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.

Executive Summary

In November 2019, NHS Greater Glasgow and Clyde (NHS GGC) was escalated to Stage 4 of NHS Scotland's National Performance Framework as a result of a continuing series of infection incidents at the Queen Elizabeth University Hospital (QEUH) and the Royal Hospital for Children (RHC). An Oversight Board was established by the Director-General of Health and Social Care in the Scottish Government and Chief Executive of NHS Scotland to address critical issues arising from the operation of infection prevention and control, governance, and communication and engagement at the QEUH and the RHC. In parallel, in her statement to Parliament on 28 January 2020, the Cabinet Secretary for Health and Sport commissioned a Case Note Review, to be undertaken by a panel of independent experts. The period defined for our review was from the time the paediatric haematology oncology service moved into the new RHC in 2015, to the end of 2019.

We were pleased to accept the invitation to undertake this work and at the heart of our report lies a shared concern for the safety of children and young people under the care of the Paediatric Haematology Oncology service at the Royal Hospital for Children in Glasgow.

Concerns about risk of infection emerged after the hospital moved from its old site at Yorkhill to the new Queen Elizabeth University Hospital in June 2015. Views about deficits in the design, commissioning and maintenance of the buildings have been aired in other publications and will continue to be addressed in the future. Our task, however, has been to determine how many children and young people with cancer, leukaemia and other serious conditions were affected by a particular type of serious infection, caused by Gram-negative environmental (GNE) bacteria, from 2015 to 2019; to decide, as far as it is possible so to do, whether these infections were linked to the hospital environment; and to characterise the impact of the infections on the care and outcome of the patients concerned. Our full Terms of Reference are set out in Chapter 2. These also include a responsibility to offer, based on our exploration of the issues, recommendations that may strengthen infection prevention and control in future.

An agreed database was used to identify patients eligible for our review. Using healthcare records provided by NHS GGC, we extracted relevant clinical and microbiological information to create an individual timeline for each episode of infection in the included patients. We supplemented this by an assessment of data relating to the location of the patient’s care; results of microbiological surveillance of the environment; detail of any building repairs or maintenance activity taking place in the vicinity; and information about additional investigations undertaken at the time. This approach allowed us to build up an individual picture of the factors that could indicate the origin of an infection, and to assess the impact it had on patient care and outcome.

For every episode, we answered a predefined set of questions to help us determine the likelihood that an infection could be linked to the hospital environment; to quantify key measures of the impact on the child/young person affected; and to do so in a structured and consistent way. We formally reviewed all infection episodes included in the Review twice and, in some episodes, more frequently until we were content that we had been able to access as much of the data we needed as was available.

We also looked at the manner in which NHS GGC had itself assessed, responded and reported the situation at the time, and looked for evidence that common themes were identified and pursued during its investigations. We critiqued the quality and adequacy the information provided to us and formed an assessment of the availability and integration of relevant data within existing NHS GGC systems.

We had access to several hundred documents. We held meetings with individuals and groups of individuals, in particular, staff at NHS GGC. We generated, received and saw many emails, all of which provided additional and/or complementary information. Although we experienced frustrations in access to NHS GGC systems, and were critical of their ability to readily provide data we considered key to our investigation of the hospital environment, our requests to them for access to documents and other data were met with courtesy and helpfulness.

A small number of families availed themselves of the opportunity to provide reflections or questions about their child’s care and, where they did, we reviewed these alongside our other assessments.

Our main findings in relation to the children and young people and the nature of their infections are as follows:

