5.1 No existing studies were identified on the cost impact of the geko device.
5.2 The sponsor submitted a de novo cost analysis using a decision tree model that estimated the cost associated with the geko device compared with no mechanical prophylaxis. The model population was patients for whom current mechanical methods of prophylaxis are impractical or contraindicated.
5.3 The decision tree structure was an amended version of the model from the NICE clinical guideline on venous thromboembolism. The model assumed that patients treated with the geko device experienced a reduction in their baseline risk of deep vein thrombosis. Of the patients who went on to experience deep vein thrombosis, most would have either symptomatic or asymptomatic deep vein thrombosis but some would progress to pulmonary embolism. A proportion of patients with deep vein thrombosis also experienced post-thrombotic syndrome, a permanent comorbidity that could generate costs over the patient's lifetime. Further, it was assumed that the patients who had a pulmonary embolism also had a risk of death. The time horizon for the decision tree was 1 year but the model also included the lifetime (15 years) cost of post-thrombotic syndrome. The External Assessment Centre stated that it believed the model structure captured the clinical pathway of care, assumptions and health states in an appropriate manner for the evaluation.
5.4 Most of the clinical parameters were based on the NICE clinical guideline on venous thromboembolism. The key assumptions for clinical parameters used in the model were:
The underlying risk of deep vein thrombosis was 29.1% with no prophylaxis (this was based on the average risk of deep vein thrombosis for all surgical-related patients according to the NICE clinical guideline on venous thromboembolism).
The proportion of deep vein thrombosis progressing to a pulmonary embolism was 10.5%.
There was a 6% chance of pulmonary embolism causing death. No other mortality cause was considered.
The relative risk of a deep vein thrombosis after treatment with the geko device was 0.39.
Post-thrombotic syndrome occurred in 25% of patients with symptomatic deep vein thrombosis or a pulmonary embolism and 15% of patients with asymptomatic deep vein thrombosis.
5.5 No evidence was available for the reduction in relative risk of deep vein thrombosis associated with the use of the geko device. The sponsor's assumption of a relative risk of 0.39 was based on the incidence of subclinical deep vein thrombosis after the use of neuromuscular electrostimulation as reported in Browse & Negus (1970). The sponsor stated that this was a conservative assumption and further justified this because the value fell within the range (0.31–0.58) identified for intermittent pneumatic compression in the NICE clinical guideline on venous thromboembolism. The External Assessment Centre disagreed with this assumption.
5.6 The cost of the geko device was £22 per pair exclusive of VAT. The cost of purchasing the device per course of 6 days, to treat both legs, was therefore £132.
5.7 In the sponsor's model, the cost per patient estimated for the geko device was £359 and for the comparator (no prophylaxis) it was £565, resulting in a cost saving for the geko device of £206 per patient. After correcting for an error in the hourly nursing cost, the External Assessment Centre calculated the cost saving per patient to be £197.
5.8 The sponsor conducted univariate, 2-way and probabilistic sensitivity analyses. The 3 factors that affected the cost analysis the most were the cost associated with post-thrombotic syndrome, the relative risk of deep vein thrombosis associated with the geko device as a form of prophylaxis, and the proportion of deep vein thromboses that are symptomatic. The probabilistic sensitivity analysis showed that the geko device remained cost saving in 99% of simulations performed, with a mean cost saving of about £200 per patient. The sponsor concluded that the geko device was cost saving compared with no prophylaxis. The External Assessment Centre, stated that, although it believed the underlying assumption on risk reduction to be flawed, the sensitivity analysis covered all the uncertain variables, was well performed and that the results supported the conclusions about cost savings from the submitted model.
5.9 The sponsor performed subgroup analysis in people for whom pharmacological prophylaxis is indicated and prescribed. An economic model was developed using values for the relative risk of deep vein thrombosis with pharmacological prophylaxis alone and with pharmacological prophylaxis plus the geko device of 0.14 and 0.02, respectively. Compared with pharmacological prophylaxis alone, the geko device in combination with pharmacological prophylaxis was cost saving for the first 2 days and cost neutral if used for 3 days. It was not estimated to be cost saving after more than 3 days of treatment, with an incremental cost of £69 after 6 days of treatment.
5.10 The Committee considered that the cost model structure was appropriate and that the sponsor had addressed some of the uncertainties in the cost model through sensitivity analyses.
5.11 The Committee discussed the relevance of studies (some conducted many years ago) that demonstrated the efficacy of neuromuscular electrostimulation in reducing deep vein thrombosis. It noted that the reduction in relative risk in deep vein thrombosis used for the geko device (0.39) in the base case was taken from a neuromuscular electrostimulation study in 1970 by Browse and Negus, and was further justified by falling within the range identified in the NICE clinical guideline on venous thromboembolism (0.34–0.58) for intermittent pneumatic compression. The Committee heard from the External Assessment Centre that the unique mode of action of the geko device introduces uncertainty about the association between the type of muscle contractions generated and a reduction in the incidence of deep vein thrombosis compared with those generated by using either neuromuscular electrostimulation or intermittent pneumatic compression.
5.12 The majority of experts advised the Committee that the increase in venous blood flow shown by other mechanical prophylaxis devices was comparable to that demonstrated for the geko device. The Committee heard a range of expert opinions about the generalisability to the geko device of the reduction in risk observed in neuromuscular electrostimulation studies. It judged, on balance, that a reduction in risk for people who are unable to receive any other means of prophylaxis was plausible. The Committee therefore considered that the assumption in the cost model that the geko device reduces the risk of venous thromboembolism compared against no prophylaxis was reasonable. The Committee noted that in the model a relative risk of 0.39 was used and so the baseline risk of deep vein thrombosis with no prophylaxis of 29% was reduced to 11% with the use of geko. This base case gave an estimated cost saving of £197 per patient for geko compared with no prophylaxis. The Committee was also aware that the geko device continued to be cost saving up to a relative risk of deep vein thrombosis of 0.76, meaning that for a baseline risk of 29%, using geko would be cost saving as long as it reduced the risk to less than 22%. The Committee concluded that even if the risk reduction associated with geko was less than that associated with intermittent pneumatic compression devices, it was unlikely that the reduction in risk would be so small that the geko device would incur costs, especially for patients who are unable to receive alternative means of prophylaxis.