Re: Cost-effectiveness of continuous subcutaneous apomorphine in the treatment of Parkinson’s Disease in the UK and Germany

We appreciate the authors’ efforts in carrying out an economic evaluation comparing continuous subcutaneous apomorphine (CSAI) to levodopa/carbidopa intestinal gel (LCIG), deep brain stimulation (DBS) and the standard-of-care (SoC) in advanced PD in the UK and Germany1. However, we would like to seek clarification regarding unit costs utilized in the model, how and why particular studies were selected to provide the evidence-base that drives the model’s results, and data not described in full.

The authors applied a unit cost for DBS surgery in the UK of £22,800 per DBS procedure when the 2014/15 National Tariff2 indicates a cost to the National Health Service (NHS) of £6090. While the £22,800 may include the cost of the DBS system, although this is not obvious, adding the addition of the system cost the correct cost would have been £18,292, £4508 less per patient than that reported. The cost of DBD replacement surgery was also over-estimated. It is currently £1063 (2014/15 National Tariff) and not £2775 as reported in Table 2.

We also noted that the cost estimated for DBS complications, permanent and temporary, was based on substantially outdated evidence. The rates of complications in the model were based on data cited by Tomaszewski and Holloway3, who in turn referenced data from a 24-patient study published 17 years ago by Limousin et al.4. However, as the DBS procedure and currently available systems have evolved significantly in 17 years, complication rates today are lower and, based on evidence from randomized, controlled trials5–10, are mostly transient. As a consequence, the rate and, therefore, the cost of DBS complications in this paper are likely to have been over-estimated. In addition, to apply the cost of DBS complications estimated by Tomaszewski and Holloway, Walter and Odin. inflated them over at least 14 years instead of sourcing more recent estimates. Considering the fact that the audience of this publication is likely to include payers and decision-makers, it would be difficult to rely on the results when costs were estimated in this manner.

The selection of clinical data used in the model appears to have been subjective rather than systematic as well as unbalanced in terms of the quality of the studies providing the data. For DBS, the authors used data from the RCT by Deuschl et al.5, but for CSAI, data were extracted from seven small case series; a limitation of the analyses from the perspective of the quality of the evidence base.

Incorrect referencing and reporting on key parameters in the model were also detected. First, Deuschl et al.5 was referenced as the source of data for Hoehn & Yahr stage improvement. However, Deuschl et al. did not report Hoehn & Yahr stage at follow-up, only at baseline. Second, when describing complication rates for LCIG (see Table 1), the authors referenced Tomaszewski and Holloway3, a cost-effectiveness analysis of DBS that did not consider LCIG. There is also a lack of clarity in the presentation of assumptions regarding the levels of OFF time reductions. These were reported without providing the unit (percentage or hours).

As accurate and transparent referencing and reporting are among the key quality criteria for reports of cost-effectiveness analyses11^,12, incorrect and misleading references significantly reduce confidence in the results. Therefore, readers should be cautious about basing decisions on the conclusions provided by Walter and Odin1.

When discussing the results of other DBS cost-effectiveness analyses in relation to their own, the authors failed to provide a full and transparent summary and discussion. They described the results of three other analyses only, Eggington et al.13, Dams et al.14, and Tomaszewski and Holloway3, but failed to discuss data reported by Valldeoriola et al.15, and the National Collaboration Centre for Chronic Conditions16 in the UK. It is important to note that the authors of all five studies reported incremental cost-effectiveness ratios (ICERs) that would be considered acceptable by decision-makers, being below or around the €40,000 per quality-adjusted life-year gained, the general threshold of acceptability.

The authors also failed to discuss two other published economic analyses that compared CSAI and DBS – cost analyses carried out in Spain (the SCOPE study, published in this journal in 201317) and in the UK18. The mean cumulative 5-year cost per patient was significantly lower with DBS (€88,014) compared with CSAI (€141,393) (p < 0.0001) in SCOPE. For the UK, the mean 5-year cost of DBS was £69,566 compared with £80,843 for CSAI, a cost saving of £11,277 per patient with DBS. DBS was cost saving compared to CSAI from 3 years onwards in the UK because the initial cost of DBS device acquisition was offset by the on-going provision of CSAI.

Transparency

Declaration of funding

No funding is declared for this Letter.

Declaration of financial/other relationships

ACW has received grant/research support from MRC. ACW and GB are both consultants/advisors to Medtronic Inc.

Acknowledgments

There were no non-author assistants, contributorships, or funding sources used in the development of this letter.