Investigation report: Safety risk of air embolus associated with central venous catheters used for haemodialysis treatment

This investigation aims to improve patient safety by supporting healthcare staff in the safe use of central venous catheters to access a patient’s blood supply. Specifically, it looks into the use of tunnelled haemodialysis central venous catheters, which are a type of central line.

A tunnelled dialysis catheter is a hollow plastic tube which is used with patients who need long term haemodialysis, which is treatment on a dialysis machine. It is usually inserted into a vein in the neck and sits in a large vein in a patient’s chest. The end of the tube that sits outside the patient’s body is separated into two lumens (tubes), each with a clamp on it and a cap on the end. One of the lumens is for the blood to flow out of the body and through the dialysis machine, where it is filtered and returned to the body via the second lumen. This process happens continuously during haemodialysis treatment.

In this investigation, the focus is on permanent haemodialysis catheters, although similar risks of an air embolism also apply to a temporary haemodialysis catheter. An air embolism occurs when air bubbles enter a patient’s vein and block it; this can cause serious harm or death. The uncapping and unclamping of a haemodialysis catheter (without attaching a syringe), by a user not familiar with the medical device, task or the equipment required, may allow air to get into a person’s bloodstream, resulting in an air embolus. A haemodialysis catheter has a wide bore (internal diameter) compared to other types of central line, to allow a high volume of blood to flow through. It is this wide bore that increases the risk of a large air embolus, compared with narrower catheters.

The investigation used a real patient safety incident, referred to as ‘the reference event’, to examine aspects of the management of, and access to, haemodialysis catheters to take blood samples. The reference event involved Joan, who had a cardiac arrest caused by an air embolus after her haemodialysis catheter was uncapped, unclamped, and left open to air. This took place during a procedure to take blood culture samples to test for a possible line infection (where bacteria or viruses enter the bloodstream).

This investigation’s findings, safety recommendations and safety observations aim to prevent the future occurrence of an air embolus following uncapped and unclamped haemodialysis catheters, and to improve care for patients across the NHS.

Joan was 75 years old at the time of the incident. She was taken to hospital from a nurse-led satellite haemodialysis unit, where she had been receiving regular treatment. Joan reported feeling unwell and was found to have low blood pressure and a suspected diabetic toe ulcer infection.

A plan was made in the emergency department to start Joan on antibiotics, and she was admitted to hospital for ongoing treatment. Following a short stay on a medical assessment unit, she was transferred to an older persons’ ward for further investigations and treatment.

Microbiologists informed the clinicians caring for Joan that bacteria had grown in her initial blood culture samples which was unlikely to have arisen from her toe ulcer. Microbiologists advised taking more blood cultures from Joan’s haemodialysis catheter to exclude an infection on her line and recommended a change in the type of antibiotic treatment.

The following day, Joan’s care was reviewed by a multidisciplinary team on the older persons’ ward, and the task of taking the blood cultures was discussed. A medical student asked if they could carry out the task, under the supervision of a junior doctor who had volunteered to take the blood cultures. This was agreed by a specialist trainee doctor, after a run-through of the steps involved in the task.

The junior doctor and medical student approached Joan to discuss taking the blood cultures. They started unwrapping the gauze covering over the end of the haemodialysis catheter, and realised it looked different to other central lines they had seen previously. After an unsuccessful attempt to seek nursing support/guidance, they decided to continue with the task. The cap of the line was removed and while the medical student was cleaning the port, the junior doctor unclamped the line. After approximately 4 minutes, Joan collapsed and suffered a cardiac arrest (her heart stopped). She was successfully resuscitated and transferred to the intensive care unit but died 2 days later from multi-organ failure and sepsis.

A subsequent coroner’s inquest concluded that a possible contributory cause of Joan’s death was the cardiac arrest, which was ‘probably secondary to air embolus’.

The Department of Health and Social Care notified HSIB of the incident relating to an ‘uncapped central line’, after it had been issued with a Prevention of Future Deaths report by HM Coroner. The matters of concern relating to this investigation included medical school training regarding taking blood samples from central lines; the assessment of doctors’ competencies to undertake procedures; and guidance on the safety risks when taking blood from central lines.

HSIB launched an initial scoping investigation in August 2021, which determined that the patient safety concern met the criteria for investigation. HSIB’s Chief Investigator authorised a national investigation.

A national investigation was undertaken to explore the factors that affect the ability of staff to safely access haemodialysis catheters. The focus was on this specific type of central line because of the increased safety risk of an air embolus associated with a wider bore line.

The national investigation focused on:

• Identifying the factors within the healthcare system that influence awareness and understanding of patient safety risks associated with the management of haemodialysis central venous catheters.
• Evaluating the current mechanisms for the mitigation of patient safety risks associated with haemodialysis central venous catheters.

The investigation found there are currently no long-term haemodialysis catheters on the UK market, or being developed, that have integrated ‘safety-valves’.

Manual clamps on haemodialysis catheters rely on people ensuring that the clamp is on before accessing the haemodialysis catheter ports and do not mitigate against design-induced error.

A review of patient safety risks associated with other haemodialysis devices (for example, fistulas) showed several mitigations which are not consistently used for haemodialysis catheters. These include a coloured patient wristband, line labelling, alert cards being carried by the patient and educating patients/family members.

The training and education of all grades of medical staff has not been consistent in relation to the risks of catheter-related air embolism.

There is currently no recognised national training or national training guidelines regarding the safe access of haemodialysis catheters.

Incidents appear to be under-reported to the Medicines and Healthcare products Regulatory Agency, due to misconceptions about ‘human error’ being the cause, rather than the design of the equipment.

The Medicines and Healthcare products Regulatory Agency, in partnership with NHS England, have explored integrated incident reporting system possibilities. While a recent funding bid to support full development for an in-service solution was unsuccessful, the organisations are committed to drive this project forward.

There is a general lack of literature on, and knowledge of, catheter-related air embolism in relation to access when the catheter is in situ (in position), rather than during insertion or removal of the catheter.

1.1.1 This investigation aims to improve patient safety by supporting healthcare staff in the safe use of central lines to access a patient’s blood supply to provide treatment or take samples. Specifically, it focuses on the use of haemodialysis central venous catheters (referred to in this report as haemodialysis catheters). This is because of the increased risk of an air embolism (air bubbles entering a patient’s vein and blocking it) associated with this type of wider bore line. However, there is a risk of air emboli with other types of central venous catheters if they are left open to the air.

1.1.2 The safety risk was brought to HSIB’s attention by the Department of Health and Social Care following the publication of a coroner’s report about the reference event, which involved an ‘uncapped and unclamped central line’ (figure 1). The reference event is detailed in section 2 of this report.

1.1.3 The correct sequence of uncapping/unclamping when carrying out a procedure such as taking blood samples is that the two lumens will remain clamped, with the connectors in situ, until the connector is removed to attach a syringe. Once the syringe is attached, the clamp can be released.

1.1.4 The following paragraphs explain dialysis treatment, haemodialysis, catheters and air emboli in more detail, and the prevalence of the associated safety risks.

1.2.1 Dialysis is used to treat people with very poor kidney function. It cleans the blood and filters out unwanted substances and fluids, which could build up to dangerous levels in the body and eventually be fatal. There are two main types of dialysis treatment: haemodialysis (which is relevant to this investigation) and peritoneal dialysis.

Haemodialysis treatment

1.2.2 Haemodialysis treatment involves passing a patient’s blood through a dialysis filter. A patient will usually need treatment three to four times a week for around 4 hours each time. Access to the bloodstream is needed so that blood can be taken from the patient, passed through the haemodialysis machine and back into the patient continuously throughout their treatment. This ‘access’ can take three different forms. These are described below in order of clinical preference (related to achieving the optimum blood flow for dialysis in addition to minimising the risks to the patient of accessing the bloodstream).

1.2.3 The optimum approach to access is the formation of an arteriovenous fistula (AVF). An AVF is created by joining together an artery and vein, usually in the patient’s non-dominant arm, during a minor surgical procedure, in advance of the patient needing dialysis. After some weeks, the fistula will ‘mature’ and blood flow through it will be sufficient to support dialysis treatment. Access is achieved by inserting two large needles to create a circuit of blood from the patient through the haemodialysis machine and back into the patient continuously throughout their treatment.

1.2.4 Where it is not technically possible to create a fistula, a graft of synthetic tissue can be used to connect an artery and vein in a similar way. This is known as an arteriovenous graft (AVG). Needles are used as described above.

Haemodialysis catheters

1.2.5 If neither AVF or AVG insertion is technically possible, or dialysis treatment is required urgently, a haemodialysis catheter may be used for access. This is usually inserted into the jugular vein in a patient’s neck and is situated in the chest, or sometimes in the groin (femoral). It is secured using stitches in the neck (for a temporary line usually used for a number of days) or on the chest wall (if the line is likely to be needed for weeks or months).

