CancerNCT07526441
Use of and Attitudes Towards Large Language Models for Health Information Among People With Cancer, the Public, and Health-care Professionals
An observational study of Cancer and Shared Decision Making, run by Technical University of Munich at 1 site in Germany.
Trial summary
This cross-sectional survey examined how people with cancer, the general public and health-care professionals use large language models (LLMs) such as ChatGPT for health information, what they expect from them, how comfortable people with cancer are with clinicians using them, and which conditions they set for their… Read more about this study
- Who is it for?
- Ages 18 years and olderHealthy volunteers can take part
- Conditions the study is for
- Cancer, Shared Decision Making, Artificial Intelligence
- How many people
- 6,632 participants
- Where
- 1 siteGermany
- Study dates
- 24 Dec 2024to 30 Mar 2025
- Phase
- Observational
- Are there results?
- None posted in the registry
Key requirements
Who can take part, as listed by the sponsor. Only the study team can confirm whether you are eligible, so take this list to your doctor.
- Ages 18 years and older
- All sexes
- Healthy volunteers accepted
- Adult, Older adult
Study population: Three groups recruited through Prolific: people with a self-reported cancer diagnosis within the past five years (analysed n=2142), adults from the United States and the United Kingdom (n=1957), and health-care professionals with patient contact (n=2714). Public recruitment used US and UK quotas.
Who may be able to join
- Age 18 years or older
- English language proficiency
- Internet access
- Registered on the Prolific research platform
- For the cancer group: self-reported cancer diagnosis within the past five years
- For the professional group: professional practice with patient contact
- For the public group: resident of the United States or the United Kingdom
Who cannot join
- Failure on embedded attention check questions (4 checks)
- Survey completion time less than 5 minutes or greater than 60 minutes
- Straight-line responding pattern detected by consistency validation algorithms
- Failure of cohort verification procedures
Trial locations and contacts
1 site in 1 country.
About this study
Official title: Use of and Attitudes Towards Large Language Models for Health Information Among People With Cancer, the General Public, and Health-care Professionals: A Multinational Cross-sectional Survey
This cross-sectional survey examined how people with cancer, the general public and health-care professionals use large language models (LLMs) such as ChatGPT for health information, what they expect from them, how comfortable people with cancer are with clinicians using them, and which conditions they set for their use. Three English-language questionnaires were deployed through the Prolific research panel between December 24, 2024, and March 30, 2025. The public sample used US and UK recruitment quotas, whereas recruitment of people with cancer and professionals also included other countries. After exclusion of blank, duplicate and repeat records and of main-survey responses from pilot participants, the analysed sample comprised 2142 people with cancer, 1957 public respondents and 2714 professionals (6632 distinct participants). Six principal comparisons of people with cancer with the public on items with identical wording were tested at a Bonferroni-adjusted alpha of 0.0083. All other analyses were exploratory. The study was registered retrospectively after data collection.
Read the detailed description
Shared decision making brings clinical information together with a patient's values and preferences when treatment options are considered. LLM-generated information can reach a cancer consultation from the patient side and from the professional side. This study measured use of LLMs for health information, frequency of general LLM use, expected improvement of health care, comfort with clinician use of LLMs, reliability and privacy concerns, stated conditions for use, and, among professionals, regular LLM use at work, use or recommendation of AI applications with patients and willingness to use LLM-based decision aids.
Content validity of the questionnaires was established by a 44-member expert panel in two rounds of Delphi review. Three pilot surveys (143 people with cancer and 147 professionals on January 18, 2025, and 143 US public respondents on December 24, 2024) used earlier versions of the questionnaires. Main-survey responses from pilot participants were excluded.
Prolific recorded 7845 submissions, of which 694 (8.8%) were rejected at the platform stage for failed attention checks, implausible completion times, straight-line responding or failed cohort verification. The 7151 approved records comprised 2316 from people with cancer, 2000 from the public and 2835 from professionals. Excluded were 73 patient records and one professional record without survey responses, 30 exact duplicate public rows, one public record with a recorded age of 13, nine later submissions of repeated professional participants and 224 responses from pilot participants.
