5 Cost considerations

5 Cost considerations

Cost evidence

5.1 The company presented 3 published economic studies on surgical procedures for prostate enlargement, 2 of which reported costs for bipolar transurethral resection of the prostate (TURP) compared with monopolar TURP. The External Assessment Centre identified 1 other observational study. The studies came from different healthcare systems (Japan, India and Singapore) where care pathways vary from those in the NHS. In addition, it was not clear whether patients had received treatment with the transurethral resection in saline (TURis) system and the studies did not directly compare monopolar and bipolar systems. The economic studies are summarised in the assessment report and are not considered further here.

5.2 The company submitted a de novo cost analysis comparing the cost consequences of procedures using the TURis system and a monopolar TURP system. The time horizon of the model was a non‑defined short time period designed to capture procedure‑related complications. Costs were modelled from an NHS perspective and a discount rate of 3.5% per year was applied. The population included in the model was men having surgical intervention for prostate enlargement. The model adopted a cost‑minimisation approach based on an assumption of no difference in the efficacy of TURis and monopolar TURP in terms of resection weight or completeness of resection. The model included the cumulative costs associated with the initial surgical procedure, complications resulting from the procedure and the need for reoperation or readmission. The sensitivity analysis also included clot retention and the need for reoperation in the event that the initial procedure was stopped before completion.

5.3 The company's model contained 3 clinical parameters: length of hospital stay, rate of blood transfusion and rate of TUR syndrome. The company used 0.52 days (95% CI 0.30 to 0.74) for reduction in the length of hospital stay, from a meta‑analysis of 3 studies. The reduction in the rate of blood transfusion was taken as 0.36 (95% CI 0.16 to 0.80) from a meta‑analysis of 3 studies. The rate of TUR syndrome was taken as zero for TURis patients and 1.14% (95% CI 0.30 to 1.98) for monopolar TURP from a meta‑analysis of 6 studies. Full details are in section 9.4.3 of the company's submission.

5.4 The equipment costs for the TURis system included capital costs and the consumable costs of the electrodes. The Olympus generator was assumed to be provided without cost. It was assumed that each hospital would need 3 complete TURis systems. The capital costs differed between hospitals that used Olympus monopolar TURP systems and those that did not since some of the components are interchangeable. The company took these costs from Olympus data on file. For hospitals with Olympus monopolar systems, the cost of purchasing a TURis system included 3 working elements and 3 saline cables at a cost of £8800. Hospitals not using Olympus equipment would additionally need 3 each of the following: a telescope, an inner sheath, an outer sheath and a light guide cable at a total cost of £26,715. These capital elements were assumed to have a mean working life of 7 years at 150 procedures a year. This resulted in a capital cost per patient of £9.68 for hospitals using Olympus systems and £29.13 for other hospitals.

5.5 The estimated cost of electrodes for each TURis procedure was based on 1 single‑use loop electrode and in 22% of procedures an additional single‑use roller electrode.

5.6 For monopolar TURP the company assumed that hospitals have an existing system and so capital costs were not considered. The cost of electrodes for a monopolar TURP procedure was estimated to be 50% of the TURis electrode costs; this came to £80.57 per procedure.

5.7 The company included a £1848 cost for TUR syndrome, assuming an additional 2 days in a high‑dependency unit and 2 days in a general ward. The company based the cost of a blood transfusion on an estimate used in a study by Varney et al. (2003), which was £920.40.

5.8 The results of the company's base case stated that the average total cost per patient of using the TURis system was £1043.57 for hospitals using Olympus systems and £1063.01 for hospitals not using Olympus systems, compared with £1177.20 for a monopolar TURP system. TURis therefore reduced costs for hospitals using Olympus systems by £133.63 per procedure and for hospitals not using Olympus systems by £114.19 per procedure.

5.9 The results of one‑way probabilistic and threshold analyses done by the company suggested that these results were robust. The key drivers of the savings in the company's cost model were the reduction in the length of hospital stay and the cost of monopolar consumables.

5.10 The External Assessment Centre considered the company's basic model structure to be appropriate. The External Assessment Centre revised the cost model parameters based on its meta‑analyses results and so used a zero difference in the length of hospital stay between TURis and monopolar TURP; a relative risk of blood transfusion for TURis compared with monopolar TURP of 0.35; and a relative risk of TUR syndrome for TURis compared with monopolar TURP of 0.18.

5.11 The External Assessment Centre considered that the company's costs for blood transfusion overestimated the true costs because several components were included that would not typically be needed. The External Assessment Centre estimated the cost of a blood transfusion to be £329, based on the cost of 2.7 units of red blood cells.

