Stimulants are the most effective class of medication in modern psychiatry, and they’re also the class that earned the worst reputation the fastest, mostly through the doing of a handful of telehealth outfits that figured out you could ship Schedule IIs to people in less time than it takes to verify their address. Both of those things are true at once. I’m going to walk through what the drugs actually do, who they fit, who they don’t, what the side effects look like in real life, and why the pill-mill scandal of the early 2020s gave the whole category a black eye that the medicine itself didn’t earn. Full disclosure up front, I’m a PMHNP and a DNP, I prescribe these regularly, and I’m personally on Vyvanse for adult ADHD. So I’m not pretending to be neutral, I’m just trying to be honest.
What stimulants actually are
Two families, basically. Amphetamines and methylphenidates. Both do the same broad thing, which is bump dopamine and norepinephrine in the synaptic gap, but they get there through slightly different doors. Amphetamines (Adderall, Vyvanse, Dexedrine) reverse the dopamine transporter so it pumps dopamine out of the cell instead of sucking it back in, and they also push more dopamine into the gap from inside the neuron. Methylphenidate (Ritalin, Concerta, Focalin) blocks the reuptake transporter so dopamine that’s already been released sits in the gap longer. Same neighborhood, different mechanism, and that’s why some patients respond beautifully to one class and feel nothing on the other… pharmacology being weird in ways the chart doesn’t predict.
The “paradoxical” framing you’ll still see in old textbooks, the idea that stimulants calm hyperactive kids by some special quirk of the ADHD brain, is wrong and has been wrong for a while. Stimulants don’t calm anything. They raise the signal-to-noise ratio in the prefrontal cortex, which is the part of the brain that handles “what am I supposed to be doing right now.” For somebody with ADHD whose noise floor was drowning the signal, the experience of having signal louder than noise feels like calm because for the first time the chaos has volume control. For somebody without ADHD, the same drug raises both at once and they just feel cranked. The drug isn’t doing anything different. The brain it’s landing in is doing something different.
The specific drugs, grouped by family
On the amphetamine side, the big three are Adderall (mixed amphetamine salts, the 75/25 dextro/levo blend that’s been the workhorse of adult ADHD since the late 1990s), Vyvanse (lisdexamfetamine, a prodrug where d-amphetamine is stapled to a lysine molecule and only gets activated when the bloodstream cleaves the lysine off), and Dexedrine (pure d-amphetamine, the old guard, still useful in specific cases). Adderall comes in IR (immediate release, four to six hours) and XR (extended release, ten to twelve hours via a bead-based delayed-second-pulse mechanism). Vyvanse only comes one way because the prodrug chemistry IS the release mechanism. Dexedrine comes as IR tablets or as Spansule capsules with a slow-release coating.
On the methylphenidate side, the workhorses are Ritalin (the original IR, three to four hours, the one your pediatrician’s pediatrician probably wrote in the 1970s), Concerta (osmotic OROS pump technology, a controlled-release pill that pushes drug out at a metered rate over ten to twelve hours), Focalin (dexmethylphenidate, the d-isomer only, slightly cleaner side effect profile for some patients, comes in IR and XR), Ritalin LA, Metadate CD, Adhansia, Jornay, and a half-dozen other long-acting methylphenidate brands that all do roughly the same thing with slightly different release curves. The release curve is the entire reason these brands exist as distinct products, and the release curve is what you’re actually picking between when you sit in front of a prescriber.
IR versus ER versus prodrug, the whole curve discussion
An IR dose climbs fast, peaks around ninety minutes, and falls off by hour four or five. You feel the kick, you feel the work window, you feel the taper. An XR dose climbs slower, plateaus longer, and falls off in the eight to twelve hour range. The kick is softer because the medication isn’t all in the gap at once. A prodrug like Vyvanse climbs even more gradually because the conversion from lisdexamfetamine to active amphetamine is enzymatically paced, your red blood cells do the work, and the enzyme doesn’t care whether you took it with a giant breakfast or nothing at all. You get a long, gentle ramp, a long plateau, and a slow taper without the bead-release lumpiness some XR formulations have around hour six.
