Page 101 - DP-2-1

Design+ AI’s role in medical history taking

following additional sections: Previous illnesses, family considered to ensure a broad perspective on data collection
history, allergies, medication, social environment, (Appendix B).
stimulants, bodily functions, stress level, and possible Small sample sizes in qualitative research allow for
expectations and fears. The questionnaire contains a total a more detailed and nuanced exploration of individual
of around 80 questions. Each answer can be marked with experiences and perspectives. In qualitative research,
27
a yellow exclamation mark by the included AI, thereby sample sizes are often determined by the point of theoretical
generating a corresponding support signal for the attending saturation – the point at which no new information or
physician. The complete anamnesis questionnaire that themes are observed in the data. This can often be achieved
was used as a template for the interviews can be found in with smaller samples. 28
Appendix A.
All interviews took place in Upper Bavaria in the spring
Throughout this process, patient anonymity is
preserved, and stringent measures are taken to ensure the of 2024. The interviews were conducted face-to-face,
with the choice of location adapted to the participant’s
confidentiality of sensitive information.
preferences, to create a relaxed interview atmosphere that
Semi-structured, guideline-based, problem-centered allowed the experts to speak freely and openly about their
expert interviews were conducted between February experiences, opinions, and concerns regarding the use of
20, 2024, and March 17, 2024. Initially, a guideline AI in medical practice. The doctors gave their initial verbal
was developed, which served as the foundation for the consent to participate when first asked.
interviews. This guideline was systematically formulated The study participants represented a wide age range,
by developing key questions that were subsequently from 33 to 59 years, with an average age of 47 years.
transformed into specific interview questions. During the The specialties included internal medicine, cardiology,
creation of the interview guidelines, careful consideration otolaryngology (including specialized ear, nose, and throat
was given to various dimensions of the topic to ensure a
holistic approach. This included questions addressing the surgery and facial plastic surgery), gastroenterology,
conservative orthopedics, orthopedics, rehabilitative and
potential consequences for medical practice, as well as physical medicine, as well as neurology and psychiatry.
an exploration of concerns, opportunities, hopes, and the One physician practiced general medicine. The
general acceptance of AI in medical history taken by the participants held various positions in their careers, such as
medical profession.
specialist, senior physician, and assistant physician. They
Before the interviews, participants were verbally worked in both clinical and outpatient settings, including
informed about the research topic and the interview group practices and shared practices. The sample included
process, and their consent was obtained. Before physicians working in clinics (five physicians) and private
commencing the interviews, the physicians were practices (four physicians). Regarding their experience
provided with a demonstration of Idana’s functions and with AI, five of the participants reported having no prior
results, which served as a foundation for the subsequent knowledge, while four had experience with AI tools. This
discussions. The combination of these two methods – similar separation path was also employed by Chaddad
interviews with physicians and the use of the Idana app et al., indicating this selection is well-experienced and
29
for patient surveys – provides a comprehensive basis for suitable for this study.
elucidating the opportunities and challenges associated
with integrating AI into medical history taking. The study’s diverse sample, covering a wide age range
(33 – 59 years), multiple specialties, various career stages,
2.2. Methodology and different practice settings (clinical and outpatient),
ensures comprehensive insights. In addition, the inclusion
For this qualitative study, a sample of nine physicians of both novice and experienced AI users further mitigates
from various specialties was recruited. The selection selection bias, providing a balanced view of AI’s impact on
was purposive to capture a diversity of experiences and
opinions from doctors in different fields and positions. health care.
The two medical conditions included in our analysis, The transcription of the interview data employed the
asthma, and diabetes, can influence the medical history in semantic-content transcription method by Dresing and
different medical fields, even if the respective specialties do Pehl. This method emphasizes the substantive content
30
not treat the specific disease. The combination of doctors of the conversations by omitting detailed pronunciation
from different medical professions, therefore makes a information, thus facilitating quick access to the
valuable contribution to this study. Factors such as age, conversation content and enhancing readability compared
gender, professional experience, and area of practice were to the gesprächsanalytisches transkriptionssystem

Volume 2 Issue 1 (2025) 4 doi: 10.36922/dp.7675