Six principal comparisons of people with cancer with the public (LLM selected as a health-information source, weekly or daily general use, comfort with clinician use, expected improvement, reliability concern, privacy concern) used linear probability models with participant-clustered standard errors and a Bonferroni-adjusted two-sided alpha of 0.0083. These comparisons were designated after data collection and before analysis. A sample of 1824 per group detects a 5 percentage point difference from a baseline of 22.7% with 80% power at this alpha. An exploratory complete-case logistic regression with 11 a priori covariates examined factors associated with health-information LLM use among people with cancer, with multiple imputation and region-stratified and country-effect sensitivity analyses. Use and comfort with clinician use were cross-tabulated descriptively. No clustering analysis was performed. The study was approved by the Ethics Committee of the Technical University of Munich School of Medicine and Health (2024-89-S-SB) and registered retrospectively on April 6, 2026.
- Conditions
- Cancer, Shared Decision Making, Artificial Intelligence
- Keywords
- Large Language Models, Health Information, Oncology, Shared Decision Making, Survey
- Medical subject headings
- Neoplasms
What participants receive
What participants receive. In a randomised study each person is assigned to one group by chance.
| Study group | Role | What they receive |
|---|---|---|
| People With Cancer | Adults aged 18 years or older with a self-reported cancer diagnosis within the past five years, recruited through Prolific from 28 countries (US 1072, UK 619, other countries 451). Analysed n=2142. | |
| General Public | Adults aged 18 years or older from the United States and the United Kingdom, recruited through Prolific with US and UK quotas. Analysed n=1957. | |
| Health-care Professionals | Practitioners aged 18 years or older with patient contact, recruited through Prolific from 37 countries. Analysed n=2714. |
What is being measured
The primary outcome is the main question the study is designed to answer.
Selection of large language models as a health-information source
Proportion of respondents selecting "Large language models (e.g., ChatGPT)" in a checklist of sources used for health information. Denominator: analysed group. Compared between people with cancer and the public.
Measured: At survey completion (single assessment, December 24, 2024, to March 30, 2025)
Weekly or daily general use of large language models
Proportion reporting use of LLMs often (weekly) or very often (daily) on a five-category frequency item. Compared between people with cancer and the public.
Measured: At survey completion (single assessment, December 24, 2024, to March 30, 2025)
Comfort with clinician use of large language models
Proportion agreeing or strongly agreeing that they would be comfortable if their health-care provider used large language models to support their care. Neutral responses classified as not endorsing comfort. Compared between people with cancer and the public.
Measured: At survey completion (single assessment, December 24, 2024, to March 30, 2025)
Expected improvement of health care by large language models
Proportion answering "will significantly improve care" or "will somewhat improve care" to the item on the expected impact of LLMs on health-care delivery. Compared between people with cancer and the public.
Measured: At survey completion (single assessment, December 24, 2024, to March 30, 2025)
Reliability and privacy concerns
Proportions selecting information reliability and data privacy in the checklist of concerns about LLM use. Compared between people with cancer and the public.
Measured: At survey completion (single assessment, December 24, 2024, to March 30, 2025)
Secondary outcomes (3)
Factors associated with health-information LLM use among people with cancer
Adjusted odds ratios from an exploratory complete-case logistic regression with 11 a priori covariates (age, sex, ethnicity, education, income, digital-device comfort, self-rated health, consultations, difficulty managing information, confidence understanding information, region).
Measured: At survey completion (single assessment, December 24, 2024, to March 30, 2025)
Cross-tabulation of LLM use and comfort with clinician use
Descriptive cross-tabulation of selection of LLMs as a health-information source with endorsement of comfort with clinician use, in people with cancer and in the public. No clustering.
Measured: At survey completion (single assessment, December 24, 2024, to March 30, 2025)
Professional use of large language models and AI applications
Proportions of professionals agreeing with regular LLM use in their professional routine, selecting use or recommendation of AI applications with patients, and agreeing that they would use LLM-based decision aids if available. Reported separately, not compared with the other groups.
Measured: At survey completion (single assessment, December 24, 2024, to March 30, 2025)
Timeline
From the start of the study to its completion; the mark is today. Estimated dates are the sponsor’s plan.
