Abstract
Prescription drug misuse (PDM), or medication use without a prescription or in ways not intended by the prescriber, is a notable public health concern, especially in the United States. Accumulating research has characterized PDM prevalence and processes, but age-based or lifespan changes in PDM are understudied. Given age-based differences in the medical or developmental concerns that often underlie PDM, it is likely that PDM varies by age. This review summarizes the literature on PDM across the lifespan, examining lifespan changes in prevalence, sources, motives and correlates for opioid, stimulant, and tranquilizer/sedative (or benzodiazepine) PDM. In all, prevalence rates, sources and motives vary considerably by age group, with fewer age-based differences in correlates or risk factors. PDM prevalence rates tend to decline with aging, with greater use of physician sources and greater endorsement of self-treatment motives in older groups. Recreational motives (such as to get high) tend to peak in young adulthood, with greater use of peer sources or purchases to obtain medication for PDM in younger groups. PDM co-occurs with other substance use and psychopathology, including suicidality, across age groups. The evidence for lifespan variation in PDM is strongest for opioid PDM, with a need for more research on tranquilizer/sedative and stimulant PDM. The current literature is limited by the few studies of lifespan changes in PDM within a single sample, a lack of longitudinal research, little research addressing PDM in the context of polysubstance use, and little research on minority groups, such as sexual and gender minorities.
Prescription drug misuse (PDM) has received increasing focus in the past decade, owing to precipitous increases in overdoses resulting primarily from opioids. PDM is defined in various ways, but a common definition is use without a prescription or in ways not intended by the prescriber. This definition underlies the nationally representative National Survey on Drug Use and Health (NSDUH) 1 and National Epidemiologic Survey on Alcohol and Related Conditions (NESARC) surveys.2,3 While other definitions separate out “abuse” from “misuse,” 4 this review will use the NSDUH/NESARC-based definition because of wider use in the literature and to avoid confusion with the DSM-IV diagnosis of substance abuse.
The increased focus on opioid PDM specifically, and PDM more generally, has not just occurred in the United States (US); evidence suggests increases in opioid-related overdose and other consequences in Canada,5,6 the United Kingdom (UK) 7 and the rest of the European Union (EU). 8 Preliminary 2018 estimates suggest that US opioid-related overdose deaths fell for the first time in over a decade, from 48 958 in 2017 to 47 608 in 2018, or by 2.8%. 9 In all, prescription opioid misuse accounted for roughly one-third of opioid overdose fatalities in 2018. 9
In addition, there were 11 537 benzodiazepine-related overdose deaths in 2017, with the vast majority of these involving opioids. 10 Overdose deaths from prescription stimulant misuse are virtually impossible to disentangle from methamphetamine, but stimulant misuse has been associated with significant substance use treatment demand and emergency department utilization. 11 Even when PDM does not lead to overdose, substance use treatment, or other healthcare utilization, it is not benign: the correlates (other variables or outcomes with which PDM is cross-sectionally linked)12-18 and consequences19-23 associated with PDM are significant and suggest a need for intervention.
Across most age groups, PDM prevalence rates trail only those of alcohol, marijuana/cannabis and nicotine/tobacco, with cocaine prevalence rates similar in specific cases (eg, versus prescription stimulants in those 26 and older). 24 The increasing focus on PDM referenced earlier has led to better characterization of prevalence in the wider US population and by different subgroups, including by age,16,17,24,25 sex,3,26-28 race/ethnicity,27-31 sexual identity/behavior,32-34 and educational attainment.35-37 PDM sources have been identified and characterized,16,18,38-42 as have motives for engagement.18,43-47 While gaps still clearly exist in the literature on PDM, our understanding of PDM has improved markedly since 2010.
