Clinical care
21. There are comprehensive national guidelines that should be considered when managing a patient with an upper gastrointestinal (upper GI) bleed. An upper GI bleed is bleeding from the top part of the digestive tract, including the oesophagus (food pipe), stomach, and the first part of the small intestine.
22. The British Society of Gastroenterology has published a care bundle for people with an upper GI bleed. This care bundle includes 19 care recommendations, 9 of which were relevant to Mr A’s care. These include: • urgent physical observations using the National Early Warning Score (NEWS), which is a model that calculates illness severity and risk • patients who are unstable should be reviewed by critical care colleagues • patient risk should be calculated using the Glasgow-Blatchform Score (GBS), which is a clinical tool used to assess risk in patients with an upper GI bleed.
23. NICE guideline CG141 (upper GI bleed in over 16s) also recommends these investigations. It also outlines recommendations on the timing of endoscopy for patients with an acute upper GI bleed. This guidance says endoscopy should be offered within 24 hours for stable patients, or immediately if a patient is unstable. The guidelines add that resuscitation should happen first before progressing to an endoscopy.
24. Resuscitation in this context does not mean cardiopulmonary resuscitation (CPR) but, rather, resuscitation that aims to restore normal blood pressure and heart rate by providing rapid IV fluids. Fluid resuscitation is an emergency intervention that helps to stabilise patients with a low volume of blood and is a crucial step when treating serious bleeds and haemorrhages. Fluid resuscitation had been commenced when Mr A was in the ambulance, and continued in the hospital’s ED.
25. Endoscopy can also present a high risk of death for patients who are unstable. NICE Quality Statement GS38 (upper GI bleeding in adults) states that people who are unstable can be at high risk of serious complications if undergoing an endoscopy before their vital signs are normalised. This risk must be balanced with the risk of delaying the endoscopy.
26. The term ‘unstable’ is specifically defined by NICE Quality Standard QS38 as patients ‘with active bleeding whose blood pressure or pulse cannot be normalised or who need rapid IV fluids’. By this definition, Mr A was initially not stable enough for an endoscopy when he arrived at Hospital A. This is because his pulse was very high, and he needed IV fluids to stabilise him. This means that whilst the endoscopy should have been considered immediately, his vital signs needed to improve before this could happen.
27. Mr A arrived at the hospital at 8.12am and was triaged by a nurse. He was first assessed by a doctor at 9.18am. The Trust identified the signs of an upper GI bleed, assessed risk factors, commenced IV fluids, monitored physical observations, and determined the need for an endoscopy. It also completed a GBS screen, which indicated Mr A was at high risk due to his bleed.
28. Mr A’s physical observations were not stable when he arrived at the ED. His blood tests showed his haemoglobin levels, on admission, were 100g/L, meaning, in line with NICE guideline CG141, a blood transfusion was not clinically indicated when he was first admitted.
29. The Trust commenced an infusion of IV fluids at 8.45am, following the IV fluids given by the ambulance staff. The fluids given by the ambulance staff had significantly improved his blood pressure.
30. Once Mr A’s condition had been stabilised, the Trust began exploring the need for an endoscopy. By 11.36am it had identified there were no staff on site available to perform this procedure and agreed a transfer to Hospital B.
31. The Trust did not arrange for a review from Critical Care colleagues, however. Our Emergency Medicine Adviser explained this was most likely because he was aiming to be transferred to Hospital B’s Critical Care team, and so there was no need to consider an admission to Critical Care at Hospital A.
32. The initial investigations into Mr A’s presentation were in line with national guidelines. However, it took some time to arrange the endoscopy, and we have considered whether this should have happened sooner.
33. The need for an endoscopy was identified early, with a doctor documenting that they were exploring the need for this investigation with a consultant at 10.51am. During this discussion the ED doctor noted that the hospital may not be able to provide an endoscopy. They were advised to contact a member of staff who may be on site and able to perform the endoscopy.
34. At 11.36am the doctor documented there were no staff in the hospital who could perform the endoscopy. They discussed this again with a senior doctor and agreed Mr A should be transferred to another hospital for this procedure. The doctor contacted Hospital B to agree the transfer and documented the transfer would be arranged by the nursing staff. The nursing staff called the ambulance service to arrange this at 12.20pm.
35. Mr A’s condition was too unstable to have an endoscopy when he first arrived at the ED. Once he had been stabilised, the Trust was actively pursuing the endoscopy. In stable patients, NICE guideline CG141 says an endoscopy should take place within 24 hours. It acted quickly to agree a transfer to Hospital B when it became apparent that there was nobody on site at Hospital A who could undertake this intervention.
