Queen Elizabeth University Hospital: case note review - overview report

This overview report examines the incidence and impact of qualifying episodes of infection in paediatric haemato-oncology patients cared for at the Queen Elizabeth University Hospital and the Royal Hospital for Children from 2015 to 2019 and the potential link to the hospital environment.

10. Summary of Findings and Recommendations For Action

This chapter is structured to answer the questions we were asked to address at the outset of our Review and offers recommendations for consideration and action by NHS GCC, and other organisations.

10.1 How many children in the specified patient population have been affected, details of when, which organism etc.?

The work undertaken to define the number of patients and infection episodes that would be the subject of our Review appears comprehensive. We are not able to ascertain with complete certainty that any patients/episodes that should have been included were omitted in error, but we have no reason to believe this to be the case. We identified only two episodes and one patient we deemed ineligible, resulting in a final population of 84 patients and 118 episodes of infection in our Review.

We found that the patients broadly represented the population of patients we would expect to be under the care of the Paediatric Haematology Oncology service at NHS GGC, given that it also houses a unit for Teenagers and Young Adults. Their ages ranged from 3 months to 18 years 10 months at the time of their first infection episode, with a median age of 5 years 11 months. There was an unexpected excess of female patients which we suggest is investigated further, but this may still be a chance finding. The great majority of the patients had a diagnosis of leukaemia (as expected, the largest subgroup) or other cancer but a minority had other forms of serious blood disease or a non-malignant condition.

Although, over three quarters of patients experienced 1 episode of infection, 10 had 2 episodes and several had 3 or more episodes, up to a maximum of 8 episodes in one patient.

10.2 Is it possible to associate these infections with the environment of the RHC and the QEUH?

We were able to conclude that bacteraemia was Unrelated to the hospital environment in only 8 (7%) episodes and, for reasons we have discussed in detail, we were not able to identify any episodes that were Definitely linked to the environment (as discussed in section 5.6). The remainder of the episodes were graded, in varying degrees, as Possible or Probable in their relationship to the hospital environment.

This is not as satisfactory a conclusion as many will have hoped we would be able to reach but we have described the standards of proof we required and discussed the complexity of attributing cause/origin in this population of patients.

It is without question, however, that our decision making was affected by the inconsistencies we encountered in the data we received from NHS GGC: data that, we had hoped, would clarify concerns about the maintenance and surveillance of the environment, the water system, and the use of typing methodologies to link different bacterial isolates.

We conclude that the difficulty the organisation had in locating, collating and presenting these data to us supports our belief that this information may not have been readily and/or consistently available in real time for their own investigations over the period of our Review. The very fact that, in late 2020, such data remained difficult to provide to us suggests that the previous years of concern and investigation of Gram-negative environmental bacteraemia had not translated into clear evidence that good quality data about the control of the environment were being sought, interrogated and stored in a retrievable format for future use.

We have, nevertheless, identified 37 (32%) of the bacteraemias in the Review as being ‘More likely’ to be linked to an environmental origin. These infection episodes are characterised by a particular excess of Stenotrophomonas spp. but do not otherwise appear to be related to any distinctive microbiological profile or to have occurred more than expected in any particular period in the years covered by our Review.

We are surprised that the evidence for an excess of GNE bacteraemia in the Paediatric Haematology Oncology patients was challenged by some within the organisation. By 2018, we suggest that simple observation should have identified a disturbing pattern characterised by the occurrence of bacteraemias caused by some very unusual microorganisms and apparent clusters of some of those more commonly encountered. The widespread contamination of the water system seems to have been accepted operationally and NHS GGC’s response, notably its decision to close and relocate an entire clinical unit in September 2018, must be interpreted as evidence of the organisation’s acceptance that the environment presented a risk of serious infection to a vulnerable group of patients. Although the investigations undertaken to that date had failed to identify a single cohesive hypothesis for the origin of many of the infections, the approach taken to surveillance thereafter did not appear to match the severity of what had already occurred.

10.3 Was there an impact on care and outcomes in relation to infection?

First, it should be recognised that infection occurs in Paediatric Haematology Oncology patients and carries risk, regardless of its likely origin. We have characterised the impact of all the GNE infection episodes experienced by the whole group within our Review and then looked separately to determine if these were different in those episodes we judged to be ‘More likely’ to be associated with the hospital environment.

We created a 5-point scale by which we defined the overall impact of each infection on the patient. This was based on a number of specific criteria which we shall highlight separately. In summary, we identified only 5% of episodes with a Negligible or Minor overall impact whilst 38% of episodes were associated with a Major or Critical overall impact.

