JGJan Griewatz

Jan Griewatz

*This is an algorithmically generated author record based on publicly available information.

Jan Griewatz has a total of 100 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Communication-Based Teaching on Childhood Obesity and the Planetary Health Diet in Medical Education: Proof-of-Concept Study Comparing 4 Information Sources

    by Teresa Festl-Wietek, Carla Schröpel, Friederike Holderried, Bernd Herrmann, Jan Griewatz, Stefan Ehehalt, Florian Junne, Stephan Zipfel, Anne Herrmann-Werner, Rebecca Erschens

    Background: Childhood obesity constitutes a complex medical and psychosocial challenge that requires both nutritional knowledge and sensitive, relationship-oriented doctor-patient communication. The Planetary Health Diet links individual health promotion with environmental sustainability and represents a relevant framework for contemporary medical education. Objective: This proof-of-concept study investigated how different information sources influence medical students’ acquisition, structuring, and application of knowledge on childhood obesity and the Planetary Health Diet within a communication-based teaching setting, including the exploratory use of artificial intelligence–based tools. Methods: A total of 359 second-year medical students participated in a mandatory communication seminar during the 2023‐2024 academic year. Following a precourse knowledge assessment and a brief theoretical introduction, students worked on a standardized counseling scenario addressing childhood obesity. In small groups, students used only 1 assigned information source (ChatGPT, Google Search, scientific papers, or instructional videos) to prepare a counseling approach. Group outcomes were assessed using a predefined scoring system based on a sample solution, complemented by thematic content analysis. Results: All information sources enabled students to acquire relevant knowledge on childhood obesity and the Planetary Health Diet. However, groups differed with regard to the depth, differentiation, and structuring of their responses. The ChatGPT group achieved the highest conformity scores with the sample solution and provided the most additional information, followed by the Google and video groups, while the paper group achieved the lowest scores. Prior to the course, students reported limited knowledge of the Planetary Health Diet and little practical experience in counseling children with obesity and their families. Conclusions: Communication-based teaching formats provide an effective framework for introducing medical students to complex topics such as childhood obesity and sustainability-related nutrition early in their training. Easily accessible digital tools, including artificial intelligence–based systems, may facilitate knowledge acquisition and elaboration; however, their use requires explicit didactic framing, critical source evaluation, and reflection on the complexity of chronic conditions to support responsible and realistic learning outcomes in future physicians.

  • Open Access

    Authors’ Reply: “Evaluating GPT-4’s Cognitive Functions Through the Bloom Taxonomy: Insights and Clarifications”

    by Anne Herrmann-Werner, Teresa Festl-Wietek, Friederike Holderried, Lea Herschbach, Jan Griewatz, Ken Masters, Stephan Zipfel, Moritz Mahling


  • Open Access

    Authors’ Reply: “Evaluating GPT-4’s Cognitive Functions Through the Bloom Taxonomy: Insights and Clarifications” (Preprint)

    by Anne Herrmann-Werner, Teresa Festl-Wietek, Friederike Holderried, Lea Herschbach, Jan Griewatz, Ken Masters, Stephan Zipfel, Moritz Mahling

    Journal of Medical Internet Research - International Scientific Journal for Medical Research, Information and Communication on the Internet #Preprint #PeerReviewMe: Warning: This is a unreviewed preprint. Readers are warned that the document has not been peer-reviewed by expert/patient reviewers or an academic editor, may contain misleading claims, and is likely to undergo changes before final publication, if accepted, or may have been rejected/withdrawn. Readers with interest and expertise are encouraged to sign up as peer-reviewer, if the paper is within an open peer-review period. Please cite this preprint only for review purposes or for grant applications and CVs (if you are the author).

  • Top Cited Papers
    Open Access

    Assessing ChatGPT’s Mastery of Bloom’s Taxonomy Using Psychosomatic Medicine Exam Questions: Mixed-Methods Study

    by Anne Herrmann-Werner, Teresa Festl-Wietek, Friederike Holderried, Lea Herschbach, Jan Griewatz, Ken Masters, Stephan Zipfel, Moritz Mahling

