Previous literature reviews on this topic have included articles from 1995 – 2006 (13), 1948 – 2011 (23) and 1948 – 2018 (14). However, none focused on the population of those who are prescribed stimulant medication for ADHD, nor performed a meta-analysis on prevalence rates. This systematic review included articles from 2012-2023. This search period was chosen to encompass 10 years of recent research in this area, noting that two previous reviews encompassed literature up to 2011 (14, 23).

Searches were limited to human studies in English and original research. Eligibility criteria were defined as follows: individuals of any age or setting who were currently or previously prescribed stimulant medication for a diagnosis of ADHD. Where ADHD diagnosis was not specified, if prescription status was assessed and the study otherwise met inclusion criteria, the study was included. This decision was taken due to the limited research in this area specifically recording ADHD diagnosis. Studies were eligible if they provided prevalence of misuse and diversion within those prescribed stimulant medication for ADHD, or sufficient data for this to be calculated.

The PECO (Population, Exposure, Comparator, Outcome) framework was used to define eligibility and guide data extraction:

The following search was utilized in PubMed, and translated for PsycInfo:

(misuse OR abuse OR nonmedical use OR inappropriate use OR illicit use OR diversion) AND (stimulant medications OR amphetamines OR methylphenidate) AND (ADHD OR attention-deficit hyperactivity disorder OR attention deficit disorder) AND (“2012/01/01”[Date - Publication]: “2023/07/01”[Date - Publication]).

Articles identified through the search strategy were uploaded into Covidence software (25) for screening and study selection. Titles and abstracts were screened independently by two authors (JF, WC), followed by full text screening of selected studies by both authors. Where a consensus was not reached, conflicts were resolved by the third author. Reasons for exclusion at full text stage were recorded. Where studies that appeared to meet inclusion criteria had data of interest that were missing, conflicting, or unclear, authors were contacted for clarification. If a response was not received within 2 weeks, the study was excluded.

Relevant data were extracted by one investigator using a template in Excel. The outcomes of interest were prevalence of misuse of prescription stimulant medication by those prescribed the medication for ADHD, prevalence of diversion of prescription stimulant medication by those prescribed the medication for ADHD, and risk factors for both. Studies were selected if they reported on one or more of these outcomes. Other data obtained included sample size and setting, study design, demographics of participants, terminology used to define the outcomes of interest, whether and how participants were assessed for ADHD, the total number of participants prescribed stimulant medication, and the number of these individuals who reported misuse or diversion. Risk factors for misuse and diversion by those with a prescription were also sought.

The 12 studies selected for inclusion were critically appraised using the JBI Critical Appraisal Checklist for Prevalence Studies (26). Two reviewers independently assessed methodology and disagreements were resolved through consensus (JF, WC).

Meta-analysis was performed utilizing prevalence rates of misuse and diversion from individual studies. JBI SUMARI software (27) was used to produce forest plots for visual representation of individual study results, to determine 95% confidence intervals of prevalence rates, and to perform statistical analysis of heterogeneity. Heterogeneity was quantified using the I2 statistic, with values over 50% indicating high heterogeneity. Risk factors for diversion and misuse of medication by those prescribed the medication were synthesized narratively. Meta-analysis was not performed for risk factors due to limited data and high heterogeneity in the studies. Studies with incomplete or missing data regarding the outcomes of interest were excluded.

Meta-analysis of diversion studies was performed separately, with sensitivity analysis using different definitions of diversion (i.e. Lifetime vs past year – Figures 1 – 3 ). Funnel plot asymmetry for publication bias was not conducted due to the low number of included studies in the meta-analysis (28).

Abstracts and titles were screened for 927 citations ( Figure 4 ). Reasons for exclusion at this stage (n = 871) were primarily due to an absence of prescription stimulant misuse or diversion information mentioned in the abstract or title. The most common reason for exclusion at full text stage (n = 24) was that misuse was not explored in the subgroup of the study who were prescribed stimulant medication. Studies frequently considered individuals who were prescribed stimulant medication to be ‘controls’ or appropriate users of stimulants. The second most common reason for exclusion (n = 4) was absence of comparison population for those who were misusing or diverting medication. This meant that prevalence rates of misuse or diversion were unable to be calculated.

