Table 2.

Data sources and assumptions

(1) Demographic data

Basic demographic data came from national bureaus [77,78]. As a model input, rounded figures of 9.72 million (for Germany, 90% of 10.8 million, reflecting population coverage by the GKV) and 8.1 million (England; with the NHS covering the total population) were used for the number of persons age 6–18 years. As an estimate of the true prevalence of ADHD, 6% was used for base case analysis (cf. below, scenarios) assuming that DSM-IV criteria [79] will be used in clinical practice. A sample of pediatricians surveyed in the German region of Nordbaden [74] had suggested that ADHD was diagnosed according these criteria, instead of ICD-10 [80] criteria often referred to in the European literature [81]. The figure of 6% came from reviews of epidemiological studies [82,83] that coincided with findings from a German mail survey of 165 parents of children age 6 to 10 years using a parent rating scale for ADHD [84]. The British Child and Adolescent Mental Health Survey described a prevalence of 2.23% (1999); its authors noted that diagnoses might be missed if information is not sought from teachers about children's functioning in school [85].

(2) ADHD diagnosis rates

Information on ADHD recognition rates could be obtained from two regional studies: A claims data analysis using a sample of 35,000 children age 3 to 15 years covered by a regional sick fund (AOK) in Hessen, Germany, indicated that the administrative prevalence of ADHD increased from 1.24% in 1998 to 2.43% in 2001 [86]. In 2003, the administrative prevalence of ADHD among children and adolescents age 7 to 19 years in the German region of Nordbaden (N = 317,520) was 2.95% [74]. In line with historic trends up to 2005, it was assumed that awareness of ADHD would continue to increase and result in recognition rates going up from approximately 50% (or 3% of the total age group) in 2003 to plateau at 70–80% or 4.5% of the population aged 6–18 years from 2010 onwards (for England 65(-70)%, in line with current trends).

(3) ADHD treatment rates

Data on the rate of patients receiving drug therapy (i.e., treatment prevalence) were derived from the regional analyses and tested for consistency with top-down estimates based on the number of prescriptions dispensed, revenues booked, and assumed treatment intensity – cf. (6), below. These top-down calculations included adjustment for an assumed 5% share of psychostimulant prescriptions for narcolepsy, gradually declining as ADHD prescriptions rise. They were also adjusted for an estimated 10% off-label prescriptions for adult patients with ADHD in 2003, as indicated by new data from the Nordbaden project [75]. – Analyses of the AOK Hessen sample (for year 2000) revealed a methylphenidate treatment prevalence of 0.52% among children and adolescents below age 20 [87]. Of children age 6 to 15 years with a diagnosis of ADHD in the AOK Hessen sample, 17% were prescribed methylphenidate in 1998; and this rate increased to 29% in 2000 ([88], p. 32). More recent data from Nordbaden indicate that in 2003, 40% (1,161 out of 2,939) of children and adolescents (age group 7–19) with a diagnosis of ADHD were treated with stimulant medication [74,89]. For modeling input, these treatment rates were adjusted for regional variation: In 2001, methylphenidate prescriptions (defined daily doses, DDDs, per population below age 20) were 17% below the national average in Hessen, while they were 23% above the average in Nordbaden [90]. Further it was assumed that the rate of ADHD patients receiving pharmacotherapy would continue to increase through 2010 (and remain stable thereafter), reflecting mounting evidence of long-term treatment effectiveness (cf. earlier, Introduction [33-37]), a growing number of alternative treatment options (cf. below), and the communication efforts by manufacturers competing for market share. Model inputs for England were derived from top-down calculations, corrected for prescriptions for adult ADHD patients; regarding narcolepsy, it was assumed that this indication would be covered by modafinil (trade name ProvigilR), which has been licensed for the treatment of daytime sleepiness associated with narcolepsy or obstructive sleep apnoea [75]. – Note that due to these adjustments model estimates for years 2001 to 2005 deviate from total market data delineated in the Introduction.

