5. The role of the hospital environment as a source of infection
Concerns about the QEUH/RHC hospital environment have been widely discussed. They were discussed in detail the Independent Review undertaken by Dr Andrew Fraser and Dr Brian Montgomery, published in June 2020, and will be further addressed by the Oversight Board whose final report is to be published at the same time as the publication of our own report.
Reported deficits in the hospital environment include (but not are not limited to) issues such as: the design and maintenance of the water system; lower than required air exchange in patient rooms and inadequate positive pressure protection of patient rooms; the lack of provision of particulate (HEPA) filtration in some higher risk patient areas; and uncertainties around the appropriate utilisation of chilled beams for temperature control in rooms used for immunocompromised patients.
The focus of the Independent Review was explicitly on the built environment of the QEUH and problems related to infection prevention and control. Its ToR state that it was charged “to establish whether the design, build, commissioning and maintenance of the Queen Elizabeth University Hospital and Royal Hospital for Children has had an adverse impact on the risk of Healthcare Associated Infection and whether there is wider learning for NHS Scotland”.
It is not the remit of the Case Note Review to revisit that objective but, in addressing our task to consider how many children in the specified patient population had been affected by the defined types of infection over the period from May 2015 to December 2019, and to answer the question whether it is possible to associate those infections with the environment of the RHC and the QEUH, it is inevitable that we have had to place our considerations in that context.
The remit of the Scottish Hospitals Inquiry now being undertaken by the Right Hon. Lord Brodie is much broader. Its overarching aim is “to consider the planning, design, construction, commissioning and, where appropriate, maintenance of both the Queen Elizabeth University Hospital Campus (QEUH), Glasgow and the Royal Hospital for Children and Young People and Department of Clinical Neurosciences (RHCYP/DCN), Edinburgh. The Inquiry will determine how issues relating to adequacy of ventilation, water contamination and other matters adversely impacting on patient safety and care occurred; if these issues could have been prevented; the impacts of these issues on patients and their families; and whether the buildings provide a suitable environment for the delivery of safe, effective person-centred care”.
We recognise, therefore, that our work and conclusions are not only informed by the findings of the Independent Review but also will be of relevance to the work of the Scottish Hospitals Inquiry.
The conclusions of the Independent Review record that there were, amongst many other findings, examples of non-compliance in the design of the water and ventilation systems at QEUH. The report also concluded that, at commissioning, there was a lack of documentation to prove the water and air ventilation systems in Royal Hospital for Children (RHC) wards 2A and 2B and QEUH 4B (ultimately to become the location of the adult bone marrow transplant (BMT) service, and currently offering accommodation for the paediatric BMT service whilst the deficits identified in wards 2A and 2B are being rectified) were compliant with specification.
In a succinct summary of the challenges identified with the water system, Drs Fraser and Montgomery wrote that ‘the water system of the hospital became, from within one year of admitting patients, the emerging source of infections that entered the blood streams of a substantial number of child patients with haematological cancers. The Health Protection Scotland report (2018) states that they were investigating a ‘contaminated water system’; the entire new hospital was affected and, after immediate local action in the vicinity of the affected patients, the remedy became a new system of additional chemical disinfection for the hospital water supply’.
A key statement made in the Executive Summary of the Independent Review, and relevant to the work of our Review, reads as follows:
‘Patients, staff and visitors who are vulnerable due to immuno-suppression, or who are in proximity to patients with certain highly infectious communicable diseases, have been exposed to risk that could have been lower if the correct design, build and commissioning had taken place’.
Nevertheless, the two high level findings reported by the Independent Review read as follows:
1. In the course of the Review, through examination of documentation, listening to witnesses, discussion with experts and input from the Review’s expert advisers, and site visits, we have not established a sound evidential basis for asserting that avoidable deaths have resulted from failures in the design, build, commissioning or maintenance of the QEUH and RHC.
2. The QEUH and RHC combined now have in place the modern safety features and systems that we would expect of a hospital of this type. The general population of patients, staff and visitors can have confidence that the QEUH and RHC offers a setting for high quality healthcare.