  • 84 children and young people between them experienced 118 episodes of infection which fulfilled the criteria set for inclusion in our review.
  • Their age ranged from 3 months to 18 years 10 months at the time of their first infection.
  • The great majority had a diagnosis of cancer or leukaemia but a small minority had other forms of serious blood disease or another condition requiring the expertise of a haematologist or oncologist.
  • Although over three quarters of patients experienced only one episode of infection, ten had two episodes and several had three or more episodes, up to total of eight episodes in one patient.
  • Using an approach that we describe in detail in our report, we determined that whilst eight episodes were unrelated to the hospital environment, and in one case we were unable to determine the relationship, of the rest 76 (70%) could possibly relate to the hospital environment and 33 (30%) probably did. We were unable to identify evidence that unequivocally provided a definite relationship between any infection episode and the environment. There are complex reasons for this which we discuss in more detail in the body of the report.
  • In the absence of a definitive link to the environment, we nevertheless felt the possibility of a link remained strong. We grouped episodes we had defined as ‘Strong Possible’, ‘Probable’ & ‘Strong Probable’ into a single group which we felt might reasonably be considered to be ‘Most Likely’ linked to the environment. This constituted 37 (34%) of all episodes and included an excess of one particular bacterium (Stenotrophomonas). There was also an increased likelihood that the infections constituting the ‘Most Likely’ group had occurred in 2018: this may well be related to the particular excess of Stenotrophomonas bacteraemias in that year.
  • We designed a framework for assessing the overall impact of an infection on a patient. This framework included consideration of various factors including the duration of hospitalisation attributable to the infection; duration of antibiotic therapy; the necessity to remove the patient’s Central Venous Line (CVL) to resolve the infection; the need for admission for intensive care (PICU); the need to modify the planned delivery of cancer treatment; and death. This allowed us to score overall impact on a five point scale from None to Critical. Only 6 (5%) of evaluable episodes were assessed as having no or minor impact whilst 44 (38%) scored as major or critical. The breakdown of these individual factors can be summarised as follows:
    • 57 (58%) episodes involved an additional hospital stay of over 2 weeks.
    • 78 (68%) episodes resulted in the removal of the patient’s CVL.
    • 12 (11%) episodes required admission to PICU.
    • 60 (48%) assessable episodes resulted in a delay to planned cancer treatment of which 12 (12%) were for more than 2 weeks.
  • We found that the deaths of 2 of the 22 children and young people who had died by the time of the publication of this report were, at least in part, the result of their infection. Both of these children also had other serious medical problems and it is our view that, even without the infection, their survival would still have been uncertain. Within the constraints necessary to protect individual patient identity, we discuss these deaths in more detail in the body of our report.

We recognise that nothing we have been able to measure can truly reflect the broader impact of these infections on the lives of the children and young people who were affected, and their families. Unplanned or prolonged admission, or both, will contribute to the already significant impact that their disease and its treatment has on their lives. It further disrupts schooling, social life, parental work, and the care of siblings or dependent relatives. It contributes to additional anxiety because families are well aware that infection is a risk, can be serious and may be life threatening; also, families are anxious about the consequences of delays to treatment. In this respect, our findings underline the very significant additional burden that these infections, whatever their cause, must have had on the children and young people concerned, and their families.

In respect of the wider issues, we identify a number of areas that have caused us concern in Chapter 8 of the report. These are summarised as follows:

  • We have documented our challenges with NHS GGC over access to data systems but, more importantly, over the time taken to provide us with data we had requested about the microbiological surveillance of the hospital environment and the extent of building, repair and maintenance work that took place in relevant clinical areas during the period of our review. This delay, and others regarding our access to the laboratory information systems, necessitated us to undertake a second complete review of the entire series of infection episodes so as to incorporate information received late in our work schedule. Perhaps most significantly, however, it raised questions for us about how NHS GGC had been able to make effective use of such data in its own investigations of the GNE bacteraemias as they had occurred.
  • We are critical that, despite over five years of experience in investigating outbreaks of GNE bacteraemia and concerns about the hospital environment, NHS GGC had not established an electronic database of microbiological typing results (a key strategy in the ability to link bacteria identified in one person or place with that from another person or place) and consequently had no ability to easily relate potentially linked bacterial isolates. We recognise that this work is now ongoing, and we also acknowledge the considerable amount of work required by NHS GGC staff to provide us with the data that was available. However, the fact that there were too many gaps in terms of which isolates were included in these analyses, together with an inconsistent approach to environmental sampling, led us to conclude that we were unable to interpret the true extent of relatedness between patient and environmental isolates, even with the provision of some data using state of the art Whole Genome Sequencing (WGS) methodology which has more recently been brought into use.
  • ICNet is an electronic patient management system used by the Infection Prevention and Control Team (IPCT) to manage patients identified with possible or confirmed infection. It relies on data being exported from Telepath (the laboratory information management system) and if a microorganism is identified as one of a pre-defined list of ‘alert’ organisms, it will automatically alert the Infection Prevention and Control Team. The National Infection Prevention and Control Manual provides a nationally-agreed minimum list of alert organisms, the purpose of which is to alert NHS Boards to situations that may require further investigation. The guidance states: “the list is not exhaustive and specialist units … will also be guided by local policy regarding other alert organisms not included within these lists”. We found little evidence, even as late as 2019, that (and despite assurance from NHS GGC) the alert list had been modified in light of the evolving experience with GNE bacteraemias. This resulted in frequent absence of alerts being triggered within ICNet, and the subsequent absence of IPCT input in some episodes of the GNE bacteraemia we reviewed.
  • We examined how possible outbreaks of infection were investigated and managed within NHS GGC. We found the process involving the PAG (Problem Assessment Group) and IMT (Incident Management Group) structure to have been inconsistent. We were particularly concerned that, despite the continuing existence of concern about GNE bacteraemias over several years, there was less evolution in the approach to the recognition of an outbreak than we might have expected. We believe there was too much emphasis on standard definitions, inappropriate reassurance from the use of SPC methodology and even an unwillingness to accept that there was a problem. All of this is further clarified in our report.
  • IMT minutes were not always easy to understand in retrospect. Action logs were rarely apparent either within the minutes or separately, which must have limited the ability to track completion or evolution of actions from one meeting to the next - either within an IMT sequence or between consecutive IMT sequences. This suggested a fragmentation of approach and we believe it limited the chance of learning for the future. We did not find final reports at the close of a series of IMT meetings despite this being mandated in the NHS GGC Standing Operational Procedure for Infection Outbreaks. This was despite the fact we saw examples of such documents from IMTs in other clinical areas within NHS GGC, raising questions about consistency in practice across the organisation.
  • Our observations suggested to us that the communication between microbiologists, the infection control doctors and the rest of IPCT may not have been as robust or cohesive as it should be. It seemed that the teams appeared to work independently and that communication between these staff groups was sometimes not as good as would be required for effective IPC.
  • We have recommended a systematic and structured approach to the investigation of all future bacteraemias using Root Cause Analysis methodology. We recognise that this approach was introduced in NHS GGC at the end of 2019 but it is hard to see why, given the experience of repeated GNE bacteraemias over five years, this was not introduced earlier.
  • In our report we have highlighted an example of upward reporting to the NHS GGC Board which, we believe, demonstrates an inconsistency in the process and purpose of reporting; it also raised our concern that this could be an organisation that promotes a focus on process (i.e. that a report was received) rather than ensuring clarity about the cause or consequences of a situation.
  • We used the Paediatric Trigger Tool (PTT) to identify Adverse Events in the care of the patients we reviewed and compared this with findings from NHS GGC’s own incident reporting system, Datix. It is quite clear that reporting to Datix is incomplete and incidents were sometimes inaccurately categorised and under scored for severity. Yet the data acquired from the use of the PTT allowed us to show that NHS GGC compared favourably with other paediatric institutions and we recommend the continuing use of the PTT in the future.
  • We identified significant inconsistencies in the way patient healthcare records were stored and organised within NHS GGC’s Clinical Portal system. This not only added to the complexity of our task but, more importantly, the management of records to ensure they are clear and easy to follow is ultimately an issue for patient safety.