36. We recognise how distressing it must have been to learn that there was nobody on site to perform the endoscopy. Sadly, this is not uncommon in hospitals across England, and transfers for this reason are routinely undertaken. Tens of thousands of these transfers happen each year, which includes patients who are critically ill. We cannot reasonably conclude this amounts to service failure because this issue arises from capacity across the country, not failings on the part of the Trust. However, we understand why the family has such serious concerns about this.
37. With regards to whether the Trust delayed in transferring Mr A to Hospital B and whether it failed to advise the ambulance service of the urgency of his transfer, this relates to what happened after the decision to transfer Mr A’s was agreed from approximately 11.36am.
38. In considering what happened when arranging Mr A’s transfer, we must consider the different roles of the ambulance service and the Trust. The Trust was responsible for placing the 999 calls. The ambulance service was responsible for attending Hospital A within national target timeframes. We cannot consider the actions of the ambulance service because a complaint about this service has not been made to us. Our consideration focuses solely on what the Trust did and makes no comment on the appropriateness of the ambulance service’s response times.
39. NHS England’s guidance on inter-facility transfers (national framework for inter-facility transfers) is aimed at what ambulance services should do to triage 999 calls for urgent transfers. It is clear that the ambulance service is responsible for triaging the type of response (including the urgency of the response). The Trust does not decide how urgent the call is. However, in order to do this accurately, the ambulance service relies on the clinical information being accurately communicated to it.
40. We have considered whether the information provided to the ambulance service aligned with the clinical picture in the ED, and whether this aligned with the Trust’s policy on transfers between hospitals. To do this, we have listened to the 999 calls made from the ED.
41. The first 999 call was made at 12.20pm. At this point, Mr A’s condition was stable. The Trust’s nurse stated Mr A had a GI bleed and was ‘actively bleeding’. When the call handler asked if the Trust needed an ‘immediate lifesaving intervention’ the staff member answered ‘no’. This was correct because Mr A condition was stable at that time.
42. The call handler then asked whether there was a need for an immediate intervention that could not be carried out at the current facility and the patient was at immediate risk of death, the member of staff answered ‘no’. This was correct because Mr A was not, at that time, at immediate risk of dying. His condition was stable.
43. The call handler then asked if Mr A required additional clinical management at a different hospital, and the member of staff answered ‘yes’. This was correct, Mr A did require a procedure at Hospital B.
44. When the call handler asked if the member of the Trust’s staff wanted anything further adding into the notes, they replied ‘no’. However, when the call handler repeated the information back to the nurse, the nurse asked the call handler to make sure they documented Mr A was still actively bleeding and said the transfer was ‘quite important’. The call was categorised as a Category 3 inter-facility transfer.
45. Our Emergency Medicine Adviser confirmed it was clinically accurate to answer ‘no’ to the questions about whether Mr A needed an immediate life-saving intervention, and whether he was immediate risk of death. This is because, at this time, Mr A was stable and did not require care under the Major Haemorrhage Protocol.
46. This approach also aligned with the Trust’s policy on transfers between hospitals. A Category 3 inter-facility transfer is for patients requiring an intervention or investigation not available at current location. There was no clinical reason, at that stage, to challenge the categorisation made by the ambulance service.
47. A second 999 call was placed by the Trust at 2.52pm, immediately after the Major Haemorrhage Protocol was activated. This call was placed by a doctor who explained Mr A was ‘deteriorating’ and they were ‘keen’ to get them transferred to the other hospital ‘as soon as possible’. The call handler advised the ambulance service’s computer system had gone down and they were taking paper notes. They explained they could not see where the previously requested ambulance was up to.
48. The doctor pushed for the ambulance service to chase up the request. The call handler agreed to raise a new job because the computer system was down and it could not find the previous request. The call handler also did not have access to the triage questions for the inter-hospital questions because these are generated by computer software.
49. The doctor told the ambulance call handler that Mr A was vomiting blood, his heart rate was 140bpm, and he was becoming more unstable. They stated he could only get the treatment he needed at Hospital B. The doctor confirmed he was bleeding seriously and this was an emergency. The call was coded as a Category 1 inter-facility transfer. This is the highest level of priority. The ambulance arrived within nine minutes.
50. Our Emergency Medicine Adviser listened to the 999 call and explained the clinical information was appropriately relayed to the ambulance service. Following this 999 call, the request was re-categorised as the highest level of priority. This means Mr A’s need for an ambulance would have bypassed all other emergency calls. There is nothing further the doctor could have added at that time that would change the priority of this call because the call was now being handled as the most pressing emergency.
51. The ambulance service’s computer systems going down may have had an impact on how long it took for the ambulance to arrive to facilitate the transfer. The ambulance service’s investigation report indicates the call was responded to within three hours and 17 minutes in total. The target time for a Category 3 call is two hours, though many ambulance Trusts have an average response time of three to four hours due to demand exceeding resources.
