Results

1. New products in development (and further assumptions)
The market for ADHD treatment has attracted pharmaceutical companies to invest heavily in new product development. Results of literature searches and the database analysis [49] are summarized in Table 3.
Table 3  New products in development for treatment of ADHD in children and adolescents: overview of compounds not yet available in England and Germany  Only projects in phase III of clinical development or products already marketed in the United States. DaytranaR was formerly known as "MethypatchR", SparlonR as "AttenaceR". Data source: InnoVal-HC, 2006 [49].  A number of further pharmaceutical preparations and variants of methylphenidate are in phase III of clinical development in Europe, some of which have already received marketing authorization in the United States. These include a transdermal system (TDS, approved as DaytranaR in the United States, April 2006) of methylphenidate with duration of action of 12 hours (wear time, 9 hours). At present, in the absence of head-to-head trials against established oral formulations, the main advantage of this product seems to be convenience-related [49-51]. For modeling, it has therefore been assumed that it would capture no more than 10% of methylphenidate prescriptions. This implies the expectation that there will be no significant problems associated with skin sensitization.
Dexmethylphenidate, the chirally pure active isomer of methylphenidate, has been licensed as FocalinR (Novartis) in the US in May 2005, also as an extended-release formulation [49,52]. As there is currently no evidence of superiority in terms of efficacy or tolerability [49,53], it has been assumed that FocalinR will become available in Europe without a price premium over branded immediate-release (RitalinR, also Novartis) or modified-release methylphenidate (ConcertaR, Janssen-Cilag), respectively. Given its investment in FocalinR, it has been considered unlikely that Novartis would launch RitalinR LA (a long-acting preparation of methylphenidate available in the United States and some other markets) in England and Germany.
At the time of writing, few scientific data were in the public domain about lisdexamphetamine mesylate (NRP104), a new chemical entity that is an inactive prodrug of amphetamine believed to have a reduced potential for abuse and overdose compared to other ADHD drugs. The reason is that the amphetamine is conjugated to a specific amino acid and is activated only when metabolized in the gastrointestinal tract. Shire filed the product with the Food and Drug Administration (FDA) in December 2005 for the treatment of ADHD in children aged 6 to 12. Phase III study results presented at the Annual Meeting of the American Psychiatric Association in Toronto, May 2006, indicated similar efficacy and tolerability compared to mixed amphetamine salts (AdderallR) and duration of action of 12 hours [49]. A preliminary review of NRP104 by the U.S. Drug Enforcement Administration indicated that the compound would not be subject to controlled-substance scheduling [54], although this initial judgment has been revised since. The compound was launched by Shire in the United States end of July 2007 under the trade name "VyvanseR". For modeling it was assumed that the product will be introduced to the English and German markets in 2008 and (except for the "Extremely Low Case" scenario) overtake modified-release methylphenidate products in terms of prescriptions by 2012. Obviously, this rests on the assumption that clinical advantages of NRP104 (regarding abuse potential, possibly tolerability profile, as well as label and summary of product characteristics) will be confirmed.
Modafinil has been licensed as "ProvigilR" as a wake-promoting agent for patients with narcolepsy and shift work sleep disorder in the United States and England. Its manufacturer, Cephalon, has reformulated modafinil for children as once-daily 85 mg film-coated tablets, which it claims to be smaller and easier to swallow. Of note, this change of formulation should protect the new indication from generic competition [49]. While its mechanism of action is not fully understood, modafinil is classified as nonstimulant [55-57]. Similar to atomoxetine [58], a selective norepinephrine reuptake inhibitor, improvement of core symptoms had an effect size on core symptoms (school version of ADHD-RS) of 0.69 [59] to 0.76 [60]. This compares to effect sizes around 1.0 typically achieved with stimulants [58]. In August 2006, Cephalon announced discontinuation of modafinil development for ADHD due to safety concerns raised by the FDA, and the company intends to replace modafinil by its successor compound in development, armodafinil [61]. It has been assumed for modeling that armodafinil (like atomoxetine [62,63]) would become a second line treatment option after stimulants.