As noted above, there has been increasing research on PDM by age group, with evidence suggesting important age-based differences. There are two key reasons to suspect that examining PDM differences by age group (or using a lifespan/life-stage perspective 48 ) is warranted. First, PDM prevalence varies by age groups, with increases through adolescence 49 and a peak in late adolescence or young adulthood.17,50-52 From there, however, the slope of the decline across the lifespan in PDM prevalence depends somewhat on medication class. Older adults have very low rates of prescription stimulant misuse, relative to their opioid or benzodiazepine misuse prevalence, which are still much lower than in other age groups.24,53-55 For substance use disorder (SUD) from prescription misuse, rates are lowest in those 45 and older, with peaks for sedative and stimulant misuse in the 30 to 44 year range. 56
Second, the medical conditions or developmental concerns underlying PDM vary considerably by age. For opioid medication, chronic pain and prevalence of significant surgical procedures increase with aging,57-61 furthermore, older adults have greater access to opioid medication, given the frequency with which they are prescribed such medication. 62 Increasing access, via prescription, to benzodiazepine medication also occurs with aging,17,62 despite practice guidelines advising against benzodiazepine use in older adults. 63 Sleep disturbances 64 and significant anxiety65,66 are common in older adults, with benzodiazepines commonly employed as treatment. 67 On the other hand, prescription stimulant misuse is often driven by academic and cognitive enhancement motives,38,68,69 and academic demands tend to be concentrated in adolescence and young adulthood. Elevated rates of prescription stimulant misuse in young adult college students and recent graduates supports this idea.35,37
Since PDM varies across the lifespan, our primary aim is to address how PDM prevalence rates, sources, motives, correlates and consequences vary by age. As noted, the literature on PDM across the lifespan is growing, and accumulating evidence strongly suggests differences in medication sources, motives, and, to a lesser extent, correlates across the lifespan. As the current literature on age-based differences in PDM is somewhat limited, this review is intended to be a narrative review, not a systematic review; we do not feel that the literature would support a systematic review, but it is developed enough to highlight key age-based differences in PDM. Finally, we will use the limitations of the current literature to propose topics for future research.
Search methodology
We searched PubMed/MEDLINE, PsycInfo, Scopus, Web of Science (including a cited reference search) and Google Scholar, with the literature search conducted over the November to December 2019 timeframe. Searches used one of the medication class terms: “opioid or analgesic, opioid,” “stimulant or psychostimulant or amphetamine,” or “barbiturate or benzodiazepine or sedative or tranquilizer or anti-anxiety drugs or z-drugs or hypnotic.” All searches also used the term “misuse or nonmedical or abuse” and term for the specific focus: “prevalence,” “source or sources,” “motive or motives or reason or reasons,” “correlate or correlates or consequence or consequences or predictor or predictors or risk factor or risk factors or outcome or outcomes.” Thus, a search for opioid misuse sources would be “opioid AND (source or sources) AND (misuse or nonmedical or abuse).”
The searches produced over 4500 potential articles for screening, with the screening process outlined in Figure 1. We excluded duplicates (ie, two or more search results flagging the same article in different databases), articles not written in English, articles that were not original research (eg, review articles, commentaries, letters to the editor), articles that focused on non-PDM topics (eg, animal/pre-clinical studies of drug effects, studies of “minor tranquilizers,” which is a term of some antipsychotic medication), and then articles focused only on PDM across the population. We also excluded articles without a clear definition of PDM included. We considered any articles from 2000 to the present (December 2019) for inclusion in the review. In the cases where articles from the general population were retained (eg, Compton and colleagues 18 ), these reports were used to establish population-wide baseline estimates for comparison to specific age groups. This left 117 articles and reports for inclusion in the narrative of the review (see Figure 1). Table 1 includes the 117 articles, summarizing the sample, dataset, design, research objectives and key results; it is organized in order of citation number, in the first column.
Flow diagram of records identified, screened, and included.
Research studies included in the narrative review (organized by citation number).
Notes: The table is organized by the first column of citations, by citation number (in superscript).
Prescription opioid misuse across the lifespan
Prevalence
For any past-year prescription opioid use, rates increase with aging per retail prescription data and nationally representative US data across adults 18 and older,70,71,162 peaking at roughly 40% in those 50 and older. 45 Conversely, past-year and past-month prevalence rates of opioid PDM increase through adolescence (12-17 years, unless otherwise noted) into young adulthood (18-25 years, unless otherwise noted) and decline with aging, per nationally representative data from the NSDUH.37,55,70 Work examining local college undergraduate samples 72 and birth cohorts of adolescents and young adults via nationally representative US data21,73 indicates a consistent pattern in both past-year opioid PDM and prescription opioid use disorder (P-OUD): peak opioid PDM occurs between 18 to 21 years, plateaus and declines in the early 30s;72-74 for P-OUD, increases occur from young adulthood steadily into the early 30s. 73 Data from a large school-based sample of adolescents in southeastern Michigan suggests that incidence peaks in late adolescence, 51 implying an increase in level or frequency of use to the early adult prevalence peak.