- First posted13 Apr 2026
- Study start24 Dec 2024
- Primary completion30 Mar 2025
- Study completion30 Mar 2025
Resources
- Official record on ClinicalTrials.govThe registry entry this page is built from, with its full change history.
- Other trials for CancerStudies of the same condition, open and under way.
- What the trial phases meanHow a treatment moves from first tests in people to approval.
More details
For doctors and researchers: how the study is designed, who runs it, and its registry IDs.
Study design
- Study type
- Observational
- Observational model
- Other
- Time perspective
- Cross-sectional
- Enrolment
- 6,632 participants (actual)
Sponsor and investigators
- Lead sponsor
- Technical University of Munich
- Sponsor type
- University, hospital or other
- Responsible party
- Sponsor
IDs, dates and publications
- ClinicalTrials.gov ID
- NCT07526441
- Sponsor's protocol ID
- LLM-SDM-2025
- Status last verified
- Oct 2026
- First submitted
- 6 Apr 2026
- Last updated
- 15 Oct 2026
- Results posted
- No
- FDA-regulated drug
- No
- FDA-regulated device
- No
- Individual data sharing
- Yes: De-identified participant-level data, the data dictionary, the analysis code, the study protocol and the questionnaires will be deposited in Zenodo and made available without access restrictions on publication of the primary results.
Study at a glance
From the registry record, last updated by the sponsor on 15 Oct 2026. Recruitment status is as the sponsor reported it: confirm with the study team.
- Official title
- Use of and Attitudes Towards Large Language Models for Health Information Among People With Cancer, the General Public, and Health-care Professionals: A Multinational Cross-sectional Survey
- ClinicalTrials.gov ID
- NCT07526441
- Acronym / sponsor ID
- LLM-SDM-2025
- Conditions
- Cancer, Shared Decision Making, Artificial Intelligence
- Keywords
- Large Language Models, Health Information, Oncology, Shared Decision Making, Survey
- Study type
- Observational
- Recruitment status
- Completed
- Enrolment
- 6,632 participants (actual)
- Lead sponsor
- Technical University of MunichUniversity, hospital or other
- Study start
- 24 Dec 2024 (actual)
- Primary completion
- 30 Mar 2025 (actual)
- Study completion
- 30 Mar 2025 (actual)
- First posted
- 13 Apr 2026 (actual)
- Results
- Not posted in the source record
Clinical summary
The research question as the sponsor registered it. This describes what the study plans to test, not what it found.
This cross-sectional survey examined how people with cancer, the general public and health-care professionals use large language models (LLMs) such as ChatGPT for health information, what they expect from them, how comfortable people with cancer are with clinicians using them, and which conditions they set for their use. Three English-language questionnaires were deployed through the Prolific research panel between December 24, 2024, and March 30, 2025. The public sample used US and UK recruitment quotas, whereas recruitment of people with cancer and professionals also included other countries. After exclusion of blank, duplicate and repeat records and of main-survey responses from pilot participants, the analysed sample comprised 2142 people with cancer, 1957 public respondents and 2714 professionals (6632 distinct participants). Six principal comparisons of people with cancer with the public on items with identical wording were tested at a Bonferroni-adjusted alpha of 0.0083. All other analyses were exploratory. The study was registered retrospectively after data collection.
Detailed description
Shared decision making brings clinical information together with a patient's values and preferences when treatment options are considered. LLM-generated information can reach a cancer consultation from the patient side and from the professional side. This study measured use of LLMs for health information, frequency of general LLM use, expected improvement of health care, comfort with clinician use of LLMs, reliability and privacy concerns, stated conditions for use, and, among professionals, regular LLM use at work, use or recommendation of AI applications with patients and willingness to use LLM-based decision aids.
Content validity of the questionnaires was established by a 44-member expert panel in two rounds of Delphi review. Three pilot surveys (143 people with cancer and 147 professionals on January 18, 2025, and 143 US public respondents on December 24, 2024) used earlier versions of the questionnaires. Main-survey responses from pilot participants were excluded.
