3 Evidence

NICE commissioned an external assessment centre (EAC) to review the evidence submitted by the company. This section summarises that review. Full details of all the evidence are in the project documents on the NICE website.

Clinical evidence

The clinical evidence comprises 8 studies, including 1 peer-reviewed randomised controlled trial

3.1 Eight studies were relevant to the decision problem in the scope:

  • 1 randomised controlled trial (Markowitz et al. 2018)

  • 1 before-and-after study (da Silva et al. 2021)

  • 1 case series (Turan et al. 2012)

  • 2 company reports (Zalut 2007, which is an unpublished case series, and Zillich et al. 2014, which is a non-peer-reviewed randomised study)

  • 2 conference abstracts or posters (Ikinger et al.'s 2007 randomised controlled trial, Nagy et al.'s 2011 comparative study)

  • 1 clinical trial report (Shenfeld and Haris's 2010 unpublished randomised controlled trial).

3.2 Of these, 3 studies were peer reviewed (da Silva et al. 2021, Markowitz et al. 2018, Turan et al. 2012) and only 1 study was done in the UK (da Silva et al. 2021).

The evidence base is heterogenous in population and duration of use

3.3 Patient populations varied across studies. The clinical evidence included 4 studies on short-term (28 days or fewer) catheterisation (Ikinger et al. 2007, Shenfeld and Haris 2010, Zalut 2007, Zillich et al. 2014), 3 studies on long-term (more than 28 days) catheterisation (da Silva et al. 2021, Markowitz et al. 2018, Nagy et al. 2011), and 1 study of unknown duration (Turan et al. 2012).

3.4 The studies' treatment setting varied. Three studies evaluated use in the community (da Silva et al. 2021, Markowitz et al. 2018, Zalut 2007), while 2 studies were based in hospital (Shenfeld and Haris 2010, Turan et al. 2012). The treatment setting was not clearly reported in 3 studies (Ikinger et al. 2007, Nagy et al. 2011, Zillich et al. 2014).

For full details of the clinical evidence, see section 4 of the assessment report.

The evidence base for UroShield is limited in quantity and quality

3.5 The EAC formally appraised only 2 studies (da Silva et al. 2021, Markowitz et al. 2018) because the remaining studies lacked details about study design and methods. It assessed Markowitz et al. (2018) as having an overall low risk of bias. But the study had a small sample size and statistical multiplicity in the data analysis, which may have increased the risk of a type 1 error. The EAC said da Silva et al. (2021) reported limited detail of participants and study methods. It considered that the evidence for the benefit of UroShield in people with short-term catheters is very limited and cannot be used to definitively support any clinical benefit at this time.

Evidence suggests that UroShield may significantly reduce bacteriuria

3.6 Three studies showed that using UroShield resulted in significantly less bacteriuria than comparators (Markowitz et al. 2018, Nagy et al. 2011, Zillich et al. 2014). But 2 studies reported no statistically significant difference (Ikinger et al. 2007, Shenfeld and Haris 2010). The most significant improvement was in people with long-term indwelling urinary catheters.

3.7 The company did a fixed effects meta-analysis of 4 studies to estimate the effect of UroShield on bacteriuria (Markowitz et al. 2018, Nagy et al. 2011, Shenfeld and Haris 2010, Zillich et al. 2014). The risk ratio for bacteriuria was 0.25 (95% confidence interval 0.11 to 0.57) in favour of UroShield, indicating a potential 75% reduction in bacteriuria with UroShield compared with comparators. The EAC reran the meta-analysis using both a fixed effects and random effects approach and got similar results to the company (fixed effects: risk ratio 0.27, 95% confident interval 0.12 to 0.57; random effects: risk ratio 0.34, 95% confidence interval 0.17 to 0.71).

Evidence suggests long-term use of UroShield may reduce UTI

3.8 Three studies reported urinary tract infection (UTI) as an outcome. People using UroShield had fewer new UTIs requiring antibiotics than those using the sham devices (Markowitz et al. 2018) and had fewer UTIs after approximately 12 weeks of use compared with baseline (da Silva et al. 2021). Nagy et al. (2011) found no symptomatic UTIs in either treatment arm.

Evidence suggests UroShield may reduce catheter-related complaints and improve quality of life

3.9 Da Silva et al. (2021) found significantly fewer catheter blockages and unplanned catheter changes after long-term use of UroShield. Five studies also reported improvements in patient-reported complaints with UroShield compared with baseline or comparators. These included lower levels of pain, discomfort, spasm, and itching and burning. Da Silva et al. (2021) and a patient survey reported that UroShield is easy to use and associated with positive outcomes, including fewer catheter-associated UTIs, more time between catheter changes, and improved quality of life.

Cost evidence

The company's cost model shows UroShield to be cost saving in all hospital settings and in community patients with recurrent UTI

3.10 The company submitted 2 simple decision tree models, which compared the costs and health outcomes associated with using UroShield as an addition to standard care in hospital and community settings. The settings and populations considered were:

  • all hospital patients

  • hospital patients with short-term catheterisation (28 days or less)

  • hospital patients with long-term catheterisation (more than 28 days)

  • patients in the intensive care unit (ICU)

  • all community patients

  • community patients with a recurrent UTI.

