Reply to Naudet and Colleagues. Cost-Effectiveness in Health Technology Assessment-A Case in Alcohol Dependence

Abstract

We would like to thank Drs Naudet, Granger and Braillon for their interest in our work, and for opening a discussion that is of interest here: the consideration of complementary evidence for health technology assessment (HTA) decision-making. By ‘complementary evidence’, we refer to clinical trials, patient’s preference and quality of life, clinical guidance, current clinical practice, and public health, cost-effectiveness and budget impact assessments. As such, it should be understood that a cost-effectiveness analysis, such as the one discussed here, represents an element of evidence that is weighted in relation to a continuum of complementary evidence when assessed by HTA decision makers.

Dr Naudet and colleagues have raised several points that we would like to comment on. Firstly, the results and conclusions of the meta-analysis by Palpacuer et al., (2015) should be interpreted with caution, especially from an HTA decision-making point of view. This meta-analysis did not consider the European Medicines Agency (EMA) indication for treatment with nalmefene. Nalmefene is indicated for reduction of alcohol consumption in adult patients with alcohol dependence who have a high drinking-risk level (DRL), who do not have physical withdrawal symptoms and who do not require immediate detoxification. Nalmefene should be initiated only in patients who continue to have a high DRL 2 weeks after initial assessment. The meta-analysis by Palpacuer and colleagues did not consider only patients within nalmefene’s indication but also those with low and medium DRLs at treatment initiation. Having evaluated different populations on the basis of the nalmefene clinical trials’ data, the EMA identified patients with high and very high DRLs as the optimal population for treatment with nalmefene, representing greater public health benefits and better investments of public resources from a healthcare payer point of view than the wider population considered in the meta-analysis by Palpacuer and colleagues.

Secondly, regarding Dr Naudet and colleagues’ concern of the a posteriori assessment of the nalmefene’s indicated population, this was previously addressed by Dr van den Brink and colleagues in a response to commentary letters by Dr Braillon and by Dr Des Spence, published in the BMJ in 2014 (van den Brink et al., 2014). In addition, this response by Dr van den Brink and colleagues provided clarification to Dr Braillon’s concern about attrition. This was handled by multiple imputation methods applied to the nalmefene trials’ data. Nevertheless, as further reassurance for the current case and with real-life clinical practice in mind, our cost-effectiveness analysis of nalmefene used observed case data, and the treatment withdrawals were modelled conservatively without any maintenance of effect from nalmefene.

Finally, to respond to Dr Naudet and colleagues’ concern in term of the place in therapy for nalmefene, our cost-effectiveness analysis reflects current clinical practice and guidance in the United Kingdom (UK), with nalmefene being offered in accordance with its licensed indication to patients who have failed brief intervention, and with non-responders to nalmefene having the opportunity to beneficiate from assisted alcohol withdrawal and naltrexone or acamprosate for maintenance of abstinence. Nevertheless, Dr Naudet and colleagues’ point about directly comparing nalmefene with naltrexone is a valuable one. When the National Institute for Health and Care Excellence (NICE) appraised nalmefene as a possible treatment for alcohol-dependent patients, naltrexone was considered as a potential direct comparator ‘for reduction of craving from alcohol’. While naltrexone is not explicitly licensed for reducing alcohol consumption, the NICE assessment group recognized that naltrexone could be used off-label within specialist services to reduce alcohol consumption for certain patients, even if it was not an established treatment in current clinical practice (National Institute for Health and Care Excellence (NICE)). Hence, was presented to NICE a comparative assessment of nalmefene versus naltrexone for reduction of alcohol consumption in non-abstinent patients, assessment confirming the relevance from a UK perspective of the approach taken for our cost-effectiveness analysis, which assessed in sequence the treatment with nalmefene or naltrexone.

To conclude, our cost-effectiveness analysis was based on conservative assumptions against nalmefene including for the modelling of treatment withdrawals, applied probabilistic analyses that took into account data uncertainty for all parameters included in the model, and presented scenario analyses that evaluated key assumptions. These provisions are considered best practice for health economic modelling in HTA and should provide reassurance of the robustness of the model conclusions.