Prolific recorded 7845 submissions, of which 694 (8.8%) were rejected at the platform stage for failed attention checks, implausible completion times, straight-line responding or failed cohort verification. The 7151 approved records comprised 2316 from people with cancer, 2000 from the public and 2835 from professionals. Excluded were 73 patient records and one professional record without survey responses, 30 exact duplicate public rows, one public record with a recorded age of 13, nine later submissions of repeated professional participants and 224 responses from pilot participants.
Six principal comparisons of people with cancer with the public (LLM selected as a health-information source, weekly or daily general use, comfort with clinician use, expected improvement, reliability concern, privacy concern) used linear probability models with participant-clustered standard errors and a Bonferroni-adjusted two-sided alpha of 0.0083. These comparisons were designated after data collection and before analysis. A sample of 1824 per group detects a 5 percentage point difference from a baseline of 22.7% with 80% power at this alpha. An exploratory complete-case logistic regression with 11 a priori covariates examined factors associated with health-information LLM use among people with cancer, with multiple imputation and region-stratified and country-effect sensitivity analyses. Use and comfort with clinician use were cross-tabulated descriptively. No clustering analysis was performed. The study was approved by the Ethics Committee of the Technical University of Munich School of Medicine and Health (2024-89-S-SB) and registered retrospectively on April 6, 2026.
Study design
- Observational model
- Other
- Time perspective
- Cross-sectional
- Number of arms
- 3
| Arm | Type | Receives |
|---|---|---|
| People With Cancer | – | –Adults aged 18 years or older with a self-reported cancer diagnosis within the past five years, recruited through Prolific from 28 countries (US 1072, UK 619, other countries 451). Analysed n=2142. |
| General Public | – | –Adults aged 18 years or older from the United States and the United Kingdom, recruited through Prolific with US and UK quotas. Analysed n=1957. |
| Health-care Professionals | – | –Practitioners aged 18 years or older with patient contact, recruited through Prolific from 37 countries. Analysed n=2714. |
Who may qualify
- Ages
- 18 years and older
- Sex
- All
- Healthy volunteers
- Accepted
- Age groups
- Adult, Older adult
- Study population
- Three groups recruited through Prolific: people with a self-reported cancer diagnosis within the past five years (analysed n=2142), adults from the United States and the United Kingdom (n=1957), and health-care professionals with patient contact (n=2714). Public recruitment used US and UK quotas.
Key inclusion criteria (7 listed)
- Age 18 years or older
- English language proficiency
- Internet access
- Registered on the Prolific research platform
- For the cancer group: self-reported cancer diagnosis within the past five years
Key exclusion criteria (4 listed)
- Failure on embedded attention check questions (4 checks)
- Survey completion time less than 5 minutes or greater than 60 minutes
- Straight-line responding pattern detected by consistency validation algorithms
- Failure of cohort verification procedures
Eligibility is decided by the study team after screening; this overview does not establish that a patient qualifies. Full inclusion and exclusion criteria.
Primary outcome
Selection of large language models as a health-information source
Time frame: At survey completion (single assessment, December 24, 2024, to March 30, 2025)
Proportion of respondents selecting "Large language models (e.g., ChatGPT)" in a checklist of sources used for health information. Denominator: analysed group. Compared between people with cancer and the public.
Weekly or daily general use of large language models
Time frame: At survey completion (single assessment, December 24, 2024, to March 30, 2025)
Proportion reporting use of LLMs often (weekly) or very often (daily) on a five-category frequency item. Compared between people with cancer and the public.
Comfort with clinician use of large language models
Time frame: At survey completion (single assessment, December 24, 2024, to March 30, 2025)
Proportion agreeing or strongly agreeing that they would be comfortable if their health-care provider used large language models to support their care. Neutral responses classified as not endorsing comfort. Compared between people with cancer and the public.
Expected improvement of health care by large language models
Time frame: At survey completion (single assessment, December 24, 2024, to March 30, 2025)
Proportion answering "will significantly improve care" or "will somewhat improve care" to the item on the expected impact of LLMs on health-care delivery. Compared between people with cancer and the public.
Reliability and privacy concerns
Time frame: At survey completion (single assessment, December 24, 2024, to March 30, 2025)
Proportions selecting information reliability and data privacy in the checklist of concerns about LLM use. Compared between people with cancer and the public.
