Introduction: Large language models, in particular ChatGPT, have showcased remarkable language processing capabilities. Given the substantial workload of university medical staff, this study aims to assess the quality of multiple-choice questions (MCQs) produced by ChatGPT for use in graduate medical examinations, compared to questions written by university professoriate staffs based on standard medical textbooks. Methods: 50 MCQs were generated by ChatGPT with reference to two standard undergraduate medical textbooks (Harrison’s, and Bailey & Love’s). Another 50 MCQs were drafted by two university professoriate staff using the same medical textbooks. All 100 MCQ were individually numbered, randomized and sent to five independent international assessors for MCQ quality assessment using a standardized assessment score on five assessment domains, namely, appropriateness of the question, clarity and specificity, relevance, discriminative power of alternatives, and suitability for medical graduate examination. Results: The total time required for ChatGPT to create the 50 questions was 20 minutes 25 seconds, while it took two human examiners a total of 211 minutes 33 seconds to draft the 50 questions. When a comparison of the mean score was made between the questions constructed by A.I. with those drafted by humans, only in the relevance domain that the A.I. was inferior to humans (A.I.: 7.56 +/- 0.94 vs human: 7.88 +/- 0.52; p = 0.04). There was no significant difference in question quality between questions drafted by A.I. versus humans, in the total assessment score as well as in other domains. Questions generated by A.I. yielded a wider range of scores, while those created by humans were consistent and within a narrower range. Conclusion: ChatGPT has the potential to generate comparable-quality MCQs for medical graduate examinations within a significantly shorter time.
Aim: The global pandemic of COVID-19 has led to extensive practice of online learning. Our main objective is to compare different online synchronous interactive learning activities to evaluate students’ perceptions. Moreover, we also aim to identify factors influencing their perceptions in these classes. Methods: A cross-sectional, questionnaire-based study focusing on clinical year medical students’ perceptions and feedback was conducted between February 2021 –June 2021 at the University of Hong Kong. Online learning activities were divided into bedside teaching, practical skill session, problem-based learning (PBL) or tutorial, and lecture. A questionnaire based on the Dundee Ready Education Environment Measure (DREEM) was distributed to 716 clinical year students to document their perceptions. Results: One hundred responses were received with a response rate of 15.4% (110/716, including 96 from bedside teaching, 67 from practical skill session, 104 from PBL/tutorial, and 101 from lecture). For the mean score of the DREEM-extracted questionnaire, online PBL/tutorial scored the highest (2.72 ± 0.54), while bedside scored the lowest (2.38 ± 0.68, p = 0.001). Meanwhile, there was no significant difference when we compared different school years (p = 0.39), age (p = 0.37), gender (p = 1.00), year of internet experience (<17 vs ≥17 years p = 0.59), or prior online class experience (p = 0.62). When asked about students’ preference for online vs face-to-face classes. Students showed higher preferences for online PBL/tutorial (2.06 ± 0.75) and lectures (2.27 ± 0.81). Distraction remains a significant problem across all four learning activities. A multivariate analysis was performed regarding students’ reported behavior in comparison with their perception through the DREEM-extracted questionnaire. The results showed that good audio and video quality had a significant and positive correlation with their perception of online bedside teaching, practical skill sessions, and PBL/tutorial. It also showed that the use of the video camera correlated with an increase in perception scores for lectures. Conclusion: The present analysis has demonstrated that students’ perception of different online synchronous interactive learning activities varies. Further investigations are required on minimizing distraction during online classes.
Introduction This is a prospective study on the quality of multiple-choice questions (MCQs) generated by the language model ChatGPT for the use in medical graduate examination. Methods 50 MCQs were generated by ChatGPT with reference to two standard undergraduate medical textbooks (Harrison’s, and Bailey & Love’s). Another 50 MCQs were drafted by two university professoriate staffs using the same medical textbooks. All 100 MCQ were individually numbered, randomized and sent to five independent international assessors for MCQ quality assessment using a standardized assessment score on five assessment domains; namely, appropriateness of the question, clarity and specificity, relevance, discriminative power of alternatives, and suitability for medical graduate examination. Results The total time required for ChatGPT to create the 50 questions was 20 minutes 25 seconds while it took two human examiners a total of 211 minutes 33 seconds for drafting the 50 questions. When a comparison of the mean score was made between the questions constructed by AI with those drafted by human, only in the relevance domain that the AI was inferior to human (AI: 7.56 +/- 0.94 vs human: 7.88 +/- 0.52; p = 0.04). There was no significant difference in question quality between questions drafted by AI versus human, in the total assessment score as well as in other domains. Questions generated by AI yielded a wider range of scores while those created by human were consistent and within a narrower range. Conclusion ChatGPT has the potential to generate comparable-quality MCQs for medical graduate examination within a significantly shorter time.
The incidence of pregnancy-associated breast cancer (PABC) is increasing. Its tumor characteristics and overall survival compared with those in nonpregnant patients remain controversial. While there have been suggestions that PABC patients have a 40 % increase in the risk of death compared to non-pregnant patients, other studies suggested similar disease outcomes. This study aims to review our local experience with PABC. Twenty-eight patients diagnosed with PABC and twenty-eight patients diagnosed at premenopausal age randomly selected by a computer-generated system during the same period were recruited. Background characteristics, tumor features, and survival were compared. Among the twenty-eight pregnant patients, seventeen were diagnosed during pregnancy, and eleven were diagnosed in the postpartum period. Compared to the non-pregnant breast cancer patients, they presented with less progesterone receptor-positive tumor (35.7 % vs. 64.2 %, p = 0.03). Although there was no statistically significant difference in tumor size (p = 0.44) and nodal status (p = 0.16), the tumor tended to be larger in size (2.94 +/- 1.82 vs 2.40 +/- 1.69 cm) and with more nodal involvement (35.7 % vs 25.0 %). There was also a trend of delayed presentation to medical attention, with a mean duration of 13.1 weeks in the PABC group and 8.6 weeks in the control group. However, the overall survival did not differ (p = 0.63). PABC is increasing in incidence. They tend to have more aggressive features, but overall survival remains similar. A multidisciplinary approach is beneficial for providing the most appropriate care.