Patients land on different formulations for different reasons that are mostly about how they live their day. A nurse working twelve-hour shifts wants the long, smooth profile and would resent feeling a peak. A college kid who needs three productive hours to write a paper and then wants to go to the gym wants the IR kick and the predictable taper. A trades guy whose work is physical and unpredictable might do better on a long-acting plus a small IR booster for the late afternoon when he’s writing invoices. None of this is set in stone, and most patients try two or three formulations before they land on the one that fits the actual shape of their week.
The first stim conversation
For an adult walking in with a clean ADHD picture, no cardiac history, no active substance issues, and no bipolar in the background, I usually start with Vyvanse or with a long-acting amphetamine salt formulation like Adderall XR. The smoothness of the prodrug curve is what most adults want once they’ve actually tried both, and the lower abuse-spike profile of Vyvanse (you can’t snort it, you can’t shoot it, the enzymatic step doesn’t care what route you took) is a real point in its favor for patients who are nervous about being on a Schedule II at all. The conversion math is rough but workable, thirty milligrams of Vyvanse is in the ballpark of ten milligrams total daily Adderall, fifty is in the ballpark of fifteen to twenty, seventy is in the ballpark of twenty-five to thirty.
Methylphenidate goes to the front of the line in a few specific cases. Kids and adolescents, because the appetite suppression is meaningfully less than amphetamines and that matters when somebody is actively trying to get taller. Adults with significant baseline anxiety, because amphetamines tend to crank anxiety harder than methylphenidate does in most patients (not all, but most). Adults who tried an amphetamine and felt like the volume got turned up on everything including the stuff they didn’t want louder. Patients with a history of stimulant misuse where the family is going to feel safer with the lower-euphoria option. Older adults where the cardiovascular margin is thinner and the slightly gentler pressor effect of methylphenidate is worth picking up.
Stimulants aren’t a diagnostic test
This needs to be said because the field has gotten sloppy about it. If we’re being honest, anyone with a functioning prefrontal cortex gets a productivity boost from a stimulant. That’s pharmacology, not a diagnostic finding. The grad student without ADHD who borrows their roommate’s Adderall and writes a great paper has not just confirmed they have ADHD. They’ve confirmed that dopamine in the synaptic gap helps brains focus, which has been known since the 1930s. Anyone telling you the trial of a stimulant is itself the diagnostic test is mixing up symptom relief with disease confirmation, and those are different categories of question.
Anyone with a functioning prefrontal cortex gets a productivity boost from a stimulant. That’s pharmacology, not a diagnostic finding.
What the stimulant trial actually tells you is whether this particular patient gets enough benefit to justify staying on the medication, at what dose, with what side effects, and on what formulation. That’s a treatment question, not a diagnostic one. The diagnosis has to come from somewhere else, which is history, observed pattern, ruling out the conditions that look like ADHD but aren’t (sleep apnea, thyroid issues, severe anxiety, depression, trauma history, substance use, the unmedicated bipolar patient whose attention problems are mood-driven). Skip the diagnostic work and just titrate the stimulant until the patient feels good and you’ve signed up for the entire pill-mill critique, which we’ll get to.
Side effects, the honest version
Appetite suppression is the universal one. Most patients lose a few pounds in the first month, then plateau. The fix is structural, not pharmacological… eat a real breakfast before the medication kicks in, set a phone alarm for lunch because you won’t feel hungry, and make dinner the biggest meal of the day because you’ll be hungry by then anyway. Patients who get this wrong end up underfed, which makes the side effect picture worse on every other axis (sleep, mood, headaches, the afternoon crash). Drink water like you actually like it, because stimulants are mildly dehydrating and most of the headaches I see are dehydration headaches in disguise.
Sleep disruption is the next universal. Take the dose too late and you’ll be staring at the ceiling at midnight. The half-life math matters. Vyvanse taken at noon is still working at ten PM for a lot of patients. Adderall XR taken at ten AM is mostly out of the system by eight PM. IR taken at three PM is gone by eight. Patients who fight their sleep on stimulants are usually fighting their dosing schedule, not the medication. Move the dose earlier and the sleep usually fixes itself.