1.2.6 During haemodialysis treatment, one line (red) takes blood from the body to be ‘cleaned’, and the blood is returned to the body through the second (blue) line (see figure 2). On each line is a clamp and at the end of the lumen (tube) are ports (openings), with removable caps/connectors (figure 1). The clamps are opened to access the patient’s blood supply during haemodialysis or to access for another reason, for example to take a blood sample. The line should not be unclamped until connected to the dialysis machine, or a syringe attached (if taking blood), because of the risk of an air embolus.

Peritoneal dialysis treatment

1.2.7 Peritoneal dialysis uses the lining of a patient’s abdomen (the peritoneum) as a filter for the removal of dangerous substances from the blood. Access is via a tube inserted through the abdominal wall into the peritoneal space.

1.3.1 A venous air embolism (which is the type that occurred in the reference event outlined in section 2 of this report) occurs when air enters the circulation and is carried by blood flow to the right side of a person’s heart. It then moves to the pulmonary artery, which is the blood vessel that carries blood from the right side of the heart to the lungs for oxygenation. Here the air can block the flow of blood to the lungs, causing strain on the right side of the heart. This also compromises the return of blood from the person’s lungs to the left side of the heart. Large volumes of air over a short period can cause catastrophic right-sided heart failure and cardiac arrest (when the heart stops beating), which may be fatal (Intensive Care Society, 2019).

1.3.2 Toung et al (2001) indicate that the volume of air likely to be fatal is 200 ml to 300 ml. However, it was unclear whether 200 ml represents the minimum volume of air considered to be lethal to healthy adult humans. The investigation was told by a consultant physician that a smaller volume in a bubble, should it enter a coronary artery, might in theory have fatal outcomes.

1.3.3 According to Wong et al (2017), for a venous air embolism to form ‘three essential elements’ must be in place:

• a source of air (the atmosphere)
• the connection between a vein and the source of air (the catheter)
• a pressure gradient that allows air to enter the person’s vein (when the pressure in the heart is lower than the atmosphere). This may be caused by breathing in deeply; a decrease in the volume of circulating blood in the body; and a semi-upright, seated, or upright posture, which leads to pressure changes in the venous system because of the effect of gravity.

1.4.1 Almost 30,000 people in the UK are on dialysis, and there are over one million treatments a year using dialysis catheters in the UK.

1.4.2 Air embolism when using haemodialysis catheters is considered to be a rare but ‘avoidable’ incident. A review of the research literature associated with this risk found that studies have focused on air embolism during the placement and removal of haemodialysis catheters, rather than on the management of catheters while they are in place (in situ), which is also a safety risk.

1.4.3 There are different types of vascular access for haemodialysis. Most patients on chronic (long-term) haemodialysis treatment have a fistula in situ; this carries a lower risk of an air embolism. Not all hospital trusts provide dialysis, and where they do not, staff are less likely to have the required skills, competence, and correct equipment for management of haemodialysis catheters.

1.4.4 Between April 2017 and November 2022 there were 14 air embolus incidents reported on NHS England’s serious incident reporting system (the Strategic Executive Information System) where central lines had been found uncapped and unclamped. These included three occurrences (including the event examined in this report) relating specifically to haemodialysis catheters. During the same period, incident data from the national database of patient safety incident reports (the National Reporting and Learning System) identified 70 occurrences where central lines had been found uncapped and unclamped. These included six occurrences relating specifically to haemodialysis catheters, in which there was a risk of air embolism.

1.5.1 For readability, haemodialysis central venous catheters (CVCs) are referred to as haemodialysis catheters in this report. ‘Central line’ and ‘line’ may also be used, as haemodialysis catheters are a type of central line.

1.5.2 The term ‘design-induced’ error refers to use-related design issues, as outlined in Medicines and Healthcare products Regulatory Agency (2021b).

1.5.3 Figure 3 shows the medical training programmes in the UK, relevant to the investigation. This is the terminology used to describe the medical staff posts in the report.

3.1.1 This section explores the communication of the task of taking blood cultures from the haemodialysis catheter from microbiologists to clinicians at both Acute Trust 1 and Acute Trust 2, working in parallel. In particular, the divided responsibilities are explored between the two Trusts: Trust 1 with a specialised renal facility and Trust 2 without.

3.1.2 The service level agreement between the two acute trusts included Acute Trust 1 providing a renal consultant service to patients at Acute Trust 2, including those at Hospital 2. Joan therefore received the specialist renal support while at Hospital 2. The entry in the medical record on Day 3 at 15:43 hours by the renal consultant from Hospital 1 summarised his virtual review of Joan. It included a request to discuss the positive blood cultures with the diabetes foot team, to decide whether Joan needed a course of intravenous antibiotics.

3.1.3 Following the renal consultant’s virtual review, a junior doctor on the acute medical unit (AMU) contacted microbiology service by email to seek advice and added an entry in the record to reflect they had done this. The renal consultant told the investigation that they had not received any relevant communication to inform them of this. When covering Acute Trust 2, the renal consultant had access to view and add new entries to Trust 2’s electronic patient record, but only while on site.

3.1.4 On Day 4 at approximately 10:30 hours, Joan was seen for the first time by the acting consultant physician during the routine ward round on the older persons’ ward. The documented history of her ‘complaints/issues’ focused on her infected toe, and the overall impression was an infected toe ulcer, with possible osteomyelitis (infection of the bone). The plan outlined in the medical records to have a discussion with the diabetes team about the relevance of the blood culture test result and Joan’s infected toe appears to follow from the request made by the renal consultant the day before (see 3.1.2). There is no reference by the acting consultant physician to repeating the blood cultures, as the focus was on Joan’s infected toe and not any alternative sources of infection.

3.1.5 The discussion between the renal consultant and consultant microbiologist 2 (Hospital 1) during the morning of Day 4 focused on the growth of the bacteria staphylococcus hominis. It was considered by them to be unusual for this to be causing a toe ulcer infection, and so the source of the infection needed to be identified. A wound swab of the infected toe ulcer had grown a mixture of different bacteria, one of which was a type that could cause serious infections.

3.1.6 Consultant microbiologist 2 suggested broadening antibiotic cover with another medication and to repeat blood cultures in view of the haemodialysis catheter being in situ. The renal consultant’s plan reflected this advice: ‘Please repeat blood cultures. Dialysis at [Hospital 1] today then to return to [Hospital 2].’ There was no apparent linkage between these two tasks. There is also no evidence of communication to Hospital 1 about repeating the blood cultures during dialysis treatment or to the acting consultant physician about the plan, although it is documented in Joan’s record.

3.1.7 During the afternoon of Day 4, the junior doctor on the AMU copied the email response from Acute Trust 2 into Joan’s electronic record. This was the advice from consultant microbiologist 1 (Hospital 2) about repeating cultures from the lumens (tubes) of the haemodialysis catheter, in view of concerns about line infection. The junior doctor did not appear to have seen the renal consultant’s entry in the electronic patient record at 11:20 hours. This advice was similar in terms of repeating the blood cultures, but the antibiotic treatment being recommended differed between the two microbiologists.

3.1.8 The communication of the task to repeat blood cultures from the haemodialysis catheter therefore originated from two separate sources. The first was from consultant microbiologist 2 at Hospital 1, which was communicated to the renal consultant during a discussion. The second source was the email from the microbiology registrar at Acute Trust 2 to the junior doctor at Hospital 2. There is no evidence of any correlation between the two sources of information, or direct communication of the task between the recipients of the advice and those responsible for completing the task.

3.1.9 A review of the chronological entries in the electronic record indicates that by the time the microbiology advice was added by the junior doctor on the AMU, Joan had been transferred from the older persons’ ward to Hospital 1 for dialysis treatment. The blood cultures were not repeated at the time of dialysis treatment at Hospital 1, since the task had not been communicated to the dialysis unit staff. Additionally, Joan’s electronic patient record from Acute Trust 2 could not be accessed by the dialysis unit staff at Acute Trust 1; it could only be accessed by the consultant staff rostered to provide a renal service.

3.1.10 The investigation was told that there was an opportunity for the blood cultures to have been taken from the dialysis treatment at Hospital 1, by specifically trained and competent dialysis nurses, in a dedicated renal facility with the required equipment.

3.1.11 The investigation found that it was not clearly defined who should carry out the task of taking blood cultures from Joan’s haemodialysis catheter and where it should occur. This was because the task was communicated by microbiologists to different specialties at different Trusts, with no overall responsibility assigned and no oversight. The task was at the time considered in isolation from the dialysis treatment, and from the renal consultant’s involvement.

3.2.1 The investigation was told that there was no urgency for the blood cultures to be taken on Day 5, from a clinical perspective. The investigation found no evidence that the medical staff caring for Joan questioned whether another team should carry out the task, either on the older persons’ ward or in a different location. The staff involved in Joan’s care reflected to the investigation that the events that occurred made sense to staff at the time, given the information, knowledge, and expertise that they believed they had.

3.2.2 The ward-level communication of the blood cultures task originated from the board round. The board round involved a discussion between the clinical team, not involving direct patient contact. It took place in a room away from the main ward area, using computer screens to access patient details. Included in the board round on this day was the foundation year 1 doctor (FY1), who was on the ward for a ‘shadowing’ day, to observe and support the team in a supernumerary capacity (that is, in addition to the required staffing cover). This was so that they could familiarise themselves with the environment and learn about the next clinical rotation that they would move on to as part of their medical training.

