Correspondence to: Dr Frank Sanfilippo, School of Population Health, University of Western Australia, 35 Stirling Highway, Crawley, Western Australia 6009. Fax: (08) 6488 1188; e-mail: Frank.Sanfilippo@uwa.edu.au
Objective: To investigate to what extent Western Australian (WA) patients with Attention Deficit Hyperactivity Disorder (ADHD) received prescribed stimulants in excess of their notified dose from WA pharmacies in 2004 (termed ‘discrepancy’).
Design and data sources: Analysis of administrative data about all people aged two years and older who were prescribed stimulants for the treatment of ADHD in WA, and had at least one stimulant prescription dispensed from a WA pharmacy during the period 1 January to 31 December 2004.
Outcome measures: Discrepancies were identified using minimum and maximum estimation methods (MinDE, MaxDE). We calculated for both methods the discrepancy prevalence by age and sex and annual surplus of stimulant accrued by age.
Results: Of the 15,190 ADHD patients who comprised the study population, 5.4% to 19.0% received stimulants surplus to requirement in 2004, with peak prevalences in 6-8 year-olds (MaxDE 20.1%) and 25-34 year-olds (MaxDE 27.6%; MinDE 10.5%). The amount of stimulant dispensed surplus to requirement was highly skewed, with median annual values that ranged from one to 4.1 standard bottles (100 tablets) of dexamphetamine 5 mg for the MinDE and MaxDE methods, respectively.
Conclusion: It is difficult to definitively estimate to what extent WA ADHD patients accrued excess stimulant medication using routine administrative data. Improvements to the WA Stimulant Regulatory Guidelines are recommended in the interests of patient safety, public transparency, methodological rigour and encouraging good prescribing practices.
Attention Deficit Hyperactivity Disorder (ADHD) affects approximately 3-5% of school-age children;1 of these, 40% persist with symptoms into adulthood.2 Stimulant medication is often the first-line intervention,3 despite its use being controversial for several reasons.4 These include: (i) ethical objections to treating children; (ii) difficulty in objectively quantifying the benefits of stimulant use; and (iii) lack of evidence about the effectiveness of stimulant medication over extended periods, its side-effects, and the long-term effects on physical and neurological development.4 Notwithstanding these concerns, legal use of stimulants has increased in several Western countries over the past 20 years, with the United States (US), Canada and Australia leading the way.5–7
In Australia, there is a shortage of population-based information about the use of prescribed stimulants by people with ADHD. Western Australia (WA) has received intense media attention, with interest fuelled by an estimated prevalence of stimulant use that is 3.5-times higher than the national average.8 Additionally, in 2004 a WA parliamentary report raised concerns about the misuse of prescribed stimulants, based on anecdotal evidence from expert witnesses.9
In WA, the prescription of stimulants is regulated by the WA Department of Health (WADoH), which introduced the Stimulant Regulatory Guidelines (SRG) in August 2003.10 In brief, WA medical practitioners must apply for a Stimulant Prescriber Number from the WADoH and provide standardised documentation (hereafter called notification) for each patient who is prescribed stimulant medication, including details of patient demographics, diagnosis and the maximum daily dose of stimulant authorised. Prescriber registration is restricted to certain medical specialists depending on the diagnosis (i.e. ADHD, brain damage, narcolepsy or depression). Paediatricians and psychiatrists are the most common prescribers of stimulants for ADHD. Maximum daily doses for patients with ADHD must not exceed 60 mg of dexamphetamine, 120 mg of methyl-phenidate, or 12 tablets per day if a combined dexamphetamine/methylphenidate treatment plan is used. Prescribers for patients with ADHD must apply to the Chief Executive Officer of Health for ‘Outside the Guidelines’ approval to exceed these maximum doses, to treat children aged two to four years, and to treat patients with certain co-morbidities (i.e. history of psychosis, bipolar disorder, and/or significant sustained substance abuse).10
As yet, in Australia there has been no formal evaluation of whether the maximum dose authorised by prescribers (i.e. notified or expected use) is equivalent to the amount of stimulant medication subsequently dispensed to patients (i.e. proxy for actual use). The purpose of this study was to investigate the extent to which ADHD patients received prescribed stimulants in excess of their notified dose from WA pharmacies in 2004 (hereafter termed ‘discrepancy’). Two aspects of the discrepancy were investigated: (i) the proportion of ADHD patients who were dispensed stimulant medication surplus to the expected amount (discrepancy prevalence); and (ii) the magnitude of the discrepancy.
