1Confirm it's actually a failure

Before anything is changed, clinicians typically re-examine three things: whether the original diagnosis still fits, whether the dose was ever pushed high enough or given long enough to judge fairly, and whether the person actually took the medication as prescribed. Many chronic conditions — depression, hypertension, chronic pain, heart failure — need weeks, not days, before a drug's full effect shows up. A specialist center that runs a dedicated clinic for people whose mood disorders haven't responded to standard care generally won't label a case "treatment-resistant" until someone has been through at least two adequate, well-run trials — meaning the dose and duration were genuinely sufficient, not just attempted.

2Rule out what's fixable first

A large share of apparent treatment failures trace back to something mechanical rather than the drug itself being wrong. Clinical pharmacology teams that specialize in medication safety describe systematically checking for interactions between multiple prescriptions, dosing errors relative to kidney or liver function — especially in older adults — and the buildup of side effects that get mistaken for the underlying illness getting worse. This is also where structured medication reviews earn their keep: research in primary care has found that when a pharmacist systematically reassesses an older patient's full drug list, the number of inappropriate medications and fall-risk drugs measurably drops, often without a single new prescription being added.

• Interaction check — does a newer or over-the-counter medication blunt the effect of the primary treatment?
• Dose and duration — was the ceiling dose reached, and was it given long enough to fairly judge?
• Adherence, honestly discussed — cost, side effects, or a complicated schedule are common, unspoken reasons a drug "doesn't work."
• Organ function — kidney or liver changes can silently reduce or amplify a drug's effect over time.

3Move up the escalation ladder

Once the fixable causes are ruled out, treatment typically escalates in a fairly predictable order rather than jumping straight to something drastic. In psychiatry, this usually starts with dose optimization or switching within the same drug class, then moves to combining medications with different mechanisms, then to adding an agent specifically to boost the first one's effect — a strategy known as augmentation, using options such as lithium, thyroid hormone, or a low-dose second agent alongside the primary treatment. If those steps still don't help, newer options such as fast-acting nasal or infusion therapies, or non-drug approaches like magnetic brain stimulation, become part of the conversation. Cardiology follows a comparable logic: when the standard four-drug combination for heart failure isn't enough on its own, guidelines allow for additional medications or device-based therapies layered on top, rather than replacing the foundation.

4Know that "not working" doesn't always mean "doesn't work"

It's worth separating a genuine drug failure from a case where the evidence behind a medication was thin to begin with. A major review of pain medications prescribed to children found the evidence base was startlingly small compared with adults — a handful of pediatric trials against hundreds of thousands of adult participants studied for similar drugs — meaning some "failures" in children may reflect an evidence gap rather than the drug being biologically wrong for that person. At the other end of the spectrum, some medications get validated for entirely new uses once large trials catch up: a fatigue-easing antidepressant was recently shown in a large randomized trial to meaningfully help people with long-COVID fatigue, a use nobody had proven a few years earlier. And when doubts circulate publicly — such as the theory that antidepressants only work through a placebo effect triggered by noticing side effects — dedicated re-analyses of trial data have specifically tested and rejected that explanation, finding the drugs outperform placebo independent of whether side effects occur.

5What a person can realistically do right now

Nobody has to wait for a crisis point to start this process. Bringing a full, current medication list — including anything over-the-counter or occasional — to every appointment gives a clinician the clearest shot at spotting an interaction early. Asking directly whether a dose has actually reached the top of its recommended range, and for how long it's realistically been given a fair trial, turns a vague sense of "it's not working" into a specific, answerable question. And treating a plateau as information rather than failure — a signal to escalate carefully, not to give up — is consistently what separates people who get unstuck from people who quietly stop treatment altogether.