Mood and anxiety bumps are real but variable. Some patients feel calmer because the chaos has volume control. Some feel more anxious because everything is louder. The amphetamine-versus-methylphenidate switch matters here, and so does dose. A patient who’s anxious at fifty milligrams of Vyvanse might be fine at thirty. If we’re cranking the dose to chase performance and the anxiety is climbing in parallel, the dose is too high, full stop. Blood pressure usually nudges up by five to ten points systolic and a few points diastolic, which is fine for most patients and a real conversation for anyone already on antihypertensives or anyone with a borderline picture. Get baseline vitals, recheck at every visit, get an EKG if there’s any reason at all in the history. Any prescriber who tells you stimulants are completely safe in someone with a real cardiac substrate is a damn liar, and any prescriber who refuses to write the script over a normal resting heart rate in a healthy thirty-year-old is being Chicken Little about a risk profile that’s been pretty well characterized.
Tics are the one people worry about more than they should. Stimulants can unmask existing tic disorders in patients who were going to develop tics anyway, and the appearance of new tics on a stimulant is a real clinical event that deserves a conversation, but the old fear that stimulants cause tics in children who otherwise wouldn’t get them is mostly not supported by the data. Sexual side effects show up sometimes, mostly as decreased libido or delayed orgasm, less commonly as ED, and they’re usually dose-related and reversible. The late-afternoon crash is the IR signature and the reason a lot of patients move to long-acting in the first place. Long-acting formulations crash less but don’t crash zero, and the dip is usually irritability and brain fog rather than the full IR-style collapse.
Tolerance, adaptation, and the “I need more” conversation
This is where the field gets fuzzy and prescribers get sloppy. True pharmacological tolerance to stimulants exists but is less common in therapeutic dosing than the internet thinks. What’s much more common is that the patient’s environment changed (new job, more demands, less sleep, more stress) and the medication is doing the same work it was always doing but the work it’s being asked to do got harder. The other common pattern is loss of the felt onset over time, which patients sometimes read as “the medication stopped working” when what actually happened is they got used to feeling focused and the contrast with baseline got smaller. The medication is still working. The novelty is gone.
Real tolerance does happen and looks like this. The patient has been at a stable dose across several stable years, life hasn’t meaningfully changed, and the same dose that worked at month three and month eighteen is doing demonstrably less now. In that case, the moves are a short washout (a week or two off, usually possible to time around vacation), a class switch (amphetamine to methylphenidate or vice versa), or a modest dose adjustment with a real plan to revisit it. The move I want to avoid is the slow chase upward where the patient is on seventy milligrams of Vyvanse plus twenty milligrams of IR plus another booster at four PM, and nobody has stopped to ask whether the underlying picture changed or whether we’re just keeping up with a rising tolerance curve we created.
Weekend holidays and the dogma around them
The old pediatric model of skipping stimulants on weekends came from a specific concern about growth velocity in kids and a softer concern about appetite recovery. For adults, the calculus is different. If your ADHD is the kind that makes Saturday morning a disaster (can’t get the laundry done, can’t decide what to do, end up in bedrot until two PM and then mad at yourself), skipping the weekend dose is going to make your weekend worse, not better. If your ADHD is mostly a work-performance issue and your weekends are fine without it, skipping is reasonable and gives your appetite and sleep architecture a chance to reset. Neither answer is universally right. The dogma that everyone should take weekend holidays comes from people who don’t understand that ADHD doesn’t take weekends off.
The pill-mill problem and why telehealth stims earned the bad name
The Done indictment is the one to look at. The DOJ charged Done Health’s founder in 2024 with conspiracy to distribute controlled substances and healthcare fraud, accusing the company of building a system that prescribed Adderall and other stimulants to patients after intake visits as short as thirty minutes, with minimal follow-up, structured around aggressive subscription auto-renewal and a pay-per-prescription compensation model for the contracted prescribers. Cerebral got similar scrutiny earlier, stopped prescribing Schedule II medications in 2022 after major pharmacy chains refused to fill its controlled-substance scripts, and later paid a multimillion-dollar federal settlement over the same prescribing practices. These were not paranoid prosecutions. They were responses to clinical practices that, if we’re being honest, no responsible prescriber would have signed up for, and the fact that the model worked financially is a separate question from whether the medicine was being practiced responsibly.
The legitimate critique that came out of all of that is that fast-track stimulant prescribing without real evaluation, without real follow-up, without a real prescriber-patient relationship is a problem regardless of whether the technology is telehealth or in-person. The illegitimate version of the critique, the one that says telehealth itself is the problem, is wrong and is mostly being used by people who want the regulatory pendulum to swing back to the point where ADHD adults who don’t live near a psychiatrist can’t get treated at all. Telehealth done right (real evaluation, real follow-up, real history-taking, real conversation about side effects and tolerance and the cardiovascular picture) is just psychiatry with the geography barrier removed. Telehealth done wrong is what Done was doing… the medium is not the problem, the clinical practice is the problem.
