A cost-effectiveness analysis of lisdexamfetamine dimesylate in the treatment of adults with attention-deficit/hyperactivity disorder in the UK

Attention-deficit/hyperactivity disorder (ADHD) is a neurodevelopmental disorder with symptoms of inattention, hyperactivity, and impulsivity that present in multiple settings [1]. ADHD persists into adolescence and adulthood in 50–60% of childhood cases of ADHD [2]. ADHD is estimated to affect 2–5% of the adult population worldwide, depending on country, and choice of ADHD diagnostic criteria [3‐6]. Unlike childhood ADHD, gender ratios tend to be fairly equal in studies of adult ADHD [7‐9]. Persistent inattentive ADHD symptoms in adulthood are significantly related to an increased risk of long-term work disability [10]. In the United States (US), the economic burden of ADHD in adults is estimated to be from $105 billion to $194 billion per year, of which the largest cost is productivity and income losses, ranging from $87 billion to $138 billion per year [11].

Management of ADHD usually includes psychotherapy, medications, or a combination of both psychotherapy and pharmacotherapy. In the United Kingdom (UK), methylphenidate (MPH), either extended release (ER) or immediate release (IR), is recommended by the National Institute for Health and Care Excellence (NICE) Guidance CG72 to be tried first; if MPH is ineffective or unacceptable, atomoxetine (ATX) or dexamfetamine (DEX) may be tried [12]. ATX or DEX should be considered in adults unresponsive or intolerant following an adequate trial of MPH (usually approximately 6 weeks) [12]. Caution should be exercised when prescribing DEX to those likely to be at risk of stimulant misuse or diversion [12].

Lisdexamfetamine dimesylate (LDX) is a prodrug; following absorption, LDX undergoes hydrolysis to DEX and lysine. LDX has received marketing authorizations for the treatment of ADHD in Australia, Brazil, Canada, Denmark, Germany, Ireland, Israel, Mexico, Norway, Spain, Sweden, Switzerland, the UK, and the US. LDX has been shown to be effective in reducing the symptoms of ADHD in a randomized controlled trial in adults [13], using the ADHD Rating Scale IV (ADHD-RS-IV) [14]. During the open-label phase of a modified analogue classroom study of adults with ADHD, LDX was associated with improvements from baseline in executive function behavior, using the validated, self-reported Brown Attention-Deficit Disorder Scale [15, 16]. In the modified analogue classroom study, LDX treatment demonstrated efficacy in adults with ADHD who had significant impairments in ADHD core symptoms and executive function, as well as efficacy in quality of life as assessed by Adult ADHD Impact Module (AIM-A) [17]. In a 10-week randomized, placebo-controlled trial of LDX in adults with ADHD and clinically significant executive function deficits, LDX improved AIM-A multi-item domain scores versus placebo [18]. A randomized, double-blind, placebo-controlled study using a validated driving simulator paradigm showed that LDX may reduce driving risks in young adults with ADHD [19]. Postmarketing survey data suggest that the rate of non-medical use of LDX is lower than that for short-acting stimulants and lower than or equivalent to long-acting stimulant formulations [20]. In a randomized, double-blind, placebo-controlled, two-way crossover study conducted in a simulated adult workplace environment, LDX significantly improved the Permanent Product Measure of Performance scores versus placebo and maintained improvement throughout the day from the first (2 h) to last (14 h) postdose time points versus placebo in adults with ADHD [21].

The objective of the present study was to estimate the cost-effectiveness of LDX compared with MPH-ER or ATX in the treatment of adults with ADHD. The results of the analysis are presented as the total costs and total quality-adjusted life years (QALYs) for each drug, as well as the incremental costs and QALYs for LDX when compared with MPH-ER or ATX. In addition, the incremental cost-effectiveness ratios (ICERs) for LDX relative to MPH-ER or ATX are presented and evaluated against an established cost-effectiveness threshold of £20,000 per QALY [22]. A comparison of the costs and health outcomes predicted by the model is intended to aid physicians and health care decision makers as they make decisions about efficient use of drugs indicated for adults with ADHD.

