Depression is the leading cause of disability worldwide, yet it remains widely misunderstood. At its clinical core is major depressive disorder, a diagnosable condition characterised by persistent low mood, loss of interest, and a cluster of physical and cognitive symptoms that last at least two weeks and impair daily functioning. According to the World Health Organization’s 2023 fact sheet, more than 280 million people live with depression globally. [who-depression-factsheet] The good news is that effective, evidence-based treatments exist and, for most people, recovery is a realistic outcome with the right support.
This overview draws on current clinical guidance, including the 2022 NICE guidelines and the Global Burden of Disease Study 2019 analysis in Lancet Psychiatry, to explain what depression is, how it is diagnosed, and what the evidence says about treating it. Whether you are experiencing symptoms yourself or supporting someone who is, the aim is to give you a clear, accurate starting point.
How clinicians define depression in 2026
In 2026, clinicians define depression primarily through standardised diagnostic criteria. The DSM-5-TR requires a depressed mood or loss of interest for at least two weeks, accompanied by at least four additional symptoms from a list that includes changes in appetite or weight, sleep disturbances, fatigue, poor concentration, feelings of worthlessness or guilt, and in severe cases, recurrent thoughts of death. The ICD-11, used more widely in Europe and by the WHO, takes a similar but slightly broader approach.
The 2022 NICE guideline NG222 underlines that diagnosis should be based on a clinical interview, not a questionnaire score alone, because standardised tools such as the PHQ-9 are screening instruments rather than diagnostic ones. [nice-ng222-2022] Cultural context matters too: in some populations, depression presents primarily through somatic symptoms (pain, fatigue, physical discomfort) rather than the explicit low mood described in Western classification systems. Clinicians who probe only for “sadness” can miss the diagnosis. Ferrari and colleagues’ 2022 Global Burden of Disease analysis in Lancet Psychiatry confirmed that depression remains one of the most burdensome conditions in every world region, reinforcing why accurate diagnosis is the essential first step. [ferrari-2022-gbd]
Symptoms, severity, and the diagnostic interview
The symptoms of a depressive episode span mood, physical functioning, and cognition. In the mood domain, people describe persistent low mood, an inability to feel pleasure (anhedonia), irritability, or a general emotional numbness rather than overt sadness. Irritability deserves particular attention because it is often what the person and those around them notice first, especially in men, and it is easily read as a temper problem rather than as depression: see our guide to anger and what drives it for how to tell the two apart. Physically, sleep is almost always disrupted, whether as insomnia, early-morning waking, or hypersomnia. Appetite and weight commonly change, and many people report a pervasive fatigue that does not lift with rest. Cognitively, concentration and decision-making deteriorate; memory complaints are common and, in older adults, can be mistaken for early dementia. The direction of the history matters here: concentration problems that arrived with the low mood point to depression, while difficulty that has been there since school is one of the missed signs of ADHD in adults, and the two are confused in both directions.
Severity is usually rated on a spectrum. NICE NG222 groups presentations as less severe (fewer, milder symptoms with modest functional impairment) and more severe (multiple symptoms causing significant disability). Riera-Serra and colleagues (2023) conducted a systematic review confirming that suicidal ideation is a clinically important symptom that requires explicit assessment in every diagnostic interview, not only in severe presentations. [riera-serra-2023-suicidality] The diagnostic interview itself should cover symptom duration, onset, and any prior episodes, as recurrence is common: Buckman and colleagues (2018) found in a meta-synthesis that a previous episode is among the strongest predictors of relapse, making history-taking central to both diagnosis and treatment planning. [buckman-2018-recurrence]
Causes: biology, environment, and life events
No single factor causes depression; current models point to a convergence of biological, psychological, and social contributors. Genetically, depression runs in families, and a landmark 2000 meta-analysis by Sullivan and colleagues estimated the heritability of major depression at about 37% (95% CI 31-42%), meaning environment accounts for the larger share. [sullivan-2000-heritability] Neurobiologically, disruptions in serotonin, noradrenaline, and dopamine signalling are well-documented, though the old “chemical imbalance” framing oversimplifies a picture that also involves the hypothalamic-pituitary-adrenal (HPA) stress axis, neuroinflammatory processes, and structural changes in areas such as the hippocampus and prefrontal cortex.
Psychologically, cognitive models emphasise that negative beliefs about the self, the world, and the future create a vulnerability that life stressors can tip into a full episode. Adverse childhood experiences, including neglect, abuse, and early parental loss, substantially elevate lifetime risk. The WHO’s 2022 World Mental Health Report underlined that poverty, unemployment, social isolation, and systemic discrimination are also powerful drivers at the population level. [who-2022-mental-health] Life events, from bereavement and relationship breakdown to physical illness, are common proximal triggers, though the same event will affect different people very differently depending on their biological and psychological resilience. Bereavement is the trigger most often mistaken for the condition itself: grief and depression share surface features but move differently, and our guide to the stages of grief covers what the famous model gets wrong and when grief needs assessing as depression in its own right. Understanding this multifactorial picture matters clinically because it shapes what treatment strategies are most relevant for each individual.
