Excess mortality after mental disorder diagnosis within and outside specialist psychiatric care: nationwide cohort study

For decades, a stark statistic has haunted the field of mental health: people diagnosed with mental disorders tend to die much younger than the general population. Estimates often suggest a gap of seven to fifteen years in life expectancy. This reality has driven a urgent public health focus on understanding why these lives are cut short. However, much of what we know about this shortened lifespan comes from studying patients in specialized psychiatric hospitals or treatment centers. These facilities see the most severe cases, the most complex conditions, and the most intense medical needs. It has been a reasonable assumption that what is true for these severe cases applies to everyone with a mental health diagnosis. But as mental health care has expanded, millions more people are receiving diagnoses in their local doctors' offices or community clinics, often for conditions that are common and manageable. The question remains: does the same grim prognosis apply to this much larger, broader group of people, or does the risk of early death depend heavily on the specific context of the diagnosis and where a person receives care?

A new nationwide study from Finland offers a nuanced answer that challenges the idea of a single, uniform risk for everyone with a mental disorder. Researchers followed more than 5.5 million people over a twelve-year period, tracking their health records from the moment they entered the study until 2023. They looked specifically at individuals who did not have a recorded mental disorder at the start of the study and watched to see who received a diagnosis later. By linking these health records with national death registries, the team could calculate exactly how much more likely people with specific diagnoses were to die compared to those without them. They paid close attention to where the diagnosis was made—whether in a primary care doctor's office or a specialist psychiatric hospital—and whether the person also had a diagnosed substance use disorder, such as an alcohol or drug problem.

The study found that the risk of early death is not a fixed number attached to a diagnosis like schizophrenia or depression. Instead, the risk changes dramatically depending on the clinical setting and the time that has passed since the diagnosis. For a very large group of people—specifically those aged 5 to 64 who were diagnosed with depression, anxiety, or stress-related disorders in primary care and did not have a substance use disorder—the risk of early death was high immediately after the diagnosis but dropped quickly. Within two years, the risk for these individuals fell to a level that was statistically indistinguishable from the general population. In fact, nearly half of all people diagnosed with a mental disorder in this age group fell into this category. For them, the "excess mortality" was a temporary spike, likely related to the acute crisis of the illness, rather than a permanent sentence of a shortened life.

This pattern stood in sharp contrast to other groups. People who received their diagnoses in specialist psychiatric care, or those who had a co-occurring substance use disorder, continued to face a significantly higher risk of death years after their diagnosis. The risk remained elevated and persistent for these individuals. The study also highlighted that the type of disorder mattered. While anxiety and stress-related disorders often showed this transient risk pattern in younger people, conditions like schizophrenia and other primary psychotic disorders carried a persistent risk of early death regardless of whether the person was treated in a hospital or a community clinic. Furthermore, for older adults aged 65 and above, the pattern shifted again; even for those with depression treated outside of specialist care, the risk of early death remained elevated and did not fade away with time.

The researchers also looked at the absolute number of years of life lost. For the large group of younger people with common mental disorders treated in primary care without substance use issues, the loss of life expectancy over a ten-year period was measured in days rather than months or years. For men with depression in this group, the loss was about 0.09 years; for women, it was 0.05 years. In contrast, for those with substance use disorders receiving specialist care, the loss was substantial, amounting to more than half a year for men and nearly half a year for women with depression alone. These findings suggest that the widely cited "mortality gap" in mental health is not a single, monolithic truth. It is a complex landscape where the prognosis depends entirely on the specific circumstances of the patient. A diagnosis of depression in a community clinic for a young person without substance issues tells a very different story about their future life expectancy than a diagnosis of the same condition in a specialized hospital for someone with a substance use disorder.

This distinction is vital for how doctors and patients understand the future. The study indicates that for a vast portion of the population diagnosed with common mental health conditions, the fear of a drastically shortened life may be unfounded once the initial period of the illness has passed. The elevated risk is real, but it is often temporary. However, for those with severe conditions, substance use issues, or those requiring specialist intervention, the risk remains a serious, long-term concern that requires sustained attention. By recognizing these differences, healthcare providers can offer more accurate and hopeful prognoses to the millions of people managing mental health conditions, moving away from a one-size-fits-all warning toward a more precise understanding of who is at risk and for how long.