Members of the “Millennial” generation (birth years 1979-1996) had higher lifetime opioid PDM rates than members of the “Generation X” (birth years 1964-1979) or “Baby Boomer” (birth years 1949-1964) cohorts, with increasing likelihood of opioid PDM as the first substance used in “Millennials,” per nationally representative US data. 75 Similar work using difference age cohorts and different nationally representative data across adults 18 to 57 years also found increased odds of opioid PDM in members of younger generations. 76 Longitudinal data from US high school seniors indicated that the likelihood of persisting opioid PDM (ie, PDM at multiple longitudinal timepoints) increased in younger generations, and persisting opioid PDM was associated with problematic other substance use behaviors. 77 Despite lower overall rates, both NSDUH data in adults 50 and older 53 and US nationwide poison call center data78,79 suggested that opioid PDM increased in adults 50 years and older between 2002 and 2014,53,78,79 with increases in PDM among older adults for reasons related to suicidality 79 and increases in serious medical outcomes related to older adult opioid PDM. 78
Medication sources for prescription opioid misuse
As with opioid PDM prevalence, evidence strongly suggests that the opioid sources vary by age group. Older NSDUH data suggests that young adults are more likely than other age groups to purchase or steal opioid medication from friends and family and to purchase opioids from strangers (including dealers), while adults 50 years and older were more likely to use multiple physicians. 16 Per more recent NSDUH data across the population, young adults have the highest rates of prescription opioid purchases and multiple source use, while adolescents have slightly higher rates of theft. 55 Conversely, use of physician sources (ie, one’s own medication) increases with aging, with the highest rates in adults 65 years and older.55,80 Regardless, obtaining an opioid for free from friends or family was the most common source, except in those 65 and older. 55
Across ages, research consistently indicates that purchases and use of multiple sources are associated with higher odds of concurrent other substance use, like binge alcohol use, marijuana use, prescription opioid use disorder or any SUD. Conversely, use of physician sources/one’s own medication or obtaining opioid medication from friends or family for free is generally associated with lower relative risk. These findings are consistent in US nationally representative data across adolescents,40,81,82 young adults 36 and older adults (50 and older), 55 though physician source use in older adults may not be associated with lower relative risk. 55 Adolescents have relatively high rates of multiple source use, with 20.9% and 44.2% using multiple opioid sources in two recent reports across adolescents 41 and in high school seniors. 40 A key difference in these studies is that the report with the higher prevalence 40 separated friend from family sources, allowing for a greater number of overall sources.
Recent research on young adults suggests that when the friend or family member’s source (ie, the source for the person who gives the respondent medication for free) is from a purchase or from multiple sources, odds of other substance use problems in the respondent are also elevated. 83 Having a family member on long-term opioid therapy increases odds of longer post-surgical use in adolescents and young adults, 163 and having a family member with any opioid prescription increases the likelihood of developing an opioid use disorder. 164 Finally, Rigg and colleagues 84 examined a local, southeastern Pennsylvania sample of adults 20 to 62 years and found that the pattern of sources at the initial opioid misuse episode differed from the past-year pattern in respondents, with somewhat higher rates of obtaining from a friend or family member and of purchases from a dealer at the later assessment.
Motives for prescription opioid misuse
Across the population, the most common motive for opioid PDM is pain relief. In adults 18 years and older, NSDUH data indicate that 63.4% of those engaged in past-year opioid PDM endorsed pain relief as their key motive, followed by to get high (11.6%), and to relax/relieve tension (10.9%).45,85 While any opioid PDM was associated with higher rates of other substance use and suicidal ideation, opioid PDM for pain relief was generally associated with the lowest rates of these correlates. 85 Ashrafioun and colleagues 86 linked opioid PDM motives to suicidality in adults 18 and older, with pain relief only and other motives each linked to higher rates of suicidal ideation and planning (versus no PDM). Combined pain relief and other motives, however, were linked to higher rates of attempts, with evidence that rates of ideation and planning were higher in the combined motive group than the pain relief only group. 86