A decision-tree model was developed in Microsoft Office Professional 2013 (Excel version 15) to evaluate the cost-effectiveness of LDX, compared with MPH-ER or ATX, from the perspective of the UK’s National Health Service (NHS). Drug treatment is recommended for adults with ADHD with either moderate or severe levels of impairment [12]. The selection of comparators was guided by the clinical guidelines published by NICE [12] and validated by a UK-based clinical expert. The ER formulation of MPH was chosen because it is used more commonly in clinical practice in the UK than the IR formulation (Shire data on file; IMS Database, March 2015) and MPH-ER has a broader clinical evidence base, reflected by a larger number of clinical trials in adults (see Online Resource 1). Non-pharmacological interventions were included in the analysis as part of the non-drug costs and were assumed to vary for responders and non-responders to drug therapy. There is no evidence to suggest that any of the drug treatments would result in a reduction in the amount or type of non-pharmacological or behavioral intervention required for those responding to therapy or for those not responding to therapy.

The target population for the cost-effectiveness analysis was adults with ADHD, which reflects the anticipated therapeutic licensed indication for LDX in both adult continuers (i.e., adults whose ADHD was diagnosed during childhood and adolescence) and de novo adult patients (i.e., adults with ADHD not diagnosed during childhood or adolescence). The health outcomes included were “tolerate”, “unable to tolerate”, “response” and “non-response” (Fig. 1). The impact of using LDX as an alternative to MPH-ER or ATX in terms of costs and health outcomes was estimated in the model based on the number of patients who achieved response to treatment and those who did not, including those who discontinued due to adverse events. Costs and utilities corresponding to the patients’ health states were assigned to each of these patients.

The base-case analysis evaluated direct medical costs and health-related quality of life associated with 1 year of treatment, including the initial 28-day drug titration period. A time horizon of up to 5 years was examined as a sensitivity analysis with an annual discounting rate of 3.5% applied to both costs and benefits. The modeling framework and key assumptions, including the 1-year time horizon, were adapted from the health technology assessment model used in an earlier assessment of ADHD drugs by NICE [23].

The uncertainty in the ICER estimate was explored by one-way sensitivity analysis and probabilistic sensitivity analysis (PSA). Quality control of model programming and verification of all input data with original sources was performed according to a prespecified test plan by health economists who were not involved in the model development. Key model assumptions were assessed for face validity by a UK clinician with extensive experience in treating and research in adults with ADHD, and by an independent health economics expert.

The following assumptions were made in the base-case analysis.

Adult continuers or de novo adult patients enter the model when they initiate a course of treatment with LDX, MPH-ER or ATX.

Patients begin titration lasting 28 days, during which the optimal dose of treatment is reached (ATX may require a longer titration period, which is tested in the sensitivity analysis).

Patients who experience intolerable side effects discontinue treatment in the middle of the titration period (i.e., after 14 days on treatment). For patients who discontinue treatment during the titration period, utilities and costs during the titration period (28 days) are represented by a 50%/50% mix of the responder and non-responder utility values, and a 50%/50% mix of the responder and non-responder non-drug costs, respectively. This approach was based on the assumption that, on average across different treatments, patients who discontinue responded half way through the first month (consistent with the assumption by King et al. [23] in the UK Health Technology Appraisal of ADHD drugs in children and adolescents). Alternative assumptions were explored in which these patients were assumed to have the same utility during their titration period as responders and as non-responders.

Patients who discontinue treatment due to intolerable side effects do not initiate additional pharmacological treatment. Those who discontinue, the same as non-responders, are assumed to receive behavioral therapy. This assumption was made largely due to lack of relevant clinical data for follow-up therapies and the fact that in the model, these therapies would be the same in both the LDX and the comparator arms, hence not adding any differentiation to the model results. The patients who discontinue are, therefore, assumed to have the same utilities and costs as non-responders for the remainder of the 1-year model time horizon.

At the end of the titration period, non-responding patients discontinue treatment without initiating any further pharmacological treatment and are assigned the non-response costs and utilities for the titration period and throughout the model’s remaining time horizon.

Patients who respond to treatment at the end of the titration period remain on treatment throughout the remainder of the model’s time horizon, maintaining their level of response.

Patients who responded to and tolerated treatment are assumed to be adherent and persistent on treatment over the time horizon of the model, as was generally observed in the pivotal trials. This assumption is consistent with that made in the health technology assessment model presented by King et al. [23].

This dichotomous response framework was adapted from the model developed by King et al. [23] as part of the UK Health Technology Appraisal of ADHD drugs in children and adolescents.

Clinical input parameters applied in the model were estimated via a Bayesian network meta-analysis (NMA) of trials in adults with ADHD.

Costs and outcomes are not discounted in the base-case analysis, given the time horizon of 1 year.

