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.
3.1 The company submitted 7 studies from its literature search, including 3 full-text papers (Brunelli et al. 2018, Hymes et al. 2017 and Weiss et al. 2021) and 4 abstracts (Glennon et al. 2020, Li et al. 2019, Nitz et al. 2021 and Sibbel et al. 2020). The EAC accepted 6, excluding Nitz et al. because the intervention and outcomes did not match the scope and because it felt it did not add to the decision problem. The 3 full-text papers included a total of 10,757 participants. For full details of the clinical evidence, see section 4 of the assessment report in the supporting documentation on the NICE website.
The 2 prospective cluster randomised controlled trials are the most relevant to the decision problem
3.2 Brunelli et al. and Hymes et al. were the most relevant to the decision problem. Both were prospective, multicentre, open-label cluster randomised controlled trials, which included 40 sites each. Outcomes varied. Brunelli et al. compared ClearGuard HD caps with Tego (needleless connectors) plus Curos (disinfecting caps). Hymes et al. compared ClearGuard HD caps with standard central venous catheter (CVC) caps.
3.3 The remaining full-text paper (Weiss et al.) described a large retrospective analysis but was considered methodologically weak. The abstracts (Glennon et al., Li et al. and Sibbel et al.) were all retrospective analyses with limited detail and did not add much more to the decision problem. The studies had largely homogenous populations. The proportion of men ranged from 51% to 53% and mean ages were 61.1 years to 62.8 years (except for Glennon et al., which studied children).
3.4 Most of the studies reported bloodstream infection rates, although they did not always use the same terminology. Positive blood cultures were the primary outcome in the randomised controlled trials. They both reported significantly lower positive blood culture rates for the ClearGuard group than for the comparator group. The incidence rate ratio (IRR) in Hymes was 0.44 (p=0.01) and in Brunelli 0.37 (p=0.001).
3.5 Rates of hospital admission were lower for the ClearGuard group in 3 studies (Brunelli et al., Hymes et al. and Sibbel et al.), although this was not significant across the studies. Not many of the studies reported length of hospital stay or rates of mortality. None of them reported intravenous antibiotic use or staff time.
3.6 There are 9 records of adverse events on the US Food and Drug Administration's (FDA) MAUDE (Manufacturer and User Facility Device Experience) database. Two reported that the caps came off for 1 patient. Six reported that the caps became detached while the patients were asleep. One reported that the rod broke loose in the catheter. No patients were injured. None of the full-text papers reported adverse events. For full details of the adverse events, see section 6 of the assessment report in the supporting documentation on the NICE website.
3.7 An abstract by Glennon et al. estimated total annual costs per patient for ClearGuard HD caps of £7,078, compared with £18,050 for antimicrobial lock solutions. This was estimated from 4 high-risk children having dialysis. The EAC noted that the authors did not do any sensitivity analysis to assess how robust the cost and rate were. It also noted that the applicability of the results to an adult setting was not certain.
3.8 The model included a decision tree that looked at cost savings with ClearGuard HD caps against 4 comparators:
standard CVC caps plus alcohol wipes for disinfection
standard CVC caps plus antimicrobial lock solution and alcohol wipes for disinfection
Tego needleless connectors on their own, with manual decontamination of the catheter hub with alcohol wipes.
The model had a 1‑year time horizon for cost and health outcomes. For full details of the cost evidence, see section 4 of the assessment report in the supporting documentation on the NICE website.
The EAC's minor amendments to the model and parameters address mortality, comparators and disinfection protocols
3.9 The EAC agreed that the overall structure of the model, time horizon, population, most comparators, outcomes, and assumptions were acceptable and appropriate for the assessment. The EAC excluded the mortality branch of the model, saying that the cost of caps and cost of treating catheter-related bloodstream infections (CRBSI) was adequate without the need for the mortality branch. One comparator (Tego needleless connectors) was not considered appropriate by the EAC because it is a connector alone and therefore out of scope. The EAC provided additional analysis in the ClearGuard HD caps arm around disinfection protocols when using ClearGuard HD caps. This was based on discussions with clinical experts, who expected these disinfection protocols to still be used.
Sensitivity analysis shows cost savings are from baseline incidence rate of infection with ClearGuard, comparators and the cost of treating CRBSI
3.10 The EAC recommended that all parameters not validated by clinical data should be varied up and down by 50% in the sensitivity analysis. The EAC's results were similar to the company's. The parameters that had the largest impact on cost results were:
baseline incidence rate of infection associated with the comparator
the IRR associated with ClearGuard
the average cost of treating CRBSI.
Cost savings remain even when the CRBSI incidence rate is increased in the ClearGuard group and decreased for comparators
3.11 The company did 4 'worst case' scenario analyses, in which the base-case baseline infection rate associated with each of the 4 comparators was at the lower end of the value range. The IRR of CRBSI with ClearGuard was at the upper end of the value range. For these scenarios, based on clinical expert opinion and varying clinical estimates from published studies, the EAC recommended varying the parameters up and down by 50%, or by a range informed by the evidence (rather than up and down by the 25% suggested by the company). ClearGuard remained cost saving in all the scenarios. Another scenario reduced the cost of antimicrobial lock solution; ClearGuard remained cost saving by £418.
Threshold analysis of baseline infections shows ClearGuard HD caps are still cost saving at infection rates that are clinically unlikely
3.12 The scenario analysis results were also supported by the EAC's threshold analysis. This reported cost-neutral break-even points for different CRBSI incidence rates per 1,000 CVC days:
0.228 with standard caps (baseline rate was 0.7)
0.000001 with antimicrobial lock solution (baseline rate was 0.598)
0.111 with standard caps (baseline rate was 0.63).
3.13 Experts and committee members flagged discrepancies in agency costs of standard caps compared with the cost model, which were likely to be because of volume discounts in practice. The EAC input the reported value of £0.03 for the cost of standard caps into the model. ClearGuard remained cost saving compared with:
standard caps and wipes by £351 (from £387)
standard caps, antimicrobial lock solution and wipes by £1,096 (from £1,132).
3.14 The final results showed ClearGuard HD caps were cost saving compared with all the 4 comparators. The company submission reported cost savings per patient of:
£408 compared with standard caps and wipes
£1,167 compared with standard caps, antimicrobial lock solution and wipes
£590 compared with Tego needleless connector and Curos disinfecting caps.
The EAC's revised base-case cost savings, with added disinfection costs in the ClearGuard HD caps arm, showed cost savings per patient of:
£387 compared with standard caps and wipes
£1,132 compared with standard caps, antimicrobial lock solution and wipes
£568 compared with Tego needleless connector and Curos disinfecting caps.