3.2.3 During the board round, the acting consultant physician verbally communicated that “line cultures” [the haemodialysis catheter] needed to be taken from Joan. The acting consultant physician told the investigation: “I did not delegate it to any particular one. It was not in my wildest dream that [the FY1] was going to do this test. But I anticipated that [the GP specialist trainee 2 (ST2)] was going to do it.”

3.2.4 During the ward round, the ST2 accessed Joan’s electronic patient record. The record was reflective of a detailed review being undertaken before and during the ward round. Of note, the advice from microbiology at Acute Trust 2 was included (see 3.1.1) and a request to ‘send blood cultures from all line ports [lumens]’. It is evident that the ST2 had seen the emailed advice which the junior doctor on the AMU had copied into Joan’s electronic record.

3.2.5 In accordance with the advice from microbiology, the acting consultant physician communicated the need ‘to send some line cultures to exclude the line infection’. This was documented by the ST2 during the ward round as ‘send blood cultures from all line ports [lumens]’. The medical student overheard the discussion and asked if they could take the blood samples. Then followed a discussion which involved the FY1, who had volunteered to carry out tasks delegated by the medical team while the ward round was in progress.

3.2.6 The indirect communication of the blood cultures task therefore resulted in it being attempted by the most junior members of the team, without the knowledge of the safety risks associated with haemodialysis catheters.

3.3.1 This section explores the understanding of the task of taking the blood cultures and how this was interpreted by the different members of the medical team involved.

3.3.2 The ST2 told the FY1 and medical student that the blood cultures “weren’t normal ones, they weren’t just like peripheral ones …”. The medical student had received training in taking blood cultures, but not from central lines. The medical student’s recollection was that once they found out that the blood cultures needed to be taken from a central line, they wanted to continue to complete the competency requirement.

3.3.3 When the FY1 realised that the blood cultures were to be taken from a ‘central line’, their perception was that the principles of taking the blood cultures was the same across all types of central lines. They reflected to the investigation: “I had slight niggles of doubt but you go through that thing and you think it will be fine, it’s just taking bloods …”

3.3.4 The FY1’s experience included taking blood cultures from different types of central line during their previous surgical rotation. The FY1 had received nurse-led training for taking bloods from different types of central lines during this rotation. Before attempting the task, the FY1 was not aware that Joan’s haemodialysis catheter differed from the other central lines they had accessed. They were also unaware of the increased risk of air emboli from uncapped and unclamped wide bore haemodialysis catheters (see 1.1.2).

3.3.5 The FY1’s recollection was that the ST2 gave them permission to take the blood cultures, following the communication of the plan. The FY1 described the task to the ST2 (and the medical student) based on their previous practical experience of different types of central lines, as follows:

3.3.6 The investigation learned that the task as described by the FY1 (as they understood it) was not in line with Hospital 1’s policy, which included taking blood cultures from a haemodialysis catheter. This was in relation to:

• point 2, unclamping the line, before first attaching a syringe
• point 3, discarding the initial blood collected in the syringe
• point 5, ‘locking’ the line with a different solution (to prevent infection and blood clots) following the saline flush.

This is discussed further in section 3.5.8.

3.3.7 Following a discussion about the task, the ST2 gave the FY1 permission to take the blood cultures. In relation to the medical student, the ST1 said: “… so he had gone with [FY1] to watch [them] do it …” The ST2’s perception was that the medical student was going to observe the FY1 carrying out the task but would not be undertaking it. The medical student’s recollection was that the FY1 had agreed to supervise them performing the procedure and they were told by the FY1 what to do. This also appeared to be the FY1’s perception.

3.3.8 The FY1 told the investigation that from their perspective, “you go and just try it out, that happened like a lot in surgery, and it was okay I thought I was helping the SHO [ST2] out”. The FY1 told the investigation that they wanted to make a good impression during their shadowing day on the older persons’ ward, as this would be where they were going to work on their next rotation. They wanted to come across as being helpful and to ensure that they were competent in the role. Since the incident, the FY1 said they were cautious, and it had changed the way they practised.

3.3.9 From ST2’s perspective, the fact that the FY1 had previously undertaken a similar task influenced their decision-making and provided assurances that it was safe for the FY1 to take the blood cultures. The ST2 said: “I wasn’t worried about it being a job that was out of [the FY1’s] comfort zone because [they] had said that [they] had already done it in [their] previous job, but also I knew that I had done it at [their] level in my previous job, so I just assumed that it would be okay.”

3.3.10 The investigation observed that a combination of assumptions and misunderstandings resulted in the most junior (and newest) member of the medical team, assisted by a medical student, undertaking a task that required specialist training and an assessment of competence.

3.4.1 This section explores the medical team’s understanding of the specialist training and knowledge required to carry out the blood culture task safely, mindful of the potential safety risk of catheter-related air embolism.

3.4.2 The task was communicated and accepted as a standard blood culture task, similar to taking blood cultures from a generic central line. This was summarised by the FY1:

“… everyone just knows that if you don’t know what you’re doing then don’t do it. But I didn’t know what I didn’t know at the time.”

3.4.3 In terms of training for taking bloods from central lines, the ST2 could not recall receiving any type of formal training as a more junior doctor. The ST2 told the investigation that they were aware that the handling of dialysis lines was restricted “because the lines are precious and there’s only limited amounts of lines that people will be able to have”. They thought the rationale for this was solely based on keeping the lines patent and preserved, without recognising the safety risk of an air embolism.

3.4.4 The FY1 said that they “feel that dealing with central lines should be considered a skill and because all those skills, we are talking about catheterisation, we’re talking about cannulation, if we can call this a skill, then – and this is not very invasive and this is not life threatening, so why not central line?”. This highlights the FY1’s view that accessing all types of central line should be managed in a similar way to other skills, as a competence requirement requiring sign-off once accomplished.

3.4.5 The FY1 also told the investigation that they were acutely aware of differences between professional groups, in terms of assessing competence in particular skills. They said:

“Because I understood nurses need to be signed off. Because if they are, then why not us …”

3.4.6 When the FY1 queried the requirement for a medical student to take blood cultures, they were advised that this was one of several practical procedures on a skills list that, as a medical student, they needed to carry out and get signed off. However, the investigation learned that this requirement was for taking blood cultures from standard peripheral lines (into a smaller vein close to the surface of the skin) and not central lines (into a large vein close to the heart). In the reference event, the first time the medical student had seen a central line of any type was when they saw Joan’s haemodialysis catheter.

3.4.7 The medical student told the investigation they were:

“… a little bit pressured at the time to do a blood culture because essentially our placements are in 12-week blocks, so I was nearing the end of my time at [Acute Trust 2], probably in my 10th or 11th week there, and afterwards we switch to GP placements, and essentially, I really needed to get a blood culture signed off, otherwise I can’t progress to 4th year.”

3.4.8 At the dialysis unit, only specifically trained registered nurses were able to access haemodialysis catheters, following Acute Trust 1’s policies. Training and competency were assessed during induction programmes, in line with the private provider’s training policy. The renal consultant’s perspective supported this:

“… tunnelled dialysis lines should be handled only by the dialysis nurses unless in the case of life-threatening emergency.”

3.4.9 In the reference event, there was a focus on the known risk relating to the maintenance of the line to enable dialysis to continue. However, there was no knowledge of the safety risk of an air embolus. Volunteering for the task and having the external time pressure to ‘sign off’ this procedure as part of their medical training, contributed to the risk perception (Paek and Hove, 2017), and the motivation to continue with the task.

3.4.10 The task of taking blood cultures from Joan’s haemodialysis catheter was not delegated to a specific member of the team. Taking blood cultures formed part of the plan, but there was no indication of urgency, where the task would be carried out, or who would undertake it.

• 3.5.1 In relation to taking blood cultures from the haemodialysis catheter, this section explores:
• the clinicians’ understanding in respect of how to take a blood sample
• the specific complexities relating to this task
• the issues within the healthcare system impacting on the reference event (to provide context).

3.5.2 Acute Trust 2 shared with the investigation its guidance entitled ‘Insertion, management and removal of central venous access devices (CVAD)’. This included guidance on taking blood cultures from CVAD, but not from haemodialysis catheters specifically. The investigation learned that this was because Acute Trust 2 did not provide renal services. However, this meant that for inpatients with a haemodialysis catheter in situ (such as Joan) who required an intervention, there was no appropriate guidance for clinicians.

3.5.3 Acute Trust 1 provided renal services on site and via satellite dialysis units. Detailed and comprehensive guidance on all steps of the patient pathway was outlined in its ‘Guidelines for the care of central venous access devices (CVADs)’. This included a procedure for taking blood samples and taking blood cultures from a haemodialysis catheter.