In looking at the individual components of the impact assessment, we identified that 87% patients experienced a hospital admission of more than 7 days directly as a result of their infection, and this was greater than 14 days in 50%. Seventy-five (68%) of infection episodes required removal of the central line to control the infection; this is a striking finding because it also conveys an additional risk of general anaesthesia, first to remove the line and then (in almost all cases), to insert another one. This also carries a significant logistic and resource cost in the additional operating theatre utilisation required.

Twelve patients (11% of those evaluable) required admission to the intensive care unit solely or principally because of their infection, of whom the majority (75%) could be discharged to the ward within 3 days. This statistic tells its own story in relation to how sick these patients may become and illustrates again the resource burden that GNE bacteraemia imposed.

Treatment disruption was, as we have discussed, more difficult to characterise but we estimated that treatment delays of more than one week were seen in 29% patients (and for more than 2 weeks in 12%). It is not possible to ascribe clear significance to such observations because many other factors are involved and delays in treatment are common during cancer care. We believe most clinicians would accept that, under most circumstances, a delay of 1 week is very unlikely to be significant in terms of patient outcome. However, it also seems logical to accept that the longer the delay beyond that point, the more likely there could be an impact on disease control.

These are not trivial findings and indicate the scale of the impact of GNE infections within the whole group. When we looked separately at the 37 episodes we deemed ‘More likely’ to be associated with the hospital environment, the pattern of impact was generally similar except for an increase in risk of admission to intensive care. This may link to the excess of Stenotrophomonas spp. infections seen in this group.

We measured AE using two different approaches – first by exploring incident reporting through NHS GGC’s Datix system, and second by using the Paediatric Trigger Tool (PTT). Although many of the triggers identified by the PTT relate to expected complications of chemotherapy or represent other support measures commonly required by this group of patients, the incidence of adverse events identified in this way far exceeds the evidence available from Datix reports. Furthermore, it was apparent that when incidents recognised as adverse events were entered into Datix, there was a clear possibility that the situation might be misclassified and/or its risk underestimated. The principal lesson here is that, used appropriately, the reporting of events into Datix could provide a valuable tool for auditing patient safety in this group of high risk patients, as it is intended to do.

The work using the PTT also provided an opportunity to compare the overall incidence of adverse events in these patients at NHS GGC with paediatric populations in other hospitals: our conclusions are that, when comprehensive data were used, NHS GGC performed in line with that of other comparable institutions.

We found that the deaths of 2 of the 22 patients who had died by the time of the publication of this report were, at least in part, the result of their infection. Both also had other serious medical problems and it is our view that, even without the infection, their survival would still have been uncertain. In one child, who died in PICU 6 days after the last positive culture, sepsis had been implicated at the time as the principal cause of death and was recognised as such on the death certificate issued by NHS GGC. The second child died in PICU at a longer interval (36 days) after the last positive culture and a number of other contributory factors were present. We decided that the bacteraemia was contributory to the cause of death and this was reflected in the death certificate issued by NHS GGC. In both cases we had determined that the infections were both Probably related to the hospital environment and fell within our ‘Most likely’ to be related to the environment group. Of the remainder, 19 had died of their underlying disease (all leukaemia/cancer related); and 1 from other causes unrelated to infection.

10.4 What recommendations should be considered by NHS GGC – and, where appropriate, by NHS Scotland, more generally – to address the issues arising from these incidents to strengthen infection prevention and control in future?

In our work in undertaking this Review, we have explored data pertinent to an understanding of the nature of each infection and to the factors at play in determining its likely origin, subsequent management and influence on patient outcome. We also reviewed the IPC processes in place, and the approach taken to the investigation of these infections when identified internally as an infection incident or outbreak. We identified specific concerns that we have discussed in Chapter 8.

NHS GGC should take immediate steps to ensure greater consistency in the way the organisation monitors and investigates GNE infections in Paediatric Haematology Oncology patients. The approach hitherto has been fragmented and incomplete. In responding to this report and our recommendations, NHS GGC should assure patients, families and staff of a new approach. It is particularly important that it does so before the Paediatric Haematology Oncology service returns to Wards 2A and 2B. In this way, it will be seen that change has been implemented and that risk will be effectively monitored in the return to the upgraded environment.


1. Overall Management of Gram-negative environmental infection in Paediatric Haematology Oncology

1.1 Every GNE bacteraemia occurring in a Paediatric Haematology Oncology patient at NHS GGC should be comprehensively investigated using RCA methodology, whether or not it is considered at the outset to be related to the hospital environment or thought to be part of a potential outbreak. This will ensure that future consideration of the underlying issues can be informed by consistent, comprehensive and prospectively collected data.

