The deadliest part of schizophrenia is the care people don’t get.
People with severe mental illness die 15 to 20 years earlier than everyone else — and mostly from ordinary, preventable physical illness. It isn’t the diagnosis that’s fatal so much as the chasm in care around it. This page holds the whole picture: the gap, where medication fits, and what good care actually looks like — honestly, and without scaring anyone off treatment.
People with severe mental illness in the UK die, on average, 15 to 20 years earlier than everyone else — and the gap is mostly down to preventable physical illness, not the illness itself.
Around 87,000 people with severe mental illness died from preventable physical causes between 2020 and 2022 — roughly 120 every day, and about two-thirds of all deaths in this group.
Diabetes, heart disease, lung disease and obesity all run at roughly twice the rate seen in the general population — and they are caught and treated later.
In one study of hospital staff, 77% remembered a case where a physical illness was missed or delayed because symptoms were blamed on the patient’s mental illness — what clinicians call "diagnostic overshadowing".
The widest health gap in the NHS
Long before any conversation about medication, there is a bigger scandal: people with severe mental illness get worse physical healthcare, later — and it is costing lives.
Twice as ill, half as cared for
How much more common major physical conditions are vs the general population
Stigma drives people away from care
Scottish Mental Illness Stigma Study (2022) — self-reported experiences
The barriers above are self-reported experiences of people living with mental illness — lived experience, not population rates.
The widest health gap in the NHS
People with severe mental illness die 15 to 20 years earlier than the general population. Around 87,000 died from preventable physical illness in just three years (2020–22) — about 120 a day. Roughly two-thirds of these early deaths are from preventable physical conditions like heart disease, diabetes and cancer, not from suicide. This is a physical-health emergency hiding inside a mental-health one.
Royal College of Psychiatrists (2024); gov.uk SMI physical-health inequalities.
Diagnostic overshadowing
When someone with a psychiatric diagnosis describes chest pain or breathlessness, it is too often filed under "it’s the mental illness". In a study across four London hospitals, 77% of staff recalled a case where a physical illness was missed or delayed because of the psychiatric label — and some of those cases were fatal. A diagnosis should open doors to care, not close them.
Shefer et al. (2014), BMC Psychiatry.
People stop coming to care
Being dismissed once teaches you not to come back. In a large Scottish study, people living with mental illness reported avoiding help-seeking (87%), being denied help for a physical problem (72%) and even stopping their medication (52%) because of how they had been treated. These are self-reported experiences — but they show how stigma quietly drives people away from the care that would keep them alive.
Scottish Mental Illness Stigma Study (2022), self-reported.
It starts young
The gap is not just an old-age problem. People aged 15 to 34 with severe mental illness are five times more likely to have three or more physical health conditions than their peers. Around 40% smoke — three to four times the rate in the general population — and the risks compound over a lifetime. Early, joined-up care is where lives are saved.
gov.uk SMI physical-health inequalities briefing.
The medicines do real good
Antipsychotics are part of closing that gap, not widening it. Across the largest long-term studies, staying on them is linked to living longer — not shorter.
Across a 20-year Finnish cohort of 62,250 people, those taking antipsychotics had roughly half the all-cause mortality of those taking none (adjusted hazard ratio 0.48). The medicines, on balance, keep people alive.
About one in four long-term users develops these involuntary movements — far more on older "first-generation" drugs (30%) than newer ones (20%). A real, lasting cost that deserves honesty.
Stopping suddenly triggers relapse in around 77% of people. Tapered slowly over more than ten weeks, that falls to 31%. How you come off matters enormously.
The drug with the worst reputation had the best survival — 15.6% mortality over 20 years, versus 46.2% for those on no antipsychotic at all.
Living longer, not shorter
Cumulative 20-year mortality by treatment (FIN20 cohort, N=62,250)
Protective across the board
Adjusted hazard ratio for death vs no antipsychotic — below 1.0 is protective
The survival paradox
It is the central, uncomfortable truth: the medicines that carry these long-term risks also roughly halve the risk of dying. In a nationwide cohort followed for 20 years, all-cause mortality was 46.2% for people taking no antipsychotic, 25.7% for those taking any, and just 15.6% for those on clozapine. Untreated psychosis is not the safe option.
Tiihonen & Taipale et al., FIN20 cohort, N=62,250.