Two studies examined adolescent samples, finding lower endorsement of pain relief motives than in the general population. In a US regional adolescent sample (11-18 years of age), over 80% of those engaged in opioid PDM endorsed pain relief as a motive, though other motives could be selected as well. 87 Conversely, nationally representative data from high school seniors completing the Monitoring the Future series of surveys suggests that other motives (to relax, experiment or get high; all above 50%) were more commonly endorsed than pain relief (45.5%). 88 A local sample of college undergraduates was similar in that pain relief only was endorsed by less than half of participants, with particularly low rates (29.7%) in those with past-year opioid PDM. 89 Opioid PDM to relieve pain markedly increases with age, peaking in those 65 years and older, per both nationally representative NSDUH data across the US population80,90 and data from Florida in adults aged 60 and older. 91 In contrast, there are decreases in opioid PDM motivated to get high or experiment, which peak either in adolescence or young adulthood.80,90 Adolescent and young adult opioid PDM for non-pain relief motives is consistently associated with poorer outcomes, mirroring the general population.87-89,92 Furthermore, analyses of data from Florida found elevated prevalence rates of SUD and suicidal ideation in adults 60 and older engaged in opioid PDM for non-pain relief motives. 90
Correlates and consequences of prescription opioid misuse
While opioid PDM prevalence rates, sources and motives vary significantly by age group, the correlates and consequences of such misuse are more invariant. Across age groups, prescription opioid misuse is associated with higher rates of concurrent substance use, including binge alcohol use, marijuana use, other illicit drug use, SUD symptoms from prescription opioid misuse and SUDs from any substance.16,81,93-97 Furthermore, polysubstance rates of 90% or higher appear to be common in adolescents and young adults (aged 15-21 years) engaged in opioid PDM, 98 with nationally representative data across the population indicating significantly higher rates of opioid PDM with other substance use (including other polysubstance use) in adolescents and young adults, versus those 26 and older. 99 Earlier initiation of opioid PDM is associated with elevated odds of later opioid use disorder 100 and with opioid use disorder symptoms at the age of 35 years.21,23 Limited evidence from the NSDUH suggests that links between opioid PDM and other substance use vary by age group, 101 with evidence both that past-year benzodiazepine PDM is less likely with aging among those engaged in opioid PDM but that frequency of past-year benzodiazepine PDM is somewhat higher in those 26 and older (versus young adults). 102
Findings in individual age groups are consistent with the population-wide findings that link psychopathology and suicidality to opioid PDM.3,103-111 In particular, the links between major depressive disorder, depressive symptoms and/or suicidality and opioid PDM are robust and well-established in adolescents,16,41,81,112 young adults16,113-115 and older adults (50 years and older),16,94,97,116,117 with a variety of different data sources used. The psychopathology-opioid PDM link may be mediated by race/ethnicity 118 or age, 119 with those 65 and older having weaker psychopathology-opioid PDM associations.
The age-based relationship of sociodemographics and opioid PDM is less clear. In bivariate associations across age groups, white, non-Latino/a males consistently have higher rates of opioid PDM, with unemployed persons, and those with lower family incomes and/or lower educational attainment also often linked.3,16,35,37,45,81,120 In some cases (eg, adolescents, per Schepis and Krishnan-Sarin 81 ) these sociodemographic differences cease to be significant once mental health and/or other substance use correlates are included in models. Mowbray and Quinn 16 found limited between age group differences in opioid PDM correlates via NSDUH data across the population, with the previously mentioned sociodemographic correlates of opioid PDM still significantly associated in most age groups. Their research, however, often did not find associations in older adults. Finally, Day and Rosenthal 121 found that being either divorced/separated or never married (but not widowed) was associated with higher odds of combined past-year opioid and benzodiazepine misuse in those 50 years and older.