    Hospital settings had a time horizon of the duration of catheterisation or the duration of treatment for catheter-associated UTI. Community settings were presented as a rolling 30‑day model with the same costs and benefits every 30 days. The company's base case showed UroShield saved £1.65 to £42.05 per person in hospital and £7.77 per person in community patients with recurrent UTI. The company model found UroShield incurred costs of £39.95 per person in the all-community patients population. This was because of the low cost of treating community-based catheter-associated UTIs and the relatively low base rate of infection.

    For full details of the cost evidence, see section 9 of the assessment report.

The EAC accepted the assumptions in the company's model and identified additional assumptions

3.11 The company included several assumptions in its model, which are outlined in table 11 of the assessment report. The EAC accepted these assumptions but changed driver use from 100% to 80% because it considered it unlikely that it would be used every day in its lifespan. The EAC also identified and accepted additional assumptions in the model. This included the key assumption that the reduction in significant bacteriuria in the meta-analysis can be extrapolated to symptomatic catheter-associated UTIs requiring treatment. The models also assume that the definition of 'recurrent' UTIs in community patients can be applied to catheter-associated UTI.

The EAC amended the risk of catheter-associated bloodstream infection and the rate of failure of first-line treatment in the community

3.12 The EAC agreed with most of the values of clinical parameters used in the company's model. It amended the risk that catheter-associated UTIs would progress to a catheter-associated bloodstream infection (CABSI) to 6.6% from the 4.8% used by the company. The EAC also changed the proportion of infections in the community that do not respond to first-line treatment to 14%.

The EAC updated costs for treating catheter-associated UTIs in community and hospital settings

3.13 The average cost of treating catheter-associated UTIs in the community was £386.72 in the company's model. The EAC amended this to £453.54 to account for treatment failures and CABSI. In the hospital model, the EAC updated the cost of a bed day in the ICU to the 2019 to 2020 reference costs (£1,620). The mean cost per catheter-associated UTI was calculated from the cost per catheter-associated UTI per patient, plus the cost of CABSIs for the proportion of patients who have them. The company calculated the mean cost per catheter-associated UTI in hospital as £2,131 and in the ICU as £2,964. The EAC calculated these values as £2,192 and £4,436, respectively.

The EAC's changes to the model did not change UroShield from cost saving or cost incurring in any population

3.14 The EAC's base case showed UroShield to save between £2.40 and £70.13 per person in hospital, and £16.63 per person in community patients with recurrent UTI. UroShield continued to be cost incurring by £39.34 in the all-community patients population. The greatest difference to the company's model was in the ICU, with the EAC's model finding UroShield was £40.64 more cost saving than the company's base case.

The effectiveness of UroShield and the risk of catheter-associated UTIs are the key cost drivers

3.15 The company submitted a one-way sensitivity analysis that varied parameters by the ranges taken from the source evidence or by 25% less than or 25% more than the base values. Its results suggested that the effectiveness of UroShield is the key cost driver in all models. The EAC also did one-way and two-way sensitivity analyses in all 6 populations. For hospital settings, the parameters with the largest impact on cost savings were the effectiveness of UroShield, the rate of catheter-associated UTI, and the cost of treating them. Changes in any of these parameters could convert the base case to cost incurring in populations with small cost savings (all-hospital and short-term use). For the all-community population, only a risk of catheter-associated UTIs greater than 25% (compared with a base rate of 8.5%) independently converts the base case to cost saving. The EAC noted that these rates may occur in nursing homes. In the recurrent UTI community group, UroShield's effectiveness is the only parameter that can independently convert the base case to cost incurring.

The reduction in catheter blockages independent of catheter-associated UTIs may reduce the costs of using UroShield in community settings

3.16 The EAC did an additional two-way sensitivity analysis (see the assessment report addendum on blockages and bacteriuria threshold), altering the risk of catheter-associated UTIs and the risk of catheter blockage. This was done to explore how reducing catheter blockages independent of catheter-associated UTIs may affect the base case. The EAC assumed an equivalent effectiveness for UroShield on blockages and catheter-associated UTI (as suggested by da Silva et al. 2021). Based on this, the risk of blockage at which the base case for UroShield becomes cost-neutral is 1.87 blockages per patient per 30 days. For patients at high risk of catheter-associated UTI, UroShield is always cost saving. UroShield was also found to be cost saving for people who do not get catheter-associated UTIs but who have more than 3 blockages per 30 days that require a catheter change.

Increasing staff time for catheter changes does not change whether UroShield is cost saving or incurring in community settings

3.17 The EAC explored the effect of increased nursing time for unscheduled visits on the cost modelling for the community populations. Results showed that if nurse visit time was increased to 45 minutes, as suggested by the experts, the cost savings increased in the population in the community with recurrent infections. UroShield continued to be cost incurring in the all-community population. Further details and the results of a threshold analysis are in the assessment report addendum on increased staff time for catheter change.

  • National Institute for Health and Care Excellence (NICE)