We suggest that these two, more positive conclusions stand in some contrast with the immediately previous statement we have quoted, and with the considerable detail of adverse findings in the hospital environment highlighted elsewhere in the Independent Review. This places the relevance of our work into sharper focus and, whilst we acknowledge that considerable work has been undertaken within NHS GGC to address or mitigate the risk associated with the environmental concerns described by the Independent Review, we are aware that some staff expressed a view that these concerns remained unresolved even as late as the second part of 2019.
This chapter provides our observations on the maintenance of the hospital environment and its microbiological surveillance, and on the inferences we derive for the risk of environmentally acquired infection.
5.2 The built environment and its maintenance
Regular maintenance and repair of the building, its equipment and fixtures and fittings is a normal, and essential, part of the life of any hospital. Nevertheless, the nature and frequency of interventions by Facilities department maintenance staff or other contractors provides the potential for environmentally acquired infection, despite the fact that any work of this nature must be risk assessed and mitigated in compliance with HAI-SCRIBE requirements,. Furthermore, the nature of any incident they are called to resolve may itself be evidence that a potential source of infection exists in the environment (for example, the risk posed by a blocked sink or shower drain).
We have therefore undertaken a retrospective review of a large database of logs and documents provided by NHS GGC that offered data related to the maintenance of the clinical environment with a particular focus on Wards 2A and 2B and 6A and 4B.
This has not been straightforward. Initially we found it difficult to interrogate the large amount of data related to facilities management because of the way this information was structured and presented. Nor did the initial data submissions from NHS GGC allow us to readily link a maintenance action to a specific clinical location; frequently these initial records identified only the ward and not the individual room. They also did not provide the precise date work was undertaken, more often indicating a range of days between the requisition and the completion of the work. This experience suggested to us that the data systems used within NHS GGC to record facilities maintenance activity are better designed to manage workload than to provide information of potential relevance in the management of clinical situations, particularly IPC events.
Latterly, further work by NHS GGC to clarify the data and reformat its presentation provided a more workable solution to better allow us to investigate links between patients and maintenance activity in their care environment. However, even the later database did not always reflect the location of work undertaken with sufficient detail for the information to be useful. Subject to these constraints, however, we found very few examples where work undertaken in close temporal and physical relationship to the care environment of a patient could be linked to the occurrence of a specific infection, or to potential outbreaks of infection.
Overall, however, it was apparent to us that there were large numbers of requisitions for Estates and Facilities department interventions in the Haematology Oncology wards and that those relating to plumbing and drainage seemed particularly evident (although we have no suitable comparative data with which to compare these observations). These problems include blocked toilets or drains; leaking showers and taps; and the management and maintenance of chilled beams following reports about leaks or condensation, or both, and where additional cleaning was required for control of dust.
We have not been able to ascertain with clarity what planned programme of inspection and preventative maintenance existed or was actually undertaken on a routine basis, particularly with regard to the chilled beam system (outside that suggested as part of actions agreed at IMT).
5.3 Cleaning and Standard Infection Prevention and Control Measures (SICP)
Effective cleaning and IPC practice make a significant contribution to ensuring patient safety within the hospital environment. A cycle of audit and subsequent improvement in practice contributes to the ethos of a learning organisation.
To investigate the potential link between cleaning standards, infection prevention practice and the incidence of bacteraemia, we reviewed the IPC data and the relevant national and local policies commensurate with the time period of our Review.
5.3.1 IPC audits
Infection prevention safe practice in acute care audits looks at a wide range of factors including environment, isolation, equipment, hand hygiene, personal protective equipment, linen, waste and indwelling devices, including intravenous lines. Where suboptimal practice is identified, remedial action should be instigated through a systematic action/implementation plan, work execution and recording of completion.
The National Infection Prevention and Control Manual (NIPCM) specifies standards for infection prevention and control and includes an audit tool for each of the SICP, which should be performed monthly by the Senior Charge Nurse (SCN). Non-compliance with SICP audits should be resolved locally by the SCN working with their team. On occasion SICP audits may be performed by the IPCT during incidents or outbreaks to ascertain practice against national guidance. Any non-compliance should be recorded and an action plan implemented for improvement. NHS GGC used an audit tool based on the national guidance in place when QEUH/RHC opened.
We reviewed reports of IPC audits and SICP audits undertaken at NHS GGC between 2016 and 2019. These were based on the NHS GGC IPC audit tool as part of a planned audit cycle. SICP audits formed part of an audit cycle that appeared to have commenced in 2017. The overall score from an IPC audit then defines when a re-audit is due and the report generates an action plan for any non-compliance or if standards are not met.