3 secondary outcomes
Factors associated with health-information LLM use among people with cancer
Time frame: At survey completion (single assessment, December 24, 2024, to March 30, 2025)
Cross-tabulation of LLM use and comfort with clinician use
Time frame: At survey completion (single assessment, December 24, 2024, to March 30, 2025)
Professional use of large language models and AI applications
Time frame: At survey completion (single assessment, December 24, 2024, to March 30, 2025)
Results and safety
The study is completed, but results are not available in the source record. That does not mean the treatment did or did not work.
Research links
- Official record on ClinicalTrials.govLast updated by the sponsor 15 Oct 2026.
- Individual participant dataSharing planned: Yes. De-identified participant-level data, the data dictionary, the analysis code, the study protocol and the questionnaires will be deposited in Zenodo and made available without access restrictions on publication of the primary results.
Related trials for Cancer
Recruiting
MRD-Guided Postoperative Surveillance and Adaptive Treatment for Resectable Lung Cancer
Lung Neoplasms, Neoadjuvant Therapy
Active, not recruitingPhase 2
Tamoxifen Citrate or Z-Endoxifen Hydrochloride in Treating Patients With Locally Advanced or Metastatic, Estrogen Receptor-Positive, HER2-Negative Breast Cancer
Recurrent Breast Carcinoma, Stage III Breast Cancer AJCC v7
Not yet recruitingPhase 2
A Phase II Clinical Study Evaluating the Efficacy and Safety of SXRN Injection in Patients With Cancer Cachexia
Cachexia-Anorexia Syndrome, Cachexia; Cancer
RecruitingPhase 2
Adebrelimab ±Induction Chemotherapy Followed by Concurrent Chemoradiotherapy for Unresectable Locally Advanced ESCC
Esophageal Squamous Cell Carcinoma, Immunotherapy
Not yet recruitingPhase 1/2
A Study of Sonrotoclax (BGB-11417) in Children With Relapsed or Refractory Acute Myeloid Leukemia and B-cell Acute Lymphoblastic Leukemia
Relapsed Acute Myeloid Leukemia, Refractory Acute Myeloid Leukemia
Referral and contact
AzTrials does not enrol participants or take referrals. Contact the study team or a site listed below; they decide who can take part. This study is not recruiting at the moment.
The record lists no contact. Use the official record to reach the sponsor.
1 site in 1 country: see every site and its status.
Questions about this trial
What is this clinical trial studying?
An observational study of Cancer and Shared Decision Making, run by Technical University of Munich at 1 site in Germany.
The full registered description is in the About this study section and the official record.
Who may be eligible to take part?
The sponsor lists 7 inclusion and 4 exclusion criteria. Eligibility is decided by the study team after screening, so read the full criteria and contact a listed site.
Is this clinical trial currently recruiting?
The record says: finished, as of the sponsor's last update on 15 Oct 2026. Status can change between updates, so confirm with the study contact before making plans.
Where is the study taking place?
The record lists 1 site in Germany. Whether a site is taking part right now should be confirmed with that site. See every site.
What are the possible benefits and risks?
They depend on the study and the treatment, and no study can promise a benefit. The study team explains the known risks during informed consent, before anyone joins; the record itself does not list them.
Are results available for this study?
The study is completed, but results are not available in the source record. That does not mean the treatment did or did not work.
How can a doctor refer a patient or contact the research team?
The record lists no contact. The official record names the sponsor, who can be asked directly. AzTrials does not enrol participants or take referrals.
What does the primary outcome measure?
Selection of large language models as a health-information source (time frame: At survey completion (single assessment, December 24, 2024, to March 30, 2025)). The study lists 5 primary outcomes in all. It is what the study was designed to measure, not a finding.
Where can I read the original study information?
Does a listing mean the treatment is approved or proven to work?
No. A listing means the study is registered. It does not show that the treatment is safe, works, or is approved by a regulator.
How often is this information updated?
AzTrials copies changed records from ClinicalTrials.gov every night and reloads the whole registry every week. The sponsor last updated this record on 15 Oct 2026.