    Background: Large language models such as GPT-4 (Generative Pre-trained Transformer 4) are being increasingly used in medicine and medical education. However, these models are prone to “hallucinations” (ie, outputs that seem convincing while being factually incorrect). It is currently unknown how these errors by large language models relate to the different cognitive levels defined in Bloom’s taxonomy. Objective: This study aims to explore how GPT-4 performs in terms of Bloom’s taxonomy using psychosomatic medicine exam questions. Methods: We used a large data set of psychosomatic medicine multiple-choice questions (N=307) with real-world results derived from medical school exams. GPT-4 answered the multiple-choice questions using 2 distinct prompt versions: detailed and short. The answers were analyzed using a quantitative approach and a qualitative approach. Focusing on incorrectly answered questions, we categorized reasoning errors according to the hierarchical framework of Bloom’s taxonomy. Results: GPT-4’s performance in answering exam questions yielded a high success rate: 93% (284/307) for the detailed prompt and 91% (278/307) for the short prompt. Questions answered correctly by GPT-4 had a statistically significant higher difficulty than questions answered incorrectly (P=.002 for the detailed prompt and P<.001 for the short prompt). Independent of the prompt, GPT-4’s lowest exam performance was 78.9% (15/19), thereby always surpassing the “pass” threshold. Our qualitative analysis of incorrect answers, based on Bloom’s taxonomy, showed that errors were primarily in the “remember” (29/68) and “understand” (23/68) cognitive levels; specific issues arose in recalling details, understanding conceptual relationships, and adhering to standardized guidelines. Conclusions: GPT-4 demonstrated a remarkable success rate when confronted with psychosomatic medicine multiple-choice exam questions, aligning with previous findings. When evaluated through Bloom’s taxonomy, our data revealed that GPT-4 occasionally ignored specific facts (remember), provided illogical reasoning (understand), or failed to apply concepts to a new situation (apply). These errors, which were confidently presented, could be attributed to inherent model biases and the tendency to generate outputs that maximize likelihood.


  • Top Cited Papers
    Open Access

    A Generative Pretrained Transformer (GPT)–Powered Chatbot as a Simulated Patient to Practice History Taking: Prospective, Mixed Methods Study

    by Friederike Holderried, Christian Stegemann–Philipps, Lea Herschbach, Julia-Astrid Moldt, Andrew Nevins, Jan Griewatz, Martin Holderried, Anne Herrmann-Werner, Teresa Festl-Wietek, Moritz Mahling

    Background: Communication is a core competency of medical professionals and of utmost importance for patient safety. Although medical curricula emphasize communication training, traditional formats, such as real or simulated patient interactions, can present psychological stress and are limited in repetition. The recent emergence of large language models (LLMs), such as generative pretrained transformer (GPT), offers an opportunity to overcome these restrictions Objective: The aim of this study was to explore the feasibility of a GPT-driven chatbot to practice history taking, one of the core competencies of communication. Methods: We developed an interactive chatbot interface using GPT-3.5 and a specific prompt including a chatbot-optimized illness script and a behavioral component. Following a mixed methods approach, we invited medical students to voluntarily practice history taking. To determine whether GPT provides suitable answers as a simulated patient, the conversations were recorded and analyzed using quantitative and qualitative approaches. We analyzed the extent to which the questions and answers aligned with the provided script, as well as the medical plausibility of the answers. Finally, the students filled out the Chatbot Usability Questionnaire (CUQ). Results: A total of 28 students practiced with our chatbot (mean age 23.4, SD 2.9 years). We recorded a total of 826 question-answer pairs (QAPs), with a median of 27.5 QAPs per conversation and 94.7% (n=782) pertaining to history taking. When questions were explicitly covered by the script (n=502, 60.3%), the GPT-provided answers were mostly based on explicit script information (n=471, 94.4%). For questions not covered by the script (n=195, 23.4%), the GPT answers used 56.4% (n=110) fictitious information. Regarding plausibility, 842 (97.9%) of 860 QAPs were rated as plausible. Of the 14 (2.1%) implausible answers, GPT provided answers rated as socially desirable, leaving role identity, ignoring script information, illogical reasoning, and calculation error. Despite these results, the CUQ revealed an overall positive user experience (77/100 points). Conclusions: Our data showed that LLMs, such as GPT, can provide a simulated patient experience and yield a good user experience and a majority of plausible answers. Our analysis revealed that GPT-provided answers use either explicit script information or are based on available information, which can be understood as abductive reasoning. Although rare, the GPT-based chatbot provides implausible information in some instances, with the major tendency being socially desirable instead of medically plausible information.


