Gemma Marshall

Dear Mr Eccleston,

Royal College of Radiologists Response to Regulation 28: Prevention of Future Deaths report issued on 2nd January 2025 in relation to the death of Ms Gemma Suzanne Marshall. I was very sorry to read about the death of Ms Marshall and I would like to express my deepest condolences to Ms Marshall’s family. We take the matters raised in your report very seriously and I hope this reply will be helpful in outlining how we are committed to learning from the report and supporting our members and Fellows to develop and maintain excellent medical care. The Royal College of Radiologists (RCR) is a charity which works with our members and Fellows to improve medical care across the specialties of Clinical Radiology and Clinical Oncology. We promote excellence in professional practice within our specialties and we produce a range of publications, including standards for the delivery of high-quality radiology services. Annually, the RCR publishes a Clinical radiology census report and a recurring theme each year is the shortfall in UK radiologists and the widening gap between the capacity of our radiology workforce and the escalating demand for imaging services. The data in our census is intended to inform Government, NHS and trust and health board leaders so that they can take meaningful action to grow and support the radiology workforce. Managing the imaging of a patient with abdominal pain, including a patient who may have a complication of a gastric band, is within the scope of the Clinical Radiology Specialty Training Curriculum although gastric bands are not specifically listed in the curriculum document. This is because it is a rare condition, and it is not possible to provide a comprehensive list of every rare medical condition or implantable device which might be encountered or which might cause harm to a patient. Similarly, while the complications of gastric bands are also well covered by several educational papers published in the radiology literature, it is not an

area which is covered by bespoke guidance from the RCR, nor from any of the affiliated Special Interest Groups in radiology within the UK. I am also not aware of any bespoke guidance from any other similar organisation in other countries around the world. This does not reflect a lack of importance but because it is not possible to issue detailed guidance covering every individual condition and clinical scenario which might arise. The RCR produces guidance on Standards for interpreting and reporting of imaging investigations which recommends that reporters should only work within their scope of practice and competence. All radiologists have a mixture of general and subspecialty radiology skills although their area and degree of specialism will vary. Most hospitals have a general diagnostic on-call rota that is covered by a wide variety of radiologists, all of whom are competent at reporting general imaging (like a CT of a patient’s abdomen in the context of abdominal pain, as in this case) but who also have one of many specialist areas of practice. In the on-call setting radiologists will therefore more often be reporting scans that are outside their specialist area of interest than within it. Given the current state of staffing and IT provision in the NHS it is not currently possible for all emergency imaging to be reported by a local radiologist with sub-specialty expertise in the relevant area. This is as true of other sub-specialty areas of radiology (chest, neurology, musculoskeletal, gynaecology, paediatrics etc) as it is of abdominal imaging. The fact that the reporting radiologist had a specialist interest in a different area does make it inevitable they will be less expert at identifying very rare pathology in the abdomen. If the treating team have ongoing concerns, then in most imaging departments there is typically the opportunity to discuss the imaging with a local radiologist with greater relevant subspecialist knowledge during the working week. Some places may have an informal mechanism to seek immediate peer support from colleagues with different specialist interests but formal arrangements out of hours tend to map to referral pathways for the patient themselves and are variably available. The RCR supports the development of clinical networks, and it is conceivable that in future an integrated network of linked hospitals could provide specialty reporting for a group of hospitals. However, the current IT infrastructure available remains a significant barrier in most places across the UK. Although imaging is very important patients must always be treated according to their clinical condition and not on the presumption the imaging is infallible. The RCR’s Standards for interpreting and reporting of imaging investigations guidance makes clear the importance of communication, including the communication of uncertainty where necessary. It outlines that a radiology opinion is informed by any given clinical history and in turn should guide further clinical assessment and management. A specific difficulty around imaging gastric bands is that this has become a less common procedure, with other procedures like gastric sleeve operations now being more frequently used. There has also been centralisation of NHS bariatric services so many receiving clinicians and also radiologists will not have regular experience of assessing these cases, even if they are a specialist gastrointestinal radiologist as some referrals may go beyond even a regional centre.

Gastric bands do sometimes move out of position. Slippage of a gastric band, however, is not in and of itself a surgical emergency unless accompanied by clinical features which indicate serious complications. These features are more usually obstructive symptoms rather than symptoms of gastric infarction. Given the circumstances of Ms Marshall’s death and because she was considered well enough to be self-caring and to be discharged home at the time of the CT, even if it had been recognised that the gastric band had slipped on the CT, surgical intervention may well not have taken place within the two-day time window between her initial presentation and subsequent death. Therefore, tragically, interpretation of the CT may not have been the only factor that required to change in order for Ms Marshall’s death to have been prevented. The use of outsourcing companies is becoming increasingly frequent in the UK, including out of hours. There are many reasons for this including to purchase additional reporting capacity when a trust does not have this locally or where it wishes to create an environment felt more likely to retain the few radiologists it currently has. There are also many reasons relating to the relative working environments within the NHS and within an outsourcing company which are causing radiologists to choose to do such work. A properly staffed and funded radiology service within an acute hospital is vital to support patient care but outsourcing is commonly used as a supplementary provision. That her scan was outsourced is unlikely to have made the misinterpretation of Ms Marshall’s CT any more likely. If the scan had been reported in-house by a local radiologist, it is no more likely that it would have been allocated to have been reported by a radiologist with specialist expertise in abdominal imaging and equally likely that the scan would have been reported by an inhouse radiologist with specialty expertise in some other non-abdominal area. Many outsourcing companies will report a much larger volume of work than a single trust and although it did not happen on this occasion, they may be in a better position than a single trust to attempt to allocate work by special interest area in an emergency setting because they will have multiple radiologists reporting simultaneously rather than a single person in each place individually “on call” only for their trust. There may be less clinical contact with a radiologist in an outsourcing environment and they may not have an ongoing clinical working relationship with referrers, but the expected standard of reporting is not lower and the large majority of radiologists doing this work are NHS consultant radiologists who additionally do some outsourcing work. It is internationally accepted that even the best trained radiologists working in an optimal environment will occasionally issue reports which are subsequently shown not to be completely accurate. Discrepancy is a complex area in radiology and there are many potential causes which are not all due to lack of observation or error in interpretation. At least 4% of radiology reports, of which there are many millions in the UK every year, are discrepant to some extent. It is therefore important that radiology is interpreted in clinical context and with appropriate safeguards. While ideally a service would be constructed to minimise the number of such reports they can never be fully eliminated and RCR guidance places emphasis on learning from these events as described in the Standards for radiology events and learning (REAL) meetings and allied structure around the REAL process. The RCR does publish educational material including anonymised cases and I have asked the

relevant editor to consider this case theme and signpost a suitable anonymised CT (from a different patient). Most outsourcing contracts will include some specification of who will be acceptable to interpret imaging, within what timescale and there is also typically some quality assurance written into the agreement. Outsourcing radiology companies typically have greater double- reporting and better metrics on the radiologists who work for them than NHS trusts. There exists a tension that NHS departments require a minimum radiology provision to keep the service running but that the more challenged a department becomes the less attractive it is as a place of work which further increases any requirement to outsource. There is limited data on whether outsourced reports actually do have a higher level of discrepancy, and this is a contentious area with strong views on either side. When examined in the National Emergency Laparotomy Audit (most recent data on this metric was up to November 2020) there was an approximately 2% difference between in house and outsourced reports within the context of ten-fold higher variation in rate between individual hospitals. I am grateful to you for bringing these matters of concern to our attention and for giving us the opportunity to respond. Once again, I express my deepest condolences to Ms Marshall’s family and loved ones.