1.2 A multi-professional group, with a defined and consistent membership representing all appropriate skills and backgrounds, should be established with responsibility for continuing oversight of these data: for assessment of its quality, and completeness, and for its analysis and reporting. The intent is that this group, which should have external representation, will grow in collective expertise and knowledge; have a shared understanding of the history and challenges encountered since the opening of the new QEUH/RHC site; and will be able to define and guide the organisation’s response to future concerns about environmentally acquired infection in this group of patients. The group should report directly to the IPC Manager and Lead Infection Control Doctor and its findings form a standard part of upward reporting of IPC issues within NHS GGC.

2. Demographic profile of patients

Given the unexplained but significant excess of female patients in the Case Note Review, the Paediatric Haematology Oncology service should audit all bacteraemias for a sufficient period either to reassure that there is no real gender effect, or to investigate further if this proves to be the case.

3. Environmental surveillance

3.1 The data systems used to document facilities maintenance activity in clinical areas need to consistently capture the exact location of the work done; the date(s) on which the work was actually done; and be accessible to inform the IPC process, including the investigation of clusters and outbreaks.

3.2 The frequency with which facilities maintenance activities occur in specific ward areas should be reported on a regular basis in a way that informs wider awareness of the vulnerability of the environment and tracks changes in the pattern of such activity.

3.3. The precise location of any swab or water sample taken for microbiological surveillance, and the date on which it was obtained, must be recorded and the results made accessible to inform the IPC process, including the investigation of clusters and outbreaks.

3.4 When a suspected infection outbreak is being investigated, the plans agreed for environmental sampling of the relevant area must demonstrate a systematic approach appropriate to the circumstances of the investigation.

3.5 When the Chair of an IMT (or similar future structure) identifies that environmental samples are required to inform an investigation, these should be taken, reported back promptly and evidenced in the IMT minutes.

4. Water testing

4.1 A systematic, fit for purpose, routine, microbiological water sampling and testing system is required to provide assurance going forwards. How the results from such sampling/testing are recorded, accessible and used to highlight concerns should be reviewed, including to ensure that investigations of possible links between clinical isolates and water/environment sources can be informed in a timely way. In addition, investigations of possible links between clinical isolates and water/environment sources should consider whether (short or medium/long term) changes to the routine microbiological water sampling and testing system are required.

4.2 NHS GGC should ensure that the SOP for Minimising the Risk of Pseudomonas aeruginosa infection from water explicitly states whether this also applies to high risk areas other than adult and paediatric intensive care units and neonatal units.

5. Infection Prevention Control Practice and Audits

5.1 NHS GGC should review the current approach to IPC audit: a) to ensure that the component elements are addressed individually and that the RAG rating is not determined only by an overall score; and b) to show that the governance and assurance process relating to improvement action plans can demonstrate if interventions have been effective. Quality improvement methodology should be used to drive and sustain improvement.

5.2 The current status of IPC audit should form a routine and documented component of IMT assessment.

5.3 Greater effort should be made to ensure that deficits identified by IPC audits are remedied, re-audited, linked to measures of ongoing quality improvement/compliance, and clearly documented.

5.4 Greater attention should be paid to the evidence for benefit from Enhanced Supervision by demonstrating sustained improvement in standards where this approach is introduced to a clinical area.

5.5 The validity of Hand Hygiene audits should be strengthened by ensuring the staff sample audited is sufficiently representative in terms of numbers and types of staff; and that effectiveness of the interventions are monitored to demonstrate sustained improvement.

5.6 The frequency of Hand Hygiene audits should be increased when there are concerns about infection rates potentially related to the environment

6. Infection Prevention Control Communication

NHS GGC should ensure better communication between the Microbiology and IPC teams. We recommend a forum by which sharing of information and actions occurs in real time to support and improve quality of care to patients, maintain progress and discuss action for any potential change in a patient’s condition or linked infections.

7. ICNet Alerts

NHS GGC should review the ICNet alert organism list to ensure that, at a minimum, it reflects the advice in the Scottish NIPCM and to ensure that it is further updated to reflect experience with GNE bacteraemias.

8. Infection Incident and Outbreak Policy

8.1 NHS GGC should review its Standing Operating Procedure regarding the use of the term HAI to make it clear whether this includes all Healthcare Associated Infections. This is a specific issue in the context of patients who, like those in Paediatric Haematology Oncology, frequently and repeatedly attend the hospital as outpatients, day patients and inpatients and for whom the distinction between Hospital Acquired Infection (HAI) and Healthcare Associated Infection (HCAI) is unlikely to be useful.