  • Open Access

    A Generative Pretrained Transformer (GPT)–Powered Chatbot as a Simulated Patient to Practice History Taking: Prospective, Mixed Methods Study (Preprint)

    by Friederike Holderried, Christian Stegemann–Philipps, Lea Herschbach, Julia-Astrid Moldt, Andrew Nevins, Jan Griewatz, Martin Holderried, Anne Herrmann-Werner, Teresa Festl-Wietek, Moritz Mahling

    BACKGROUND Communication is a core competency of medical professionals and of utmost importance for patient safety. Although medical curricula emphasize communication training, traditional formats, such as real or simulated patient interactions, can present psychological stress and are limited in repetition. The recent emergence of large language models (LLMs), such as generative pretrained transformer (GPT), offers an opportunity to overcome these restrictions OBJECTIVE The aim of this study was to explore the feasibility of a GPT-driven chatbot to practice history taking, one of the core competencies of communication. METHODS We developed an interactive chatbot interface using GPT-3.5 and a specific prompt including a chatbot-optimized illness script and a behavioral component. Following a mixed methods approach, we invited medical students to voluntarily practice history taking. To determine whether GPT provides suitable answers as a simulated patient, the conversations were recorded and analyzed using quantitative and qualitative approaches. We analyzed the extent to which the questions and answers aligned with the provided script, as well as the medical plausibility of the answers. Finally, the students filled out the Chatbot Usability Questionnaire (CUQ). RESULTS A total of 28 students practiced with our chatbot (mean age 23.4, SD 2.9 years). We recorded a total of 826 question-answer pairs (QAPs), with a median of 27.5 QAPs per conversation and 94.7% (n=782) pertaining to history taking. When questions were explicitly covered by the script (n=502, 60.3%), the GPT-provided answers were mostly based on explicit script information (n=471, 94.4%). For questions not covered by the script (n=195, 23.4%), the GPT answers used 56.4% (n=110) fictitious information. Regarding plausibility, 842 (97.9%) of 860 QAPs were rated as plausible. Of the 14 (2.1%) implausible answers, GPT provided answers rated as socially desirable, leaving role identity, ignoring script information, illogical reasoning, and calculation error. Despite these results, the CUQ revealed an overall positive user experience (77/100 points). CONCLUSIONS Our data showed that LLMs, such as GPT, can provide a simulated patient experience and yield a good user experience and a majority of plausible answers. Our analysis revealed that GPT-provided answers use either explicit script information or are based on available information, which can be understood as abductive reasoning. Although rare, the GPT-based chatbot provides implausible information in some instances, with the major tendency being socially desirable instead of medically plausible information.

  • Open Access

    Assessing ChatGPT’s Mastery of Bloom’s Taxonomy Using Psychosomatic Medicine Exam Questions: Mixed-Methods Study (Preprint)

    by Anne Herrmann-Werner, Teresa Festl-Wietek, Friederike Holderried, Lea Herschbach, Jan Griewatz, Ken Masters, Stephan Zipfel, Moritz Mahling

    BACKGROUND Large language models such as GPT-4 (Generative Pre-trained Transformer 4) are being increasingly used in medicine and medical education. However, these models are prone to “hallucinations” (ie, outputs that seem convincing while being factually incorrect). It is currently unknown how these errors by large language models relate to the different cognitive levels defined in Bloom’s taxonomy. OBJECTIVE This study aims to explore how GPT-4 performs in terms of Bloom’s taxonomy using psychosomatic medicine exam questions. METHODS We used a large data set of psychosomatic medicine multiple-choice questions (N=307) with real-world results derived from medical school exams. GPT-4 answered the multiple-choice questions using 2 distinct prompt versions: detailed and short. The answers were analyzed using a quantitative approach and a qualitative approach. Focusing on incorrectly answered questions, we categorized reasoning errors according to the hierarchical framework of Bloom’s taxonomy. RESULTS GPT-4’s performance in answering exam questions yielded a high success rate: 93% (284/307) for the detailed prompt and 91% (278/307) for the short prompt. Questions answered correctly by GPT-4 had a statistically significant higher difficulty than questions answered incorrectly (P=.002 for the detailed prompt and P<.001 for the short prompt). Independent of the prompt, GPT-4’s lowest exam performance was 78.9% (15/19), thereby always surpassing the “pass” threshold. Our qualitative analysis of incorrect answers, based on Bloom’s taxonomy, showed that errors were primarily in the “remember” (29/68) and “understand” (23/68) cognitive levels; specific issues arose in recalling details, understanding conceptual relationships, and adhering to standardized guidelines. CONCLUSIONS GPT-4 demonstrated a remarkable success rate when confronted with psychosomatic medicine multiple-choice exam questions, aligning with previous findings. When evaluated through Bloom’s taxonomy, our data revealed that GPT-4 occasionally ignored specific facts (remember), provided illogical reasoning (understand), or failed to apply concepts to a new situation (apply). These errors, which were confidently presented, could be attributed to inherent model biases and the tendency to generate outputs that maximize likelihood.

