Diagnosed Mental Illnesses

According to health insurance billing data, 40.4% of the population had at least one diagnosed and documented mental disorder in 2023, compared to 35.0% in 2012.
This marks a relative increase of 14%, with men seeing a significantly stronger rise (+7.9 percentage points / +22.7%) compared to women (+4.6 percentage points / +11.8%).

Sick Leave Due to Mental Disorders

Data from Techniker Krankenkasse (TK) show that from 2006 to 2024, sick leave due to mental health conditions increased by 159%.
In 2024, mental disorders were the second most common cause of sick leave, averaging 3.74 sick days per employed person, up from 3.59 days in 2023.
For comparison, in 2006, mental illnesses caused only about 1.4 sick days per person.
The BKK umbrella association also reported a 48% increase in days of sick leave due to mental health diagnoses between 2011 and 2021.

Increase in Demand for Treatment

This rise does not necessarily reflect a true increase in the actual prevalence of mental illnesses.
Rather, it reflects:

• Increased utilization of mental health services
• Destigmatization
• Normalization of seeking help
• Greater awareness among physicians
• Improved public education on mental health

Shortage of Mental Health Services

The growing demand is not being met by an expanding care system.
This leads to significant shortages, particularly in outpatient care.
According to the BARMER Doctor’s Report 2020, between 2009 and 2018, demand for psychotherapy rose by 41%.
In Germany, the average waiting time for a psychotherapy spot (from initial consultation to therapy start) is around 4.7 months (142 days).

• Initial consultation: ~6–7 weeks
• Between consultation and therapy start: ~19.6 weeks (4.5 months)

Shift in Types of Mental Health Crises

There has been a fundamental shift in the types of crises requiring hospitalization:

• Substance-related disorders are declining
• Affective (mood) and anxiety disorders are rising sharply

For example:

• Hospital treatments for depression (F32/F33) rose by 6% from 2022 to 2023, totaling 261,200 cases.
• Within this group, recurrent depressive disorders (F33) reached a new high of 160,500 cases in 2023.
• Over 20 years, this represents an increase of +76.8%.
• Conversely, hospital treatments for alcohol abuse declined by 5% over ten years until 2023.

Alarming Trends in Adolescents

A special analysis by DAK-Gesundheit (2019–2022) found a 35% increase in hospitalizations due to anxiety disorders among 15–17-year-old girls.
Data from IKK classic showed a 37% rise in diagnosed anxiety disorders from 2013 to 2022.

In 2023, 12.4 million such mental health cases were recorded, up from:

• 8 million in 2021
• 4 million in 2020 (pandemic year)

This massive increase highlights the growing strain on psychiatric care systems.

Mental Health in Emergency and Rescue Services

• A nationwide survey of emergency room doctors found that 15% of patients were mentally conspicuous, and 9%were in need of psychiatric diagnosis/treatment
• A study analyzing emergency services found that 4% of calls involved patients with psychiatric or psychosocial symptoms
• At the Charité Berlin emergency room, 9% of cases were psychosocial emergencies

Suicide as a Public Health Indicator

Completed suicide is the most severe consequence of untreated psychiatric crises.
It is estimated that 50–90% of all suicides are due to mental illnesses.

• Between 2014 and 2020, the suicide rate per 100,000 people declined from 6 to 11.0
• However, this positive trend was broken in 2022, with a 10% increase in suicides compared to the previous year
• In 2022, there were 10,100 suicides, the first time since 2015 that the number exceeded 10,000
• From 2022 to 2023, there was a further increase of 1.8%

Artificial Intelligence (AI) in Psychiatry: Between Data and Trust

The mental health field is changing. New technologies such as artificial intelligence (AI) are playing an increasingly important role in diagnosis, therapy, and care. This is particularly true in psychiatry, where complex datasets, including images, voice, behavior, and even social media usage, can be analyzed to improve understanding and treatment of mental illnesses. At the same time, there are concerns: What does this mean for the relationship between doctors and patients? Who is responsible for decisions made with or by AI? And how do we ensure that such systems are fair and non-discriminatory?

AI Systems in Mental Health

AI is already being used in psychiatry in various forms:

• Diagnosis Support: AI can recognize patterns in data that are not visible to the human eye. For example, speech analysis may provide clues to depression or psychosis. MRI data can be used to identify changes in brain structure associated with certain disorders.
• Therapy and Monitoring: AI-based applications help patients track their emotional state, offer low-threshold support, and enable doctors to monitor their condition remotely. In some cases, chatbots are used to accompany therapies or provide information.
• Predictive Systems: AI is also being researched for its ability to predict relapse risks or the onset of mental illnesses. This involves continuous data collection from wearables or smartphones.

