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 EAC assessed 4 studies including 332 people with diabetic foot ulcers (DFUs). One study was a randomised controlled trial (RCT; n=266) and 3 were case series, 1 of which was published as an abstract (the case series included 44, 5 and 17 people). Two further studies identified by the company were not included by the EAC because these were not relevant to the decision problem. For full details of the clinical evidence, see section 3 of the assessment report.
3.2 The Game et al. (2018) RCT was considered to provide the best available data on the use of 3C Patch in relation to the decision problem. This was because it is a UK-based RCT that included people whose ulcers had a less than 50% reduction in area after 4 weeks of standard care (described as 'hard-to-heal' ulcers by the study authors). The trial also measured clinically relevant outcomes and the EAC judged it to have a low risk of bias. However, the EAC noted some issues with the generalisability of the results to current clinical practice. Expert advice indicated that, following the publication of NICE's medical technologies guidance on UrgoStart for treating diabetic foot ulcers and leg ulcers, UrgoStart has become the standard of care. As the Game et al. study took place between 2013 and 2017, only 1 person had UrgoStart in the run-in period. Other protease modulating dressings (classified by the BNF) were used by 2% of people during the run-in. Additionally, clinical experts had different opinions on whether 3C Patch would be continued if there was an infection and advised that the treatment will be at least temporarily halted to evaluate the infection severity. The EAC noted that 3C Patch treatment was continued while ulcers were infected in the Game et al. trial in most cases, in line with the trial protocol. The EAC concluded that although the trial was well conducted, some aspects of the study design may not reflect NHS practice.
3.3 The EAC noted the way the intervention was delivered in the trial did not align to the company's proposed treatment pathway. The company stated that 3C Patch use should be reviewed after 4 to 6 weeks and stopped if adequate progress in healing has not been seen, such as a reduction of 50% or more in ulcer area. This stopping rule was not followed in the clinical trial because everyone in the treatment group had 3C Patch until healing or up to 20 weeks. Clinical experts stated that a 50% ulcer area reduction rule to mark adequate healing progress is not routinely used in practice to judge response to treatment. The EAC considered this an important limitation of the evidence base.
3C Patch increases the proportion of people with complete epithelialisation or healing at 20 weeks in the trial population
3.4 RCT evidence (Game et al. 2018) found that 34% of ulcers (45 out of 132) in the intervention group had complete epithelialisation or healing at 20 weeks compared with 22% (29 out of 137) in the standard care group (odds ratio 1.58; 95% confidence interval [CI] 1.04 to 2.40; p=0.0235). In the case series, healing rates at 20 weeks were 52% (23 people out of 44) and 61.9% (13 ulcers out of 21; Löndahl et al. 2015 and Katzman et al. 2014, respectively).
3.5 RCT evidence (Game et al. 2018) found that 3C Patch reduced time to healing compared with standard care over 20 weeks (hazard ratio 1.709; 95% CI 1.071 to 2.728; p=0.0246). In the subgroup that had healed at 20 weeks, the median time to healing was 72 days (interquartile range [IQR] 56 to 103) in the 3C Patch group compared with 84 days (IQR 64 to 98) in the standard care group (difference 12 days; p=0.0343). This study also found a statistically significant decrease in ulcer area over a 20‑week period in the 3C Patch group (p=0.0168).
Evidence does not support 3C Patch reducing the risk of amputation or ulcer infection and direct clinical evidence for the other company-claimed benefits is limited
3.6 Game et al. (2018) found no significant difference in those with a new infection within 20 weeks, visits reporting infection (as a proportion of total visits) or total days of antibiotic therapy. The study also found no significant difference in new minor or major amputations affecting the index or contralateral limb. However, the study was not powered to detect differences in these parameters. The EAC further noted that although there was a reduced time to healing seen, no data on the demand for care across NHS settings (outpatient, community, primary and inpatient) were presented. Any improvement in quality of life was uncertain as these measures were only reported in an abstract for a small subgroup of people (10 people in the 3C Patch group and 8 people in the standard care group, all with ulcers extending into tendons; Löndahl et al. 2019).
The company's cost model uses a Markov model comparing 3C Patch with standard care in those with hard-to-heal DFUs
3.7 A Markov model was used to estimate costs and quality-adjusted life years associated with the use of 3C Patch plus standard care compared with standard care alone. It took into account the impact of each treatment option on the likelihood of healing, re-ulceration, major amputation, minor amputation and death over a 2‑year time horizon. The population included in the model were those with hard-to-heal DFUs, which aligned with the population included in Game et al. (2018). For full details of the cost evidence, see section 4 of the assessment report.