The medium isn’t the problem. The clinical practice is the problem.
Who stimulants don’t fit
Untreated or undiagnosed bipolar is the big one. Stimulants can push a bipolar patient into mania, and the rate of bipolar disorder being missed and called “treatment-resistant ADHD” is higher than the field admits, especially in young adults whose first manic episode hasn’t happened yet. If somebody walks in describing what sounds like ADHD but the family history has bipolar in it, or the sleep history has periods of three-hours-a-night with massive energy, or the mood picture has episodic highs that look more than usual, the workup needs to slow down before any stimulant gets written.
Uncontrolled hypertension and certain cardiac conditions are real contraindications, not soft ones. Anyone with a known arrhythmia, structural heart disease, or unstable blood pressure needs cardiology input before a stimulant gets started, and anyone with a family history of sudden cardiac death in a young relative needs a careful EKG read at minimum. Active substance use disorder is a careful conversation, not an automatic no. A patient in stable recovery from stimulant use disorder is a hard conversation and usually a no until non-stimulants have been tried. A patient in stable recovery from alcohol or opioid use disorder with clear ADHD is usually fine on a long-acting stimulant with reasonable monitoring, and the calculus has to be made one patient at a time.
Severe baseline anxiety often gets worse on stimulants before it gets better, and sometimes it just gets worse and stays worse. The patient who walks in with primarily anxiety symptoms and a secondary attention complaint is usually better served by treating the anxiety first and seeing what attention picture is left over, because attention problems driven by anxiety usually fix themselves when the anxiety quiets down. Patients with active psychotic symptoms or a clear schizophrenia spectrum picture should not be on stimulants without specialty input, full stop, because the dopamine push can destabilize the psychotic process.
Non-stimulant alternatives
The non-stimulant ADHD medications are real options and they get underused. Strattera (atomoxetine) is a selective norepinephrine reuptake inhibitor, takes four to six weeks to reach full effect, doesn’t have the abuse potential of stimulants, and is FDA-approved for ADHD in adults and kids. It works, but the effect size is smaller than stimulants on average, and the side effect profile (GI upset, fatigue, sexual side effects, rare hepatic issues) is different. Qelbree (viloxazine) is a newer norepinephrine reuptake modulator approved in 2021 for kids and 2022 for adults, broadly similar in effect size to Strattera, with a slightly different tolerability profile. Both are reasonable when stimulants are contraindicated or unwanted.
Guanfacine (Intuniv in the extended-release form) and clonidine are alpha-2 agonists, originally blood pressure medications, that work on the noradrenergic side of the attention system. They’re particularly useful in patients where the ADHD comes with significant emotional dysregulation or trouble winding down at night, and they pair well with stimulants in patients who need a stimulant for daytime focus and an alpha-2 for evening regulation. Bupropion (Wellbutrin) is off-label for ADHD, works on the dopamine and norepinephrine side, and is a reasonable option for adults whose ADHD comes with a depression picture and who want one medication doing two jobs. Modafinil and armodafinil are sometimes used off-label, work for narcolepsy more reliably than ADHD, and are worth considering in specific cases.
Do I have to take it forever
The honest answer is most adults who get real benefit from a stimulant stay on it for the long term, because adult ADHD isn’t something you grow out of and the structural fixes (better sleep, better habits, better systems) are easier to build when the medication is doing its half of the work. Some adults do come off, usually because life circumstances changed (retirement, a job with less demand on executive function, a partner who took over the parts of life the patient was failing at), and that’s fine and reasonable. The framing that staying on a medication is failure and coming off is success is wrong and is a holdover from a moral model of psychiatry where medications were the embarrassing fallback when willpower hadn’t worked. Willpower has nothing to do with whether your dopamine transporter is doing its job.