We have reviewed the domestic and estates facilities management tool audits made available from 2015 to 2019. All the audits we reviewed demonstrate high compliance to the standards set in the National Cleaning Services Specification.
Compliance against an audit resulted in a RAG + Gold rating according to criteria shown in Table 5.1
|RAG + Gold score||% compliance obtained||Re-audit interval|
During 2017 IPC audits in ward 2A were undertaken monthly from May to September and then twice monthly in October and December of the same year. We noted that whilst a score may be classified as Gold, the highest rating, some elements may have less satisfactory compliance. For example, an audit might score 91% overall and yet the environment score could be 67% and equipment 75%. The Gold outcome would indicate that a re-audit was not required for 12 months despite there being obvious areas for improvement: in such situations we would expect to see a focused plan for improvement in areas that were not compliant. Significantly the guidance within the NIPCM about audit includes a statement about the use of RAG scores: ‘…although RAG status can be useful; where it is used there should also be structures in place which weights the risk associated and not necessarily concentrates on the percentage score’.
In 2018, there were monthly audits for Ward 2A (until it closed in September and patients were transferred to Ward 6A). We noted again that an overall Gold rating could be achieved but with some sections (usually environment and equipment) achieving non-compliant scores, demonstrating no sustained improvement. As an overall Gold standard was reached, the next scheduled audit would not have been required for 12 months. This is not indicative of a culture that was thinking carefully enough about quality improvement and we are not convinced that the data shown in Table 5.2 are sufficient to tell the whole story.
There is insufficient evidence from documentation we have reviewed to assure us that the improvement actions were robustly and continuously undertaken and we were unable to ascertain the governance process underpinning action plans.
5.3.2 Enhanced Supervision
A process of Enhanced Supervision was used by NHS GGC to support ward 2A to monitor and drive improvement with IPC. The aim of the supervision was to support staff and provide real time education to the clinical teams. The process involves review of areas such as equipment, cleaning, clinical wash hand basins, PPE and hand hygiene. If standards were not adequate, the issue was referred to the nursing manager for action. It is not clear to us, from the documents we have received, how actions were pursued or how improvement and learning was shared and sustained.
During 2017, there were six such interventions in Ward 2A but the Enhanced Supervision appears to have ended prior to assurance that all the standards had been achieved.
During 2018, Enhanced Supervision was undertaken from March to December (the period from late September relating to Ward 6A) and we observed that standards were often under achieved.
Enhanced Supervision was undertaken again in Ward 6A during 2019, and yet our observations were that the standards were again often not compliant. This leaves us to question whether this approach offered a reliable improvement intervention and we are uncertain where the accountability lay for the assurance it provided in relation to IPC.
5.3.3 Hand Hygiene
Hand hygiene is considered an important practice in reducing the transmission of infectious agents that cause healthcare associated infections. It is one of the core SICPs. Hand hygiene refers not only to hand washing using the established technique but also to the appropriate use of alcohol based hand rubs at point of use. The NIPCM for Scotland has a framework for hand hygiene to support a safe and clean care audit programme. We reviewed hand hygiene audit results undertaken by the NHS GGC hand hygiene coordinators. The information we received for audits between 2015 and 2019 did not appear have a consistent frequency, and it was unclear to us how a lower compliance score triggered an improvement response and re-audit.
The audit is measured as a percentage of opportunities taken for hand hygiene and compliance with correct procedure. It then provides a combined score to give an overall indication of hand hygiene practice. From the data we have seen, it is not clear how many hand hygiene opportunities were observed for each audit or which staff groups were represented in the audit, although circumstances relating to non-compliance were occasionally described in IMT minutes.
Regarding the use of improvement plans for improving hand hygiene, we saw, for example, that in 2017 there was a programme of ward-based hand hygiene education, but we were unable to link the impact with subsequent improvement in compliance or any effect on the incidence of infection episodes. We also saw data pertaining to Enhanced Supervision of ward 2A during 2017 but only one question related to hand hygiene. Where inconsistencies or non-compliance were observed, the ward manager was informed but we have not been able to identify records of improvement actions.
In addition, data provided by hand hygiene audits were also included as a part of the SICP audit programme. This audit records only a yes/no response and from the data we received we were unable to identify how regular hand hygiene audit was used as a tool to contribute to sustainable improvement in the provision of care.
