Is adjunctive pharmacotherapy in attention-deficit/hyperactivity disorder cost-effective in Canada: a cost-effectiveness assessment of guanfacine extended-release as an adjunctive therapy to a long-acting stimulant for the treatment of ADHD

This economic evaluation indicates that, compared to long-acting stimulants as monotherapy, GXR as an adjunctive therapy to long-acting stimulants is a cost-effective alternative among children with ADHD with a suboptimal response to stimulants. Results of comprehensive sensitivity analyses confirm the robustness of the base-case results.

This is the first Canadian economic evaluation of GXR as an adjunctive therapy in the treatment of ADHD. The results of a previous CEA performed from a US third-party payer perspective suggested that GXR as an adjunctive therapy to long-acting stimulants in the treatment of children and adolescents with ADHD who had a suboptimal response to stimulants was cost-effective according to a willingness to pay threshold of $US50,000/QALY, with an ICER of $US31,660/QALY [17]. The present economic evaluation is a Canadian adaptation of the US study published by Sikirica et al. [17]. In the US study, the estimated ICER was higher than that calculated for the Canadian adaptation ($CA23,720/QALY from a MoH perspective). These differences could be explained by the higher treatment cost of GXR in the US study and the higher weekly medical costs. Nevertheless, the findings of both studies were similar and thus strengthen the conclusion that GXR as an adjunctive therapy to long-acting stimulants is a cost-effective alternative compared to long-acting stimulants as monotherapy.

This economic evaluation has several strengths. First, the model provides results in terms of cost per responder, which gives further evidence on the cost-effectiveness of GXR as an adjunctive therapy to a long-acting stimulant in the treatment of children with ADHD. Moreover, the analysis accounted for AEs associated with treatments as well as productivity losses associated with disease severity, thus allowing a broader perspective and perhaps a more representative assessment of all the impacts of the disease and intervention. Lastly, the comparison of GXR as an adjunctive therapy to short/intermediate-acting stimulants with short/intermediate-acting stimulants alone allowed considering the fact that a significant proportion of patients with ADHD in Canada may receive short/intermediate-acting stimulants because of the reimbursement criteria and specific conditions required by Canadian provinces for the coverage of long-acting stimulants [32].

However, this economic evaluation also has some limitations. The difference in number of QALYs between the two treatments is small, but the difference in number of patient-weeks with response also indicates a gain in efficacy with the adjunctive therapy. As for any model-based analysis, in the absence of data, assumptions were made that may increase the uncertainty of the results. First, although the two-stage model was consistent with an approach used in a health technology assessment conducted by the National Institute for Health and Care Excellence, it was assumed that patients discontinued their treatment if they transitioned into a moderate or a severe state during the second stage of the simulation (weeks 9–52) [22]. Although this may not always be observed in clinical practice, the latter assumption may be realistic because most patients are considered to achieve optimal dosing within the first 8 weeks of treatment and because a significant proportion of patients discontinue their medication by the end of 1 year anyway [34, 35]. Furthermore, it was assumed that patients who discontinued treatment did not receive subsequent therapy. Although this may not be representative of the real-world clinical setting, there was insufficient clinical evidence or consensus in treatment guidelines on the management of patients who were non-responsive to adjunctive therapy. Another limitation of the current study involves the availability of data regarding cost parameters. As medical costs and productivity losses associated with disease severity have not been reported, some assumptions were made based on existing studies in order to estimate these parameters. Although costs associated with productivity losses usually represent the main cost component specific to a societal perspective, other costs, such as “out-of-pocket” expenses can be also considered with a societal perspective. In this study, only the addition of costs associated with productivity losses was considered for the societal perspective. Moreover, cost of long-acting stimulants was based on daily dose and number of pills according to Canadian data for children aged 0–12 years, and therefore, specific data for patients aged 6–12 years, which corresponded to the target population in the study, were not available. However, patients aged less than 6 years would be on much lower doses in general. As lower doses used by very young children may contribute to lower the mean daily cost of comparators, this would potentially be a conservative assumption. Similarly, because the OCCI Costing Analysis Tool divided age into three categories (0–17, 18–69 and 70+ years), some cost parameters associated with health care resource utilization did not precisely reflect costs specific for the target age of the population in the present study. Despite these limitations, findings of this cost-utility analysis are robust according to the base-case and confirmed by the robust one-way and probabilistic sensitivity analyses. Lastly, one other limitation should be mentioned. As the Wilens et al. study was designed and powered collectively for only those long-acting stimulants described previously, future research may need to study whether there are differences in efficacy among the individual long-acting medications combinations when GXR was added to them, or those medications that are on the market within Canada but were not in the US clinical trial.