Severe mental illness in Europe: 2.6 times the death rate, mostly physical disease

That people with severe mental illness die younger than everyone else is one of the best-replicated findings in psychiatry. What has been much less clear is which physical diseases account for the gap, and whether the answer depends on how you ask the question. A study in The Lancet Public Health, built from national health registers in five European countries, shows that it does [s1].

The design

The researchers used national health registers to identify people diagnosed with severe mental illness — defined as schizophrenia spectrum disorders, bipolar disorder or major depressive disorder — between the ages of 15 and 65, in Denmark, Finland, France, Poland and Sweden, over the period 2004 to 2023 [s1].

Between 1 January 2004 and 31 December 2023, the cohort comprised 4,861,795 people with severe mental illness, among whom there were 561,903 deaths from any cause [s1].

Excess mortality was measured before age 75 against country-specific general population mortality, with causes of death classified by ICD-10 [s1]. Estimates were pooled across countries by random effects meta-analysis, and reported three ways: sex- and age-standardised mortality ratios for relative inequity, standardised excess death rates per 10,000 person-years for absolute burden, and potential years of life lost before 75 for the severity of premature death [s1].

Using three metrics rather than one is the methodological point of the paper, and it is what produces the interesting result.

The headline gap

All-cause mortality was 2.6-fold higher among people with severe mental illness than in the general population, with a pooled standardised mortality ratio of 2.64 and a 95% confidence interval of 2.25 to 3.11 [s1].

Nothing about that is new. What follows is.

Where the deaths are, versus where the inequality is

By absolute excess — the number of additional deaths per unit of population time — cardiovascular disease dominated across all three diagnostic groups. The standardised excess death rate was 22.08 per 10,000 person-years for schizophrenia spectrum disorders (95% CI 8.77 to 35.39), 8.36 for bipolar disorder (2.96 to 13.75), and 9.82 for major depressive disorder (3.18 to 16.45) [s1].

By relative excess — how many times higher the rate is than in the general population — the ranking changes. Respiratory disease came first: a standardised mortality ratio of 6.49 for schizophrenia spectrum disorders (5.52 to 7.64), 2.72 for bipolar disorder (2.04 to 3.63), and 3.48 for major depressive disorder (2.63 to 4.62) [s1].

Endocrine and metabolic disease followed, with ratios of 5.09 for schizophrenia spectrum disorders (4.29 to 6.04), 2.60 for bipolar disorder (2.26 to 2.98), and 3.07 for major depressive disorder (1.84 to 5.12) [s1]. Then gastrointestinal disease, at 3.47 for schizophrenia spectrum disorders (2.58 to 4.67), 2.34 for bipolar disorder (2.00 to 2.74), and 3.48 for major depressive disorder (2.63 to 4.62) [s1].

So cardiovascular disease kills the most people, but respiratory disease is where the gap between people with severe mental illness and everyone else is widest — more than sixfold in schizophrenia spectrum disorders [s1].

This is the study's actual argument. A health system prioritising by death counts invests in cardiology. A health system prioritising by inequality invests in respiratory care, smoking cessation, and metabolic monitoring. The authors' position is that considering both dimensions informs priorities that are not apparent from either measure alone, or from focusing on cause-specific numbers of deaths [s1].

What it does not show

This is registry data, and registry data describes association. The study establishes that people diagnosed with severe mental illness in these five countries died of these causes at these rates; it does not identify why. Smoking prevalence, antipsychotic metabolic effects, physical inactivity, poverty, diagnostic overshadowing in emergency care, and reduced access to routine screening are all plausible contributors, and this design cannot apportion them.

Diagnosis in registers also means diagnosis by contact with services. People with severe mental illness who never reached a specialist system are not in the cohort, and their outcomes are unknown. Countries differ in how readily major depressive disorder is coded in national registers, which is one reason the pooled confidence intervals are as wide as they are — the interval on the all-cause ratio, 2.25 to 3.11, reflects genuine heterogeneity between the five countries [s1].

The five countries are also not Europe. Denmark, Finland and Sweden have unusually complete registers and comparatively well-resourced health systems; France and Poland differ from them and from each other. Results should not be read as a continental average.

The study was funded by the 2024 European Partnership on Transforming Health and Care Systems [s1].

What follows

The authors conclude that reducing premature mortality will require an integrated public health approach combining universal strategies with targeted interventions, addressing both high-burden causes of death and those with the greatest relative inequalities [s1].

That is a conclusion about system design rather than about clinical practice, and it is the honest one for this evidence. The specific thing to watch is whether European mental health services begin reporting physical-health outcomes — respiratory function, metabolic screening completion, cardiovascular risk assessment — as core performance measures rather than as an adjunct. The mortality gap has been documented for decades; measuring the physical care that would close it is the step that has not consistently happened.

Sources

• [s1] Cause-specific mortality due to physical illness in severe mental disorders in Europe: a population-based multi-country cohort study — The Lancet Public Health, published online 21 July 2026. https://doi.org/10.1016/S2468-2667(26)00116-7