(4) New product profiles and availability

An extensive literature and database search was conducted to obtain information on the expected availability of new products and their likely therapeutic profiles. In addition to standard Medline searches, presentations at recent psychiatry and child and adolescent psychiatry congresses in the United States and Europe were screened for reports on ADHD treatment. To identify new products on the horizon, further research on drug development programs in the field was conducted using the pharmaceutical databases of Scrip World Pharmaceutical News and Therapeutics-Daily [49,61]. For key findings, see Results, below, and Table 3. As a rule, it was assumed that new products would become available in Europe (with no difference between Germany and England) 18–24 months later than in the United States, reflecting the predominant strategy of pharmaceutical companies to choose the United States (currently accounting for more than 90% of global ADHD drug sales) as lead market [49].

(5) Market diffusion rates

Diffusion rates and market shares were modeled separately for each alternative preparation, both by category and by product. Each of the quantitative model inputs was informed by findings of the literature and database review (cf. above, 4, and Results, below), and supported by expert consensus derived from semi-structured interviews with experienced pharmaceutical market specialists. For immediate-release preparations of methylphenidate (MPH-IR), generic market penetration was assumed to reach 70–75% towards the end of the projection period (branded MPH-IR, RitalinR, 10%, and dexmethylphenidate, FocalinR, 20% – cf. also Results); this applied similarly to the projections for England and Germany.

(6) Treatment intensity

Treatment intensity was expressed by a single index representing the average number of days on treatment with one defined daily dose (DDD) per day. (For atomoxetine, in a certain number of patients twice daily dosing may be required [91,92]; this was assumed to be the case in 20% of patients on this drug.) On the basis of a methylphenidate (MPH-IR) prescribing analysis in the German region of Hessen, it was estimated that the median duration of treatment had been about 120 days in 2000 [87]. This figure may underestimate the actual duration of treatment since no adjustment for data edge effects was made in this study. Treatment persistence with modified-release products should be higher owing to improved treatment compliance. This expectation has received support from two independent Medicaid claims data studies from California and Texas that found a 37% increase in uninterrupted duration of initial MPH-MR treatment compared to MPH-IR [93,94]. Between 2000 and 2003, mean duration of ADHD treatment was 158 days with MPH-MR and 128 days with MPH-IR [93]. It was further assumed for some scenarios (cf. below, Table 3) that average treatment intensity would tend to increase reflecting findings of the NIMH MTA study [35-37].

(7) Product acquisition costs

Acquisition costs per defined daily dose were calculated for each product from the perspectives of the NHS (England) or the GKV (Germany), respectively. For marketed products, data for large pack sizes (Germany: N2, typically containing 28 to 30 single doses) were retrieved for March 2006 (England) or July 2006 (Germany) from standard sources [75,76]. Ex-pharmacy prices were not corrected for co-payments since the vast majority of patients age 18 years or younger have been exempt from cost-sharing in both England and Germany. It was generally assumed that there would be no price increases during the projection period. Pricing assumptions for new products are described below (cf. also Tables 1 and 3).

(8) Prescribing patterns

Data indicating physician-specific prescription patterns were available for Germany only. At the same time, such data may be most relevant to German physicians, since they are subject to individual drug budget regulation, which differs from the system of Primary Care Groups in England [95]. Pediatricians treated most (5,605 of 11,245 below age 20 years, or 50%) of these patients in Nordbaden (2003), followed by child# and adolescent psychiatrists (3,369 or 30%) [74]. For comparison, in the smaller Hessen sample from 1998–2000, pediatricians accounted for 44% of methylphenidate prescriptions in 2000 [90]. Within physician groups, care for patients with ADHD was highly concentrated in Nordbaden: the top-50% of child and adolescent psychiatrists accounted for 92.1% of ADHD patients treated by their group, and the top-20% of pediatricians accounted for a share of 66.2%. Since budgetary impact of ADHD medications will be of relevance only to those physicians involved in care for these patients, the average impact was determined for the top-50% of child and adolescent psychiatrists and the top-20% of pediatricians, respectively. Total prescription drug spending caused by pediatricians (including drugs for ADHD) was €92,000 per physician in 2004 (2002: €83,000) [20]. Total drug spending of each of the top-50% child and adolescent psychiatrists was estimated at €51,000 in 2004, assuming that 76% of their psychotropic drug expenditures in this year were due to ADHD treatment, derived from data on psychotropic prescriptions for children age 15 or younger [20]. This estimate was consistent with prescribing data obtained from Nordbaden [96], when allowing for regional variance described above. Further assuming non-ADHD expenditures to grow at an annual rate of 5%, the impact of projected future ADHD spending on individual physicians could be estimated.