Homelessness among psychiatric inpatients: Prevalence, clinical characteristics, and patterns of inpatient care in a multiyear electronic health record cohort

In the bustling cities of Europe, a growing number of people find themselves without a home, a situation that experts recognize as one of the most severe forms of social exclusion. This lack of shelter is not just a housing crisis; it is a profound health crisis. People living on the streets or in temporary shelters face a heavy burden of illness, carrying a higher risk of chronic physical diseases and mental health struggles than the general population. The relationship between losing a home and losing one's mind is a two-way street: mental illness can make it harder to keep a roof over one's head, while the stress and danger of homelessness can worsen mental health and make recovery nearly impossible. Because of this deep connection, hospitals often become the primary place where these individuals receive care. However, for a long time, doctors and policymakers have struggled to see the full picture of how many people experiencing homelessness are actually inside psychiatric wards, or how their treatment differs from those with stable housing. Without clear data, it is difficult to design care that truly meets their needs.

To solve this puzzle, researchers at Charité–Universitätsmedizin Berlin turned to the hospital's own digital records. Instead of relying on small surveys or manual checks of paper charts, they used a computer algorithm to scan the electronic health records of nearly 5,400 adult patients admitted to the psychiatric ward between 2020 and 2025. The team built a tool to search for specific clues that indicate a patient has no fixed address. This included looking for placeholder numbers in the address field, such as a string of zeros, and scanning the free-text notes written by doctors and nurses for words like "homeless" or "without fixed residence." By combining these structured data points with the narrative details in medical files, they were able to identify a group of patients that other methods might have missed.

The results revealed a startling reality: nearly 29 percent of the psychiatric inpatients in this major Berlin hospital were experiencing homelessness. This is a significantly higher proportion than previous studies in Germany had suggested, likely because the new method of searching through detailed clinical notes caught cases that standard administrative forms overlooked. When the researchers compared this group to patients who had a home, clear patterns emerged. The patients without housing were, on average, younger, more likely to be men, and more often born outside of Germany. Their medical profiles were also distinct; they were far more likely to be diagnosed with substance use disorders or psychotic conditions, while diagnoses like mood disorders were less common in this group compared to their housed counterparts.

The way these patients moved through the healthcare system told an even more urgent story. People experiencing homelessness were much more likely to arrive at the hospital through the emergency department rather than through a planned appointment. Once inside, they faced a higher likelihood of being subjected to coercive measures, such as physical restraints or involuntary medication, which are used when a patient is at immediate risk of harming themselves or others. Furthermore, their time in the hospital was often shorter, yet they were far more likely to return. Within one year of being discharged, nearly 59 percent of the homeless patients were readmitted to the psychiatric ward, compared to less than 46 percent of those with stable housing. Even when the researchers accounted for age, gender, and country of birth, homelessness remained a strong, independent predictor of these intense and often crisis-driven care patterns.

One of the most significant findings was the link between homelessness and the use of coercive measures. While previous research has often tied these restrictive interventions to the severity of a patient's symptoms or the circumstances of their admission, this study showed that simply being homeless was enough to increase the odds of such measures being used. The researchers suggest this might be because the crisis of homelessness often coincides with untreated illness and a lack of community support, pushing patients toward emergency services where restrictive interventions are more common. However, they also note that the data cannot prove that homelessness directly causes the use of restraints; it only shows that the two happen together frequently.

This study does not offer a simple fix, but it provides a crucial map of where the system is failing. The authors argue that housing status should be treated not just as a social detail to be noted on a form, but as a vital clinical indicator of vulnerability. By routinely and reliably identifying patients who are homeless, hospitals could better prepare for their discharge, connecting them with housing support and community care before they are released back onto the streets. The research suggests that without these targeted interventions, the cycle of emergency admissions, short stays, and rapid readmissions will continue. Ultimately, the work highlights that for a large portion of psychiatric patients, the path to recovery is blocked not just by illness, but by the absence of a safe place to call home.