There were 12 studies included in the systematic review and meta-analysis ( Tables 1 – 3 ). Whilst the search criteria allowed for several study designs, all eligible full-text studies were cross-sectional surveys. Eight studies were online surveys, one study was paper based (29) and two studies did not specify if the survey was electronic or paper based (30, 31). There was no uniform or validated questionnaire employed across studies. Survey content typically included items on demographics, ADHD diagnosis, stimulant prescription status, and misuse and/or diversion. The full survey instruments were not typically published, and no study reported use of standardized tools or methods to control for social desirability bias.

These 12 studies varied in survey design, methodology and terminology. Terminology included ‘misuse’, ‘medical misuse’, ‘non-medical use’, ‘diversion’ and ‘distribution’. Some definitions referred solely to misuse of one’s own prescribed medication (the focus of this study), while others encompassed the use of nonprescribed stimulant medication. Due to this inconsistency in terminology across studies, Tables 1 - 3 provide a detailed description of the terminology employed in each study.

Studies variously explored misuse and diversion over the past 30 days, past year, or lifetime. Stimulant prescription could refer to holding a current prescription, having been prescribed a stimulant in the past year, or prescribed a stimulant in one’s lifetime. The time period assessed in each study is outlined in Tables 1 - 3 .

Study populations ranged from 88 respondents to 10,000 respondents. Eleven studies were from the United States (US), and one was from Austria. Eight studies sampled university or college students, three sampled patients of primary care practices, one surveyed school students, and one surveyed adults across the US. Ten studies sampled adults aged 18 and older. Two studies sampled undergraduate students but did not specify ages. Adolescents were sampled in two studies.

Eleven studies explored the presence of ADHD in some way: six asked participants if they had a diagnosis of ADHD, two assessed symptoms of ADHD using a self-report scale, and three confirmed a diagnosis of ADHD with the treating healthcare practitioner or medical records.

The Joanna Briggs Institute (JBI) critical appraisal tool was used ( Table 4 ). Methodological shortcomings identified in the studies included inadequate sample size and lack of comprehensive assessment for ADHD. Few (17%) studies had an adequate sample size regarding the subgroup of interest, as the number of individuals prescribed stimulant medication tended to be low. Methods used to determine ADHD diagnosis were assessed in question 6: ‘Were valid methods used for the identification of the condition?’. JBI guidance states: ‘If the outcomes were assessed based on existing definitions or diagnostic criteria, then the answer to this question is likely to be yes. If the outcomes were assessed using observer reported, or self-reported scales, the risk of over- or under-reporting is increased, and objectivity is compromised.’ (26) Two studies used self-report scales only to explore ADHD symptomatology. Although the use of standardized self-report scales are valuable in assessing ADHD, particularly when used alongside thorough clinical observations, these were assessed as invalid methods in accordance with the guidance. One study did not explore ADHD diagnosis or symptoms in any way and therefore was marked as an invalid method of identifying the condition (32). Nine (75%) of the studies were considered to use valid methods, asking participants if they had a diagnosis of ADHD and/or confirming with health records or provider.

Six studies reported the prevalence of misuse of stimulant medication over the past year, while one study reported lifetime misuse of stimulant medication. The total number of current or previous prescription stimulant holders who responded to questions assessing misuse was 1,507. Prevalence rates of misuse by those prescribed stimulant medication ranged from 11% to 33% in the past year. The one study assessing lifetime misuse reported a prevalence of 57.6% (29).

The six studies reporting past year prevalence were included for meta-analysis ( Figure 5 ). Heterogeneity was high with I2 = 90.2 (X2 64.98). This may be explained by the differing terminology used to define misuse, demographics of participants, setting, and sample size. Random effects model was utilized in the context of high heterogeneity. Meta-analysis found an average prevalence of past year prescription stimulant medication misuse of 22.6%, with 95% confidence interval of 0.143-0.321 ( Figure 5 ).