Prescription tranquilizer/sedative and benzodiazepine misuse across the lifespan
Prevalence
Data from nationally representative US surveys and administrative prescription databases find increased prevalence of benzodiazepine or tranquilizer/sedative (a survey-specific term capturing benzodiazepine or Z-drug medication) use with increasing age, though the 50-64 and 65 and older groups are equaivalent.17,29,62,122 For PDM, nationally representative US data indicate that young adults have the highest rates of past-year (5.8%) and past-month (1.8%) misuse, and adults aged 26 to 34 years evidenced the highest rates of lifetime rates (8.1%) and the second highest rates of both past-year and past-month tranquilizer/sedative PDM (4.0% and 1.4%, respectively). 17
Prevalence of tranquilizer/sedative PDM has changed differentially by age group in the past 20 years. In adults 50 and older, lifetime tranquilizer/sedative PDM
Medication sources for prescription tranquilizer/sedative misuse
Adolescents are most likely to obtain prescription tranquilizer/sedative medication for free from a friend or relative, followed by purchases.40,41,82 The most common source for young adults is also from friends or relatives for free, especially among full-time college students or recent graduates; the second most common young adult source is from purchases. 36 Adults 18 to 49 years were more likely than those 50 and older to obtain benzodiazepines at their most recent PDM episode from friends or family for free (55.4% versus 45.1%) or purchase them (23.1% versus 8.7%), while adults 50 and older were more likely to use physician sources (40.3% versus 14.9%). 124 Physician sources for tranquilizer/sedative medication misuse increase from young adulthood (9.9%) to a peak in those 65 years and older (38.2%), while purchases decrease from young adults (25.9%) to those 65 and older (2.3%). 125 Across age groups, use of multiple sources and purchases is associated with higher relative rates of problematic substance use and SUD,36,40,41,55,82,125 though even misuse of one’s own medication (ie, physician sources) is associated with elevated risk above those without PDM. 70 Notably, use of physician sources may not be linked with lower relative risk in older adults. 125
Motives for prescription tranquilizer/sedative misuse
Motives underlying prescription tranquilizer/sedative misuse often align with the medication’s FDA indication, such as to induce sleep or relieve anxiety. 165 In a study on tranquilizer/sedative PDM motives in a southeastern Michigan sample aged 12 to 18 years, the most commonly endorsed reasons for misusing to help decrease anxiety, to help sleep, and to get high. 126 Nationally representative data in high school seniors suggest that the most common motives were to relax (66.0%), get high (53.3%), and experiment (49.5%); misuse to sleep was not assessed. 88 In undergraduate college student US samples, conflicting findings suggest that self-treatment is less common than are recreational or combined self-treatment and recreational motives, 89 and conversely that self-treatment was the key set of motives. 47 Data in adults 18 years and older suggests that those 50 and older were more likely than those 18 to 49 years of age to misuse benzodiazepine medication to help with sleep (41.7% and 22.4%, respectively) and less likely to do so to get high (3.0% versus 13.9%); the most common motive, to relax (47.1% overall), did not differ by age. 124
Correlates and consequences of prescription tranquilizer/sedative misuse
Across the population, white/non-Latino/a individuals have elevated tranquilizer/sedative or benzodiazepine misuse, with inconsistent findings for sex differences.29,165,127,128 Consistently, though, those engaged in tranquilizer/sedative or benzodiazepine PDM have higher rates of other problematic substance use and SUD, psychopathology and suicidality.17,106,107,116,124,129-131,162,165 Adolescents, those not in school, those living in socially disorganized neighborhoods, and those with delinquent behavior have higher rates of tranquilizer/sedative PDM.37,131-133 Young adults not in college also had higher rates of tranquilizer/sedative PDM. 37 Research over a three-year period found that while only a small percentage (4.3%) of adults 18 and older engaged in tranquilizer/sedative misuse develop a tranquilizer/sedative-specific SUD, a much larger percentage (45%) developed an SUD from any substance. 134
Only two studies have directly compared tranquilizer/sedative PDM across age groups (12 and older, both using NSDUH data), with one examining any tranquilizer/sedative misuse, 17 and the other examining zolpidem misuse specifically. 25 In both cases, sociodemographic correlates showed a weaker pattern of age-based association with PDM than substance use and SUD correlates.17,25 Generally speaking, substance use/SUD and psychopathology variables were more weakly associated (or not significantly associated) with tranquilizer/sedative PDM in adults 50 and older,17,25 especially in the 65 and older cohort. 17
Prescription stimulant misuse across the lifespan
Prevalence
As with other medication classes, incidence of stimulant PDM peaks between 17 and 19 years of age. There is a clear female peak at 18 years and a bimodal distribution in males with peaks at 17 and 20 years, per NSDUH data in those 12 to 21 years of age. 135 Other research in adolescents 10 to 18 years points toward increasing prevalence across this age window,136,137 corresponding with increases in receipt of diverted stimulant medication. 138 Young adults have the highest stimulant PDM point prevalence rates,36,37,52 with evidence that current full-time college students and recent college graduates have the highest rates within young adults.35,37 Limited evidence suggests increased lifetime stimulant PDM in adults 50 to 64 years of age and 65 years and older between 2002/03 and 2012/13, but prevalence rates are both much lower than younger groups and quite low overall. 53 Notably, while the pattern of stimulant PDM prevalence mirrors other medication classes, nationally representative US data and multi-national administrative pharmacy data indicate that stimulant use patterns differ, with increases through childhood and adolescence to a peak in young adulthood, followed by declines.18,139
Medication sources for prescription stimulant misuse
Data from across US adults 18 and older indicates that obtaining stimulants for PDM from friends or family for free was the most common recent source, at 56.9%; 21.8% purchased or stole the stimulant from a friend or relative, while 11.1% used physician sources. 140 Older nationally representative data in adolescents indicated that use of friends or family to obtain stimulant medication for free was somewhat less common (47.2%) and purchases were somewhat more common (29.7%) than in adults. 82 More recent nationally representative adolescent data indicated that 18.1% of adolescents used multiple sources. 136 The greatest difference between the older (which did not include multiple sources) and newer data was in use of friends or family to obtain stimulant medication for free (29.0% versus 47.2%), 41 suggesting that adolescents obtaining stimulant medication from friends or family for free may have high rates of multiple source use.