We would have expected to see more frequent hand hygiene audits in the ward environments, particularly during the periods where continuing concerns regarding the increased occurrence of bacteraemia were under investigation by an IMT. Example 5.1 provides one situation to illustrate our concern:
The minutes of a PAG meeting in early June 2019, called because of 2 recent Stenotrophomonas infections, and 2 further GNE isolates in May that year, document that the last hand hygiene audit on Ward 6A had been held in October 2018 and the last Infection Control audit in November 2018.
The infrequency of these audits seems surprising, as was a statement that enhanced supervision of environmental cleaning was discontinued in April 2019 on the basis that practice observed was of a consistently high standard.
Given the continuing focus on a possible link between bacteraemia (particularly due to GNE bacteria) and the hospital environment and its water supply, we cannot find consistent reference to IPC audits in the IMT process.
The documentation we have reviewed does not assure us there was a robust enough culture of continuous improvement for IPC within the organisation during the period of our Review or that the Enhanced Supervision process for IPC had sustained impact.
We were unable to determine a strong governance and assurance process for IPC and formed a view that the focus of the organisation appeared to be directed more towards the task of audit than to the achievement of quality improvement outcomes.
5.4 Environmental microbiological surveillance
In contrast to water sampling (section 5.5), we recognise that routine microbiological sampling of so called ‘hard surfaces’ offers little to routine IPC practice but we consider it relevant in the investigation of outbreaks of specific or unusual infection providing it is undertaken systematically.
We have had access to a database, provided by GGC, of ‘hard surface’ samples taken during the period of our Review. This also included samples taken from drains. Initially these data were subject to the same limitations as those for facilities maintenance in that the information supplied frequently failed to link samples to a recognisable clinical location. Later in the Review, reprovision of the data allowed us to investigate links more readily between the location of patient care and environmental microbiology samples. In reality, however, it proved difficult to link environmental samples taken from patient rooms to dates of specific bacteraemia, not least because samples (we had the results for both positive and negative samples) were infrequent and, when taken, seemed not to be taken in a systematic way. It was also often not clear to us which microorganisms had been sought/identified during laboratory processing of samples.
There were, however, occasions when samples requested by the IMT were reported positive for an organism under investigation. A good example of this would be the identification of Enterobacter in drains on ward 6A during a cluster of Enterobacter spp. bacteraemias in 2019. Even here, however, positive samples came from different areas of the ward and were not specifically found in the rooms previously occupied by the patients who developed bacteraemia. This does not, in our view, diminish the argument that the environment was the potential likely source but limits our ability to strengthen the observation that it was.
The further specific significance of microbiological typing to consolidate a relationship between isolates from different sources is discussed in Chapter 8, section 8.3.1.
In other IMT records, where actions recorded that environmental samples should be taken, evidence was not always available to confirm that this had been done (and whether this was a single sampling exercise or was repeated) or, if it had been done, the outcome had not been recorded.
Overall, we were unable to conclude that the organisation had a systematic approach to environmental sampling in the context of either a specific, unusual infection or an outbreak of a more commonly seen infection.
5.5 Water safety
5.5.1 Water testing policies and practice
Following increasing evidence relating to outbreaks and incidents of Pseudomonas aeruginosa in augmented care units, and notably a cluster of infections in a neonatal unit in Belfast, the Department of Health (England) published ‘Water sources and potential Pseudomonas aeruginosa contamination of taps and water systems: advice for augmented care units’ in 2012. An addendum to Health Technical Memorandum 04-01 was also published in 2013 and superseded the 2012 document.
This guidance is concerned with controlling/minimising the risk of morbidity and mortality due to P. aeruginosa associated with water outlets. It provides guidance on: assessing the risk to patients when water systems become contaminated with
P. aeruginosa or other opportunistic pathogens; remedial actions to be taken when water systems are contaminated; protocols for systematic sampling, testing and monitoring of water for P. aeruginosa; and forming a Water Safety Group and developing water safety plans. The guidance is aimed at Estates and Facilities departments and IPC teams and is directed towards healthcare organisations providing patient care in augmented care settings. These include patients:
- who are severely immunosuppressed because of disease or treatment: this will include transplant patients and similar heavily immunosuppressed patients during high-risk periods in their therapy;
- cared for in units where organ support is necessary, for example critical care (adult paediatric and neonatal), renal, respiratory (may include cystic fibrosis units) or other intensive care situations; and
- those patients who have extensive breaches in their dermal integrity and require contact with water as part of their continuing care, such as in those units caring for burns.