Six studies reported prevalence of diversion across the lifetime, and four reported prevalence of diversion in the past year. The total number of current or previous prescription stimulant holders who responded to questions assessing diversion was 1,813. Meta-analysis of all 10 studies reporting diversion (past year or lifetime) found an average estimated prevalence of 18%, with 95% confidence interval of 0.073-0.320 ( Figure 1 ). Heterogeneity was high with I2 = 97.9 (X2 = 422.87). Meta-analysis of studies reporting past year diversion found an average prevalence of 18.2% with a 95% confidence interval of 0.088-0.299 ( Figure 2 ). Meta-analysis of studies reporting lifetime diversion found an average estimated prevalence of 17.9% with confidence interval 0.026-0.420 ( Figure 3 ).

Three studies explored risk factors for stimulant misuse in those with a diagnosis of ADHD and/or a current or previous prescription, and four studies explored risk factors for diversion. Risk factors for misuse included having an amphetamine-based stimulant prescription, comorbid depressive and anxiety symptoms, and beliefs that misuse was unlikely to be associated with risk of harm.

Ramachandran et al. surveyed students who were filling prescriptions for stimulants at a campus pharmacy. They found that 97% of those reporting misuse of their prescribed stimulant were filling a prescription for amphetamine-based stimulants, whilst only 3% were filling a prescription for methylphenidate (33). They found that those reporting misuse of their medication were more likely to report depressive and anxiety symptoms, and were more likely to believe that nonmedical use was associated with no or slight risk (41% vs 31%).

Hartung et al. assessed personality factors of perfectionism and ‘sensation-seeking’ and found no significant association with misuse (34). They also explored the contribution of ADHD symptoms with an 18-item self-report measure and found no significant association with misuse (30).

In terms of diversion, there were conflicting findings regarding the impact of gender, substance use, and misuse of medication. Three studies found no increased risk of diversion associated with gender (29, 35, 36). In contrast, one study of university students found male gender was significantly associated with increased risk of diversion (X2 = 13. 061, df = 1, p = 0.0001) (37). One study found that substance use including alcohol, marijuana, ecstasy and painkillers was significantly associated with diversion (37). Interestingly, students who used cocaine were less likely to divert medication. The authors hypothesized that these students may have been using their stimulants as a substitute for cocaine and therefore had less medication available for diversion.

Gallucci et al. found that individuals who reported misusing stimulant medication were five times more likely to divert (OR 4.967, CI 2.149–11.48, p < 0.001) than those with no misuse history. DeSantis et al. found that students who misused their own medication were 2.51 times (SD = 0.38) more likely to divert medication (37). In contrast, Holt et al. found no increased risk of diversion when individuals reported misuse of their own medication (35).

Having excess medication available, and exposure to peers who were misusing stimulant medication, emerged as risk factors for diversion. DeSantis et al. found that those who diverted medication were taking their stimulant less frequently than those who did not divert (4.7 days per week vs 5.3 days per week) (37). Holt et al. found that having others attempt to persuade the person to share or sell medication was associated with a higher risk of diversion (OR 1.888, CIs [1.132, 3.148]) (31). DeSantis et al. found that diverters reported that 50% (SD = 25.6) of their friends were using stimulant medication without a prescription, compared to only 26% (SD = 26.21) of friends of those who were not diverting (37).

Three studies reported that participants frequently experienced being approached to divert medication (requested to give away or sell). Gallucci et al. found that of 151 prescription holders, 94% of whom reported a diagnosis of ADHD, 58.9% had diverted during their lifetime, and 76.8% had been approached to divert (29). Ramachandran et al. surveyed 88 undergraduate students waiting to fill their prescription for stimulant medication at a university pharmacy, 92% of whom reported a diagnosis of ADHD. 18% reported diversion in the past year, while 59% were approached to divert (33). Epstein-Ngo et al. (38) sampled adolescents from 5 public schools, 179 of whom reported a diagnosis of ADHD for which they were prescribed stimulants in the past 12 months. 10% reported diverting their stimulant medication in the past year, while 20% reported being approached to divert. This suggests that rates of diversion are often much lower than rates at which individuals are being requested to share or sell medication.

There was a high degree of heterogeneity (I² > 90%) between the 12 studies included in meta-analysis. The high heterogeneity is attributable to variation in methodology, sample population, sample size, terminology and definitions. Sample size varied from 20 to 827. Definitions varied, and the differences in definitions and terminology are described in Column 4 of Tables 1 - 3 . Survey questions differed across studies and were not routinely available as published data. Sampling strategies varied, and populations varied in terms of age and setting. The high heterogeneity warrants cautious interpretation of the pooled prevalence estimates, as the applicability of the findings may be limited in diverse contexts.