Finally, mixed amphetamine salts are marketed successfully in the U.S. (AdderallR, AdderallR XR, by Shire) but have been licensed neither in England nor in Germany. It is believed that these products will not become available in Europe. Further compounds in phase II clinical development for ADHD include selective GABA-B receptor antagonists (SGS742, by Saegis, a privately held company with Novartis among its investors) and ampakine molecules (by Cortex). These projects have been excluded from the present study in light of their inherent uncertainty; statistically, attrition rates of compounds in clinical phase II are as high as 62% [64].

2. Projected budgetary impact from the perspectives of the German Statutory Health Insurance (SHI, GKV) and the National Health Service (NHS) in England
First, the model was calibrated using data on ADHD-related prescriptions and expenditures from 1998 to 2005 (cf. above). Besides a base case for projection through 2012 (Figure 3), four additional scenarios were defined to address the uncertainty surrounding critical assumptions (for details, see above and Table 4). Two scenarios (upper and lower bounds of base case) represented plausible variants, assuming different rates of treatment prevalence and intensity. Two further scenarios reflected extremes, the lower one assuming no price premiums for the new products and low treatment intensity, while the high extreme combined the effects of intense treatment with higher price premiums for new products.
Table 4  Key assumptions underlying scenarios (base case and extreme cases)  1Abbreviations: MPH: methylphenidate; IR: immediate-release formulations (RitalinR, branded generics [EquasymR, MedikinetR], generics; FocalinR); MR: modified-release formulations (MPH-MR12: ConcertaR XL; MPH-MR08: EquasymR XL, MedikinetR retard, FocalinR XR; MPH-TDS: transdermal system (patch, DaytranaR); LisDEX: lisdexamphetamine (NRP104); Nonstimulants: ATX, atomoxetine (StratteraR), ARM, armodafinil (NuvigilR); DEX: dexamphetamine (England only)
Figure 3  Projected prescription drug expenditures for ADHD in children and adolescents, 2001 – 2012 (base case). a, b: Defined daily doses (DDDs) p.a.; c, d: expenditures by category p.a.; e, f: total (cumulated) expenditures p.a.; left: England; right: Germany. MPH: methylphenidate; IR: immediate-release formulations (RitalinR, branded generics [EquasymR, MedikinetR], generics; FocalinR); MR: modified-release formulations (ConcertaR XL, EquasymR XL, MedikinetR retard, FocalinR XR; MPH-Patch: transdermal system (DaytranaR); LisDEX: lisdexamphetamine (NRP104); Nonstimulants: atomoxetine (StratteraR), armodafinil (NuvigilR); DEX: dexamphetamine (England only).   While all scenarios will be described briefly below, Tables 4 and 5 have been limited to a more detailed account of the base case and the extreme scenarios. Interested readers may retrieve the full details from our Institute's website [97].
Table 5  Projected expenditures by scenario: ADHD pharmacotherapy for children and adolescents in England and Germany, 2012  1For abbreviations, see legend to Table 4. 2(share of) market for psychotropics (D) or CNS drugs (UK), calculated assuming a growth rate of 5% p.a. (figures in brackets represent one of the sensitivity analyses, indicating market shares assuming no growth of non-ADHD market segment); for comparison: share of market segment in 2002 was 0.77% (NHS, England) and 1.8% (GKV, Germany), respectively.  The base case scenario, which is believed to reflect the most probable course of future events, implies an increase of drug spending for children and adolescents with ADHD for year 2012 by a factor of 12 (GKV in Germany) or 10 (NHS in England) over 2002. Assuming an annual growth rate of 5% for drug expenditures between 2005 and 2012 (except for ADHD in children and adolescents), then the projected ADHD treatment costs (from the NHS perspective, £78 million) would make up 3.8% of NHS spending for CNS drugs in 2012, compared to just 0.77% in 2002. For the German GKV (projected spending, €311 million), the corresponding figures would be 12.9% of total spending for psychotropic drugs in 2012, compared to 1.8% in 2002. This increase is driven by the multiplicative effects of increasing awareness and recognition of ADHD, growing rates of pharmacotherapy, somewhat increased intensity (dose and duration) of treatment, and the shift to novel, more expensive products (cf. Figure 3). The upper and lower bounds of the base case can be interpreted as sensitivity analyses with regard to treatment intensity; these indicate a plausible range of spending estimates from €259 to €380 million in Germany and from £63 to £101 million in England in 2012. The differences between both jurisdictions reflect the effects of differences in population size, unit costs, available products, and a lower baseline and less dynamic increase in England compared to Germany.