Data sources and population selection
The WADoH provided de-identified data about all WA people aged two years and older who were authorised to receive stimulants (immediate or slow-release formulations) for the treatment of ADHD and had at least one stimulant prescription dispensed during the period 1 January to 31 December 2004. These data are maintained by the WADoH in two population-based databases known as the Stimulant Notification Record Database (notification database) and the Dispensed Stimulant Medication Database (dispensing database), which were introduced as part of the SRG and are described in detail elsewhere.11,12 In brief, the notification database provides a measure of expected or intended stimulant use based on daily stimulant dose prescribed, whereas the dispensing database provides a proxy measure of actual stimulant use based on the quantity and strength of stimulant medication dispensed.
Conversion of stimulant doses to dex-equivalents
As the outcome of interest was the amount of stimulant dispensed overall, rather than drug-specific stimulant use, doses and quantities are reported in terms of dexamphetamine (dex) equivalents. Methylphenidate hydrochloride was converted to its dex-equivalent dose (mg) by halving the authorised daily dose (mg/day) from the notification database or halving the quantity dispensed (mg/dispensing record) from the dispensing database.11 The dose and quantity of dexamphetamine sulphate remained unchanged. Accordingly, the authorised daily dose and the amount dispensed reported here represents the sum of all dexamphetamine sulphate and/or dex-equivalent methylphenidate hydrochloride authorised per day (mg/day) or per dispensing record, respectively.
Definition of ‘discrepancy’: minimum and maximum estimates
It was impossible to produce a single, unbiased discrepancy estimate using WA data, because the SRG allows prescribers some discretion in altering the dose without having to update these details in the notification database. Specifically, prescribers must re-notify the WADoH only when “… the quantity of drug required for a new dose will result in an additional standard pack to be supplied per month” (SRG subsection 5.3).10Table 1 illustrates this rule using dexamphetamine as an example. It shows that an ADHD patient could have their daily dose of dexamphetamine increased from one to three tablets (5-15 mg), four to six tablets (20-30 mg) or seven to 10 tablets (35-50 mg) by a prescriber without this change being updated in the notification database. Hence, the notification database does not contain an exact description of the amount prescribed but rather the limitations imposed on prescribers.
Table 1. Interpretation of WA Stimulant Regulatory Guidelines regarding dosage increases (expressed in dexamphetamine equivalents).
Maximum authorised dose recorded by WADoHa
New notification required if dose increased tob
Dosage increase allowed without new notification
(a) WADoH = WA Department of Health.
(b) Notification = mandatory documentation sent by the prescriber to the WADoH advising the maximum stimulant dose a patient is authorised to receive. A new notification is required if the dosage increases to the values shown in the column from the dosage specified in the first column of the table.
(c) Standard pack size = 100 tablets of dexamphetamine 5 mg.
(d) Bottle = standard pack of dexamphetamine 5 mg.
It was therefore considered necessary to estimate a minimum and maximum prevalence at which surplus occurred in the study population, with the true population prevalence lying between these estimates. At the request of the WADoH, we defined our minimum discrepancy estimate (MinDE) as the equivalent of at least one standard bottle of dexamphetamine above the expected monthly requirement, consistent with SRG subsection 5.3.10 This estimate assumed that: (i) all stimulant increases within the SRG-defined thresholds (i.e. where no notification update is required) resulted from prescriber discretion rather than patients refilling their prescriptions too quickly; (ii) prescribers did not authorise doses higher than they intended (in addition to the prescriber discretion already permitted); (iii) prescribers did not decrease doses (which requires no additional notification unless stimulant treatment is terminated); and (iv) most WA pharmacies dispensed stimulants in their standard pack sizes (100 and 30 tablets for short-acting and long-acting formulations, respectively) rather than in split-packs (exact quantities).