In terms of behaviors that carry the most risk, only two studies asked about route of administration of stimulant medication e.g. snorting, smoking or injection (16, 39) with one study reporting that up to 6% of individuals who reported misuse were injecting medication (16). The absence of inquiry about route of administration in the other surveys highlights the variability of the definitions utilized, and the difficulty in interpreting the level of risk accompanying misuse. Similarly, frequency of misuse and diversion was not explored in the majority of studies.

It is possible that studies that found higher rates of misuse and/or diversion were more likely to be published than those that found the risk was lower or negligible. This potential for publication bias could have contributed to an overestimation of the prevalence of stimulant medication misuse or diversion in our review. Although we conducted a comprehensive search of multiple databases to minimize the exclusion of unpublished studies, the small number of included studies precluded the use of a funnel plot to formally assess publication bias. Therefore, it is possible that publication bias impacted the results, contributing to higher estimates than would be found in comparable settings where publication is not a factor in distribution of results.

Studies rarely described how they supported participants to correctly identify ‘stimulant medication’. One study used photographs to aid the correct identification of stimulant medication (16), one study utilized health records to confirm prescription history (30), and the remainder either assessed prescription status by giving participants brand names of stimulant medications, or did not describe methods used to help participants correctly report stimulant medication history.

Few studies explored the exact stimulant that a participant was prescribed, or associated the stimulant with the participant’s response and risk of misuse or diversion. This limited the possibility of disaggregation of the results by stimulant. As a result, it was not possible to obtain meta-analysis results regarding which stimulants were associated with higher risk of misuse and/or diversion.

A notable limitation of the current body of literature, and consequently this review, is the limited geographical and demographic representation of the included studies. Most research has been conducted within the United States and predominantly involves college student populations. 11 out of 12 studies were US-based, with one study from Austria. Eight of the 12 studies focused on university/college students. The predominance of studies in college students has been noted in previous reviews (13, 23). This narrow focus restricts the generalizability of the findings to diverse populations and sociocultural environments.

There was considerable heterogeneity in how misuse and diversion were measured across studies. None of the included studies used validated or standardized tools to assess these behaviors, and most relied on author-developed questionnaires, the content of which was not always fully reported. Timeframes and definitions also varied, complicating cross-study comparisons. Furthermore, no study explicitly addressed the risk of social desirability bias, such as through anonymity assurances, social desirability scales, or indirect questioning methods. These methodological limitations may have contributed to underreporting of sensitive behaviors and may limit the generalizability and reliability of prevalence estimates”.

This study sought to examine misuse and diversion of stimulants amongst individuals of all ages. However, only two of the included studies explored misuse and diversion among children and adolescents. Limited information was available regarding the ages of participants in a number of studies. The data that was available regarding age of participants is described in column 2 of Tables 1 - 3 . The small number of studies with children and adolescents, and the limited information about participants of older ages, limits the ability to generalize findings to individuals across the age spectrum.

Small sample sizes present an additional limitation. Although the overall sample size of the study may have been large, the subpopulation of interest tended to be relatively small, ranging from 20 to 827. These smaller sample sizes contributed to wide confidence intervals within the individual studies and the meta-analyses.

This study sought to focus on the specific population of individuals with a diagnosis of ADHD. However, methods used to explore or diagnose ADHD differed. Only three of the 12 studies utilized health records or confirmed the diagnosis with a clinician. Not all studies explored ADHD diagnosis in any form. Six studies relied on self-report of diagnosis. Two studies involved a self-report rating scale of ADHD symptoms, without explicitly addressing if the person had a diagnosis of ADHD. One study did not assess history of ADHD in prescription holders either through self-report or self-assessment. Self-reports could be influenced by recall inaccuracies or misunderstanding of diagnostic criteria, leading to possible misclassification. Assessment scales, whilst useful screening tools, cannot be substituted for a specialist diagnosis using international standards such as the DSM-5.