There were of course positive findings, in particular:

  • We found that clinical records kept by the medical and nursing teams were detailed and comprehensive; that there was good communication between the microbiologists and the haematology oncology team about the diagnosis and management of infections; and that communications with parents were generally well documented and of a high standard, despite some parents raising concerns in this respect.
  • We particularly commend the work achieved by the Quality Improvement Group established in 2017 to drive down (the then very high) central line associated line infection (CLABSI) rates. The latest data we have seen show these to have fallen to low levels consistent with international best practice. We should emphasise in this context that a substantial reduction in CLABSI rates does not negate the possibility of an environmental risk for GNE bacteraemia and that continuing surveillance is required.

Our report makes 43 recommendations within 15 separate themes. We recognise that work has already commenced in some areas, some of which represent themes highlighted in previous reports including the November 2019 HPS report. Most of our recommendations apply to NHS GGC but some may have wider relevance to NHS Scotland and to the Managed Service Network for Children and Young People with Cancer.

Overall, we urge NHS GGC to take immediate steps to ensure greater consistency in the way it monitors and investigates GNE infections in Paediatric Haematology Oncology patients as the work to date 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.

We recognise that some families will be disappointed at our ability to identify a links between their child’s infection and the hospital environment with greater certainty than has been possible. This not only represents the limits of a retrospective review and the shortcomings we have described in the data we were able to access, but also highlights the fundamental challenge of identifying a specific source in all such infections. However, the purpose of continuing to try to do so is to further reduce risk to patients in the future.

Whilst it is not our task to determine whether the environment at NHS GGC is now safe from the risk of hospital acquired infection for these patients, we wish to acknowledge the steps the organisation has taken to date to respond to what was an extremely challenging situation.

We would like to thank our Review Team for their outstanding work; the Oversight Board for its guidance; and the many individuals within and without NHS GGC who played their part in informing our Review and in the preparation of this report.

Michael Stevens
Emeritus Professor of Paediatric Oncology, University of Bristol

Gaynor Evans
Formerly Clinical Lead for the Gram-negative Blood stream Infection Programme, NHS Improvement England

Mark Wilcox
Professor of Medical Microbiology, University of Leeds and Leeds Teaching Hospitals.

March 2021


Email: CNOD_Admin@gov.scot

Back to top