Since all but the last of these MinDE assumptions were unknown, we also developed a maximum discrepancy estimate (MaxDE). The MaxDE was defined in terms of an ideal situation in which: (i) the notified dose was the same as the prescribed dose (i.e. no prescriber discretion for dosage increases/decreases); (ii) there was complete patient compliance, including that prescriptions were not dispensed until all of the previous supply had been used; and (iii) pharmacists dispensed split-standard packs of medication if necessary to ensure no excess medication was accumulated throughout the course of a year. Hence, the MaxDE included any amount of stimulant that was dispensed above the expected/notified requirement.
The terms discrepancy and surplus are used interchangeably throughout this paper. We have intentionally avoided using terms such as stockpiling, overuse, abuse, misuse, diversion and the like, because these require knowledge of why people received medication in excess of requirement, which cannot be determined from the WADoH data used in this study.
Calculation of actual, expected and discrepancy in stimulant use
The actual amount (mg) of stimulant dispensed in 2004 (in dex-equivalents) was calculated by multiplying the drug strength (in dex-equivalents) by the quantity dispensed for each dispensing record and then aggregating the amounts for each patient. Dividing this annual figure by 500 mg (one standard pack) converted the unit of measurement to annual bottles of dex-equivalent medication. To ensure the SRG were appropriately applied, an additional step was needed to calculate the MinDE. Specifically, the annual figure was divided by 12 and then rounded up to the nearest dex-equivalent standard pack size to obtain an average monthly amount. This process produced a degree of rounding error, the magnitude of which varied depending on the daily dose. It was virtually negligible for daily doses of 16 mg or less (∼38% of the study population), but increased sharply thereafter. For example, a patient who was dispensed a total of 7,300 mg of dex-equivalent stimulants in 2004 (dose/20 mg/day) received on average two standard packs of dex-equivalent medication per month (i.e. 24 bottles annually) when rounded up, rather than the 14.6 bottles annually when no rounding was applied. In contrast, the actual amount of stimulant dispensed was not rounded when calculating the MaxDE.
Two different methods were used to calculate the expected amount of stimulant required for the MinDE and MaxDE respectively. By definition, it was initially necessary to calculate the expected MinDE by month. The first step for the MinDE involved calculating the expected number of bottles of dex-equivalent medication per month (i.e. 30 days x maximum authorised daily dose/500 mg), which was then rounded up to the nearest dex-equivalent standard pack size. For example, a patient with a maximum authorised daily dose of 20 mg required 1.2 bottles of dex-equivalent medication per month, which rounded up to two standard packs. A monthly surplus to requirement variable was calculated by subtracting the expected number of bottles from the average actual number of bottles dispensed monthly (also rounded up), and a surplus of 1 bottle (standard pack of dexamphetamine) was included in the MinDE. This monthly discrepancy was then converted to an annual figure of dex-equivalent standard packs (12 months x 500 mg) to allow comparison with the MaxDE.
For the MaxDE, the expected amount of dex-equivalent medication required annually was calculated by multiplying the highest authorised daily dose by 365.25 days. A surplus to requirement variable was constructed by subtracting the expected amount from the actual amount of stimulant dispensed in 2004, and all discrepancies 5 mg were included in the MaxDE.
The discrepancy prevalence for each estimation method was calculated overall and separately for five child age groups (2-5, 6-8, 9-11, 12-14, 15-17 years) and five adult age groups (18-24, 25-34, 35-44, 45-54, 55+ years). The denominator for the prevalence estimate was all WA people aged two years and older who were authorised to receive stimulants for the treatment of ADHD and had at least one stimulant prescription dispensed during the period 1 January to 31 December 2004. Fisher's exact test (two-sided) was performed for each age group to examine sex differences in the prevalence estimates. The size of the stimulant discrepancies was described for each age group as medians, because the results were highly skewed.
