5.1 The committee discussed the current management of suspected gastroenteritis in the NHS. It heard from clinical experts that the condition is usually self-limiting and does not need further investigation or treatment. The committee noted that there may be specific groups of people in whom gastroenteritis could have a greater clinical impact, for example, people who are younger, older, are immunocompromised or who have pre-existing health problems. It heard that these groups of people may benefit from treatment and hospitalisation, and also that there are specific pathogens that may need treatment, such as Salmonella, Campylobacter and Shigella. The committee concluded that although gastroenteritis is generally a self-limiting condition, for certain groups of people or clinical scenarios, the identification of gastrointestinal pathogens is of substantial clinical benefit.
5.2 The committee considered the effect of gastroenteritis on infection control practice. It heard from the clinical experts that people in hospital who have diarrhoea are often nursed in isolation until their symptoms resolve or the presence of infectious organisms has been excluded. It also noted that infection control teams often have a low threshold for isolating people because of the risks associated with infectious pathogens causing hospital outbreaks; that is, people who have symptoms that may be associated with an infection are isolated before the cause of the symptoms has been investigated. The committee also heard from a patient expert that being nursed in isolation may not be a positive experience for the patient, especially because they are prevented from having contact with other people on the ward. This can be particularly damaging for children who also miss school attendance during isolation. It also heard that patients and carers can feel anxious while waiting for test results to confirm whether or not an infection is present, particularly when they are not fully informed of the possible effect of the results on their treatment plan. The committee concluded that because the integrated multiplex polymerase chain reaction (PCR) tests are intended to give results quicker than current methods, they could potentially benefit infection control practices by allowing earlier decisions to stop isolation and so improve the experience for patients, who are being cared for by the infection control team.
5.3 The committee reviewed the evidence available on the xTAG Gastrointestinal Pathogen Panel, the FilmArray GI Panel and the Faecal Pathogens B assay. The committee noted that most of the included studies reported data on diagnostic accuracy only and further, that no subgroup analyses had been possible because of a lack of data. It also noted that no data were available for the Faecal Pathogens B assay and concluded that it could not consider this test further in its discussions.
5.4 The committee discussed the accuracy of the xTAG Gastrointestinal Pathogen Panel and the FilmArray GI Panel. It noted that because of the absence of a reference standard, the external assessment group (EAG) had not been able to calculate traditional measures of test accuracy such as sensitivity and specificity. Instead, it had presented the data as measures of both positive and negative agreement. The committee heard from the clinical experts that this was an acceptable approach because the tests used in current practice (culture, single PCR tests, enzyme immunoassays and microscopy) are unlikely to be 100% accurate. It noted that the results of the meta-analyses suggested that there was good agreement between the new tests and current methods, but that there were exceptions to this. One study (Pankhurst et al. 2014) reported low (0.455) positive agreement for Salmonella with the xTAG Gastrointestinal Pathogen Panel when compared with culture. The committee considered that the reasons for this were not clear, but heard from the clinical experts that detecting Salmonella with current methods is known to be highly accurate because enrichment broth is routinely used in the culture process, reducing the likelihood of false-negative results. It also noted that a further study (Gu et al. 2015) showed low positive agreement for adenovirus with both the xTAG Gastrointestinal Pathogen Panel (0.130) and the FilmArray GI Panel (0.130) when compared with PCR for adenovirus. The committee heard from the clinical experts that this is probably because the PCR test used as the comparator was designed to detect multiple serotypes of adenovirus, some of which are not common causes of gastroenteritis. The committee concluded that the tests showed reasonably good agreement with the tests used in clinical practice, but that there was substantial uncertainty around their diagnostic accuracy because of the absence of a reference standard.
5.5 The committee discussed the clinical significance of discordant results for the new tests compared with current methods of detecting gastrointestinal pathogens. It noted that only limited data on verifying discordant results were available from the systematic review because studies that tried to resolve discordant results often used a PCR test that would show bias in favour of the new tests. The reason for the bias, particularly for bacteria and parasites, was because they use the same analytical approach. The committee concluded that this introduced substantial uncertainty into an assessment of diagnostic accuracy, particularly regarding the proportion of true- and false-positive results and the new tests' accuracy for clinically significant disease.