Evidence-based treatments and what they actually do
Current guidelines converge on a stepped-care model. For less-severe presentations, NICE NG222 recommends starting with a low-intensity psychological intervention, such as guided self-help based on CBT principles, or group CBT. For moderate-to-severe depression, the first step is typically an antidepressant, a course of individual CBT, or both together.
On the medication side, Cipriani and colleagues’ landmark 2018 network meta-analysis in The Lancet, covering 522 trials and over 116,000 patients, confirmed that all 21 antidepressants studied were more effective than placebo. [cipriani-2018-antidepressants] SSRIs such as sertraline and escitalopram had the most favourable balance of efficacy and tolerability, which is why they remain the default first choice. Antidepressants generally take two to four weeks to produce a meaningful response, and the full effect can take six to eight weeks, a timeline patients often need to hear clearly at the outset. Several newer options work through different mechanisms and arrive with claims about speed; our guide to what is actually changing in antidepressant treatment sets out how they differ from the SSRIs and what to raise with a prescriber.
On the psychological side, Cuijpers and colleagues’ 2023 meta-analysis of 409 trials in World Psychiatry found CBT to be substantially more effective than control conditions and broadly comparable to pharmacotherapy. [cuijpers-2023-cbt] Combined treatment (medication plus therapy) tends to outperform either alone, particularly for severe or recurrent episodes.
For people who do not respond adequately to a first treatment, options include switching antidepressant, augmenting with lithium or an atypical antipsychotic, adding structured psychotherapy, or in severe cases, electroconvulsive therapy. Buckman and colleagues (2018) noted that the risk of relapse rises with each episode, which is why NICE recommends continuing antidepressants for at least six months after remission, and longer for those with recurrent depression. Mindfulness-based cognitive therapy was built for exactly this relapse problem, and NICE lists it as an option for people with recurrent episodes who are currently well; our guide to what mindfulness actually does covers the evidence and its limits. Readers who have not responded to multiple treatments can find more detail in our article on treatment-resistant depression.
Two weeks, and what else has moved
Tick anything that has been true nearly every day for two weeks or more. This is a reflection prompt rather than a test, and it produces no diagnosis.
0 of 6 ticked
The first two items are the ones a diagnosis requires at least one of, and "nearly every day for two weeks" is the actual clinical bar rather than a cautious approximation of it. The screener below uses PHQ-9, the questionnaire most clinicians use, so it gives you a number and a vocabulary to arrive with.
Loss of interest is the one people are least likely to report and most likely to explain away as being busy. Our guide to [anhedonia](/depression/anhedonia/) covers it, including why it responds least reliably to the first thing usually offered.
Nothing here matched. Sadness that tracks something sad, and lifts, is not depression, however heavy it is at the time.
One question comes up often enough to answer directly: imaging research does find average differences between depressed and non-depressed brains, and it still cannot diagnose an individual. Our piece on what brain scans show in depression covers why, and why some clinics sell them anyway.
When to seek help and what to expect
The most important signal is persistence. Feeling low for a few days after a difficult event is a normal response; a low mood that lasts two weeks or more, and that is starting to affect work, relationships, or basic self-care, is worth discussing with a doctor or primary-care clinician. Other prompts to seek help sooner include thoughts of self-harm or suicide, a marked inability to function at work or at home, significant weight loss or gain, or a sense that life holds no meaning.
At a first appointment, a patient can expect the clinician to take a full history of symptoms, ask about any prior episodes, screen for other conditions (including hypothyroidism, which can mimic depression), and ask about current medication and any substance use. NICE NG222 is explicit that the assessment should cover suicidal ideation and the presence of any psychotic symptoms, and that this should be done sensitively but directly. Riera-Serra and colleagues (2023) found that untreated suicidal ideation is a strong predictor of subsequent attempts, reinforcing why a frank conversation at the first appointment matters. [riera-serra-2023-suicidality]
Treatment decisions should be shared. Patients have the right to understand what is being proposed and why. It is reasonable to ask a clinician about the expected timeline for response, what to do if a first treatment does not work, and how long any medication course is likely to run. A short daily record of mood, sleep and what happened, the low-effort core of journaling for mental health, gives those conversations something concrete to work from, which beats reconstructing a fortnight from memory in the waiting room. Browsing the topics page can help build the vocabulary needed to have these conversations, and our anxiety hub covers the substantial overlap between anxiety and depression, which frequently coexist and require integrated management.
Recent research we have covered
- A German trial gave 144 people with treatment-resistant depression 25mg of psilocybin, 5mg, or a placebo alongside psychotherapy. Response rates were 17, 12.5 and 10.6 per cent, a gap that was not statistically significant, and the authors themselves call the trial inconclusive.
- A trial of esketamine nasal spray in 147 hospitalised adolescents at imminent suicide risk met its depression endpoint only once two doses were pooled together. On the severity of suicidality, every group improved by a similar amount, including the one given an active placebo.
- Eighty people with depression and non-suicidal self-injury were randomised to DBT skills training or a support group for 13 weeks. Both helped. DBT looked clearly better only among the 13 to 17 year olds, and that was a subgroup rather than the comparison the trial was built to make.