In young adults, data from local samples strongly suggest that peers are the most common stimulant source, whether the medication is purchased or given for free.44,141,142,166 This pattern did not change appreciably in a four-year longitudinal college student sample. 38 Nationally representative young adult data indicated that prevalence of obtaining stimulants from friends or family for free was higher in young adults than adolescents, though not for young adults not in college. 36 The prevalence of multiple sources was similar between adolescents and young adults, though young adult college graduates had much lower rates (9.3%). 36 Chen and colleagues 143 found higher relative rates of physician source in adolescents and higher relative rates of friend/relative or illegal (eg, theft) sources in young adults, versus adolescents. Those aged 26 and older were not different than adolescents or young adults. 143
Again, use of purchases or multiple sources was associated with higher odds of other problematic substance use and psychopathology in adolescents and young adults.36,41,82 Theft was also associated with somewhat higher rates of these correlates. 82 Finally, use of physician sources (ie, one’s own medication) was associated with higher odds of SUD from stimulant PDM in young adults. 36
Motives for prescription stimulant misuse
Most adults (18 and older) engaged in past-year stimulant PDM did so to increase alertness or concentrate (56.3%), with significant proportions engaged to help study (21.9%) or for recreational purposes (eg, get high, experiment; 15.5%). 140 Motives are unexplored in age-based subgroups of adults aged 26 years and older, with all published research in older adolescents or young adults. In high school seniors, stimulant PDM is most commonly motivated to increase energy or to experiment (56.8% and 54.6%, respectively; multiple motives could be selected), with large subgroups motivated to stay awake (47.7%), feel good (47.4%), or lose weight (35.5%). 88 Academic enhancement motives (eg, to concentrate or study) were not specifically assessed, but a second study in high school seniors suggested that stimulant PDM to study was endorsed by roughly half (48.7%) of the sample, though stimulant PDM solely motivated to study was somewhat rare (7.4%). 144
In contrast, academic enhancement motives are key in young adult college undergraduates. Regional undergraduate samples suggest that endorsement of self-treatment motives only (eg, to concentrate or study) were found in 25.4% to 78.3% of respondents, perhaps due to inclusion of weight loss and increasing alertness as self-treatment motives in study with the higher rate.47,89 In both cases, recreational-only motives for stimulant PDM were less common, at 3.5% 47 and 10.0%. 89 Longitudinal research indicates that study-related motives increase through four years of college (73.8% to 91.5%), while stimulant PDM to experiment or for curiosity decreases (18.7-1.3%). 38 Other research in undergraduates finds that academic enhancement is the key motive, with much higher scale score ratings than those for recreational motives.26,46
Correlates and consequences of prescription stimulant misuse
Both US local and nationally representative samples of adolescents and young adults engaged in stimulant PDM find consistent associations with other substance use, including problematic alcohol use, illicit drug use, and other forms of PDM.136,145-149,167 Initiation of stimulant PDM before high school graduation is associated with higher rates of problematic substance use at 35 years of age.22,150 Stimulant PDM is also associated with neuropsychological impairment 151 and psychopathology, particularly with ADHD symptoms and diagnosis in college undergraduate samples.44,152 Stimulant PDM is less consistently linked with depressive and anxiety symptoms in adolescents (10-18 years) and college undergraduates,136,153-155 mirroring findings from the adult population 18 and older, where stimulant PDM was only inconsistently associated with such psychopathology. 106 In college undergraduates, non-oral stimulant PDM and more frequent PDM were linked to higher depressive symptom scores, 156 and college students engaged in stimulant PDM had high rates of irritability, insomnia and headaches associated with such PDM. 44
As stimulant medications have appetite suppressing effects, research has examined potential relationships between stimulant PDM and weight suppression. This work suggests links between disordered eating behaviors or weight suppression efforts and stimulant PDM in both adolescents and college undergraduates,140,157 with weight loss motives more common in female college students. 158 Nonetheless, stimulant PDM motivated to lose weight is typically not a common motive, even in younger groups. 140
Summary, limitations and future directions