Prevalence and characteristics of ADHD patients who received surplus stimulant medication
Of the 15,190 people with ADHD who were dispensed stimulant medication in 2004, 818 (5.4%) and 2,889 (19.0%) were classified as receiving medication surplus to requirement according to the MinDE and MaxDE, respectively. The prevalence distribution for the MinDE was unimodal, and peaked at 25-34 years (10.5%, n=150). The MaxDE had a bimodal distribution, with peaks occurring at 6-8 years (20.1%, n=398) and 25-34 years (27.6%, n=393) (see Figure 1). The MinDE prevalence was higher for females aged 18-24 years compared with males (p=0.037), and the MaxDE prevalence was higher for females aged 12-14 years (p=0.034). No other significant sex differences were found.
Surplus of stimulants dispensed
The amount of stimulant dispensed in surplus to requirements was skewed across all ages for both estimation methods (see Figures 2a and 2b), but most noticeably for ages 18-44 years. The median annual surplus for the MinDE remained constant across all ages (500 mg or the equivalent of one standard bottle of dexamphetamine 5 mg). According to the MinDE, few ADHD patients received in excess of 500 mg above their annual requirement (n=52, 6.4%); the most extreme case received 5,500mg (11 bottles of dexamphetamine 5 mg) above that notified by their treating physician.
In comparison, the MaxDE showed the median surplus for children ranged from 698 mg of dex-equivalent stimulants for 2-5 year-olds (n=71) to 1,195 mg for 15-17 year-olds (n=255), or the equivalent of 1.4 and 2.4 bottles of dexamphetamine 5 mg, respectively. For adult ADHD patients, the median MaxDE annual surplus ranged from 1,238 mg of dex-equivalent stimulants for patients aged 18-24 (n=518) to 2,053 mg for patients aged 55 years and older (n=28), or the equivalent of 2.5 to 4.1 bottles of dexamphetamine 5 mg, respectively. The top 10% of these MaxDE ADHD patients (n=290), which included 97 children (33.4%), had an annual stimulant surplus of at least 4,021 mg of dex-equivalent medication (8.0 bottles of dexamphetamine 5 mg). Half of these ‘top’ MaxDE patients had been dispensed more than 5,542 mg surplus to requirement (11.1 bottles of dexamphetamine 5 mg). The most extreme case received 63,585 mg (127.2 bottles of dexamphetamine 5 mg) above expectation. This was the same extreme case as identified in the MinDE, hence indicating an 11.6-fold difference in the size of the discrepancy/surplus calculated using the two different estimation methods, due mainly to the ‘rounding’ used in the MinDE.
To our knowledge, this is the first population-based study to examine whether ADHD patients of all ages are dispensed more stimulant medication than they are authorised to receive (termed discrepancy or surplus). Rather than presenting multiple possible models, we have presented two simplified estimation methods to provide a preliminary insight into a complex process. The first, called the minimum discrepancy estimate (MinDE), is based on the WA SRG and was included at the request of the WADoH. The second, called the maximum discrepancy estimate (MaxDE), represents an ‘ideal’ notification/prescribing scenario. It should be noted that conclusions drawn from this study should be interpreted with an appreciation of the limitations of both estimation methods and an awareness that the assumptions have not been formally evaluated. The rounding used to calculate the MinDE was not ideal and in our view attenuated the effect size. Other estimation models are possible. For example, we have also produced a single ‘surplus’ prevalence estimate of 12.2% using an alternative method that identified ADHD patients who received more than one month of stimulant in surplus to their requirement (results not reported).
