5 Cost considerations

5 Cost considerations

Cost evidence

5.1 The company identified 3 published economic studies, but the external assessment centre considered them to be outside the scope. It did not identify any further economic studies.

5.2 The company presented a de novo cost model that compared the cost consequences of using SecurAcath in people with peripherally inserted central catheters (PICCs) compared with StatLock, and in people with central venous catheters (CVCs) compared with sutures.

5.3 The model used a decision-tree structure where people entered the model at the point of having a securement device (either SecurAcath, StatLock or sutures). Both trees contained 5 outcomes after securement: no complications; catheter migration; catheter malposition; catheter occlusion; or catheter-related infection (catheter-related bloodstream infection or catheter-related thrombosis). There was an additional outcome of needlestick injuries for health professionals in the suture group.

5.4 The model was constructed with a time horizon of 25 days for PICCs and 3 days for CVCs. Other clinical parameters such as the probability of migration, malposition, occlusion, infection and thrombosis were derived from published and unpublished literature. Device and resource costs relating to the cost of placement (such as nurse time) and complications were also from published and unpublished sources.

5.5 The results of the company's base case found cost savings with SecurAcath of £41.40 per patient for PICCs compared with StatLock and £1,005.60 per patient for SecurAcath with other CVCs compared with sutures. The main reasons for StatLock's greater costs as compared with SecurAcath were device costs and differences in catheter migration rates. For sutures, the main reasons for the greater costs were differences in the rates of catheter-related bloodstream infection or catheter-related thrombosis.

5.6 The company did a 1-way deterministic sensitivity analysis, increasing SecurAcath device costs by up to 200%. It also did multiway deterministic sensitivity analyses, changing the values for each economic and clinical parameter simultaneously by ±20%. In all cases, SecurAcath remained cost saving compared with its comparator. When PICC dwell time was increased to 6 months (reflecting a typical clinical situation of a patient having cancer treatment), the cost saving with SecurAcath increased to £115.00 per patient.

5.7 During the evaluation, a new peer-reviewed article was published that provided a simple cost analysis (Zerla et al. 2017). The published cost comparison of SecurAcath with StatLock only included the cost of the devices. Few details on the analysis are given, but the 30 patients having SecurAcath were compared with a historic control population (Zerla et al. 2015) of 793 patients who had PICCs secured with StatLock. Devices were assumed to cost €30 for SecurAcath and €6 for StatLock, giving a total device cost of €900 for SecurAcath and €4,254 for StatLock. The authors concluded that SecurAcath is cost saving. The authors also report no dislodgments with SecurAcath and compare this with results from Zerla et al. (2015) in which 63 dislodgements were seen. An overall cost for reinsertion for all 63 dislodgements is estimated to be €18,710.

5.8 The NICE adoption and impact scoping report (included in the assessment report overview) describes a single centre's experience of real-world total cost savings of up to £59,000 with SecurAcath when compared with StatLock when placing 1,100 PICCs over 6 months.

5.9 NICE has published a resource impact template on SecurAcath, which can be used to calculate the local resource impact of implementing the guidance.

Additional work by the external assessment centre

5.10 Overall, the external assessment centre considered the company's model to be appropriate given the limited evidence base available, but noted some errors in the model. These included figures wrongly quoted, applying probabilities as rates and a lack of clarity on some sources of evidence. It queried several assumptions in the model: dwell times; the differential impact of securement device placement by nurses or doctors; that no extra resources are needed to place securement devices and the clinical outcomes chosen. It also queried the significant assumption in the model that outcomes were independent of time and were collected over similar dwell times. It noted there was a risk of study heterogeneity and uncertainty about variations in clinical practice and outcome measurements.

5.11 The external assessment centre regarded Yamamoto et al. (2002) as the best evidence available on the incidence of complications, because it reported rates rather than probabilities. Yamamoto et al. (2002) is a single-centre, US-based, prospective randomised controlled trial, comparing StatLock with sutures in patients with PICCs (n=170). The primary end point was catheter-related complications. Mean dwell times were 33 days for StatLock and 35 days for sutures. The risk of total complications was 49.4% and 71.7% for StatLock and sutures, respectively (p=not significant). There was no statistically significant difference in dislodgement or migration rates between the 2 groups, but a significant reduction in infections with StatLock was seen (p=0.032).

