Three weeks in, a woman is ready to quit. Her doctor told her to give it a month. She's had a headache for most of two weeks, she felt queasy the first few days, and she can't point to a single thing that's better. Her sister keeps asking if it's helping yet, and she's tired of saying no.

She's about to stop taking it. Most people in her spot do.

What she doesn't know is that she's standing right at the edge of the window where things usually start to shift, and that the side effects she's been putting up with are the part that fades first.

Here's the timeline nobody explains properly.

The short answer

Four to eight weeks for most of the benefit, with the first hints showing up much earlier than people expect.

That range frustrates people, and I understand why. But it's honest. Antidepressants don't switch anything on. They shift a system gradually, and the shift shows up in pieces rather than all at once.

Week one and two: small changes, mostly invisible

The old advice was that nothing happens for two or three weeks. That turns out to be wrong.

A meta-analysis in Archives of General Psychiatry (2006) pooled 28 randomized trials covering 5,872 people and found that SSRIs produced measurable improvement over placebo by the end of the very first week. The improvement then kept building for at least six weeks, just at a slowing pace.

The catch is that early improvement is small and easy to miss. It rarely feels like mood lifting. What people usually notice first is sleeping through the night, or getting through a work meeting without their thoughts scattering, or snapping at someone less than they would have.

The side effects, meanwhile, arrive right away. Nausea, headache, restlessness, changes in appetite, and trouble sleeping tend to show up in the first week or two and then ease off. So the first stretch feels backwards: you get the downsides before the upside. That's the point where most people quit.

If side effects are severe or you feel worse rather than flat, call your prescriber rather than stopping on your own. Some medications need to be tapered.

Weeks four to eight: where most of it happens

This is the real window, and it's longer than most people are told.

In the STAR*D trial, the largest real-world antidepressant study ever run in the U.S. and funded by NIMH, the average time to remission was about 47 days. Nearly seven weeks. Roughly half of the people who eventually got well didn't get there until after week six, and about 40% of them needed eight weeks or more.

Read that again if you're at week five and discouraged. Being at week five and unsure is not the same as it not working.

Your prescriber will usually want to see you around weeks two and four. Those visits aren't formalities. They're checkpoints to decide whether the dose needs to go up, since a common reason a medication "fails" is that it was never raised to a dose that could do anything.

When the first one isn't the right one

Sometimes eight weeks pass and the answer really is no.

STAR*D found that about a third of people reached full remission on the first medication they tried. That number gets used to argue antidepressants don't work, which isn't the right conclusion. Most of the people who didn't remit on the first one improved on a later attempt with a different medication, a different dose, or medication combined with therapy.

What it does mean is that needing a second try is normal, not a personal failure and not a sign that nothing will help. The CDC reports that 11.4% of U.S. adults took prescription medication for depression in 2023. A large share of them are not on the first thing they were handed.

What slows the timeline down

A few things quietly push the whole process back, and one of them comes up constantly.

Alcohol. If you're drinking to take the edge off while you wait for a medication to work, you're working against yourself. Alcohol disrupts the deep stages of sleep, and poor sleep drives low mood and anxiety, which is usually the thing you're trying to treat. Some medications also shouldn't be combined with heavy drinking at all.

STAR*D found that people with concurrent conditions, particularly anxiety disorders and substance use, had lower remission rates than people without them. That isn't a reason to give up. It's a reason to make sure the person prescribing knows the full picture, because a treatment plan built on incomplete information takes longer to get right.

Other things that stretch the timeline: skipping doses, an untreated thyroid or sleep problem, and dose increases that never happen because nobody followed up.

When and how to get help

If you've been on a medication for eight weeks at a full dose and you feel no different, that's worth a conversation, not a quiet decision to stop.

Book sooner if you feel worse, if side effects are unbearable, if your drinking has climbed since starting, or if you started having thoughts about not wanting to be here.

Call your prescriber first. If you don't have one, ask your primary care doctor for a referral, use the behavioral health number on your insurance card, or contact a practice that provides ongoing medication management and counseling and tell the intake staff where you are in the process.

If you're having thoughts of hurting yourself, call or text 988 in the U.S. It's free and available around the clock.

Before you stop anything, ask three questions: is this dose as high as it should go, has it been long enough, and is anything else in my life canceling it out. A lot of medications get abandoned at week three that would have worked at week seven.

Sources

Taylor MJ, Freemantle N, Geddes JR, Bhagwagar Z (2006). Early Onset of Selective Serotonin Reuptake Inhibitor Antidepressant Action: Systematic Review and Meta-analysis. Archives of General Psychiatry, 63(11), 1217–1223 — https://jamanetwork.com/journals/jamapsychiatry/fullarticle/668229
Trivedi MH, Rush AJ, Wisniewski SR, et al. (2006). Evaluation of Outcomes With Citalopram for Depression Using Measurement-Based Care in STARD: Implications for Clinical Practice.* American Journal of Psychiatry, 163(1), 28–40 — https://psychiatryonline.org/doi/full/10.1176/appi.ajp.163.1.28
Warden D, Rush AJ, Trivedi MH, et al. What Did STARD Teach Us? Results From a Large-Scale, Practical, Clinical Trial for Patients With Depression.* Psychiatric Services — https://psychiatryonline.org/doi/10.1176/ps.2009.60.11.1439
Centers for Disease Control and Prevention (CDC), National Center for Health Statistics, Characteristics of Adults Age 18 and Older Who Took Prescription Medication for Depression: United States, 2023, NCHS Data Brief No. 528 — https://www.cdc.gov/nchs/products/databriefs/db528.htm