Challenges and Ethical Questions

The use of AI in psychiatry raises complex ethical and social questions:

• Transparency: AI decisions are often opaque. Patients and professionals need to understand how results are generated.
• Bias: Algorithms can reinforce social inequalities if trained on unbalanced datasets.
• Trust and Relationship: Psychiatry thrives on empathy and the therapeutic relationship. The integration of AI must not undermine this trust.
• Privacy and Data Protection: Mental health data is particularly sensitive. Who has access? How is it anonymized? How is consent obtained?

Research Projects and Networks in Germany

The BMBF is funding multiple projects to explore the opportunities and risks of AI in psychiatry. These include interdisciplinary collaborations between data scientists, clinicians, and ethicists. Some projects include:

• KIP-SDM (AI in Psychiatric Shared Decision Making)
  Development of AI-based decision aids for use in clinical everyday life, with emphasis on patient participation.
• MindAhead
  Use of machine learning to detect early signs of psychiatric disorders based on speech and behavior.
• PsyKI
  Examines the legal and ethical implications of AI in psychiatry and develops guidelines for responsible use.
• BIPoC-AI
  Investigates how bias in AI systems can affect people with migration backgrounds or minorities.

Public Involvement and Transparency

For AI to be successfully used in psychiatry, the public must be informed and involved. The BMBF promotes exchange formats such as citizen dialogues, public discussion events, and interactive platforms that explain AI in a comprehensible way and allow critical discourse.

Outlook: AI as a Complement, Not a Replacement

AI has the potential to improve mental healthcare and make it more individualized. However, it must be clear: AI systems are tools that support but do not replace the judgment of professionals. Human empathy, ethical reflection, and a respectful relationship remain at the core of psychiatry.

Artificial Intelligence in Psychiatry: Data and Scientific Evidence

Growing Market and Increasing Demand: The global market for chatbots in mental health and therapy is expected to grow significantly—from USD 1.3 billion in 2023 to approximately USD 2.2 billion in 2033—at a compound annual growth rate (CAGR) of 5.6%. This growth is driven by increasing demand for mental health services, intensified by events such as the COVID-19 pandemic, which led to a significant rise in depressive symptoms and anxiety worldwide.

Shortage of Mental Health Professionals: There is a severe global shortage of mental health professionals—only about 13 professionals per 100,000 people. In the U.S., over 169 million people live in areas with limited access to such professionals, and the shortage is even more critical in developing countries such as India. This gap underscores the need for new solutions, and AI emerges as a potential tool to support and enhance mental health services.

Deployed AI Technologies: AI in mental health includes various technologies:

• Machine Learning (ML) and Deep Learning (DL): Dominating the technology space with a 58.7% market share in 2023, these are crucial for interpreting external data, learning from it, and adapting to achieve goals.
• Natural Language Processing (NLP): Enables AI to understand human language through speech recognition, text analysis, and conversation.
• Conversational Agents (CAs): Often referred to as chatbots, they simulate human conversations using AI techniques. There are different types:
  • Chatbots: Software programs that simulate conversation via text or voice.
  • Embodied CAs (ECAs): Digital characters simulating verbal and non-verbal human communication.
  • CAs in Virtual Reality (VR): Provide controlled exposure in virtual environments for therapeutic purposes.
  • Avatars: Digital representations used in therapy to interact with patients, e.g., in treating auditory hallucinations.

Diagnosis:

• AI tools are accurate in detecting, classifying, and predicting the risk of mental illnesses, as well as in forecasting treatment success and monitoring disease progression.
• Accuracy rates in diagnostic studies range from 51% to 97.54%.
• Frequently used algorithms include Support Vector Machine (SVM) (e.g., 95% accuracy for anxiety disorders, 95.8% for depression) and Random Forest.
• Predictors often include demographic information, socioeconomic data, clinical history, physiological data, psychometric data, medical imaging data (e.g., MRI, EEG), and semantic content.
• AI can detect early warning signs of mental crises with high accuracy (89.3%)—on average 7.2 days before human professionals.

AI-assisted interventions, particularly chatbots, have shown statistically significant short-term effects in improving the following symptoms (Hedges’ g):

• Depressive symptoms (g = 0.29)
• General anxiety symptoms (g = 0.29)
• Specific anxiety symptoms (g = 0.47)
• Quality of life/well-being (g = 0.27)
• General psychological distress (g = 0.33)
• Stress (g = 0.24)
• Symptoms of mental disorders (g = 0.36)
• Psychosomatic illness symptoms (g = 0.62)
• Negative affect (g = 0.28) – compared to control groups.

Personalization and empathic responses are crucial factors for AI effectiveness. AI offers immediate 24/7 support, which is critical for crisis management. Users can interact with chatbots conveniently from home via smartphones or web-based applications.