Within studies that did explore ADHD history, not all individuals with a prescription for stimulant medication reported a diagnosis of ADHD. In Australia and the US, stimulants may be approved for both ADHD and narcolepsy, and off-label prescribing may expand the potential indications. Narcolepsy is a relatively rare condition in comparison to ADHD, with estimated prevalence of 0.079% (40) versus estimated ADHD prevalence of 4-11% (41, 42). It is likely that very few individuals in the included studies were prescribed stimulant medication for reasons other than ADHD, however, this remains a possibility and may have influenced the findings.

The variability in methods used to explore ADHD, and the presence of participants who were prescribed stimulant medication but did not report a history of ADHD, may affect the internal validity of the findings.

This study highlights the importance of risk assessment for misuse and diversion prior to prescription of stimulant medication for individuals with ADHD. Possible risk factors identified in this review were substance use, peer misuse of stimulants, personal beliefs about the risks of misuse and diversion, having excess medication at home due to missed doses, and past history of misuse of stimulant medication.

Notably, there is evidence that the risk assessment process itself has the potential to mitigate the risk of misuse and diversion. One study found that individuals were less likely to divert medication if their doctor frequently asked if they ever ‘share’ their medication with others, and if they understood the dangers associated with sharing stimulants with nonprescribed individuals (37). This aligns with findings that when healthcare providers delivered brief interventions aimed at reducing diversion risk, patients were less likely to report an intention to divert (31).

Comorbid substance use is currently recognized as a risk factor for misuse and diversion of stimulant within existing guidelines (5, 6). Considering the findings in this review, risk assessment should also include exploration of an individual’s history of stimulant misuse, peer misuse, and beliefs about potential consequences of misuse and diversion of stimulants. When enquiring about adherence with treatment, prescribers might also ask about leftover medication or medication ‘stockpiles’, which may increase the risk of the individual giving away or selling medication. Education of individuals about the risk of serious adverse outcomes should oral medication be administered intravenously, either by the individual or a recipient of diverted medication, is recommended. Regular assessment and open communication between patients and healthcare providers can play a pivotal role in reducing these risks, acting as a brief intervention, as well as risk assessment. These practical recommendations are echoed by the studies included in this review. Further elaboration of these suggestions is highlighted in several studies (31, 34, 35, 37). A standardized checklist could be useful as an opportunity for brief intervention at clinical encounters. We propose a brief intervention that utilizes the pneumonic ‘FOCUS’ ( Box 1 ).

This review found that few studies have explored risk factors for misuse and diversion within the population of those who are prescribed stimulant medication for a diagnosis of ADHD. There is a paucity of data relating to individuals prescribed stimulant medication for ADHD who are not in higher education. There is also limited data relating to children and adolescents with ADHD who are prescribed stimulants. Where young children and adolescents are prescribed stimulant medication, their parent or carer is likely to be managing stimulant medication, and likely to be the individual most at risk of diversion and misuse. Further studies involving children and parents would strengthen the current knowledge base around risks in this demographic.

Further research that utilizes standardized clinical evaluations to identify individuals with a diagnosis of ADHD would strengthen the reliability of conclusions drawn. Future research should prioritize including a wider variety of populations across different age ranges and cultural backgrounds. Research in diverse geographic and demographic populations will inform a more nuanced and culturally informed risk assessment for misuse and diversion of stimulant medication.

Future research would be strengthened by a more uniform definition of key terms such as misuse and diversion. We recommend that future studies use the terminology adopted in this review to promote consistency across research. Misuse is defined as the use of prescribed stimulants in a manner not intended by the prescriber. This includes consumption in excess of the prescribed dose, administration via non-oral routes (e.g., intranasal or injection), use for purposes other than the treatment of ADHD, or concurrent use with alcohol or other substances. Diversion is defined as the transfer of stimulant medication prescribed for ADHD from the individual for whom it was prescribed to another individual without a prescription, typically through giving away or selling the medication. These definitions encompass the full range of potentially problematic or concerning behaviors associated with the use of stimulant medications outside the parameters of medical guidance. Study designs that explicitly differentiate between various forms of misuse and diversion would enhance the clinical applicability of the findings. Studies that specifically explore intravenous use of stimulants, which is associated with substantial morbidity and mortality, would address an important and under-studied area of concern.