52. When Mr A deteriorated and became critically unwell, a doctor promptly contacted the ambulance service again and requested an emergency response. The call was escalated to a Category 1 call, and the ambulance was in attendance within nine minutes. The target time for a Category 1 call is within 15 minutes, with an average response time of eight minutes.
53. After Mr A deteriorated, the Trust could have done more to expedite this transfer, and its actions were in line with national guidelines and its policy on inter-hospital transfers.
54. We recognise the incredibly tragic and distressing circumstances surrounding Mr A’s death. His family has lost a much-loved husband and son, and this has been front of mind when considering their complaint.
Complaint handling 55. Our NHS Complaint Standards (our Standards) set out what good complaint handling looks like in NHS services and aims to support organisations to deliver this. In line with our Standards, the Trust should have:
• given a fair and balanced account of what happened, based on established facts • been open and honest where things may have gone wrong • ensured the process was responsive to the needs of Mr A’s family.
56. Mrs A raised a complaint about her son’s care on 12 February 2024. There appears to have been some confusion around whether the complaint was received by the Trust and Mr A’s widow sent the complaint again on 14 February. The Trust acknowledged this the same day.
57. On 26 June the Trust issued its first complaint response. We have reviewed the accuracy of the explanations of when care was provided. The response was mostly factually accurate as it aligned with Mr A’s medical records. There is some information that did not align with the medical records:
• the Trust said his observations were taken half-hourly. This was correct for most of the morning; however, these happened hourly between 11.38am and 2.15pm • the time when the Trust says it established there was no specialist cannot clearly be corroborated from the medical records. At 10.50am the Trust documented a plan to establish if the there was a specialist on site. The next entry is at 11.36am, documenting a number of discussions with colleagues and the decision to transfer to the second hospital. This may have given the impression of a longer period of time between the decision being made and contacting the ambulance service, though not a significant period of time • the ambulance was a Category 3 following the first 999 call, rather than a Category 2, as outlined in the complaint response. A Category 3 allows much longer for an ambulance to attend. A Category 2 call has a target time of between 18 and 40 minutes. A Category 3 has a target time of within two hours. This likely gave the family the impression of a much longer delay in comparison to the target response times • a minor inconsistency in the time when the ambulance arrived at Hospital B, referenced as 4.30pm at one point, then referenced as 4.42pm later in the response. This had no impact on the otherwise accurate account of the clinical narrative.
58. We have found the majority of the first complaint response was factually accurate and, overall, the narrative of the events which took place was correct. The Trust should have paid more attention to ensuring the category of the ambulance service’s response was correct. It also should have paid more attention to accurately relaying when it decided Mr A should be transferred. This does not amount to service failure, but this inconsistency likely gave the impression of a greater delay and was, understandably, concerning for Mr A’s family on receiving this response.
59. We have seen no inconsistencies between Mr A’s medical records and the meeting transcript and final response. There are also no inconsistencies between these responses and the first response. The areas that were not related to the clinical narrative, for example the need to query Mr A’s alcohol use, were also factually correct. It is well-established that alcohol, particularly prolonged use, damages the mucosal lining of the upper gastrointestinal tract and can cause bleeding.
60. With regards to whether the Trust initially refused to meet with the family, this is difficult to reach a view on. Whilst there is no specific requirement for a Trust to offer a meeting, our Standards require Trusts to respond to the needs of complainants flexibly. This would reasonably include offering to meet with the complainant(s) to discuss outstanding concerns.
61. The Trust’s first complaint response was issued on 26 June 2024. The first documented reference to a meeting was on 30 August, in an email to Mrs A. The email of 30 August indicates the Trust believed the family wanted a written response rather than a meeting. The email says this agreement arose from a telephone conversation but does not say when this took place. There are no records of emails between the Trust and Mrs A between 26 June and 30 August.
62. The next email is dated 19 November and references a further telephone call about the meeting. It offers dates for a meeting to take place.
63. We cannot account for what happened between 26 June and 30 August, and between 30 August and 19 November. It appears the communication took place via telephone, and the Trust did not retain a written record of these calls. With hindsight, it may have been advantageous to maintain a record of telephone calls by emailing Mrs A about the actions discussed and agreed on these calls. That said, not doing so does not fall short of our Standards. We also acknowledge Mrs A’s view that this meeting was initially refused.
64. We are unable to reach a view on the balance of probabilities whether the Trust initially refused to meet with Mrs A and her family.
65. We have not upheld this complaint. We recognise the circumstances of this complaint were incredibly tragic and distressing. It is understandable that Mr A’ family had such serious concerns about the care he received, especially given the time that elapsed between identifying the need for an endoscopy and the ambulance arriving at Hospital A.