3.5.4 Within Acute Trust 1’s guideline was a list of the equipment required for taking blood cultures from a haemodialysis catheter. This included a blood collection set (collection tube with adapter), a locking agent (see 3.3.6) and new (sterile) Luer lock caps to apply to both lumens of the haemodialysis catheter after the procedure.

3.5.5 At Acute Trust 2, on the older persons’ ward, the only equipment available was the blood culture bottles and a standard sterile dressing pack, since patients on this ward rarely had a central line (of any type) in situ. In Joan’s case, the FY1 and the medical student had antiseptic wipes to clean the port, some sterile gauze and a selection of Luer lock syringes to insert on to the end of the central line in order to take the required blood sample. There was no sterile pack, locking solution or sterile cap to put back on after the procedure.

3.5.6 The investigation visited the dialysis unit where Joan had dialysis treatment three times a week. The purpose of the visit was to understand more about dialysis treatment at the unit and the task of taking blood cultures from central lines. The dialysis unit used Acute Trust 1’s guidelines.

3.5.7 During the site visit, the investigation observed a simulation of the task of taking blood cultures from a haemodialysis catheter and a patient procedure was observed. This entailed taking a patient off the dialysis machine at the end of treatment and comprised some of the steps included in the blood cultures task. The steps included the attachment of a syringe to the dialysis line once the cap was removed, before the line was unclamped. Integral to the various procedures outlined in Acute Trust 1’s guidelines, it was stated that registered nurses must be assessed as competent and assessment criteria were included.

3.5.8 In relation to the specific sequence of steps for the task of taking blood cultures from a haemodialysis catheter, the investigation identified that across several of Acute Trust 1’s guidelines the following steps were included in figure 5:

3.5.9 When the FY1 and medical student went to Joan’s bedspace to take the blood cultures from her haemodialysis catheter, she was sitting in a chair and remained there for the duration of the procedure. The investigation found that Joan’s position (upright in a chair) was not the safest option when uncapping and unclamping the lumen of the haemodialysis catheter for blood sampling (see 1.3.3). The investigation heard this would have been outside the expected knowledge of any of the medical team on the older persons’ ward.

3.5.10 When the FY1 subsequently attempted to access Joan’s haemodialysis catheter, and discovered it was covered in gauze, they asked the medical student to seek help as it looked different from other lines. The medical student attempted to speak to the nursing team to ask them why the line was wrapped and how to handle it. Once it was apparent that all the nurses were busy and unable to assist immediately, the FY1 said: “Okay, I’ve done these before so we’ll carry on, and I got told by the SHO [ST2] this is fine.” They did not attempt to seek advice from the medical team on the ward round or the on call renal team at Hospital 1.

3.5.11 Concern was expressed by the FY1 about the gauze wrapping around the lumens of the haemodialysis catheter, primarily because they had never seen it before. The medical student did not know what to expect or recognise that the haemodialysis catheter looked different to standard central lines, since they had not seen one before.

3.5.12 The FY1 told the investigation they fetched a pair of scissors to cut the wrapping around the haemodialysis catheter and recalled saying aloud: “Let’s be really careful with this because it’s a dialysis line.” The gauze was removed and the FY1 saw two lumens of different colours, which alerted them that the haemodialysis catheter was unfamiliar (figure 6).

3.5.13 The medical student was instructed by the FY1 to remove the cap from one of the ports of the haemodialysis catheter and to start cleaning it for 30 seconds. The FY1 proceeded to unclamp the same lumen, which they recall was the blue one. They were initially not sure which lumen to access, but chose the blue one, as the colour red had negative connotations. The FY1 did not at the time appreciate the risk of air entry when unclamping the line without attaching a syringe first (see 1.3.3).

3.5.14 The investigation was told that, in comparison with a standard central line, there was an increased risk of significant air embolus when a haemodialysis catheter was uncapped and unclamped, due to the large lumen. The larger lumen is required to allow high blood flows for efficient haemodialysis treatment. An air embolus can occur when any type of central line is uncapped and unclamped. However, the risk increases in haemodialysis catheters due to the greater volume of air and the speed with which it can enter a person’s circulation.

3.5.15 The FY1 told the investigation about their actions at the time and their reflections following the coroner’s inquest. They said: “Now I know … the right way to do it. I know that you should only take the clamp off once you’ve got the syringe on …” Joan had a cardiac arrest before the syringe was connected to the line to take the blood cultures.

3.5.16 The investigation observed that there was no barrier or safeguard (control) to prevent the FY1 and the medical student from removing the Luer cap and unclamping the line without attaching a suitable device (syringe/vacutainer). Therefore, mitigation of the risk of air embolus from an equipment perspective was minimal (see figure 6). Administrative barriers that rely on people following a process are weak in comparison to engineered barriers that physically mitigate a risk (see figure A1). This will be explored further in section 4.

3.5.17 The mitigation of the risk of air embolus when taking blood cultures from a haemodialysis line relies upon staff as a control measure to follow a defined procedure. Furthermore, it also relies upon staff being aware that there was a defined process to follow, which was not evident in the reference event.

3.5.18 It was evident that the FY1 and medical student’s decision to proceed with the task of taking blood cultures from a haemodialysis catheter was based on limited awareness of the specific requirements and of the associated risks. Acute Trust 2 did not have a policy for this procedure, and its staff did not have access to Acute Trust 1’s guidelines. Staff were not aware of the specific equipment required or the availability of it on the older person’s ward. The investigation was advised by Acute Trust 2 that only the intensive care unit at Hospital 2 would have the required equipment for the task.

3.5.19 Additionally, the controls (clamps) in place in the reference event were not effective in mitigating the risk of air embolus during the task of taking blood cultures from a haemodialysis catheter. The controls in place relied upon staff following a defined procedure; however, the staff involved were not aware that this was a requirement, and therefore the procedure (control) was ineffective.

3.6.1 Based on the evidence described in this section, the investigation identified a range of issues, both those specific to haemodialysis catheters, and those applicable more widely across the healthcare system in various care settings.

3.6.2 The investigation found that at the time Joan’s blood cultures were taken, there was limited evidence of effective controls (safeguards and barriers) to mitigate the risk of catheter-related air embolism. The controls that were in place were administrative in nature (see figure A1).

3.6.3 Key findings from the reference event included blood cultures from a haemodialysis catheter being taken in an inappropriate location, without the required equipment available, and without trained staff with the required competencies. Analysis of the reference event findings identified the following issues which are applicable to wider healthcare settings, some of which are explored further in section 4:

• the communication of a task or plan to the most appropriate staff/team
• awareness of the risks associated with tasks outside of individual knowledge/training/speciality
• the assessment of competence in relation to new tasks carried out by undergraduate and postgraduate medical staff differs from other clinicians
• the culture of attempting new and unfamiliar tasks is widespread and challenging to address, especially in respect of junior medical staff
• differing competence requirements across medical schools, often between those in the same region.

Lines of enquiry

4.1.1 Starting at the most effective level of the hierarchy of hazard control – that is, eliminating the hazard by physically removing it or substituting it for a less hazardous one – the investigation found that it was not feasible to eliminate or substitute the hazard. Certain patients require long-term haemodialysis catheters, of the type Joan had in situ (as outlined in section 1). These patients will require the device to be intermittently accessed for haemodialysis treatment or other interventions, such as taking blood samples.

4.1.2 In relation to the next level of the hierarchy of controls – engineering controls – the investigation considered that the use of ‘equipment to prevent or separate [manage] the hazard’ was relevant to this investigation, based on the design risks identified in the analysis of the reference event. This section of the wider investigation therefore focuses on engineering controls (see bullet points 3-5) and the factors that contribute to their impact and effectiveness (see bullet points 1-2):

• a review of the relevant equipment regulatory requirements
• post-market monitoring and surveillance
• exploration of the feasibility of engineered safety controls for haemodialysis catheters
• the application of human factors and usability engineering to medical devices
• medical device innovation and design.

Overview of the equipment regulatory requirements

4.1.3 The investigation engaged with the Medicines and Healthcare products Regulatory Agency (MHRA) to better understand the equipment regulatory requirements. This was to provide some context ahead of exploring whether there was an engineering control that could mitigate the risk of catheter-related air embolism.

4.1.4 The MHRA advised the investigation that a dual lumen/port haemodialysis catheter would be classified as a Class III (high risk) medical device, with the strictest control measures assigned to it. This was in line with the Medical Devices Regulations 2002 (SI 2002 No 618, as amended) (Medical Devices Regulations 2002).

4.1.5 There are requirements arising from this legislation for healthcare organisations (Medicines and Healthcare products Regulatory Agency, 2021a) and manufacturers. The manufacturer has a responsibility to ensure that the device is compliant with the Medical Devices Regulations 2002. This includes the device being assessed by a ‘Approved Body’ (for example, the British Standards Institution) before the UK Conformity Assessed (UKCA) marking is added to the device (previously the CE marking). The haemodialysis catheter used in the reference event was CE marked prior to 26 May 2021. Therefore, the Directive 93/42/EEC on medical devices (Council Directive 93/42/EEC) is applicable and the manufacturer has to comply with the requirements of the Directive 93/42/EEC in order to place its products on the UK market (Medicines and Healthcare products Regulatory Agency, 2020).

