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

Reply to: http://informahealthcare.com/doi/pdf/10.3111/13696998.2015.1066797

Original Article: http://informahealthcare.com/doi/abs/10.3111/13696998.2014.979937

The basic objective of any health economic evaluation is to assist rational and unbiased decision-making for the provision of the best and affordable therapies. Therefore, evaluation studies have to meet the criteria of methodological soundness, retrieval of transparent input data, and their sources and connectivity to relevant studies on the same topic. Debating studies and their analytical procedures as well as results is usually helpful, provided the rules of the game are observed. In this vein, the comments of Barnett and Williams1 (hereafter B&W) are welcome in principle, although the purpose of their Letter to the Editor, namely to contribute to clarification and transparency, is harmed by submitting non-verifiable cost data and by referring to studies which have a differing coverage. Their harsh final criticism that readers of our study ‘should be cautious about basing decisions on the conclusions provided’ strikes back to their comments.

Let us take up the major points in the order they have been raised by B&W.

The aim of our health economic evaluation was to compare the cost-effectiveness of all available therapies in advanced Parkinson’s disease (aPD), continuous subcutaneous apomorphine (CSAI), levodopa/carbidopa intestinal gel (LCIG), deep brain stimulation (DBS), and the standard-of-care (SOC). In contrast, only one study, carried out by Valldeoriola and colleagues2^,3 has compared all three therapies, CSAI, LCIG, and DBS in a detailed cost analysis spanning over a 5-year time horizon. However, even this study did not consider outcomes, which was one aim of our article as clearly mentioned in its introduction. The published studies include only two alternatives, usually comparing one therapy vs SOC4–7.

The major part of B&W pertains to cost. Regarding UK costs the following should be noted: Cost data for DBS surgery in our study were derived from 2013–2014 National Tariffs. This was the only source at the time of publishing the article (2014). The 2014–2015 National Tariffs, mentioned by B&W, were naturally not available at the time of carrying out our calculations. Turning to the calculation of total surgery costs it is explained in the cost section of our article that they include not only the procedure costs (£7131; Code: AA072Z), but also pre-operative assessment (£641; Code: AA25Z), implantable pulse generator device price (£8326; UK list price), electrode (£1530), extension lead (£1786), patient controller (£560), accessory kit (£90), and re-hospitalization. These detailed costs are made transparent in the article, whereas the sum of the more recent 2014/2015 costs of £18,292 given by B&W seems to be a lump-sum, and the components which it contains remain unknown. It may be noted also that, for the UK cost calculations, we followed an accepted and comparable approach by Eggington et al.4. Finally, total cost of DBS replacement surgery must also be considered and include different cost components like battery cost and hospital stay.

As to the evidence of cost and probability of DBS complications, B&W criticized our reference of Tomaszewski and Holloway6 for being outdated. As mentioned, Tomaszewski and Holloway have documented overall DBS surgery complications to be 35% (first and subsequent cycles), including temporary complication (28.2%), permanent complication (5.4%), and operative death (0.7%). Comparing this publication with more recent studies, it appears that these complication rates are still quite similar. For example, the recent cost-effective analysis of Eggington et al.4 has used AE/complication rates per cycle (6 months) derived from Deuschl et al.8 and Weaver et al.9. During the first cycle (6 months), 19.8% of patients exhibit temporary complications (mainly infections and lead dislodgement) and 9.2% in the subsequent cycles (29% first year vs 28.2% which we have used). Therefore, the derived cost estimates from Tomaszewski and Holloway are still relevant and can be used today. Obviously they had to be inflated, which is a well-established standard approach used in the majority of published analyses. Again, comparing our cost data of temporary complications (£9120) with the temporary events of Eggington et al. (£9825 average cost first year) costs are similar and, therefore, representative for temporary complications.

Although, as mentioned above, the 2014/2015 costs were not available at the time of publication, it is an interesting exercise to redo our calculation on the basis of the figure provided by B&W, i.e., £18,292, despite its lack of detail. Keep in mind that also patients with CSAI and LCIG switch to DBS, which impacts also lifetime costs of the comparator strategies. The re-calculation of costs for DBS surgery would show the following results: Discounted life-time costs after DBS would decrease from £87,730 to £83,222. Per-patient costs for aPD treatment with CSAI are estimated at £77,516. The incremental costs difference would be reduced to £5706. Hence, even with these more recent data, CSAI remains the dominant strategy.

Concerning our selection of studies it should be noted that a series of studies confirm that DBS and CSAI are effective strategies in the management of motor fluctuations and dyskinesias in patients with aPD10. Evidence from 19 open-label studies including more than 400 patients confirm that CSAI is able to reduce OFF time by more than 50% (40–85%)11. The availability of RCTs is not a pre-condition for a health economic evaluation. This fact was mentioned as a limitation in our discussion. In order to select clinical data of CSAI for our health economic evaluation, a literature search was undertaken using the databases Medline, Embase, and SciSearch in February and March 2014. Sixteen clinical studies have evaluated the efficacy of CSAI. All these studies were also presented together in a review paper by Ceballos-Baumann11 [written in German]. When calculating the percentage reduction in OFF time from baseline in time intervals, studies with a follow-up time up to 12 months were used (n = 151 patients)10^,12–17 to avoid over-estimation of the effect. For the subsequent cycles the long-term study from Gracia Ruiz et al.18 was selected.

Turning to the supposed ‘incorrect referencing’, the following clarification can be made: The randomized trial from Deuschl et al.8 compared DBS treatment plus medication with medical management. In the randomized-pairs trial 156 patients were enrolled with aPD and severe motor symptoms. The primary end-points were changes from baseline to 6 months in the quality-of-life and the severity of symptoms without medication, according to the Unified Parkinson’s Disease Rating Scale, part III (UPDRS-III). With DBS the period of immobility was reduced by 4.2 h8. These improvements were used for the reduction of time in OFF time. We have not used the data from Deuschl et al.8 for the H&Y stage (see Table 1 in the original article). Our model has defined combined health states; each H&Y stage is associated with three OFF states. Our cost-effectiveness study suggests that DBS reduces OFF time to a greater extent. Also, the percentage improvement in OFF time is documented in detail in the section ‘Transition probabilities’ of our analysis.

The need for sensitivity analysis: Input data of health economic evaluations have to be drawn from a variety of sources and could always be subject for discussion. Therefore, in general, uncertainty is considered in health economic evaluations and investigated by way of sensitivity analyses. Our probabilistic sensitivity analysis (PSA) indicates that, in more than 2/3 of all cases, CSAI is cost-effective vs DBS and, in 1/3 of cases, DBS is favorable. Using the reduced cost given by B&W, DBS is the cost-effective strategy in almost 50% of cases.

Finally, our discussion section comprises consciously only full economic evaluations in aPD and cannot focus on partial—cost only—studies. In this respect, also the SCOPE study, published by Vallideoriola19 and recommended by B&W is a very detailed cost analysis, but does not include outcomes.

As each patient experiences PD in a unique way, and requires a specific plan to meet his or her needs, our analysis is a health economic evaluation stating that CSAI contributes as an additional cost-effective option to treat patients with aPD. CSAI could, therefore, be considered as a cost-effective alternative to DBS and LCIG, and they all significantly improve outcomes compared to SOC.

Transparency

Declaration of funding

No funding declared for this Letter.

Declaration of financial/other relationships

E.W., P.O have previously received lecture fees from EVER Neuro Pharma.