5.6 The committee discussed the complexity of verifying discordant results in clinical practice. It heard from the clinical experts that verifying discordant results could be problematic because of the lack of an accurate reference standard, but that clinical follow‑up may be used to establish the clinical significance of positive results. The committee also heard that the reasons for additional positive results with the new tests could include the new tests detecting a broader spectrum of pathogens than current methods (for example, non‑O157‑shiga toxin producing Escherichia coli), the new tests having greater sensitivity for pathogens, or the new tests detecting non-viable pathogens that are unlikely to be of clinical significance. The committee concluded that the clinical significance of additional positive results arising from the new tests, and the reasons for discordance need to be established before the new tests can be recommended for routine use in the NHS (see section 6.1).
5.7 The committee discussed the likely effect of the new tests on clinical practice. It noted that the systematic review had not found any evidence of clinical management being changed because of the results from the xTAG Gastrointestinal Pathogen Panel and the FilmArray GI Panel. The committee heard from the clinical experts that early detection and treatment may have substantial benefits for certain groups of patients, for example, people who are immunocompromised, or for certain pathogens, for example, Escherichia coli O157. Early identification of pathogens could also result in early discharge when a more benign pathogen that is less likely to lead to complications is identified. It also noted that negative test results that are received more quickly could benefit hospitals, because they would allow people to return to the main ward rather than being nursed in isolation, or in some instances allow discharge from hospital. However, the committee heard from the clinical experts that in practice, the reasons for de-escalating infection control measures or discharging a person from hospital are complex and also take into account the resolution of clinical symptoms or whether a person is known to be colonised with drug-resistant pathogens. Furthermore, any de-escalation decision taken using the information from the new tests would highly depend on the clinician's confidence in the tests' negative-predictive values and the breadth of pathogens included in the panel. It also heard that the effect of the new tests on resource use could vary between NHS trusts, hospitals and departments depending on the case mix of patients that they see and the isolation facilities that are available. The committee concluded that using the new tests could affect treatment and infection control decisions, but that the lack of evidence on these outcomes means that it is not possible to determine whether this would be realised in clinical practice at present.
5.8 The committee considered the cost-effectiveness analyses for the xTAG Gastrointestinal Pathogen Panel and the FilmArray GI Panel. It noted that 5 economic models were available, but heard from the EAG that they should be considered as exploratory models only because of the lack of clinical outcome data for the new tests, and the assumption that the comparator is 100% accurate. The committee noted that the models suggested that the new tests could be cost effective, but that the confidence intervals around the incremental costs were very wide, reflecting the substantial uncertainty in the models' conclusions. The committee concluded that the models were useful for investigating which parameters have the greatest effect on the model outputs, but that conclusions on the cost effectiveness of the new tests could not be made with the evidence available at present.
5.9 The committee discussed the hospital-based models and noted that the one-way sensitivity analysis for the base-case hospital-model suggested that it is highly sensitive to the assumptions made about length of stay and treatment, partly because the incremental cost and quality-adjusted life year (QALY) differences are small. Also, the hospital base-case model was sensitive to the assumption made about the proportions of false-positive results because this parameter drives unnecessary treatment and hospital stays. The committee heard from the clinical experts that it was uncertain whether the new tests would reduce length of stay in practice because it was likely that the test results would not affect the duration of symptoms. It also heard that a decision to discharge a patient from hospital or stop infection control measures may only be considered if a clinician had confidence in the negative test results. The committee therefore concluded that data on clinical outcomes and resource use in the hospital setting, particularly the effect on management after a negative test result, are needed to reduce the substantial uncertainty around these key parameters.