The company's cost model uses a stopping rule for 3C Patch treatment and makes use of data from an unplanned post-hoc analysis of the trial
3.8 The company's model included a number of assumptions that reflect the company's proposed use of 3C Patch within the DFU treatment pathway. It incorporated an assumption that 3C Patch use would be stopped if an ulcer has not reduced in area by 50% or more within 5 weeks of treatment. This stopping rule was not used in the Game et al. (2018) trial, so the company did an unplanned post-hoc analysis of the trial data to generate the following clinical inputs:
the proportion of people who would stop 3C Patch treatment at 5 weeks (57.9%)
healing rates with 3C Patch at weeks 0 to 5, weeks 6 to 20 and week 21 onwards
healing rates for people who would stop using 3C Patch after week 5 if a stopping rule had been applied.
The company's model structure is appropriate, but the EAC created a second model to consider a 'moderate or severe' infection state
3.9 The EAC judged the overall model structure and time horizon to be appropriate. However, it disagreed with some of the key clinical and cost parameters used in the company's model (see sections 3.10 to 3.13). Additionally, in light of the varying clinical expert views on whether 3C Patch use should continue when an ulcer is infected (see section 3.2), the EAC created a second model (model B) that added a 'moderate or severe' infection state. In this state, people with a moderate or severe infection stop using 3C Patch until their ulcer is no longer infected. The company's model did not have a separate infection state as it followed the protocol used in the Game et al. (2018) RCT whereby 3C Patch was not stopped while an ulcer was infected. Instead, the company's model included infections as recorded in the RCT, with their associated impact on costs and healing rates.
3.10 The EAC made amendments to the costs in the base-case model by using resource use data, when possible, from an unpublished economic analysis of the Game et al. (2018) RCT (Farr et al., unpublished). These changes included adjusting the number and length of outpatient visits and adjusting the proportion of people having inpatient procedures. Dressing costs were also changed from BNF to supply chain. Additionally, the EAC made 3 further changes to the cost inputs:
changed relative costs to absolute costs for additional care for dressing changes, done by district nurses, between outpatient consultations (in both arms of the model)
removed the cost of a district nurse to avoid double counting in the EAC model (as the EAC changed the way in which district nurse costs were included in their model when compared with the company's model) for outpatient and community care costs (in both arms of the model)
applied cost of training up front (as opposed to weekly).
These changes resulted in almost all costs in the EAC model being updated.
3.11 As stated in section 3.7, the company model included a stopping rule applied in the 3C Patch arm, which was implemented at week 5. The EAC noted that in Game et al. (2018), everyone in the treatment arm continued to use 3C Patch until healing or for up to 20 weeks. It also noted that clinical experts stated that the stopping rule used in the company model was unlikely to be implemented in clinical practice. This is because 3C Patch treatment would likely continue if any significant improvement in ulcer size is seen when compared with previous treatments. Therefore, the EAC changed the discontinuation rate to 0% (meaning everyone in the treatment arm would continue 3C Patch until healing or for 20 weeks).
The EAC revised the healing rates in the model in line with published RCT data and its preferred discontinuation rates
3.12 As noted in section 3.7, the healing rates in the company's model were based on an unplanned post-hoc analysis of the Game et al. (2018) trial data. The EAC revised these parameters in their model to reflect the healing and discontinuation rates seen in the intention-to-treat population published in the RCT (Game et al.). This was because the post-hoc analysis excluded a substantial amount of the data, particularly for healing at 6 to 20 weeks in the 3C Patch arm. This increased uncertainty in the probabilities of healing used in the model. This was important because the probability of healing with 3C Patch in weeks 6 to 20 was a key driver in the company model and an absolute reduction in healing rate of around 0.6% changed the direction of the company's cost case.
3.13 The company's base-case results showed cost savings of £191 per person over 2 years when 3C Patch is used instead of standard care. But, the EAC's base-case results found that 3C Patch is cost incurring compared with standard care. The incurred costs were £1,590 per person over 2 years when modelled without an infection state (model A) and £1,993 when modelled with an infection state (model B).
The EAC's sensitivity analysis found the cost of index ulcers and discontinuation rate to be the biggest cost drivers
3.14 The EAC's sensitivity analysis found that the biggest cost drivers in the economic model were the probability of discontinuing 3C Patch and the cost of ulcer treatment when using 3C Patch, standard care or when 3C Patch is discontinued and replaced with standard care. The EAC did a 2‑way sensitivity analysis to explore the impact of varying the probability of discontinuing 3C Patch and the probability of healing with 3C Patch in weeks 6 to 20 simultaneously. The EAC recognised that there is likely to be interaction between these variables. The results suggested that if there is no discontinuation of treatment at 5 weeks (0% discontinuation rate), and weekly healing rates after week 5 are over 4.5%, then 3C Patch would be cost saving. However, this healing rate is significantly higher than the rate used in the EAC base case (2.7%), which was aligned with the Game et al. (2018) RCT.