4.1.6 Before being placed on the market in Great Britain, the medical device must be registered with the MHRA, ‘to ensure that the supply chain for medical devices is safe and secure’ (Medicines and Healthcare products Regulatory Agency, 2022a). Manufacturers based outside the UK (as in the case of the haemodialysis catheter) must appoint a UK ‘Responsible Person’ who will assume certain responsibilities on behalf of the manufacturer, including registering the device with the MHRA. According to the Medical Devices Regulations 2002, the UK Responsible Person must: ‘cooperate with the MHRA on any preventive or corrective action taken to eliminate or, if that is not possible, mitigate the risks posed by devices’.

Post-market monitoring and surveillance

4.1.7 The MHRA told the investigation that following the government’s consultation on the future regulation of medical devices in the UK (Medicines and Healthcare products Regulatory Agency, 2021a), the purposes for which investigations will be designed, authorised, conducted, recorded, and reported will be strengthened in several areas. Relevant to this investigation, this will include verification of the ‘clinical safety of the medical device and to determine any undesirable side-effects, under normal conditions of use of the medical device’ (Medicines and Healthcare products Regulatory Agency, 2021a).

4.1.8 Once a medical device has been placed on the market in Great Britain, the manufacturer is responsible for monitoring the product and reporting serious adverse incidents to the MHRA (Medicines and Healthcare products Regulatory Agency, 2015). In the case of the reference event, the manufacturer did not report the adverse incident, as it was unaware until notified by HSIB.

4.1.9 The reference event Trust reported the serious incident via the NHS’s serious incident management system (the Strategic Executive Information System (StEIS)), and not to the manufacturer or the MHRA (see 4.1.12). This was because of the perception that the incident was caused by ‘human error’. The investigation was unable to determine how many other serious adverse events go unreported to manufacturers. Reporting of these events would afford an opportunity for manufacturers to amend their medical device risk profiles and potentially the devices they manufacture too.

4.1.10 NHS England is introducing a new incident reporting system, the ‘Learn from patient safety events service’ (LFPSE). LFPSE will replace StEIS (a database of serious incident reports) and the National Reporting and Learning System (NRLS). The new system gathers more information and greater learning from patient safety incidents, including medical device related incidents (NHS England, 2023). LFPSE also correct some limitations in the way that the NRLS collected data.

4.1.11 NHS England told the investigation that, if seeking to identify issues involving haemodialysis catheters using LFPSE, it would search the ‘device details’ fields (see figure 7). This would then provide details about what kind of medical device was involved and/or free text searches would identify other mentions of these catheters or exclude any irrelevant records. Users may also have included further details about the manufacturer, which should indicate whether the device has malfunctioned or not (see figure 7). However, the device details are optional, so may not be completed by the user initially but they can be added at a later date.

Yellow card reporting system

4.1.12 A safety concern arising from a medical device can also be reported by trusts using the MHRA’s ‘Yellow Card’ website or the Yellow Card app. The Yellow Card site was originally set up to collect information on suspected adverse drug reactions to provide an early warning of possible hazards. It now also collects safety concerns or problems associated with other healthcare products including medical devices.

4.1.13 To inform this investigation, data was requested from the MHRA regarding haemodialysis catheter incidents reported through the Yellow Card system. A review of its database from 1 January 2017 to 5 December 2022 found that the MHRA had received 1,341 reports for the following device areas:

• central venous catheters (1,146 reports)
• paediatric central venous catheters (90 reports)
• catheter dialysis (105 reports).

4.1.14 There were no reports of ‘air embolism’ or reports relating to uncapped and unclamped haemodialysis catheters. However, the MHRA advised that it is important to note that these figures do not correspond to complication rates. The inclusion of a report on the MHRA adverse incident database does not necessarily mean the events described were caused by that device but could be due to other factors.

4.1.15 The MHRA has a data sharing agreement with NHS England which allows data relating to medical devices to be sent to the MHRA from the NRLS. This agreement appears to have arisen following a patient safety alert in 2014 (Medicines and Healthcare products Regulatory Agency, 2014), which was a directive for improving medical device incident reporting and learning. It was anticipated that the development of an integrated NRLS would reduce the need for duplicate data entry by staff in healthcare organisations.

4.1.16 The alert made it clear that NHS England and the MHRA were ‘working together to simplify and increase reporting, improve data quality, maximise learning and guide practice to minimise harm from medical devices incidents’ (Medicines and Healthcare products Regulatory Agency, 2014). However, the data shared via the NRLS does not fully meet the MHRA Yellow Card standards for a full assessment. In addition, the way in which data is shared involves a resource-intensive triage by the MHRA to identify and follow up on relevant Yellow Card data.

4.1.17 A previous HSIB investigation recommended that the MHRA review its documentation to determine whether more specific guidance was required on how to incorporate human factors into post-market adverse event investigations (safety recommendation 2018/023, Healthcare Safety Investigation Branch, 2018). The MHRA’s response in February 2019 indicated that it had planned to review whether ‘sufficient weight is given to post market surveillance of devices within the guidance’.

4.1.18 The improvement of adverse events and incident reporting is a commitment made by NHS England and the MHRA, with a clear strategy and shared responsibilities (NHS England, 2021). Integrating the Yellow Card database with LFPSE could potentially prevent duplicate reporting and improve the current underreporting. The MHRA would need to be able to extract Yellow Card-related data directly from LFPSE, so that the relevant reports could be uploaded directly into the MHRA’s Yellow Card database for assessment. This could be assessed alongside the information directly reported into the Yellow Card reporting system by patients/the public.

4.1.19 The investigation was told by NHS England that the MHRA had undertaken an alpha (development) phase, in conjunction with themselves, to explore options for the integration of the Yellow Card database with LFPSE. While a recent bid for an in-service solution was unsuccessful, the MHRA has advised that both organisations are committed to work in partnership to drive this project forward.

• Engineered safety controls for haemodialysis catheters

4.1.20 There are 10 manufacturers of double-lumen haemodialysis catheters on the MHRA’s medical device register. However, the investigation found that there has been limited innovation in the haemodialysis catheter market over the last 30 years in relation to the haemodialysis catheters for long-term use. The investigation could find no evidence of planned innovation in the foreseeable future.

4.1.21 The manufacturer of the product used in Joan’s case advised the investigation that its current risk documents had ‘mitigated the risk of air embolism as far as possible, therefore, a CVC [central venous catheter] product redesign is not planned at this time’. The manufacturer further advised that it continuously ‘trend risks’ to ensure product safety and performance, and as the company developed new products, it would evaluate state-of-the-art technologies and designs to improve its devices.

4.1.22 HSIB received additional referrals relevant to this investigation in October and November 2022. In the first referral, three cases were reported in which central venous catheter lumens were left open, either through a misplaced cap or a three-way tap left open. The second referral related to a patient who died following an air embolus which entered his brain and caused a catastrophic brain injury. The incident involved a central venous catheter lumen unintentionally left open. No cap was connected to the lumen in use.

4.1.23 One of the referrers indicated that ‘New CVCs exist with integrated valved connectors eliminating the risks of embolism or haemorrhage, and we are looking to introduce these in our hospital/trust’. On further investigation, the technology being referred to was for short-term haemodialysis devices only.

4.1.24 In terms of haemodialysis catheters for short-term (temporary) use, a manufacturer has integrated an ‘automatic clamp’ to maintain a sealed system upon disconnection (see figure 8). This negates the need for external clamps to prevent air from entering the line on disconnection and mitigates the risk of catheter-related air embolism (Kimal, 2022). However, it is intended for use in attaining short-term access (less than 30 days) for haemodialysis, making it unsuitable for patients who require long-term access (permanent) for haemodialysis. The manufacturer highlighted to the investigation that their product could only guarantee 200 activations of this needle free device, hence the short-term restriction. It was discussed that further innovation and feasibility assessment could increase this.

4.1.25 There are also needle-free connectors for haemodialysis catheters on the market (see figure 9) which similarly create a closed system. However, they must be changed every 7 days and could potentially be removed by the user (in the same way as caps), so leaving the haemodialysis catheter open to air.

4.1.26 The investigation found there are currently no long-term haemodialysis catheters on the UK market, or being developed, with integrated ‘safety-valves’. Such a safety-valve would remove the need to use manual clamps and would maintain a closed system upon disconnection, so reducing the risk of air embolism. This would represent an engineering control, acting as the most effective type of barrier in mitigating the risk.

4.1.27 The guidance section of the ‘Essential principles of safety and performance of medical devices and IVD medical devices’ (International Medical Device Regulators Forum, 2018) relates to essential risk control for the development process. However, neither this guidance nor the Directives (see 4.1.5) permit the MHRA to compel manufacturers to develop new devices where residual risks have been identified. The MHRA advised that these types of risks relating to devices already on the market should be managed with appropriate mitigations such as labelling and training.