Our economic analysis used the results of a Bayesian MTC of efficacy to assess the cost-effectiveness of LDX when compared with MPH-ER or ATX in the treatment of adults with ADHD. The base-case results showed that LDX was a dominant strategy when compared with MPH-ER; however, there was some uncertainty in this result, with the PSA estimating a 61% probability of LDX being cost-effective vs. MPH-ER under a £20,000 willingness-to-pay threshold. One-way sensitivity analysis results revealed that the model, when comparing LDX with MPH-ER, was most sensitive to increase in discontinuation rates due to adverse events. From the MTC outputs, rates of discontinuation due to adverse events had fairly broad credible intervals largely translating from small sample sizes for patients who discontinue in the trials used in the MTC (Table 2 and Online Resource 1). LDX was a dominant strategy when compared with ATX. Sensitivity analyses confirmed the robustness of this result, showing an 80% likelihood of LDX being cost-effective relative to ATX when applying probabilistic methods. One-way sensitivity analysis results revealed no changes in the model results, when comparing LDX with ATX, for any of the variables examined.

The present economic analysis, to the authors’ knowledge, is the first published study on the cost-effectiveness of pharmacotherapy in the adult ADHD population. A recent systematic review published in 2012 identified no published studies on the cost-effectiveness of pharmacotherapy in the adult ADHD population, comparing stimulants, non-stimulants or adjuvant therapy [36].

Our model found that LDX is cost-effective relative to MPH-ER or ATX and would provide a good additional stimulant option. The inclusion of a new cost-effective treatment option for adults with ADHD could allow for patients’ individual needs and preferences to be taken into consideration. Having multiple stimulant and non-stimulant treatment options available is likely to contribute to a reduction in the overall burden of ADHD.

The choice of a short time frame for this analysis was driven by the lack of long-term data needed for a longer model time horizon. The chosen time frame requires minimal extrapolation of the short-term data from the trials (duration of trials used in the MTC ranged from 4 to 34 weeks), thus minimizing the uncertainty. The shorter time horizon can also be justified by the current clinical guidelines, which recommend that treatment for adults with ADHD should be reviewed at least annually [12]. The one-way sensitivity analysis and the scenario analyses showed that extending the model time horizon to 5 years and changing non-response resource use did not have much impact on the cost-effectiveness results.

This modeled assessment had some limitations. First, the MTC was limited by the number of prior studies, which led to the need for an imputation analysis and limited the precision of the estimates. Although the imputation analysis was used to produce conservative point estimates of the relative risk of CGI–I response, the estimates of variability from that analysis do not account for the additional variability due to the imputation. Hence, although the point estimates are likely to be conservative, the corresponding credible intervals and confidence intervals should be interpreted with caution. Generally, indirect treatment comparisons enable us to compare treatments not otherwise compared in a head-to-head clinical trial, but they do not equate to the same level of evidence as direct (head-to-head) comparisons. The results of the current MTC were generally consistent with the results of the pivotal trials, and no statistical evidence of heterogeneity across trials was found. Second, due to lack of data, the study did not consider any treatment for non-responders and assumed these patients discontinued drug therapy. In real-world clinical practice, such patients may receive one of the other comparators or off-label medications (e.g. bupropion, clonidine, modafinil or imipramine) or combination treatments [12]. Neither did it consider dose reduction if side effects became troublesome. Third, the utility data for intolerable side effects were not based on disutility data of individual side effects leading to discontinuation, but were estimated in the model as a 50%/50% mix of the responder and non-responder utility values. However, alternative assumptions around the utility weight for patients who discontinue during the titration period had very little impact on the ICERs. Fourth, the model did not consider real-life medication compliance, again, due to the lack of data, this time on the relationship between adherence to therapy and symptom reduction. Finally, the study was conducted from the UK NHS perspective and did not include the broader societal perspective, which is an important cost driver in the overall cost burden of ADHD. Taking into account costs associated with the societal perspective likely would result in an even lower ICER for LDX, given a potential cost-offset.

There is limited evidence on the cost-effectiveness of medication to treat ADHD in adults, as well as long-term cost-effectiveness of pharmacotherapies in ADHD. To better inform payers about the economic value of existing medications, future studies should consider identifying subgroups that may have heterogeneous responses to different treatments and expanding the time horizon to incorporate long-term outcomes [36].

This study suggests that LDX is likely to dominate both MPH-ER and ATX as a therapy for adult patients with ADHD (both previously treated and untreated patients); i.e., total costs are expected to be lower and outcomes (QALYs) are expected to be improved with LDX therapy compared with both MPH-ER and ATX therapy. These results, particularly for the comparison versus MPH-ER, must be seen in light of some uncertainty detected by the PSA. The presented model adds to the health economic information available for policymakers and to general considerations in economic modeling.