Please read the following information carefully before deciding to participate in this research study. You can keep/download this letter in case you might wish to follow-up about the research.

What is the research study about?

This research studies the effect of an artificial intelligence (AI) assistant on the diagnostic reasoning of doctors, we will present doctors & trainee doctors with hypothetical but realistic patient conditions and ask them to come to a diagnosis. Half of these participants will receive an AI chatbot assistant, half will not. We will assess the impact of the AI through the diagnostic reasoning of participants.

Who can be contacted about this research study?

This research is conducted as part of the research programme at the School of Business and Economics, Maastricht University. The research team consists of:

• Prof. Dr. Mark Levels, Professor, Maastricht University, m.levels@maastrichtuniversity.nl (primary contact)
• Dr. Marie-Christine Fregin, Project Leader, Maastricht University
• Nicholas Rounding, PhD Candidate, Maastricht University

How will the research study be conducted? 

You will be asked to participate in an experiment where you will be asked to come to a diagnosis from a hypothetical patient story and to take part in a survey. We may also ask you to take part in an interview, but another informed consent document will be provided for this.

Experiment

• We will conduct the experiment in a controlled laboratory setting, you will be asked to complete up to 4 patient vignettes. These are hypothetical patient stories with the information required to come to a diagnosis. The experiment will be executed via Qualtrics.
• The AI assistant will be randomly assigned to half of the participants. If you are assigned the AI assistant, you will also have access to an AI chatbot. This chatbot will be executed using an interface designed by researchers at Maastricht University and will use GPT-4, provided by OpenAI. No personal information will be sent to OpenAI. The chat logs will be linked to the Qualtrics data via a personal identifier.

Survey

• After the completion of the experiment, you will be provided with a survey to complete. This survey will be executed via Qualtrics. 

What are the conditions of participating? Are there potential risks?

You are taking part on a voluntary basis and may stop whenever you want without negative consequences. You do not need to answer any questions or provide any information that you do not want to provide. There are no physical, legal, or economic risks involved with participation in this research. 

How will you be compensated for your participation?         

There is no payment or compensation for taking part in this research.

What personal data do we process as part of the research and why?

Personal information for research purposes: 

• Participant ID: A randomly generated participant ID will be used to identify the survey responses, this will also be linked to any chat logs generated via the chat interface.
• Demographic characteristics: gender; age; and working experience will be collected
• Chat logs: For participants that receive AI assistance we will also collect and analyse the chat logs. This data may be retained by OpenAI and processed in line with their data regulations. More information may be found here: https://openai.com/policies/privacy-policy

We collect this information to investigate the research objectives. Any reporting of this data in scientific publications cannot be traced back to you as an individual. We will achieve this by reporting on no demographic groups with lower than 5 observations. 

Storage and data retention 

The personal information you provide will be pseudonymized. All research data is collected via Maastricht University IT infrastructure. Data in the Netherlands will be stored securely in Maastricht University’s IT infrastructure. Data. Data sent to OpenAI via the chatbot interface may be retained by OpenAI. No data submitted via Qualtrics will be sent to OpenAI. At Maastricht University, all research data and consent documentation, collected during the project will be retained for at least 10 years in agreement with Maastricht University guidelines. In accordance with Open Science practices, the researchers at Maastricht University will publish the gathered, pseudonymized data on DataverseNL for re-use. 

The results of this experiment will only be used for scientific analysis, they will not be used to contribute to any grades for any qualification provided by Maastricht University.

Data subject rights

If you have any general enquiries about your rights as a data subject, you can learn about this by going to https://www.maastrichtuniversity.nl/about-um/um-general-privacy-statement. If you want to exercise your rights as a data subjects, you contact the research team, or you can do so by going to https://www.maastrichtuniversity.nl/about-um/um-general-privacy-statement/your-rights.