Diversion, briefly
Don’t lend your script. Don’t sell your script. Don’t give a pill to your roommate the night before an exam. Federal law treats unauthorized distribution of a Schedule II as a felony regardless of whether you charged money or not, and the consequences in the system are real and durable. Beyond the legal piece, the friend you’re sharing with doesn’t have the workup, doesn’t know their cardiac picture, doesn’t have a prescriber watching for trouble, and the worst-case scenario isn’t theoretical. Real people have had real cardiac events on borrowed stimulants and the prescriber whose name is on the bottle is going to have a bad afternoon when the chart pull happens.
What the work looks like in practice
The late-diagnosis adult pattern is the most common one I see now. Say you’ve got a guy in his thirties or forties who walks in, has been white-knuckling his way through executive function for twenty years, finally got tired enough or had enough fall apart in his life that he asked the question. We do the workup, the history lines up, the rating scales line up, the family history has the usual ADHD markers, and we start a long-acting stimulant. Six weeks in he describes it as the volume getting turned down on the chaos. He can finish a task without the sixteen other tabs pulling him sideways. He’s not high, he’s not jazzed, he just has working executive function for the first time he can remember. That’s what treatment looks like when it’s working, not a personality swap or a sudden new self, just signal louder than noise on a brain that was always there.
The teenager pattern shows up too. Picture a kid who was struggling, grades were dropping, parents were exhausted, we got the diagnosis right and got him on a methylphenidate-based long-acting at the right dose. Grades climb. The parents are mostly thrilled and weirdly also a little wistful because the chaotic kid was their kid, and the calmer version is a version they’re getting to know. That’s a real conversation that often comes up in the second or third follow-up. The kid is still the kid. The chaos was getting in the way of the kid being the kid. The medication didn’t change his personality. It cleared the runway.
The recovery patient is the one that takes the most care. Say you’ve got somebody in stable recovery from a non-stimulant substance issue, clear ADHD, has been white-knuckling the attention piece since they got sober and is now wondering if the medication piece would help. The honest conversation is that stimulants are habit-forming for some people and not for others, that we can’t always predict which is which, and that going on a stimulant in recovery is a calculated risk that needs the patient’s recovery support system in the loop. With the right monitoring, with a long-acting formulation (Vyvanse pulls ahead here because of the prodrug profile), with regular follow-up and an honest agreement about what we’re watching for, this works for most patients. It doesn’t work for some, and the willingness to come off if the recovery picture wobbles has to be in the agreement from day one.
Where this leaves you
Stimulants are the most effective class in psychiatry by effect size, and they’re also the class with the most reputational baggage, most of which the medicine didn’t earn. If you’ve been thinking about an evaluation for adult ADHD and have been put off by what you read about the pill-mill scandals, that’s a reasonable instinct, and the answer isn’t to give up on treatment but to find a prescriber who’s doing the work the way it should be done. Real evaluation, real follow-up, real history-taking, real conversation about side effects and the cardiovascular picture. The medicine works. The clinical practice around it has to be worth the medicine. For a broader walkthrough of what evaluation and treatment look like, see the ADHD treatment page.
Sources
Cortese S, Adamo N, Del Giovane C, et al. Comparative efficacy and tolerability of medications for attention-deficit hyperactivity disorder in children, adolescents, and adults: a systematic review and network meta-analysis. Lancet Psychiatry 2018;5(9):727-738 (PMID 30097390). Cochrane review on amphetamines for adult ADHD (Castells X, et al. 2018, CD007813.pub3). Cochrane review on methylphenidate in children and adolescents (Storebø OJ, et al. 2015, CD009885.pub2, updated 2023, CD009885.pub3). Cochrane review on pharmacological treatment of ADHD in children with comorbid tic disorders (Pringsheim T, Steeves T, 2011, CD007990.pub2), which found stimulants did not worsen tics at the group level. FDA prescribing labels for Adderall, Vyvanse, Concerta, Ritalin, Focalin, Strattera, Qelbree, and Intuniv (current revisions). AAP clinical practice guideline for ADHD in children and adolescents (Wolraich et al. 2019). AACAP practice parameter for the assessment and treatment of children and adolescents with ADHD. APA Practice Guideline materials and the World Federation of ADHD International Consensus Statement (Faraone et al. 2021, Neurosci Biobehav Rev 2021;128:789-818). United States v. Ruthia He, Northern District of California, indictment filed June 2024 (Done Health DOJ matter). DEA telehealth controlled-substance prescribing rules and the 2023-2024 extension notices.