5.12 The External Assessment Centre could not find a rationale for the company's assumption that the cost of monopolar electrodes was 50% of the cost of the TURis electrode. Based on advice from the clinical experts, the External Assessment Centre assumed that all monopolar TURP procedures, in both Olympus and non‑Olympus cases, involved both a loop and a roller electrode. The External Assessment Centre considered that hospitals using Olympus systems obtained the generator on loan and paid the list price for monopolar TURP consumables (£137.75). Hospitals not using Olympus systems have the option to purchase a non‑Olympus electrosurgery unit generator, incurring a higher initial cost but allowing the purchase of monopolar electrodes at a lower price from NHS Supply Chain, saving money over the lifetime of the electrosurgery unit. The External Assessment Centre used a price of £66.84 for hospitals not using Olympus systems (based on the price of generic monopolar TURP consumables [£56.84] from NHS Supply Chain and a £10 per procedure electrosurgery unit cost).

5.13 The results for the base case in the External Assessment Centre's revised model found a total cost per TURis procedure in hospitals using Olympus systems of £1183.99 and in other hospitals of £1203.44. The total costs for a monopolar TURP were £1196.60 for hospitals using Olympus systems and £1125.69 for other hospitals. TURis was cost saving for hospitals using Olympus systems by £12.60, but added costs of £77.75 for other hospitals. The savings are driven by a reduction in risk of TUR syndrome and blood transfusion.

5.14 The External Assessment Centre reported an additional scenario involving readmissions for all causes, based on data from the Fagerstrom et al. (2011) study. The rate of readmission (all causes) for TURis was 5.1% and for monopolar TURP was 16.1%, giving a relative risk for TURis of 0.31, p=0.011. The External Assessment Centre estimated the cost of a readmission (all causes) as £2781, based on the NHS reference cost 2012/13 code LB20D. Results obtained when readmission from all causes was included in the model revealed that TURis saved £319.62 per procedure for a hospital with an existing Olympus monopolar TURP system and £229.27 per procedure for other hospitals.

5.15 The External Assessment Centre calculated a further revision to the model at the request of the Committee, with a change to the mean difference in hospital stay from zero to 0.19 days in favour of TURis, based on the External Assessment Centre's meta‑analysis. The results for the recalculated base case in the External Assessment Centre's revised model found a total cost per TURis procedure in Olympus centres of £1126.04 and in non‑Olympus centres of £1145.49. The total costs for a monopolar TURP were £1196.60 for a hospital using Olympus systems and £1125.69 for other hospitals. TURis was cost saving for a hospital using Olympus systems by £70.55, but added costs of £19.80 for other hospitals.

5.16 The External Assessment Centre calculated a revised result based on the meta‑analysis results for the reduction in readmissions associated with TURis, including data from the Fagerstrom et al. (2011) study at the request of the Committee. The results showed TURis was cost saving by £375.02 per procedure for a hospital with an existing Olympus monopolar TURP system and by £284.66 for other hospitals.

Committee considerations

5.17 The Committee agreed with the External Assessment Centre's conclusions that the published economic studies did not contain relevant evidence. It also agreed with the revisions suggested by the External Assessment Centre in terms of the costs of the consumables and blood transfusion costs. It heard expert opinion that patients having a blood transfusion may also have an increased length of stay in hospital and it noted that this was not included in the model. The Committee considered it was quite likely that TURis could be cost saving, but noted the uncertainties in the External Assessment Centre and company meta‑analyses for length of hospital stay. At the draft guidance meeting the Committee considered that the cost model should include the 0.19 days difference in the length of hospital stay in favour of TURis compared with monopolar TURP. Results from the revised model showed that TURis saved around £71 per patient for hospitals that already use Olympus systems and has an additional cost of around £20 per patient for other hospitals (see section 5.15). The Committee concluded that, although uncertainty remained in the cost model, the use of the TURis system is likely to generate cost savings compared with the monopolar TURP system.

5.18 The Committee noted that the data available to estimate differences in readmission rates between TURis and monopolar TURP were limited in quantity, but it received expert advice that a reduction in readmissions was likely if TURis was used, instead of monopolar TURP. From the results of the External Assessment Centre's scenario analysis based on the Fagerstrom et al. (2011) study it considered that it was plausible there would be cost savings for hospitals with TURis, attributable to fewer readmissions, whether or not the hospitals were already using Olympus equipment.

  • National Institute for Health and Care Excellence (NICE)