Across medication classes, a set of common themes emerge. First, PDM prevalence rates increase from adolescence to young adulthood, where they peak; rates decrease slightly in the 26 to 34-year-old cohort, then drop more strongly through the rest of the lifespan,17,37,53,72-74 despite higher rates of opioid and benzodiazepine prescribing to adults 50 years and older.17,62,70,71,162 Peak PDM incidence is in the 17 to 20-year age range,72,73,135 though these estimates may be biased by the young adult populations providing data. PDM frequently occurs with other substance use,70,81,98 and adolescents have high rates of co-ingestion of a controlled medication and another drug, most often alcohol or marijuana.159-161 This set of patterns is consistent with those in other drugs of abuse, 24 though age of initiation may be somewhat later for PDM. 168
Second, medication sources for PDM change across the lifespan as well. Use of physician sources (ie, one’s own medication) increases with aging, while purchases and theft decrease from a peak in young adulthood.16,36,40,41,55,125 Throughout the lifespan, obtaining medication from friends or family for free is the most common source, with only small differences between age groups.16,55,124,125 Notably, the association of specific medication sources and poor outcomes, such as substance use and psychopathology, may change over the lifespan; while use of physician sources (or one’s own medication) is generally associated with lower relative risk in adolescents and young adults,36,40,41,82 this is not true for older adults.55,125 Stimulant PDM sources are virtually unstudied outside of adolescent and young adult samples, and the extant evidence suggests differences, such as higher odds of SUD in those using physician sources. 36 Future research that separates out friend and family sources is important, given that friend sources are associated with greater odds of problematic substance use. 40
Third, motives may change over the lifespan with increased self-treatment motives with aging.88,89,91,124 Nonetheless, it is not clear whether tranquilizer/sedative motives in young adults are primarily self-treatment or recreational in nature, given conflicting findings. Motives generally remain understudied, particularly in older adults, with evidence for a relative increase in self-treatment from research on benzodiazepine motives. 124 More recent research on opioid PDM suggests also increasing self-treatment (ie, pain relief) motives with aging, with concomitant decreases in recreational motives.80,90
Fourth, PDM correlates and consequences are similar across medication classes and age groups, except for overdose, which is primarily a consequence of opioid or combined opioid-benzodiazepine misuse and concentrated in early middle adults. PDM is robustly associated with problematic substance use, SUD, and psychopathology.21,23,81,94,110,131,145,169 Limited evidence suggests that these associations for tranquilizer/sedative PDM may be weaker in older adults,17,25 but one study of opioid PDM does not concur, finding similar prevalence rates of other substance use and psychopathology in those engaged in opioid PDM across age groups. 16 The relationship between PDM and specific sociodemographic characteristics may weaken with age, 17 but as with substance use and psychopathology, this is not consistent.16,25
Limitations of the review
While there are clear limitations in the literature, addressed below, there are four limitations of the current review. First, this is not a systematic review, a choice made because of the inherent limitations of the literature, including different definitions of PDM, some inconsistencies in age groups, different samples studied, and a lack of replication for many findings. The literature is sufficiently developed to suggest age-based differences and directions for future research (below), but this review is limited by its nature as a narrative review. Second, the review is limited by the specific search terms used, though use of MESH terms and terms used in systematic reviews of PDM (eg, Votaw et al 165 ) should mitigate this. Third, the review addressed studies that combined tranquilizer (eg, benzodiazepine) and sedative (eg, Z-drug) medications in one category and studies that examined benzodiazepines only. This is an inherent limitation of some of the strongest data we have on PDM, the NSDUH, which uses the tranquilizer and sedative categories. Finally, only studies published in English and those in the listed databases were included, raising the possibility that potentially relevant studies were excluded.
Limitations of the literature and future directions
The literature on age-based or lifespan changes in PDM is limited in five major ways. Most importantly, there are very few age-based studies in the same sample on PDM prevalence, medication sources, motives or correlates. The result of this limited literature is that comparisons between age groups are often only possible using different studies, and these usually are composed of different samples (local versus nationally representative) at different times, with different survey questions and other methodology. Such comparisons are inherently limited. Understanding of age-based changes in motives and in stimulant PDM are limited by a lack of such single sample investigations, and a greater number of investigations from a single sample would greatly aid our understanding of age-based changes in PDM. Furthermore, correlates associated with PDM that could vary by age, such as employment, 120 educational attainment, 101 marital status, 52 or religiosity 170 warrant further study in a lifespan or age-based manner.