Applying human factors and usability engineering to medical devices

4.1.28 The interaction between people and medical devices is frequently identified as a contributory factor to patient safety events within HSIB’s national investigations (Healthcare Safety Investigation Branch, 2023). These contributory factors may be referred to as ‘human error’ without consideration of the context or design features of the device. In the case of the reference event, the clamps on each lumen of the haemodialysis catheter are an example of a design feature with the potential to increase the likelihood of error. As outlined in Healthcare Safety Investigation Branch, (2023) ‘Designing equipment with the potential for likely ‘use’ errors in mind is a regulatory requirement and considered one of the most effective ways to minimise risk and increase safety’.

4.1.29 The investigation consulted a subject matter advisor, who was a consultant nephrologist (kidney doctor). The subject matter advisor told the investigation that the manual clamps on the haemodialysis catheter used in the reference event represent design induced error. This is because current devices rely on people following processes to ensure patient safety, namely by ensuring that the clamp is on prior to accessing the haemodialysis catheter ports.

4.1.30 The NHS Supply Chain is one of the supply chain routes into the NHS which manages the procurement, sourcing, and supply of medical devices to care providers. They told the investigation that a high-level description of these products was either (open-ended) non-valved or valved, with the ‘valve’ essentially being a pressure-sensitive slit that remains closed unless fluids are infused. When not in use the valve is closed, thereby sealing fluid inside the haemodialysis catheter. This reduces the potential for the haemodialysis catheter to become blocked with a blood clot and reduces the risk of an air embolism. The investigation was told that haemodialysis catheters without valves will always require a clamp, designed as a safety measure to prevent a backflow of blood or air entry into the lumen(s).

4.1.31 The Medical Devices Regulation (MDR) (Medical Devices Regulation (EU) 2017), outlines the need to incorporate human factors methods into the design process, specifying usability testing to do this. The UK’s guidance on the consideration of human factors in engineering of medical devices to improve patient safety (Medicines and Healthcare products Regulatory Agency, 2021b), is a further step towards accounting for the environmental conditions in which devices are used. The guidance includes the requirements for specialised training, the lack of which increases ‘the potential for use error’ (Medicines and Healthcare products Regulatory Agency, 2021b).

4.1.32 This ‘advisory’ guidance is intended for manufacturers, developers of medical devices and UK Approved Bodies (the organisations responsible for assuring the quality of the devices). The investigation was told by the MHRA that there are currently no plans to mandate the guidance, and manufacturers can use other methods for demonstrating conformance with the requirements of the MDR 2002. The only legal requirement is that manufacturers meet the requirements of Annex 1 of the MDR 2002. Similarly, the MHRA encourages the use of ‘relevant designated standards’ (Medicines and Healthcare products Regulatory Agency, 2021b) however, these are not compulsory.

4.1.33 In addition to the safety recommendation mentioned in 4.1.15, HSIB issued a further safety recommendation to the MHRA relating to an evaluation of how the human factors guidance document was used in practice by manufacturers and Notified Bodies (safety recommendation 2018/021, Healthcare Safety Investigation Branch, 2018). The MHRA responded in February 2019 indicating that it had ‘planned to evaluate how this guidance has been received’, and to expand the evaluation to include a survey.

Medical device innovation and design

4.1.34 In Australia, there are system-wide strategies for the improvement of medical devices (Clinical Excellence Commission, 2015). The strategies include engaging industry, with a view to sourcing devices and connectors with integrated valves and other safety options (for example, an alert attached to the central venous access device).

4.1.35 A manufacturer explained to the investigation that if a similar process were to be adopted in the UK, this would not provide an immediate solution. This is because the process for getting a new product on to market is lengthy, involving scoping, building a business case, development, testing and validation. Representatives of the Association of Renal Technologists added that experience from other incidents supported the fact that an engineered solution could take a long time to achieve. It would also need the co-operation of all haemodialysis catheter manufacturers and the cost may prove to be prohibitive.

4.1.36 Where the MHRA identifies a suspected safety issue due to malfunction of a device, it gathers all available information from various sources for further assessment and internal discussions on possible actions. Such sources include Yellow Card reports, case incidents from other databases such as the NRLS, published literature, international experience, and studies.

4.1.37 The MHRA’s remit as a regulator is mainly focused on ensuring that medical devices are safe when used as intended/designed, and to take action to mitigate risks when safety is compromised. The MHRA advised the investigation that where there is a safety risk due to ‘human error’, and clinical practice or possibly lack of training are an issue, the responsibility to flag this to healthcare providers and healthcare professionals falls into the remit of national bodies such as NHS England. If there are safer designs on the market than the device in question, the MHRA may flag this to manufacturers, but does not have the legislative power to push for changes.

4.1.38 The investigation was told by NHS England that encouraging safer procurement systems through NHS Supply Chain could help to prevent error. This alongside the involvement of the MHRA to support improved devices which address the safety concern coming to market and being available to the healthcare system.

4.1.39 In summary, the investigation identified that there are short-term haemodialysis catheters in use (in some trusts) that have valves/needle-free connectors/closed circuit technology, which reduce the risk of air embolus incidents. While pursuing the commercial viability of a suitable Luer/replaceable device for haemodialysis catheters was outside the scope of this investigation, an engineering control was identified. Evidence indicates that an engineered solution to stop inadvertent opening of haemodialysis catheter lumens to air is the most robust way of addressing the safety issue.

Administrative controls: Implement procedures and change the way people work

Lines of enquiry

In relation to the next level of the hierarchy of controls – administrative controls – the investigation considered that the implementation of ‘procedures and change the way people work’ was relevant to this investigation. This section of the wider investigation therefore focused on the following administrative controls:

review of the medical education and training requirements for the safe access of haemodialysis catheters, in addition to the assessment of competence

identification and review of the current controls in place (safeguards/barriers)

review of the relevant guidance and policy for the management of haemodialysis catheters.

While this section of the report focuses on medical staff accessing haemodialysis catheters, similar principles apply to other types of central line (although the risk of an air embolus is reduced in narrower bore lines). The principles will also apply to other staff groups who may access haemodialysis catheters, such as advanced care practitioners and physician associates.

Undergraduate medical education

4.2.1 The General Medical Council (GMC) approves, monitors, and quality assures undergraduate medical education. The GMC’s powers do not extend to approving, or mandating specific content in, medical school curricula. Instead, the GMC sets outcomes for graduates of UK medical school degrees leading to entry on to the medical register (Courts and Tribunals Judiciary, 2021).

4.2.2 To be approved by the GMC, a medical school must demonstrate that it meets GMC standards. Included in these standards are ‘Outcomes for graduates’ (2018) and the accompanying ‘Outcomes for graduates – Practical skills and procedures’ (2019) list, which together describe the knowledge, skills, and behaviour graduates must show to be able to work safely as newly registered doctors in the first year of the UK Foundation Programme (Courts and Tribunals Judiciary, 2021). Medical schools develop and implement their own curricula and assessments which make sure their graduates have demonstrated these outcomes.

4.2.3 The procedure of taking blood samples from central lines is not included in either the ‘Outcomes for graduates’ (2018) or the ‘Practical skills and procedures’ (2019). The GMC wrote in its response to the Regulation 28 Prevention of Future Deaths report concerning the reference event, that procedures using central lines are not included in the outcomes, however, medical students should be taught about the general risks associated with these lines (Courts and Tribunals Judiciary, 2021).

4.2.4 In the reference event, Acute Trust 2 was affiliated with two medical schools. At the time, it was a local requirement for medical students at one of the medical schools to ‘demonstrate competence in undertaking blood sampling from a central line under direct supervision’ (Courts and Tribunals Judiciary, 2021). Only students in their final year of study were permitted to undertake this procedure and they were not able to do so unless they had first undertaken a simulated training session and participated in the pre-course learning. All final-year students were required to record that they had successfully accessed a central venous device on a minimum of three occasions in clinical practice, under direct supervision. This formed part of the requirement for all final-year students to graduate and progress into the foundation training programme. It was reviewed by the examination board.

4.2.5 The GMC wrote to HM Coroner in May 2021 (after being issued with a Regulation 28 Prevention of Future Deaths report (PFD) in March 2021), saying that it had agreed with the medical school that this requirement would cease going forwards (Courts and Tribunals Judiciary, 2021). To ensure the same action was replicated across all medical schools (not just in the region where the event occurred), the investigation engaged with the Medical Schools Council and the GMC. Through this engagement, the GMC agreed that any medical school requiring their students to demonstrate competence in taking bloods from central lines should reflect and feedback to the GMC on whether continuing with this requirement presented a risk to patient safety.

Required skills of graduates

4.2.6 The General Medical Council’s ‘Practical skills and procedures’ (2019) defines the core practical skills and procedures that graduates must be able to do when they start to practise as a newly qualified doctor, within the UK Foundation Programme. The section on ‘Diagnostic procedures’ includes taking blood cultures and the description states: ‘Take samples of venous blood to test for the growth of infectious organisms’, with the level of competence being described as: ‘safe to practise under direct supervision’ (General Medical Council, 2019).