NHS Scotland did not adopt a similar approach to water testing in augmented care units until 2018 and was provided in an addendum from HPS to advice directed at neonatal units and adult and paediatric intensive care units.
5.5.2 Water testing at NHS GGC
We set out the summary of the policy above because, whilst the timing of the guidance issued in Scotland means that water systems in Haematology Oncology wards at NHS GGC were not required to be tested for P. aeruginosa contamination, there must have been professional and managerial awareness that such guidance was in place elsewhere in the UK. This ought to have further strengthened the need for regular, systematic sampling/testing of water given the emerging concerns over this timeframe about possible environmental sources for paediatric bacteraemias. NHS GGC informed us that they had in fact implemented testing for P. aeruginosa in 2016 and we have confirmed this by reference to the risk assessment undertaken for that year. However, we found that their SOP for Minimising the risk of Pseudomonas aeruginosa infection from water is confusing: even the 2019 version is still headed ‘Applicable in all adult and paediatric intensivecare units and neonatal units’ and makes no reference to other high risk areas such as transplant units. This is important as critical control of this issue is not just about water testing but also about flushing regimes and alert surveillance.
The investigation undertaken by HFS and the findings of the Independent Review have each confirmed that there were serious issues about the design and commissioning of the water system. The response of the organisation to the point at which additional whole system chlorination was introduced, suggests that these issues were accepted. Yet we have been told that there was a lack of a robust water testing strategy from the point at which the new hospital building was commissioned, including assurance that the system was fit for purpose.
From the information with which we have been provided, it has proved difficult to understand the rationale for how water sampling/testing took place, in particular to assure the organisation that water systems/sources were not related to the observed GNE bacteraemias in children. There did not appear to be a systematic water sampling process in place, or a consistent water system related response to clusters of infections caused by (often unusual/uncommon) GNE bacteria. We are not assured that there was adequate communication about what sampling and testing occurred and the results obtained. We have been told that some key staff involved in IPC at NHS GGC were denied access to water sampling/testing information despite multiple requests. As the concerns increased about whether the bacteraemias occurring in children on the Haematology Oncology wards at NHS GGC might be related to environmental/water contamination, the lack of a clear step change in the organisation’s approach to water sampling, testing, reporting and strategy is of concern.
After repeated requests for information on what water system sampling testing took place, we were provided with data that frequently did not specify the precise location from where a sample was obtained, and/or precisely which bacteria were sought and identified in the laboratory. It is possible that water samples were examined to determine only the burdens (total numbers) of bacteria present, without formal identification of the bacteria present; conversely, samples may have been taken to look for specific bacteria (e.g. in relation to bacteraemias caused by uncommon microorganisms). Specific bacteria may have been sought in some samples, but this does not mean that all bacteria present were identified. Also, searching once or only occasionally for specific bacteria, and from only a limited number of sites, limits the confidence that a bacterium of concern was not contaminating a water point/system and thus could have been the source of one or more bacteraemias. Example 5.2 illustrates some of our concerns.
We summarise here the results provided in a file provided to us labelled ‘2018 Potable Water Master File Complete 13.11.20’.
Despite the electronic title referring to ‘potable’, the samples were actually taken from a mixture of sources including water tanks, taps and showers. The Excel spreadsheet contains detailed information about the samples (n = 2864), dates, investigations and results. At first glance it appears to represent a comprehensive set of sampling/testing information. However, for over 70% of the listed water samples, Cupriavidus is stated as the target microorganism. As such, there appears to have been limited testing for other bacteria performed on these samples.
The file contains results from multiple locations/buildings but 336 are stated as coming from water sources on Ward 2A. Of note, however, with the exception of 22 (dated during September 2018), almost all the samples were taken during one of two adjacent months (i.e. March or April 2018). For Ward 2B we see a similar time constrained sampling pattern, but for only 27 samples; 16 were in March, 2 in May and 9 in September 2018.
We emphasise that 2018 was a year of heightened concern about the possibility of contamination of water sources.