Second, longitudinal investigations are limited as well. Austin and colleagues found that any experience of emotional or physical abuse prior to the age of 18 was associated with lifetime prescription opioid misuse at a follow-up assessment at least 6 years later. 171 Also, opioid 23 or stimulant 22 PDM by 17 or 18 years of age is associated with a greater likelihood of SUD symptoms at 35 years of age. These studies suggest that early risk factors increase the risk for PDM later in life and that PDM in adolescence increases the risk for poor adult outcomes. A further example that supports the need for longitudinal research is from Rigg and colleagues, 84 who found changes over time in patterns of opioid PDM sources. Sources may change with greater duration of PDM, but given that misuse tends to be a discontinuous phenomenon, 50 lifespan concerns and motivations may play a more prominent role in source selection. Together, these studies indicate that longitudinal studies are needed in all age groups, but investigations that extend later into middle and older adulthood are particularly needed.
Third, while more work is needed on age-based changes in PDM, this is particularly true for benzodiazepine or tranquilizer/sedative PDM. There are no studies of stimulant PDM in older adults, and such studies are especially needed. While stimulant PDM has low prevalence rates in older adults and investigations may be hampered because of this, benzodiazepine PDM is much more prevalent in older adults. Older adult benzodiazepine PDM is also associated with significant consequences, 19 making its study especially important.
Fourth, PDM commonly occurs with other substance use,3,18,99,129,169 and emerging evidence suggests that P- OUD clusters strongly with other SUDs, including other SUDs resulting from PDM. 172 Rates of non-PDM SUDs are very high in those with PDM, 70 and evidence from Howard and Grigsby 99 suggests that combined opioid PDM and other substance use is more common in younger individuals, reinforcing the need to study polysubstance use among those engaged in PDM. Too often, PDM from different medication classes is investigated as a “stand-alone” phenomenon, with other forms of substance use and PDM treated as correlates. Variable-centered approaches, such as latent class or latent profile analysis, are needed to identify naturalistic patterns of PDM and other substance use in different age groups. It is also unclear whether polysubstance use involving PDM is more common or severe in different age groups, and the same is true for poly-PDM (or, PDM from multiple medication classes). In all, studies that examine PDM in the larger environment of substance use are needed. Similarly, psychopathology and PDM frequently co-occur,106,107,172 and longitudinal research is needed to clarify the nature of these relationships.
Finally, many potentially vulnerable populations are understudied, including sexual minorities, gender minorities and multi-minority individuals (eg, members of both a racial/ethnic and a sexual or gender minority group). Sexual minority individuals likely have higher PDM rates than their heterosexual counterparts,34,173-175 with one study suggesting higher rates in transgender individuals. 176 Multi-minority individuals may also have higher PDM rates, 177 but this varies by race/ethnicity and sex.178-180 Finally, the degree to which un- or undertreated ADHD symptoms motivate stimulant PDM is unclear and those with significant ADHD symptoms or an ADHD diagnosis are in need of further study. 169 Studies of these potentially vulnerable minority groups are needed both across the population and in specific age cohorts.
Ultimately, greater knowledge of age-based changes in PDM prevalence, processes, and clustering with other forms of substance use and SUD can help policymakers and clinicians make more informed decisions about prevention and treatment. To illustrate, we noted previously that self-treatment motives for PDM (eg, pain relief motives for opioid PDM) are likely to increase with aging.80,90,124 If this continues to be replicated, clinicians may need to encourage more comprehensive health assessments and treatment plans (including non-pharmacological options) for older adults to treat the pain, anxiety, insomnia or other medical conditions that motivated PDM. Such information would be a significant aid in reducing the morbidity and mortality associated with PDM across the lifespan.
Footnotes
Funding:
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the National Institute on Drug Abuse [grant numbers R01 DA043691, R01 DA42146, and R01 DA031160]. The content is the authors’ responsibility and does not necessarily represent the views of the National Institute on Drug Abuse.
Declaration of conflicting interest:
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Authors’ Contribution
All authors contributed to the conceptualization, focus, and organization of the review manuscript. TSS and DLK performed the literature review and drafted the initial version of the manuscript. All authors revised the review for content and contributed revised text, and all authors approved of the final version of the submitted manuscript.