5.12 Using this evidence and the updated list price of SecurAcath, the external assessment centre revised the model base case. It assumed clinical equivalence for all outcomes between SecurAcath and comparators, except for needlestick injury, where a reduced risk without sutures was highly likely. Therefore, base-case costs related to placement and maintenance costs over the relevant dwell time only, with needlestick injury costs included where relevant. It also considered 3 dwell times for both CVCs and PICCs: 5 days (short), 25 days (medium), and 120 days (long). Other amendments included: adding StatLock as a comparator for CVCs; varying placement and maintenance times; suturing being done by a band 6 nurse; sutures remaining throughout the dwell time and updating resource costs.

5.13 The revised base case found that StatLock was the cheapest option for PICCs for short dwell times (5 days), but that SecurAcath was cost saving for medium to long dwell times (25 days and over). For CVCs, StatLock was the cheapest securement option for short dwell times and sutures was the cheapest for medium to long dwell times. After the increase in the SecurAcath list price to £20, the external assessment centre reran the model, which increased SecurAcath costs in all analyses by £4 (see table 1 for a summary of external assessment centre results). The external assessment centre concluded that the impact of the list price change was minimal. StatLock remained the cheapest option for dwell times of 5 days and SecurAcath remained cheaper than StatLock for dwell times of 25 days and 120 days.

5.14 A one-way sensitivity analysis, reducing SecurAcath placement time to 30 seconds (as reported by the company), made SecurAcath slightly more cost saving and sutures slightly less cost saving, but did not change the base-case results (see table 1 for a summary of external assessment centre results). Another one-way sensitivity analysis assumed an insertion site maintenance time of 7.3 minutes for sutures, equivalent to the time reported for StatLock in Janssens (2016b). This changed the results for CVCs such that SecurAcath was cheaper than sutures for both PICCs and CVCs with a medium and long dwell time (see table 1).

5.15 The external assessment centre did a multivariate sensitivity analysis including differential risks of migration, dislodgement and catheter-related bloodstream infections, based on the figures reported in Yamamoto et al. (2002) and its meta-analysis. This found that StatLock was the cheapest option for short dwell times for both PICCs and CVCs, but for medium and long dwell times, SecurAcath was the most cost saving (see table 1).

Table 1 Cheapest (cost saving) device in each of external assessment centre's 5 scenario analyses

Scenario

1

2

3

4

5

CVC (central venous catheter): 5 days

StatLock (£5)

StatLock (£5)

StatLock (£5)

StatLock
(£15)

StatLock (£5)

PICC (peripherally inserted central catheter): 5 days

StatLock (£16)

StatLock (£11)

StatLock (£16)

StatLock (£18)

StatLock (£16)

CVC: 25 days

Sutures (£11)

Sutures (£6)

SecurAcath (£8)

SecurAcath (£21)

Sutures (£6)

PICC: 25 days

SecurAcath (£13)

SecurAcath (£18)

SecurAcath (£13)

SecurAcath (£9)

SecurAcath (£13)

CVC: 120 days (not clinically relevant: non-tunnelled CVCs only used short term)

Sutures (£11)

Sutures (£6)

SecurAcath (£90)

SecurAcath (£112)

Sutures (£6)

PICC: 120 days

SecurAcath (£90)

SecurAcath (£95)

SecurAcath (£90)

SecurAcath (£64)

SecurAcath (£90)

Scenarios:

1. Base case (placement and maintenance costs only; no differences in complication rates across devices).

2. One-way sensitivity analysis: assumes a SecurAcath placement time of 30 seconds.

3. One-way sensitivity analysis: assumes a suture maintenance time of 7.3 minutes.

4. Multiway sensitivity analysis including complication rates.

5. Suturing done by a consultant anaesthetist.

5.16 A threshold sensitivity analysis for dwell times using the base case indicated that SecurAcath was the cheapest option for securing PICCs for 15 days or more. For CVCs, the costs of sutures dropped below those of StatLock for dwell times of 8 days or more, but SecurAcath remained more expensive than sutures for securing CVCs over any dwell time. The increased list price of SecurAcath did not affect the threshold analysis.

5.17 For PICCs, the external assessment centre agreed with the company's conclusion that SecurAcath appears to be cheaper than StatLock over medium and long dwell times (25 days and over). Cost savings arise from shorter maintenance times with SecurAcath and the need to replace StatLock weekly. It concluded that these cost savings were robust: it found smaller savings in the base case excluding complications, but similar results in sensitivity analyses including complications.