4.2.7 It is the view of the GMC that the taking of blood samples from central lines is beyond the level of competence required for newly qualified doctors. It is a procedure undertaken by specialists, or in specialist units, due to the inherent complexity and risks. The GMC stated that:

‘A doctor in training, and especially a recently graduated Foundation Year 1 doctor, should only undertake this procedure on the specific advice of, and under the direct supervision of, a suitably qualified senior colleague. They also must receive specific authorisation and training.’
(Courts and Tribunals Judiciary, 2021)

4.2.8 That this is beyond the level of competence required for newly qualified doctors is not explicitly stated in ‘Practical skills and procedures’ (General Medical Council, 2019) and therefore agreement was reached to amend the guidance. Through this engagement, the GMC will ask that any medical school requiring its students to demonstrate competence in taking bloods from central lines, reflects and feeds back to the GMC on the mitigating actions to reduce the risk to patient safety.

Required skills of graduates

4.2.6 The General Medical Council’s ‘Practical skills and procedures’ (2019) defines the core practical skills and procedures that graduates must be able to do when they start to practise as a newly qualified doctor, within the UK Foundation Programme. The section on ‘Diagnostic procedures’ includes taking blood cultures and the description states: ‘Take samples of venous blood to test for the growth of infectious organisms’, with the level of competence being described as: ‘safe to practise under direct supervision’ (General Medical Council, 2019).

4.2.7 It is the view of the GMC that the taking of blood samples from central lines is beyond the level of competence required for newly qualified doctors. It is a procedure undertaken by specialists, or in specialist units, due to the inherent complexity and risks. The GMC stated that:

‘A doctor in training, and especially a recently graduated Foundation Year 1 doctor, should only undertake this procedure on the specific advice of, and under the direct supervision of, a suitably qualified senior colleague. They also must receive specific authorisation and training.’
(Courts and Tribunals Judiciary, 2021)

4.2.8 That this is beyond the level of competence required for newly qualified doctors is not explicitly stated in ‘Practical skills and procedures’ (General Medical Council, 2019) and therefore agreement was reached to amend the guidance. Through this engagement, the GMC will ask that any medical school requiring its students to demonstrate competence in taking bloods from central lines, reflects and feeds back to the GMC on the mitigating actions to reduce the risk to patient safety.

Cultural change regarding requirements and assessment of competence

4.2.9 In the context of the reference event, the GMC emphasised that doctors in training were expected to acknowledge the limits of their capabilities and understand the risk of performing procedures beyond their level of competence. The GMC emphasised though that there must also be effective systems, policies, and processes for determining a learner’s competency to undertake procedures, whether they have received appropriate training, and the level of supervision needed to ensure patient safety (Courts and Tribunals Judiciary, 2021).

4.2.10 This further suggests that an employer should ensure that there is an appropriate process in place to approve a doctor’s competency to undertake specialist and high-risk procedures, including opportunities to train on specific procedures in a simulation centre. To support this, the importance of good supervision is emphasised throughout the Foundation Curriculum (which underpins the training and professional development of newly graduated doctors).

4.2.11 The investigation also considered the competencies of junior doctors to undertake procedures more generally, with a focus on the criteria and training required prior to undertaking them, such as with nursing competency frameworks. The subject matter advisor told the investigation that there had been progress in this respect with the introduction of electronic portfolios, but these do not specifically include stepwise progression to competence.

4.2.12 A new edition of the GMC’s ‘Good medical practice’ (General Medical Council, 2013) is expected to be published in the second half of 2023. ‘Good medical practice’ describes what is expected of all doctors registered with the GMC. The current version recognises that a safety culture is essential, and the GMC anticipates that the relevant principles are likely to be replicated in some form in the updated version. At the time of finalising this report, the GMC was unable to confirm the content of the new edition, whilst the analysis of consultation feedback was ongoing. The principles in the current version of particular relevance are:

• recognising and working within the limits of your competence
• promoting and encouraging a culture that allows staff to raise concerns safely and openly
• being readily accessible to colleagues seeking, information, advice and support
• making sure staff you manage have appropriate supervision
• being satisfied that a colleague you delegate the care of a patient to has appropriate qualifications, skills and experience (General Medical Council, 2013).

4.2.13 ‘Achieving good medical practice: guidance for medical students’ (General Medical Council, 2016) is due to be updated following publication of the new edition of ‘Good medical practice’ (General Medical Council, 2013). This guidance shows how the principle in ‘Good medical practice’ (2013) apply to medical students. The GMC has advised that it will be able to consider the investigation’s findings as part of that work. The GMC will also consider this when reviewing its wider educational and ethical explanatory guidance documents for doctors, which support them in applying the principles in ‘Good medical practice’, and which includes guidance on delegation and referral.

4.2.14 Following discussions with the GMC, they reflected on the findings of the investigation and a concern about a culture of, or learners being encouraged ‘to have a go’, which could lead to them acting beyond their competence and without appropriate supervision. This led to the safety recommendation concerning the upcoming review of the ‘Achieving good medical practice’ for medical students’ guidance.

Foundation year doctors: Assessment of skills and competence

4.2.15 The 2016 Foundation Programme curriculum contained a list of 20 procedures that all foundation year doctors had to get signed off. Assessment of foundation year 1 doctors was at the level of the 20 foundation professional capabilities. To be signed off at the end of foundation year 1, the doctor’s knowledge, skills and behaviours must have met or exceeded the minimum expected level of performance.

4.2.16 In 2021, the list of procedures was removed from the curriculum, after much debate, so it is now less detailed and more strategic (UK Foundation Programme, 2022). Following on from this, there is currently no mandatory assessment of procedural skills, but rather doctors learn under supervision.

4.2.17 The investigation was advised by the UK Foundation Programme Office (UKFPO) that the decision to remove the 20 procedures from the curriculum was taken for several reasons, including:

• feedback from doctors and supervisors that the 20 procedures did not represent an exhaustive list, so was not completely reflective of all the procedures a junior doctor needed to undertake
• foundation year doctors not having the opportunity to carry out tasks resulting in a failure to meet the minimum expected level of performance
• issues relating to what sign-off meant in practice, as tasks could vary in difficulty depending on the patient, for example taking blood from a frail, elderly person is more challenging than taking blood from a young, physically fit person.

4.2.18 The UKFPO advised the investigation of the importance of getting the message across to doctors relating to working within their level of competence and empowering them to decline tasks they are not competent to undertake. The investigation heard that a senior clinician working in the UKFPO tells junior doctors on his clinical team that the three most important words are ‘I don’t know’.

From a clinician’s perspective

4.3.1 ‘Instructions for use’ are included in the packaging with each haemodialysis catheter, which highlight the air embolus risk as a safety-critical task. Under a heading, ‘warnings’, there is an instruction stating that ‘in the rare event that a hub or connector separates from any component during insertion or use, take all necessary steps and precautions to prevent blood loss or air embolism and remove catheter’. While the investigation was unable to verify, there is a risk that the instructions will not be retained after the insertion of a haemodialysis catheter and so would not be available for clinicians post insertion.

4.3.2 The investigation discussed safeguards and barriers with staff across the providers they visited, with the intention of comparing practice across other providers and nationally.

4.3.3 The investigation was told that a wristband is used for patients having dialysis via a fistula, which contains information regarding the associated risks. This represents a safeguard (a control that has the intention of enhancing safety and reducing the likelihood of an incident). Although the investigation found initiatives to support the use of wristbands in different parts of England, it did not identify a formal national guidance or policy. A wristband, and therefore an additional safeguard, is not afforded to patients with a haemodialysis catheter. This type of additional safeguard is an administrative control and therefore a relatively weak barrier, particularly as the wristband may not be seen on a patient’s arm when accessing a haemodialysis catheter.

4.3.4 Throughout the investigation, the investigation heard from a number of sources (including the satellite dialysis unit and members of the Association of Renal Technologists), that labelling applied to the haemodialysis catheter, or its individual lumens, could be used to highlight risks and the required competence to access.

4.3.5 A survey was distributed to the membership of the Renal Vascular Access Special Interest Group on behalf of HSIB (see appendix 2). The aim was for healthcare professionals to anonymously share their insight and experience regarding incidents related to inadvertently uncapped and unclamped central venous lines. Responses received were from various healthcare specialities, including renal nurses, haemodialysis staff, a renal doctor, a transplant surgeon, and advanced nurse practitioners. The key findings were:

When asked whether they had ever experienced an air embolus event, or near miss, when using a haemodialysis catheter or another type of central line, six incidents were noted.

When respondents were asked if they were aware of visual alerts that prompt healthcare professionals to the increased safety risks associated with in situ haemodialysis catheters, and the restricted access of this type of medical device to specially trained staff, several were discussed. These included coloured bungs on the haemodialysis catheters, coloured wrist bands (see 4.3.3), line labelling, stickers, and alert/emergency cards (see 4.3.8).

When asked if they were aware of devices that can mitigate the risk of an air embolism, several were discussed. These included needle-free connectors (9 responses), which are referred to earlier in this report (see figure 9).