The extremely high scenario indicates the sensitivity of projections to higher prices of new products (cf. Tables 4 and 5). Additional scenarios were computed and fell within the range indicated. Of note, a scenario with much less successful lisdexamphetamine (assuming no advantage over modified-release methylphenidate) produced a spending projection of €249 million or £67 million in 2012, roughly corresponding to the lower bound of the base case (Figure 4). An extremely low case was calculated resting on very conservative assumptions (cf. Tables 4 and 5) regarding diagnosis and treatment prevalence rates, absence of any further increase in treatment intensity, a disappointing clinical profile and subsequent low adoption rate of lisdexamphetamine, and denial of market access for modafinil and armodafinil. This led to an estimated spending of €170 m or £49 million in 2012, still an increase over 2002 by a factor >6 for both Germany and England (Figure 4).
Figure 4  Range of plausible projections: expenditures under extreme case scenarios, 2001–2012. a, b: "High Extreme" Case; c, d: Base Case modified "LisDEX without clinical advantage over MPH-MR"; e, f: "Low Extreme" Case; left: England; right: Germany. MPH: methylphenidate; IR: immediate-release formulations (RitalinR, branded generics [EquasymR, MedikinetR], generics; FocalinR); MR: modified-release formulations (ConcertaR XL, EquasymR XL, MedikinetR retard, FocalinR XR; MPH-Patch: transdermal system (DaytranaR); LisDEX: lisdexamphetamine (NRP104); Nonstimulants: ATX, atomoxetine (StratteraR), ARM, armodafinil (SparlonR); DEX: dexamphetamine (England only).

3. Impact on individual physicians
These projections have been related to individual German physicians involved in care for children and adolescents with ADHD. Since ADHD care is a highly concentrated phenomenon (cf. Table 2), average spending (again, from the perspective of the GKV) for ADHD pharmacotherapy was calculated for the top-50% of child and adolescent psychiatrists, and the top-20% of pediatricians (see Figure 5).
Figure 5  Projected impact of ADHD treatment for children and adolescents on individual physicians' prescription drug expenditures, 2001–2012 (base case). a: Child and adolescent psychiatrists; b: pediatricians in private practice in Germany. Expenditures expressed as €/physician and year; perspective of statutory health insurance (i.e., excluding privately health insured patients). Data represent average values for one of the upper 50% of child and adolescent psychiatrists and one of the upper 20% of pediatricians in terms of relative involvement in care for ADHD patients, respectively. Concentration of care modeled according to Nordbaden data [74]. Abbreviation: Non-ADHD-Rx: expenditures for treatment of conditions other than ADHD.   Under base case assumptions, total annual drug spending of these child and adolescent psychiatrists would grow from an average total of less than €30,000 per physician in 2002 to €282,000 in 2012, of which €264,000 would be caused by prescriptions for ADHD treatment in children and adolescents (Figure 5). (Of course, as adult patients with ADHD have not been included in this analysis, the total budgetary impact of ADHD may be greater for those specialists who also treat adults, for example parents of patients.) This almost tenfold increase is the result of (a) the projected ADHD spending in combination with (b) the high concentration of patients among a small total number of specialists and (c) the very small number of prescriptions traditionally written by this specialist group. Even for the extremely low case looked at, this overall increase of drug spending by individual physicians would result in >5-fold drug spending caused by these specialists in 2012, compared to 2002. In this most conservative scenario, €144,500 or 89% of projected expenditures of €162,600 per physician in 2012 would still be accounted for by young ADHD patients.
Pediatricians as a group cause higher drug expenditures, and each of them treats a smaller number of ADHD patients compared to specialists, even those who are among the top-20% service providers for children with ADHD among their group. Hence, the incremental impact of ADHD-related spending is smaller compared to child and adolescent psychiatrists (Figure 4). Still, under base case conditions, ADHD-related pharmaceutical expenditures would account for 39% (or €75,000 out of €193,000 per physician in 2012, assuming a 5% growth rate of non-ADHD pharmaceutical spending) of total drug costs induced by this group. These figures compare with average drug costs of €83,000 per pediatrician in 2002. Thus, ADHD-related prescriptions alone would account for two thirds of the growth in drug expenditures caused by these physicians between 2002 and 2012.