Protective across the board
The benefit is not only about suicide. Adjusted for confounders, active antipsychotic use was linked to lower all-cause mortality (hazard ratio 0.48), lower cardiovascular death (0.62) and lower suicide (0.52) — with no increase in the risk of being hospitalised for a physical illness.
FIN20 cohort; pooled global meta-analysis RR 0.57.
The sweet spot is a moderate dose
Risk follows a U-shaped curve. Both no treatment and very high doses (above 1.5 defined daily doses) carry the highest mortality; moderate, therapeutic dosing carries the lowest. The goal is the lowest effective dose — not zero, and not the maximum.
Vermeulen et al. meta-analysis; FIN20 dose-response.
And real costs
Taken for years, antipsychotics can cause lasting problems. People living with these effects deserve to have them named plainly — not minimised.
Tardive dyskinesia is the signature risk
Prevalence by drug generation (global meta-analysis, mean 25.3%)
These risks are dose- and time-related, and they differ enormously between drugs. That is precisely why they are manageable: the right drug at the lowest effective dose, with regular monitoring, keeps most of them in check — and many are reversible if caught early.
Tardive dyskinesia
Long-term blockade of dopamine receptors can cause tardive dyskinesia — involuntary movements of the face, tongue and limbs that can be permanent. Mean prevalence is about 25%, higher on older drugs (30%) than newer ones (20%), and it rises with age. It rarely resolves on its own, but newer VMAT2-inhibitor treatments can help.
Global TD meta-analysis (mean 25.3%).
Heart rhythm and sudden cardiac death
Antipsychotics can lengthen the heart’s QT interval and roughly double the rate of sudden cardiac death in current users (incidence-rate ratio about 2.0 for older drugs, 2.3 for newer), rising further at high doses. The risk is tied to current use and dose — which is exactly why ECG monitoring matters. Clozapine carries a small early risk of myocarditis, mostly in the first month.
Ray et al., NEJM (2009), Tennessee Medicaid cohort.
Hormones and bone health
Some drugs (risperidone, paliperidone, haloperidol) raise the hormone prolactin, which can disrupt periods, sexual function and, over years, thin the bones — raising fracture risk roughly 1.7-fold in current users. Others, such as aripiprazole, barely touch prolactin and can even lower it. This is one reason drug choice is so individual.
Pharmacovigilance & fracture cohort studies.
Brain volume and cognition
Long-term imaging studies link higher cumulative doses to small reductions in brain tissue volume, independent of how severe the illness is, and some cohorts link higher dose-years to poorer cognition over time. The effects are real but modest, and have to be weighed against the substantial brain harms of untreated, relapsing psychosis.
Ho et al. (2011); Northern Finland Birth Cohort 1966.
Feeling "not myself"
Beyond the measurable, many long-term users describe emotional blunting — a flatness or sense of being "not like myself" — alongside sexual difficulties, drowsiness and, for some, distressing restlessness (akathisia) that can drive low mood. In surveys of long-term users these experiences are commonly reported, and they matter just as much as any lab result.
Self-reported surveys of long-term antipsychotic users.
A trade-off, navigated together
None of this means “take the maximum forever” or “stop as soon as you can”. It means finding the lowest effective dose with a prescriber — and changing it carefully.
Never stop suddenly
Stopping antipsychotics abruptly triggers relapse in around 77% of people. Tapered slowly, over more than ten weeks, that risk falls to about 31%. Any change should be planned with your prescriber — never done alone.
How you come off changes everything
Relapse rate by speed of discontinuation (Schizophrenia Bulletin meta-analysis)
The goal is balance, not zero
Risk follows a U-shaped curve: both no treatment and very high doses carry the highest mortality, while moderate, therapeutic dosing carries the lowest. The aim is the lowest dose that keeps you well — reviewed regularly as life changes.
Never stop suddenly — this is the big one
If there is one thing to take from this page: do not stop antipsychotics abruptly. Stopping cold triggers relapse in around 77% of people; tapering very slowly, over more than ten weeks, cuts that to about 31%. Any change should be planned with a prescriber, never done alone.
Schizophrenia Bulletin discontinuation meta-analysis.