  • Open Access

    How to communicate with patients in written asynchronous online conversations: an intervention study with undergraduate medical students in a cross-over design

    by Teresa Festl-Wietek, Rebecca Erschens, Jan Griewatz, Stephan Zipfel, Anne Herrmann-Werner

    Introduction: The patient-physician encounter is the core element in the treatment of patients and the diagnosis of disease. In these times of digitalization, patient-physician communication is increasingly taking place online: patients embrace new possibilities offered digitally, and physicians are encouraged to adapt accordingly. Since a huge part of online communication is written, this study aims to investigate how medical students communicate with patients online by focusing on their written competencies and whether an intervention might improve their competencies.Methods: This study was performed in an explanatory cross-sectional manner with a cross-over design. Second-year medical students participated. An intervention was developed on how to formulate an appropriate written response to a patient’s request and integrated a longitudinal communication class. The intervention consists of education on general set-up (e.g., greetings), syntax, spelling, content and kind of communication (e.g., appreciative attitude). After meeting a patient in a simulated role play medical students received the patient’s request via a digital platform. The control group had the same simulated role play and the same task but they received the intervention on communication afterwards. Intervention and control group were statistically compared based on a checklist.Results: Twenty-nine medical students took part in the study. The results showed that the medical students had basic competencies in dealing with written communication independent if they received the intervention (CG: M = 3.86 ± 1.23 vs. IG: M = 4.07 ± 1.03; p = 0.625). Similar results were also for the emotional competency ratings (MCG = 3.36 ± 1.08; MIG = 3.67 ± 0.98; p = 0.425).The intervention was able to lead to a more appreciative response toward patient.Discussion: Intervention on basic competencies such as simple language and clear presentation might not be needed as an integral part in medical education. However, medical students should learn how to present empathic and authentic behavior in written online communication.


  • Open Access

    Online student tutorials for effective peer teaching in digital times: a longitudinal quantitative study

    by Teresa Festl-Wietek, Nils Kern, Rebecca Erschens, Jan Griewatz, Stephan Zipfel, Anne Herrmann-Werner

    Peer-assisted learning represents a favoured method of teaching in universities. The COVID-19 pandemic has necessitated transferring medical education to digital formats, and subsequently, the question has arisen of whether online tutorials might be effective. This study, thus, investigated the efficacy of online tutorials in a communication course by assessing the interaction, verbal communication, and nonverbal communication of tutors and students. Second-year medical students were invited to participate in this longitudinal quantitative study. Validated and self-developed questionnaires (e.g., Jefferson Empathy Scale) including 39 questions (rated on a 7- or 5-point Likert scale) were used to assess the different variables including interaction, verbal and nonverbal communication and students’ learning success. Out of 165 medical students, 128 took part in the study. The students as well as tutors reported that they found each other likeable (Mstudents = 4.60±0.71; Mtutors = 4.38±0.53; p > .05). Learning success increased throughout the communication course (Cohen’s d = 0.36–0.74). The nonverbal and verbal communication in the simulated patient (SP) encounter was also rated as high by all three groups (Mnonverbal = 3.90±0.83; Mverbal = 4.88±0.35). Interaction as well as nonverbal and verbal communication occurred in the online format, indicating that online tutorials can be effective. The implementation of SPs increases the efficiency of synchronous online learning as it enhances the simulation of a real patient–physician encounter. Thus, online tutorials are a valuable amendment to medical education.


  • Open Access

    Everybody is able to reflect, or aren't they? Evaluating the development of medical professionalism via a longitudinal portfolio mentoring program from a student perspective

    by Sylvia Schrempf, Lene Herrigel, Justus Pohlmann, Jan Griewatz, Maria Lammerding-Koeppel

    Introduction: Reflective competence is fundamental for responsible medical practice and must be systematically incorporated in medical training. To promote this, a longitudinal portfolio-based mentoring program was made mandatory at the Medical Faculty of the University of Tubingen in 2013. This study examines medical students' attitudes toward professional reflection and toward the program in general to draw conclusions about conditions as well as the needs-based design of the program. Method: In winter semester 2017/18, a retrospective questionnaire survey with free text fields was conducted (total sample: N=1.405; students S 1-9; response 37%; S 1-4 "Pre-clinic": n=231; S 5-9 "Clinic": n=241). Opinion trends of semester groups were identified through seven semi-structured interviews with semester speaker and peer tutors. Results: Differences in understanding and attitudes resulted in three positions: 1=approval, 2=ambivalence, 3=rejection. All three groups included individuals from pre-clinical and clinical settings with varying levels of experience. Prior experience and hidden curriculum influenced the position. Opinion trends confirmed the feedback. Conclusion: Although reflection appears in the National Competencebased Learning Objectives Catalogue for Medicine (NKLM), reflective competence is not regarded as equivalent to other study content. Motivation, commitment on the part of the mentors, and a trusting mentormentee relationship are effective. The flexibility of the portfolio in terms of content and methodology, as well as the curricular integration of the program are also beneficial.