The results suggest that between 5.4% (MinDE, n=818) and 19.0% (MaxDE, n=2,889) of WA ADHD patients who were approved to receive stimulant treatment acquired medication in excess of requirement in 2004. The difference between the two estimates is large, and it is difficult to conclude whether the results give rise to optimism or concern. Nevertheless, these equivocal findings demonstrate how different estimation methods can produce substantially different results, which could be used selectively to either fuel or reduce the political concern and negative media attention that surrounds the use of stimulants in WA. Notwithstanding this comment, a surplus prevalence of up to one in five WA ADHD patients provides sufficient evidence for the WADoH to act pre-emptively and decisively in taking stronger regulatory action. As an example of pre-emptive action in the absence of definitive data, the US Food and Drug Administration recently recommended a ‘black-box’ warning on ADHD stimulant medication to highlight the potential for cardiovascular risks.13 While it is possible for other drugs of addiction to be obtained surplus to requirements, lack of data prevented comparison with our observations in stimulants for ADHD.
The two models also differed in their estimates of the average amount of medication dispensed surplus to requirement in 2004. The results ranged from 500 mg (MinDE) to 698-2,053 mg of dex-equivalent medication (MaxDE), or 1 (MinDE) to 1.4-4.1 (MaxDE) standard bottles of dexamphetamine 5 mg, respectively. Furthermore, the MinDE showed that only 52 people (0.34% of the study population) received in excess of 500 mg of dex-equivalent medication (i.e. one standard bottle of dexamphetamine 5 mg) above their annual requirement. In contrast, the MaxDE identified 290 ADHD patients (1.9% of the study population) who received 4,021 mg of dex-equivalent medication (8.0 bottles of dexamphetamine 5 mg) in excess of requirement, of whom one-third were children (17 years).
Interpreting these results is difficult, because patients may acquire excess medication throughout the course of a year for many reasons. The lower annual surpluses possibly reflected ADHD patients who retained a spare bottle or two of stimulant medication ‘just in case’. The higher surpluses raise significant safety concerns if patients are deliberately and inappropriately exceeding the maximum daily dose authorised by prescribers. Apart from the health risks associated with excessive stimulant consumption,13 it also raises the possibility that the excess medication is not for the patient's own use. It is quite plausible that prescribed stimulants are being diverted to illegitimate black-market use, especially in light of our findings that surpluses reached as high as 63,585 mg (127.2 bottles of dexamphetamine 5 mg) in a single year. Although these extreme occurrences were rare, this should not diminish public concern about the non-medical use of prescriptions in countries such as Australia and the US, where the prescription rates are high. A recent US study found significant diversion of prescription stimulants to non-medical use, including newer long-acting (extended release) formulations.14 Diversion of stimulants was reported as the reason for cessation of authorised stimulant treatment for 15 patients in the 2005 annual report of the Western Australian Stimulant Regulatory Scheme.12
Interpretation of the results is further complicated by the Pharmaceutical Benefits Scheme (PBS) 20-day rule, which prohibits pharmacists from dispensing repeats of PBS medicines that have five or more repeats unless 20 days have passed since the previous prescription was dispensed.15 We made no allowances for the 20-day rule in our methods because it was both impracticable and meaningless to do so given the amount of ‘rounding’ required to calculate the MinDE. Furthermore, methylphenidate was not included on the PBS at the time of this study, and so the 20-day rule could not be applied consistently to our data. However, further investigation of this rule as it applies to stimulant use is warranted for several reasons. First, the rule provides significant opportunity to accrue excess stimulants, despite being designed to curtail stockpiling activity and to promote the safe use of medicines. To illustrate this 20-day rule in the context of a 30-day stimulant supply, a prescription refill would be permissible when a patient has one-third (10 days) of his/her medication remaining, thus leading to the potential annual accrual of six months of medication surplus to requirements. Second, this rule can be broken if there are legitimate reasons why a patient needs their medication earlier (i.e. the PBS immediate supply rule), although there has been no formal evaluation of why and how often this occurs. Third, it can be circumvented altogether by using prescriptions with fewer than five repeats. Other criticisms of the 20-day rule have attracted parliamentary debate and are discussed elsewhere.16