We conclude that these data do not support a systematic approach to water sampling (i.e. frequent, repeated sample collection) certainly for wards 2A and 2B and in the context of concern regarding possible environmental sources of bacteraemias.
In summary, and crucially, without any other clear account of which water points/systems were/were not sampled, when and how often sampling occurred, and which bacteria were specifically sought, we frequently could not confidently exclude these as potential point sources for bacteraemias caused by GNE bacteria that are known to be associated with such environments.
5.6 The likelihood that infections were linked to the hospital environment
Chapter 3 addresses the methodology utilised for the work of the Panel and section 3.6.6 describes the principles we used in reaching our conclusions about the likelihood of an environmental source for an infection in each episode of infection. That section also describes the cautions and limitations we had to consider in making our decisions. Table 5.3 summarises our overall findings.
|Likelihood of a link to the hospital environment||No. of Episodes||Proportion|
|Unable to Determine||1||1%|
Whilst we classified 8 episodes as being unrelated to the hospital environment, and 1 we were unable to determine, of the rest of the episodes (n=109), 76 (70%) fell into the Possible group and 33(30%) into the Probable group.
Our decisions reflected our judgements based on the balance of probability when considering all the data we had available. They also reflect the complexity of drawing such distinctions in a population of patients who, by the nature of their diagnoses and treatments, are susceptible to serious infection. Many of these infections can arise both from endogenous (within the patient him/herself) and exogenous (from the external environment) sources. Exogenous sources include not only the environment of the hospital but also all environments encountered by the patient outside the hospital.
The lack of any episodes being classified as Definite reflects the tight criteria, agreed before we started our Review, that were required to achieve this descriptor. Decisions at this level were also influenced by the inconsistency with which our investigation and evaluation could be informed by data systematically investigating the microbiological environment (section 5.4), the water system (section 5.5.2), and the likelihood that, by using typing methodologies, different bacterial isolates were linked (Chapter 8, section 8.3). Microbiological information alone was insufficient for us to reach our conclusions and we also looked carefully at clinically relevant information. Above all, the complexity of the challenge we faced was in the retrospective acquisition of adequately informative data.
The distinction between classification as ‘Strong Possible’ and ‘Probable’ was often relatively subtle, as was that between ‘Probable’ and ‘Strong Probable’, and by linking these three categories we believe we can reasonably create a group of infections with the closest likelihood of a link to the hospital environment (‘Most likely’ to be associated with the hospital environment). In total, these three groups constituted 37 (34%) of those designated as either possibly or probably related (and accounted for 31% of the whole series). Table 5.4 describes the profile of bacteria encountered in this ‘Most likely’ group of episodes, compared to that of all other episodes.
|Organism||Seen in ‘Strong Possible’, ‘Probable’ & ‘Strong Probable’ groups
‘Most Likely’ (n = 37 episodes)
|Seen in all other episodes
(n = 81)
16 and 214 episodes were polymicrobial (i.e. they involved more than one bacteria)
There is a striking excess of Stenotrophomonas spp. in the ‘Most likely’ group which is significant (Chi square test 14.80; p<0.05) but differences in the frequency of all other bacteria are less obvious. Other characteristics of the ‘Most likely group’ are discussed in Chapter 6.
We also looked at the frequency with which we identified episodes as ‘Most likely’ in relation to the year of infection. We did this in case there might have been a shift in the amount of data available to us over the era of the Review. We found that there was a substantially greater proportion of ‘Most likely’ episodes in 2018 but concluded that this probably reflected the fact that most isolates of Stenotrophomonas (11/21) occurred in that year.
In closing this chapter, we offer one further observation. Whilst we are not reassured about the adequacy of the systems in place to monitor the environment during the period of our Review, and believe that about one third of the episodes we reviewed were ‘Most likely’ linked to the hospital environment, we suggest NHS GGC must also have recognised that some links with the environment were likely to exist. Acting with the support of external advisers, they introduced significant interventions and control measures and it is difficult to consider the actions they took – such as the closing of Wards 2A and 2B (with relocation of services to Ward 6A and 4B); the addition of point of use filters for water outlets; augmented chlorination of the entire water supply; and additional decontamination of the healthcare environment - would have all taken place, primarily to address public confidence (important though that would be), if there was not also some acceptance of environmental risk.