue could not be scanned due to breast implants
+
+### Partially mixed messages
+
+In everyday language, 'partial' often refers to a process you've started and intend to finish. You partially boil potatoes, partially pay off your mortgage, or have partially mowed your lawn.
+
+However the partial mammography definition in the guidance relates to the amount of breast tissue scanned.
+
+> If the examining mammographer believes that at least half of the breast can be demonstrated on one or more views of the routine examination, these views should provide sufficient useful information to be reported as a screening mammogram.
+
+Partial mammography applies when the 'examination' is incomplete, but the 'event' has concluded.
+
+This definition is repeated in the [Ionising Radiation (Medical Exposures) Regulations 2017 (IR(ME)R)](https://www.gov.uk/government/publications/ionising-radiation-medical-exposure-regulations-2017-guidance/guidance-to-the-ionising-radiation-medical-exposure-regulations-2017). There it provides a justification for getting enough information to assess a participant while minimising their radiation dose.
+
+### The point of confusion
+
+Our research uncovered a significant gap in the current process. NBSS does not provide mammographers with a standard way to indicate when more breast tissue needs to be scanned and the participant needs to be invited back to finish their examination.
+
+Official guidance is to report a technical recall for cases where participants could come back for more images, however our research has found this is inconsistently applied in breast screening offices (BSOs). Mammographers may:
+
+- communicate with image readers through the 'Film Comment' text field suggesting they should report the case as a 'Technical recall' to trigger a follow-up appointment invite
+- use the 'Partial mammography' checkbox and add a 'Film Comment', which means the case needs to be reopened if another appointment is scheduled
+
+## Basing our service on real situations
+
+After talking through the various scenarios they come across during appointments we established that mammographers want a service that allows them to:
+
+1. record when and why they have not taken all the images they wanted to
+2. recommend that a participant should be invited back to have more images taken
+3. communicate relevant information about the mammograms to image readers
+
+Achieving point 2 will allow us to negate the need for workarounds, and pass the responsibility of requesting appointments to take more images to mammographers rather than radiologists or admin staff. As the ones who were there when the images were taken, they are in the best possible position to give a professional opinion on whether this is necessary.
+
+
+## Straightforward options, separate outcomes
+
+One of our subject matter experts neatly summarised our goal when designing this part of the service:
+
+> We want partial to be an outcome of what they're recording, not the decision itself.
+
+To help meet their needs, we've given users various things to select within an 'Additional details' section on the [image details step of the mammogram workflow](/manage-breast-screening/2025/12/recording-images-taken/).
+
+![A screenshot of the workflow step where image details are recorded](provide-image-details.png)
+
+We have avoided using the term 'partial' at this stage. It comes with a lot of baggage and learned behaviour from the current system. The language used better reflects the information the mammographer needs to record.
+
+### Imperfect, but best possible images
+
+Despite every effort, mammographers can sometimes struggle to adequately position participants with physical or mental health conditions. For example, it could take two staff members more than 30 minutes to get a set of images taken, and they may still come out as substandard.
+
+To avoid these images being marked as a technical recall, they often write "best possible images" (or "BPI") on the participant sheet to communicate to the reader that a recall may not result in any clearer views.
+
+We've formalised this with a checkbox, allowing mammographers to record their recommendation that a recall would be unproductive.
+
+The 'Notes for reader' can be used to add extra context where necessary. Image readers will still have the option to request a technical recall, but clearly presenting the opinion of the mammographer will help them to weigh up the pros and cons of doing so.
+
+![The additional details section with the 'Imperfect, but best possible images' checkbox selected](additional-details.png)
+
+### Not all mammograms taken
+
+Selecting this option reveals several fields to collect more details.
+
+We ask users to explain why not all images were taken. Multiple reasons can be selected (to account for more complex scenarios), but there is just a single text box for a more detailed explanation when necessary.
+
+![Checkboxes for each of the reasons why a mammographer could not take all the images during a mammography](not-all-mammograms.png)
+
+The mammographer can then recommend whether or not the participant should be recalled.
+
+If they select yes, details can be added on timescales, such as allowing for any temporary injuries to be resolved.
+
+![Options asking whether a participant should be recalled, with 'yes' selected](to-be-recalled.png)
+
+If they select no, a 'partial mammography' is recorded. We're retaining the current language here to reflect the scenarios outlined in breast screening guidance, but as a result of choices made rather than as a primary option.
+
+![Options asking whether a participant should be recalled, with 'no' selected](partial-mammography.png)
+
+These new options aim to keep follow-up appointments distinct from 'technical recalls'. These are primarily used by image readers to request new images when the mammograms they are reviewing are not of sufficient quality.
+
+Technical recalls are something all breast screening offices are keen to avoid. They cause hassle, anxiety and extra radiation for participants, and undue pressure on busy screening teams. Local targets are in place that aim to keep technical recalls to a minimum.
+
+As an adjustment to how the screening process works, this is something we will need to put to national policy and clinical assurance teams. The initial support from the subject matter experts we work closely with suggests this is a long overdue improvement so we are hopeful these discussions will be positive.
+
+After continuing, the detail provided is included within the [appointment summary screen](/manage-breast-screening/2025/12/checking-information-collected-during-an-appointment/).
+
+![The image details section of the appointment summary showing the 'not all mammograms taken' reason](appointment-summary.png)
+
+## What happens next?
+
+As long as one image has been taken during the mammogram, the case will still be sent to image readers for review. Even when recalls are requested, images should still be reviewed within a 14 day window just in case a follow-up appointment does not take place.
+
+Radiologists will see the images, along with a summary of any issues encountered during the appointment.
+
+_Note: We are in the early stages of developing the image reading side of our service and these screenshots are early drafts._
+
+![The image reader opinion screen showing image quality details recorded by a mammographer](review-imperfect.png "Based on the 'Image quality' note, the reader could decide they are good enough for them to give an opinion ('Normal' or 'Recall for assessment'). Or, they could choose to request a technical recall for one or more images.")
+
+![The image reader opinion screen showing details on an incomplete mammography recorded by a mammographer](review-incomplete.png 'In this scenario, the reader is asked to give their opinion on the images the mammographer was able to take, in the knowledge that more images are due to be coming from another appointment. If they deem the images that were taken to not be of good enough quality, they could still request a technical recall and specify which mammograms need to be retaken when the participant returns.')
+
+## Initial user reactions
+
+We've been met with positive reactions when showing this solution to both mammographers and image readers. Comments include:
+
+> It's really useful to differentiate between whether it's a partial mammogram, full stop, or whether it's just pending and we can retry another day.
+
+> I think it is really good as a mammographer to be able to have that ability to put forward your clinical opinion after spending the time performing the mammogram. You'll probably have the most knowledge as to whether that can be redone.
+
+One area we need to look into is whether any of the options we've created could be overused.
+
+For example, if mammographers are able to initiate follow-up appointments themselves, it's feasible for them to select this option and avoid having to deal with tricky participants. All mammographers are assessed on how many technical recalls are flagged on the images they have taken, so we would need to add in performance data on how frequently they request recalls themselves.
+
+Another consideration is the messages that are sent to participants as a result of the selections made within the service. The additional structured information we're collecting can be tailored based on whether they are being recalled for more images, or for repeat images (due to a technical recall), or if their result is based on a partial mammography.
+
+And finally, we need to determine how these options relate to statuses within the service. Things get complicated when we are deciding how to handle appointments based on certain selections, particularly regarding keeping 'episodes' open when there is a potential follow-up appointment to be arranged.