When asked if they were aware of, or have ever used, a haemodialysis catheter in which technology to mitigate an air embolism was integral/bonded, as opposed to a removable device, several were discussed. These included the bonded haemodialysis catheter for short-term (temporary) use mentioned earlier in this report (Kimal, 2022) (see figure 8).

From a patient’s perspective

4.3.6 The investigation heard that patients with kidney conditions are knowledgeable in the care of their haemodialysis catheters and could be empowered (when able) to challenge access to their lines by anyone other than dialysis unit staff or renal healthcare staff. Patients/their families are often advised not to let clinicians they are unfamiliar with who are not kidney professionals to touch their haemodialysis catheters.

4.3.7 Kidney Care UK (2022) has a patient information leaflet on its website ‘Looking after your haemodialysis line’. In this, it states ‘Do not allow anyone other than a member of your kidney team to use your line’. However, it does not currently explain why this is important which, if known, may make patient challenge more likely to happen in practice. Therefore, although there is written information for staff and for patients, neither would mitigate the risk of air emboli in their current form.

4.3.8 It was also suggested that patients could be asked to carry alert cards, highlighting the in situ haemodialysis catheter and the restricted access. Such cards appeared to be used inconsistently across providers and were not currently considered to be a reliable safeguard because of this.

Addressing the deficiency in knowledge

4.4.1 The investigation has been told that venous air embolism is a rare, but avoidable occurrence. In 2016, the PFD was issued by HM Coroner, following the inquest of Patricia Steer (Courts and Tribunals Judiciary, 2016). She died after the clamp on a central venous catheter port was briefly left open, resulting in air embolism and a stroke. The Coroner’s main concerns related to the nursing staff’s lack of awareness of the safety risk and the absence of literature or guidance relating to this.

4.4.2 The response by NHS Improvement (2016) to the Prevention of Future Deaths report in 2016 indicated that the Safe Anaesthesia Liaison Group (SALG) had ‘raised awareness’ among its members about the risk of leaving a central line uncapped and unclamped. For example, an ‘Air embolism safety alert’ was issued (The Faculty of Intensive Care Medicine and the Intensive Care Society, 2016), which included the following safety measures:

• central venous catheters should have clamps associated with each lumen of the line
• these should be used to clamp closed any lumens that are not in use.

4.4.3 It was suggested that organisations were best placed to communicate the message to the staff who were providing professional leadership, training, and supervision, in addition to the staff providing direct care to patients with central lines (NHS Improvement, 2016).

4.4.4 The safety risk of air embolus associated with the removal of dialysis catheters was reported nationally in 2018. The Renal Association, British Renal Society, and Intensive Care Society set out advice and an action plan in response. This recommended several precautions and monitoring measures and indicated national guidelines would be developed. There was no similar intervention directly relating to the safety risk associated with accessing the line for purposes other than insertion or removal.

4.4.5 In 2019, the Intensive Care Society produced guidance on the safe management of central lines which emphasised the need to cap and clamp lines when they are not in use. This did not address the Coroner’s concerns about training, competency, or local and national standard operating protocols, raised by the PFDs in 2015, and 2020 (the reference event) (Courts and Tribunals Judiciary, 2016; 2021).

4.4.6 The National Association of Medical Device Educators and Trainers (NAMDET) leads the development of medical devices training modules on the national ‘E Learning for Health’ (eLfH) training website (National Association of Medical Device Educators and Trainers, 2022). NAMDET confirmed that it contributes to the clinical device e-learning session, but the content does not specifically include the safe access of haemodialysis catheters.

Local, national and international guidance for accessing haemodialysis catheters

Local guidance

4.4.7 In relation to guidance at a local level, trusts use a variety of local guidelines for the management of haemodialysis catheters, as evidenced by the findings from a survey sent to members of the Renal Vascular Access Special Interest Group (see 4.3.5). Trusts mentioned the use of Local Safety Standards for Invasive Procedures (LocSSIPs) in this context too (NHS England, 2015), to ensure a consistent approach to the care of patients undergoing invasive procedures. In relation to haemodialysis catheters specifically, the focus was primarily on the insertion of the haemodialysis catheters as opposed to in situ access.

4.4.8 Trusts also use ‘The Royal Marsden manual of clinical and cancer nursing procedures’ (The Royal Marsden NHS Foundation Trust, 2020). In part 4 of the manual, ‘Supporting patients through treatment’, is a chapter entitled ‘Vascular access devices: insertion and management’. Complications are clearly outlined in sections on vascular access devices and skin-tunnelled catheters. The investigation has been unable to determine whether clinicians consistently refer to such guidance before carrying out a procedure.

National guidance

4.4.9 In relation to guidance at a national level, there is a clinical practice guideline on vascular access for haemodialysis (UK Renal Association, 2015). While there is a section on ‘complications of vascular access’, there is no mention of catheter-related air embolism.

4.4.10 The UK Kidney Association has told the investigation that it will communicate the message about the catheter-related air embolism safety risk in its newsletter, considering the current absence of integrated valve technology for long-term haemodialysis catheters.

4.4.11 The National Infusion and Vascular Access Society (2022) published a White Paper outlining a standardised structure and approach for the NHS to deliver vascular access services in every hospital. This was launched because of the variation in the provision of vascular access services, resulting in longer hospital stays and an increased cost to the NHS.

4.4.12 The subject matter advisor advised the investigation that the narrow bore lines that are described in the White Paper (which are more commonly used) are low risk in terms of catheter-related air embolism. The subject matter advisor’s view was that the processes for line insertion and line care could be expanded to include temporary and tunnelled dialysis line insertion by the type of dedicated team outlined in the White Paper (National Infusion and Vascular Access Society, 2022). Therefore, HSIB makes the following safety observation.

4.4.13 The Association of Anaesthetists advised the investigation that it is in the early stages of updating the ‘Safe vascular access guidelines’ (Association of Anaesthetists of Great Britain and Ireland, 2016). The guideline working group has been established and it is anticipated that the new set of guidelines will be published within 6 months. Based on anecdotal knowledge of incidents associated with haemodialysis catheters, the Association of Anaesthetists has specifically included a nephrologist and representatives from the Vascular Access Society and the National Infusion and Vascular Access Society.

4.4.14 It has been agreed that a section on haemodialysis catheters will be included in the revised ‘Safe vascular access guidelines’, and the working party was invited to share the investigation findings/recommendations to inform the content. The guidelines will include recommendations on reducing the incidence of air embolus.

4.4.15 In March 2022, the Medicines and Healthcare products Regulatory Agency published ‘Dialysis guidance’. This refers to the potential risk of air embolism during pre-use checks of equipment prior to haemodialysis treatment. However, the ‘vascular access’ section of the guidance contains no reference to the risk of air embolism when accessing the haemodialysis catheter to obtain blood samples/cultures (Medicines and Healthcare products Regulatory Agency, 2022b).

4.4.16 The MHRA agreed to carry out an internal consultation, but in principle agreed to amend the guidance to include a section on the safe handling of haemodialysis catheters.

The hierarchy of hazard control is used by those in industry who plan and implement mitigations to reduce safety risks that have been identified in the workplace. Risks should be reduced to the lowest practicable level and the hierarchy helps decision makers assess and prioritise mitigations that are more likely to be effective (Health and Safety Executive, 2019).

In the case of the reference event, there was limited evidence of effective controls (safeguards and barriers) specific to haemodialysis catheter and air embolus risks (as they were all administrative in nature – see figure A1). The investigation therefore explored effective controls and design via the hierarchy of hazard control, to determine how a haemodialysis catheter could be safely accessed in an appropriate location, by skilled practitioners, with the required equipment available. The investigation focused on haemodialysis catheters, but the findings may also be applicable to other types of central line.

A survey was forwarded to the members of the Renal Vascular Access Special Interest Group. The membership includes nurses, vascular/transplant surgeons, nephrologists, technicians, radiologists, and patient representatives.

Out of 38 members, 15 responses were received (39% response rate).

As the survey contained qualitative data, the results were collated and analysed by themes. The themes are included in section 4 of the report.

Haemodialysis central venous catheter (CVC) safety survey questions:

1. Can you briefly describe your current role?
2. How long have you worked in your current role?
3. Have you ever experienced an air embolus event, or near miss, when using a haemodialysis CVC, or another type of central line? If so, can you share the details?
4. What local/national guidance is available to support you in the management of haemodialysis CVCs?
5. If guidance is available, does it cover the access of haemodialysis CVCs, in addition to the insertion and removal? If so, can you share the details?
6. Are you aware of any type of visual alert that prompts healthcare professionals to the increased safety risks associated with in situ haemodialysis CVCs and the restricted access of this type of medical device to specially trained staff? If so, what was it? (For example, coloured wrist band, line labelling, etc).
7. Are you aware of, or use, any devices that mitigate the risk of an air embolism by maintaining a closed system? If so, what are the advantages and disadvantages?
8. Are you aware of, or have ever used, a haemodialysis CVC in which technology to mitigate an air embolism is integral/bonded, as opposed to a removable device?
9. Have you ever used a bonded haemodialysis CVC, in which the technology is integral?