Patients’ Handling of Sensitive Health Data During Everyday Use of Large Language Models for Medical Questions

Patients’ Handling of Sensitive Health Data During Everyday Use of Large Language Models for Medical Questions - TradeOff

Status

Active, not recruiting

Phases

Unknown

Study type

Observational

Source

DRKS

Registry ID

DRKS00040786

Enrollment

500

Registered

2026-06-26

Start date

2026-07-08

Completion date

Unknown

Last updated

2026-08-03

For informational purposes only — not medical advice. Sourced from public registries and may not reflect the latest updates. Terms

Conditions

Use and risks of LLM in the medical field

Interventions

Group 1: Observational group: Participants in an anonymous web-based cross-sectional survey on the use of AI chatbots or large language models for medical or health-related questions and on the handli

Eligibility

Sex/Gender

All

Age

18 Years to No maximum

Inclusion criteria

Inclusion criteria: - Patients with a current or past medical condition - Ability to give independent informed consent - Sufficient language skills to communicate and understand the study information - Willingness to take part in an anonymous survey

Exclusion criteria

Exclusion criteria: - age < 18 Jahre - Inability to give informed consent - Inadequate understanding of the study information - Acute emergencies or time-critical examinations - Acute mental or cognitive impairment that precludes participation

Design outcomes

Primary

Measure	Time frame
Descriptively assessed extent of voluntary input of sensitive personal or medical health information when using large language models or AI chatbots for medical questions in everyday life. The primary endpoint is assessed based on self-reported information on whether and which types of sensitive health data were entered when using large language models or AI chatbots, such as symptoms, diagnoses, medication, laboratory values, examination results, contents from medical letters or reports, personal or intimate health information, physical baseline data, documents, or images. In addition, the level of detail and the subjectively perceived sensitivity of the entered information are assessed descriptively.	—

Measure

Time frame

Descriptively assessed extent of voluntary input of sensitive personal or medical health information when using large language models or AI chatbots for medical questions in everyday life. The primary endpoint is assessed based on self-reported information on whether and which types of sensitive health data were entered when using large language models or AI chatbots, such as symptoms, diagnoses, medication, laboratory values, examination results, contents from medical letters or reports, personal or intimate health information, physical baseline data, documents, or images. In addition, the level of detail and the subjectively perceived sensitivity of the entered information are assessed descriptively.

—

Secondary

Measure	Time frame
Secondary endpoints include the descriptive and exploratory assessment of the current status and contextual factors of using large language models or AI chatbots for medical or health-related questions in everyday life. These include in particular: - prevalence and frequency of AI chatbot use overall and for medical or health-related questions - AI chatbots or large language models used use contexts and medical or health-related topics - reasons for not using AI chatbots for medical or health-related questions - reasons for not entering personal or health-related information - trust in medical answers provided by AI chatbots - frequency of verifying medical statements from AI chatbots using other sources - perception and assessment of data protection risks - subjective level of information about what happens to data entered into AI chatbots - willingness to enter personal or sensitive health data into AI chatbots in the future if more accurate or individualized information is expected - importance of personal contact with medical professionals in the context of increasing digitalization - exploratory associations between use behavior, data-sharing behavior, expectations of accuracy, perception of data protection risks, and sociodemographic or health-related characteristics	—

Measure

Time frame

Secondary endpoints include the descriptive and exploratory assessment of the current status and contextual factors of using large language models or AI chatbots for medical or health-related questions in everyday life. These include in particular: - prevalence and frequency of AI chatbot use overall and for medical or health-related questions - AI chatbots or large language models used use contexts and medical or health-related topics - reasons for not using AI chatbots for medical or health-related questions - reasons for not entering personal or health-related information - trust in medical answers provided by AI chatbots - frequency of verifying medical statements from AI chatbots using other sources - perception and assessment of data protection risks - subjective level of information about what happens to data entered into AI chatbots - willingness to enter personal or sensitive health data into AI chatbots in the future if more accurate or individualized information is expected - importance of personal contact with medical professionals in the context of increasing digitalization - exploratory associations between use behavior, data-sharing behavior, expectations of accuracy, perception of data protection risks, and sociodemographic or health-related characteristics

—

Countries

Germany

Contacts

Public ContactPhilipp Zehnder

Klinik und Poliklinik für Unfallchirurgie TUM Universitätsklinikum rechts der Isar

philipp.zehnder@mri.tum.de089 / 4140-8361

Outcome results

None listed

Source: DRKS (via WHO ICTRP) · Data processed: Aug 10, 2026