Slow, hyperbolic tapering
Modern guidance favours "hyperbolic" tapering — smaller and smaller dose cuts as you go down, because the final milligrams have an outsized effect on the brain. For some people the safe end point is a fraction of a standard dose, reached gently over many months.
Maudsley Deprescribing Guidelines.
Monitoring makes it safer
Most long-term harms are detectable and manageable when watched for. UK guidance recommends a thorough annual physical-health check — weight, blood sugar and lipids, prolactin where relevant, and an ECG for heart rhythm — so problems are caught early and the dose or drug adjusted. Shared, informed decision-making is the whole point.
NICE guidance on physical-health monitoring.
What good care looks like
Here is the hopeful part: almost all of this is fixable with ordinary care. The gap is not inevitable — it is a to-do list.
annual health check
The annual physical health check
The NHS Long Term Plan set out to give everyone with severe mental illness a yearly physical health check — six simple checks that catch the biggest killers early:
The annual physical health check
The single most powerful fix already exists. The NHS Long Term Plan set out to give 390,000 people with severe mental illness a full annual physical health check — six core checks covering weight, blood pressure, blood sugar, cholesterol, smoking and alcohol. Where it happens, problems are caught early and treated. The job now is making sure everyone who is eligible actually gets one.
NHS England Long Term Plan; Core20PLUS5.
Parity of esteem
Physical and mental health are meant to be treated as equally important — that principle is written into NHS policy. In practice it means keeping a register of patients with severe mental illness, monitoring physical health from the moment antipsychotics are started, and joining up GP, mental-health and physical-health teams so nobody falls through the gap.
NICE guidance; NHS England.
Small things save lives
Most of these early deaths are preventable with ordinary care: stop-smoking support, statins, blood-pressure and diabetes management — and, above all, being believed when you say something is wrong. None of it is high-tech. It is simply offering people with severe mental illness the same physical healthcare everyone else takes for granted.
Royal College of Psychiatrists (2024).
Informed, shared, monitored
Good care is a partnership. Whether it is your medication or your heart, blood sugar or lungs, ask for your annual physical health check, take someone with you if it helps, and keep asking until you are heard. If a symptom is being waved away as “just the mental illness”, it is okay to push back — and to ask for it to be written down.
Find support & servicesIf a physical symptom is worrying you, or medication is making you feel distressed, restless or unsafe, please don’t wait — talk to your GP or prescriber as soon as you can. If you need to talk to someone now, call the Samaritans free on 116 123 (24/7), or NHS 111 for urgent medical advice.
“People with schizophrenia are not dying of an untreatable illness — they are dying of treatable ones we keep failing to treat. Close that gap, and you give people their years back.”
The physical-health figures on this page were fact-checked against authoritative UK sources: the Royal College of Psychiatrists’ 2024 analysis (around 87,000 preventable deaths over 2020–22, about two-thirds of all deaths in this group), the gov.uk briefing on physical-health inequalities for people with severe mental illness (the 15–20 year life-expectancy gap, the roughly two-fold disease risks, and the 15–34 age finding), Shefer et al. (2014) on diagnostic overshadowing (77% of staff), and NHS England’s Long Term Plan and Core20PLUS5 programme (the 390,000 annual health-check ambition and the six core checks). The barriers-to-care figures come from the Scottish Mental Illness Stigma Study (2022) and are self-reported experiences of people living with mental illness, presented as lived experience rather than population rates. Some figures from the original report were deliberately left off where they conflicted with newer authoritative data. None of this is medical advice.
Every figure on this page was fact-checked against the original peer-reviewed sources — the FIN20 nationwide cohort (Tiihonen, Taipale et al.), Ray et al. (NEJM, 2009) on sudden cardiac death, Ho et al. (2011) on brain volume, the Maudsley Deprescribing Guidelines and the Schizophrenia Bulletin discontinuation meta-analysis on tapering, and the global meta-analyses of tardive dyskinesia. Where a source disagreed with the published paper the published figure was used: the FIN20 cohort is 62,250 people (not the larger figure quoted in some summaries), and the sudden-cardiac-death ratios are 1.99 and 2.26. Quality-of-life figures (emotional blunting, sexual difficulties and similar) come from self-reported surveys of long-term users and are presented as lived experience, not population rates. None of this is medical advice — never change your medication without your prescriber.
Honesty beats fear.
Share the whole picture — the gap, the medicines, and the care that closes it.