5.18 For CVCs, the external assessment centre agreed with the company's conclusion that SecurAcath was cost saving compared with sutures over short dwell times, but disagreed with the exclusion of StatLock as a comparator. Additional analysis including StatLock concluded that it was the cheapest for CVCs with short dwell times in all scenarios. For medium to long dwell times, suturing was cheaper than SecurAcath or StatLock in the base case (excluding complications). However, evidence suggested an increased risk of infection with suturing. Sensitivity analyses including complications found SecurAcath to be cheaper than suturing and StatLock over 25- and 120-day dwell times. This led the external assessment centre to conclude that SecurAcath is likely to be the cheapest option for securing CVCs over medium and long dwell times.

5.19 The external assessment centre reviewed the cost analysis in Zerla et al. (2017). The authors report shorter maintenance time for SecurAcath than StatLock, but do not include this in the cost comparison. They report median maintenance times of 10 minutes for SecurAcath and 20 minutes for StatLock, but the methodology for these estimates is unclear. Although the times reported are longer than the estimates of 4.3 minutes for SecurAcath and 7.3 minutes for StatLock reported by Janssen et al. (2016b; used in the external assessment centre cost analysis), they were in about the same ratio (that is, 1:2). The external assessment centre noted that the estimated cost of PICC reinsertion is similar to the cost of PICC dislodgement used in its cost analysis. In summary, the external assessment centre concluded that this study does not significantly change the findings of its cost analysis.

5.20 The external assessment centre used the increased maintenance times from Zerla et al. (2017) in an updated threshold analysis. The results reduced the cost-saving threshold for SecurAcath to a dwell time of 8 days or more. The external assessment centre noted that the use of maintenance times from this study instead of Janssen et al. (2016b) would not change the conclusions of its cost comparison (that is, StatLock is cheaper than SecurAcath over short dwell times and SecurAcath is cheaper than StatLock over medium and long dwell times). The external assessment centre highlighted that Janssen et al. (2016b) provides higher quality evidence than Zerla et al. (2017) on maintenance times.

Committee considerations

5.21 The committee noted that the experts disagreed with the external assessment centre's assumption that nurses would place sutures in the NHS, because they considered it would usually be done by a doctor in an operating theatre environment. Furthermore, the committee noted that when the external assessment centre had recalculated the CVC costs for using a consultant anaesthetist to place sutures, there were only minor differences in the results of the cost modelling (see table 1), with sutures becoming slightly less cost saving compared with SecurAcath in the base case.

5.22 The committee was advised by the experts that a 25-day dwell time for PICCs was an underestimate of routine clinical practice, because haematology and oncology patients usually have PICCs in place for 4 to 6 months, and even up to 1 year.

5.23 The committee concluded that while SecurAcath may take a few minutes longer than StatLock to place and remove (although the experts indicated that this difference reduces with increased experience), maintenance times with SecurAcath are a lot shorter than with StatLock.

5.24 The committee concluded that SecurAcath would not usually be used for tunnelled (Hickman) CVCs or implanted ports, but may be used for non-tunnelled CVCs. However, the committee also noted that non-tunnelled CVCs are used for short-term vascular access (usually less than 10 days). Furthermore, if SecurAcath is used for non-tunnelled CVCs, experts advised that in their experience, an adhesive device would also be placed on top of SecurAcath as an additional measure to prevent potential dislodgement. For all these reasons, the committee concluded that the cost-modelling results for non-PICC CVCs (with dwell times for up to 120 days) are unlikely to be clinically relevant.

2022 guidance review

5.25 For the guidance review, the EAC revised the model to reflect 2021 costs. Further details of the revised model are in the cost update in the review decision from July 2022.

5.26 During the review, the original economic model was corrected so that the model accounted for the weekly replacement of StatLock devices. This meant that in the original guidance, SecurAcath should have been cost saving by £12.60 to £148.54 for PICCs with medium- to long-term dwell times, respectively.

5.27 Based on the 2022 guidance review updated cost model, the EAC updated the costs included in the original model. It found that SecurAcath was still cost saving for medium- and long-term dwell times of PICC lines but that the extent of cost saving was reduced compared with the original guidance. Cost savings result from shorter maintenance times and less need for device replacement with SecurAcath.