8.2 NHS GGC should revisit how they will monitor and, if necessary, trigger concerns about future outbreaks of Gram-negative environmental infections. Relaince on SPC charts to determine if episodes of infection caused by unusual/uncommon microorganisms are significant should be re-evaluated. The process in place for much of the Review period appears to have been insensitive to identifying clusters that should have raised earlier concerns about potential for a common/environmental source of infection.

8.RCA methodology should become the standard approach to the investigation of serious infections in Paediatric Haematology Oncology patients.

8.4 NHS GGC should consider the further and consistent use of the RCA process across the organisation a) to identify evidence of common themes as a cause of infection over time; and b) what can be extracted from the RCA process for organisational learning and improvement.

8.5 NHS Scotland should consider if this approach should become a recommendation in the NIPCM.

9. IMT Process

9.1 The IPC Team should ensure IMT minutes are filed with all supporting papers so that a complete record of the discussions held, evidence presented, actions agreed and the overall report concluding the process, is available and accessible in a single place.

9.2 The IMT action log should be a continuous and evolving document throughout all meetings in an IMT series. The log should be reviewed and updated at each meeting so that there is a clear record of actions agreed, responsibility held and tasks completed. The IMT should not be closed if there are actions which have not been completed.

9.3 The absence of IMT reporting at the closure of an IMT sequence is a breach of NHS GGC’s own policy. This should be remedied so that practice complies with policy.

9.4 In addition to confirming that due process has been followed in line with organisational policy, IMT and other IPC reports intended for upward reporting within the organisation should more fully describe the scale and significance of the incident that has been investigated from the patient perspective.

9.5 NHS GGC should assure that the governance of the IMT process, its reporting and escalation to Board level, is clearly defined and followed; and that an audit trail of all evidence related to any suspected or actual outbreak is clearly documented and fully reported.

10. Bacterial typing data / Reference laboratory reports

10.1 NHS GGC must (continue to) develop a comprehensive and searchable database that allows details of microbiology reference laboratory reports to be compared between samples of the same bacteria obtained from different patients or environmental sites.

10.2 The system for integrating microbiology reference laboratory reports into the patient microbiology record needs to be reviewed and strengthened. Similarly, the system for ensuring that microbiology reference laboratory information is available to and used by the IMT process, including the investigation of clusters and outbreaks, needs to be reviewed and strengthened.

11. Patient Records

11.1 NHS GGC should undertake a review of the current effectiveness of the system for collating, storing and integrating both scanned hand written records and digitally recorded records and how this achieves an accurate, accessible and chronologically accurate health record for each patient.

11.2 NHS GGC should clarify their strategy for further evolution towards fully digital records

11.3 Consideration should be given to the integration of the microbiology recommendations regarding the diagnosis and management of infections, as currently documented in the Telepath patient notepad, into the patient clinical record.

12. Patient location coding

It should not be possible to code patient activity to a clinical area in which the patient was not present: this should be addressed.

13. Adverse Events

13.1 The Paediatric Haematology Oncology service should engage with regular reporting and analysis of adverse events. Admission to PICU is an obvious way of identifying, for audit purposes, the patients most likely to have the most serious (Category I) AE.

13.2 The PTT offers a useful tool to identify and monitor trends in the occurrence of adverse events that occur during care.

13.3 NHS GGC should assure and report consistent utilisation of the Datix system, and audit the validity of the classification and risk categorisation given to incidents by its staff.

14. Central Venous Line Care

14.1 The Paediatric Haematology Oncology service should review the practice of ‘challenging’ central venous lines in line with evidence for its risks and benefits.

14.2 When it is agreed that a central line should be removed for optimal management of a patient’s infection, operating theatre and anaesthetic resources must be made available to ensure its prompt removal (within 24 hours).

14.3 The Paediatric Haematology Oncology service should ensure that a decision not to remove a central venous line contrary to the advice of the microbiologists is always documented in the medical record.

15. Other aspects of Clinical Care

15.1 The Paediatric Haematology Oncology service should ensure that Morbidity and Mortality reports are not restricted to a review of patients who die. Future GNE infections should be used as a trigger for an M&M review; to assess management and outcome; and with the inclusion of an action plan to identify approaches to reduce risk and improve care.

15.2 International consensus guidelines have recently been published for use of antibiotic prophylaxis in Paediatric Haematology Oncology. These should be reviewed by both the service and by the Managed Service Network, and local and network policy and practice should be amended accordingly.

15.3 The Paediatric Haematology Oncology service should audit the use of antibiotic prophylaxis against the new policy once implemented.

15.4 The Managed Service Network and NHS GGC should review any changes to the use of shared care that have evolved as a result of the service disruption experienced in recent years, and ensure the structures and processes in place adequately address patient safety and staff support across the shared care network.



Back to top