5.10 The committee discussed the influence of negative test results in the hospital-based models. It heard from the EAG that around 70% of results obtained with the xTAG Gastrointestinal Pathogen Panel and the FilmArray GI Panel were negative in the studies included in the systematic review. The committee noted that the sensitivity analyses suggested that negative results were a possible driver of cost savings because they could influence resource use and length of stay. It heard from the clinical experts that false-negative results or insufficient pathogen coverage by the new test could lead to infections being missed, which may then cause outbreaks in a hospital. This may also be a significant problem for residential homes. The committee noted that these possible adverse outcomes from false-negative results had not been included in the economic models and concluded that these potential effects would need to be considered in future cost-effectiveness analyses.
5.11 The committee considered the test costs used in the economic models. It heard from the clinical specialists that the cost of the comparator could have been overestimated in the model. It noted that the cost of the comparator was from a published study (Goldenberg et al. 2015), which the clinical experts suggested may have used additional PCR tests that are not mandated by Public Health England's UK standards for microbiology investigations for gastroenteritis and diarrhoea, for example, for Clostridium difficile, and consequently are not widely used in current practice. The committee considered that the costs used may not represent standard practice. It also noted that the additional scenario analyses on test costs provided by the EAG suggested that test costs only influence the community-based models because the new tests remain cost saving in the hospital-based models under all test cost assumptions. This is because there are no hospital-stay costs and only limited treatment costs in the community models. The committee also noted the threshold analyses provided by the EAG, which suggested that when conventional testing costs drop below £37, the xTAG Gastrointestinal Pathogen Panel is no longer cost saving in the community-based models. Under all assumptions, the FilmArray GI Panel test cost was always greater than the cost of current practice. The committee concluded that the cost of conventional testing was subject to uncertainty in each of the models, and is also likely to vary between trusts in practice.
5.12 The committee considered the level of uncertainty in both the clinical- and cost-effectiveness analyses. It noted that there was substantial uncertainty around the accuracy of the new tests and also a lack of evidence on their effect on clinical management and resource use. In combination, these uncertainties mean that there is insufficient evidence to determine the cost effectiveness of the new tests at present. However, it noted that the exploratory cost-effectiveness analyses done for the xTAG Gastrointestinal Pathogen panel and the FilmArray GI Panel suggested that the new technologies could be cost saving. The committee therefore decided that there was too much uncertainty at present to recommend the new tests for routine use, but wished to encourage further research (see section 6).
5.13 The committee heard from the clinical experts that there are several modular multiplex assays that can detect bacteria, viruses or parasites, available to the NHS. It noted that these technologies fall outside of the scope for this assessment, which focused on integrated multiplex assays that are able to detect bacteria, parasites and viruses simultaneously. However, it considered that the uncertainties highlighted in this assessment, particularly regarding the clinical significance of PCR positive results, might reasonably apply to the modular multiplex assays.
5.14 The committee noted that assessing the integrated multiplex PCR tests is further complicated by the lack of a reference standard against which the new tests can be compared. It heard from the clinical experts that culture and microscopy are unlikely to be 100% accurate and that the comparison of viral PCR tests may be confounded by the serotypes detected by each assay. Nevertheless, the committee concluded that verifying discordant results was an important area of further research for the new tests and it therefore wished to encourage further methodological research on ways of assessing diagnostic accuracy in the absence of a reference standard.
5.15 The committee noted that the clinical outcomes and resource use associated with the new tests were likely to differ between population subgroups and hospital- and community-based settings. It discussed the importance of outcomes such as length of time excluded from work and school in community settings but noted that these fall outside the reference case, which takes the perspective of the NHS and personal social services. Subgroups that could be considered in further research include children younger than 5, people with a recent history of foreign travel to areas other than western Europe, North America, Australia or New Zealand, and people who are immunocompromised. The committee noted that the ongoing Integrate study, funded by the Department of Health and Wellcome Trust, aims to assess the clinical utility of molecular panel tests in managing gastroenteritis in the community. The committee concluded that future research should take into account differences that may arise between these subgroups and different healthcare settings.