Despite the difficulty of interpreting these results, this study raises several additional issues that warrant attention and focus particularly on reducing the uncertainty that exists in the SRG and associated databases. Chief among these is the degree of treatment discretion afforded to WA prescribers and the inconsistent manner in which it is applied across different medication dosages. We can find no publicly available, detailed documentation regarding the nature and design of the SRG to suggest why this level of prescriber discretion is necessary. We believe that the SRG were developed based on the expert opinion of WA prescribers. This expert consensus approach has been used elsewhere and has our support in principle.17 However, formally documented evidence is lacking, or at least not publicly available, on questions such as how widely experts were surveyed or whether these experts represented the best judgement of the entire field of ADHD prescribers. Indeed, we support the need for an independent and public review of the SRG that includes consultation with national and international experts, as suggested by Berbatis and colleagues in the 2004 WA parliamentary inquiry.9
Additionally, we recommend inclusion of data items in the dispensing database that represent: (i) the type of prescription (i.e. original or repeat); (ii) the number of prescription repeats authorised; and (iii) whether the PBS immediate supply rule was applied. Together with improved certainty that the notified dose recorded by WADoH is the same (or a close approximation) of the prescribed dose written on the prescription, these data items could be used to measure medication persistency, which represents the consistency or timeliness of refilling behaviour.18 This would allow for early detection of irregular and suspicious refilling practices and potential breaches or abuses of the PBS 20-day rule that could be investigated further by the authorised prescriber, or referred to an independent drug surveillance body. It was impractical to investigate medication persistency within the context of the present study because of the rounding used to calculate the MinDE.
Since January 2006 (i.e. after the time period of our study), there have been several WA legislative amendments to curb prescription fraud in the prescribing and dispensing of Schedule 8 drugs, to which stimulants belong.19 This includes that all repeat prescriptions are retained by the pharmacy that dispensed the original prescription, hence creating one-to-one prescriber-pharmacist and pharmacist-patient relationships as originally recommended by Berbatis and colleagues.9 Additionally, WA prescribers must now specify a repeat interval and pharmacists must obtain verbal or written approval from the prescriber to dispense a repeat before the repeat interval ends. It is too early to comment on the effectiveness of these strategies, although they do potentially overcome some limitations of the PBS 20-day rule, discussed earlier. However, we believe that further refinements, such as regulating the number of repeats and requiring prescribers to specify a repeat interval that resembles the expected usage, also warrant consideration. Lastly, the addition of Atomoxetine, a non-psycho-stimulant medication, to the PBS in July 2007 for the treatment of children with ADHD, and the increasing use of long-acting stimulants in WA since 2005,20 may also limit the potential to accrue excess stimulant medication. Future studies can confirm this by using the current study results as a baseline.
Using population-based administrative data, the intent of this study was to estimate the proportion of WA ADHD patients who were dispensed stimulant medication surplus to expectation and the magnitude of the discrepancy. We cautiously conclude that the majority of ADHD patients use their medication as prescribed. However, accrual of excess medication surplus to requirement occurred in 5.4% to 19.0% of ADHD patients during 2004. Our contribution to the paucity of information in this area was limited by design characteristics of the current WA SRG. This study highlights the need to review the design, administration and enforcement of these guidelines in the interests of patient safety, public transparency, methodological rigour and encouraging good prescribing practices. The generalisability of our results to other Australian States may be limited given that the rate of stimulant treatment in WA, at the time of writing, was 3.5-times greater than the national average.8
Statement of competing interests
The minimum estimation method reported in this paper was developed at the request of the WA Department of Health (WADoH). The WADoH asserted no other control over the authors' right to publish and is not otherwise responsible for the views expressed in this publication. The authors disclose a potential competing academic and professional interest with the WADoH. There was no financial support from the WADoH for this project, although the authors have received past financial support for analysis of WADoH ADHD data conducted to the WADoH specifications.
We thank the Pharmaceutical Services Branch at the WADoH for providing access to the data used in this investigation. We also thank Mr Andrew Milani, proprietor and pharmacist, Medina Pharmacy, WA, for helpful advice on matters related to the dispensing of stimulant prescriptions and patterns of use, as well as information on the 20-day rule of the Australian Pharmaceutical Benefits Scheme.