Are We Ready to Integrate Artificial Intelligence Literacy into Medical School Curriculum: Students and Faculty Survey.

BACKGROUND: The effects of Artificial Intelligence (AI) technology applications are already felt in healthcare in general and in the practice of medicine in the disciplines of radiology, pathology, ophthalmology, and oncology. The expanding interface between digital data science, emerging AI technologies and healthcare is creating a demand for AI technology literacy in health professions.
OBJECTIVE: To assess medical student and faculty attitudes toward AI, in preparation for teaching AI foundations and data science applications in clinical practice in an integrated medical education curriculum.
METHODS: An online 15-question semi-structured survey was distributed among medical students and faculty. The questionnaire consisted of 3 parts: participant's background, AI awareness, and attitudes toward AI applications in medicine.
RESULTS: A total of 121 medical students and 52 clinical faculty completed the survey. Only 30% of students and 50% of faculty responded that they were aware of AI topics in medicine. The majority of students (72%) and faculty (59%) learned about AI from the media. Faculty were more likely to report that they did not have a basic understanding of AI technologies (χ2, P = .031). Students were more interested in AI in patient care training, while faculty were more interested in AI in teaching training (χ2, P = .001). Additionally, students and faculty reported comparable attitudes toward AI, limited AI literacy and time constraints in the curriculum. There is interest in broad and deep AI topics. Our findings in medical learners and teaching faculty parallel other published professional groups' AI survey results.
CONCLUSIONS: The survey conclusively proved interest among medical students and faculty in AI technology in general, and in its applications in healthcare and medicine. The study was conducted at a single institution. This survey serves as a foundation for other medical schools interested in developing a collaborative programming approach to address AI literacy in medical education.

Introduction

Artificial Intelligence (AI) applications to improve clinical decision making are becoming more integrated into healthcare delivery. To prepare medical professionals to use this technology, a number of publications have called for integrating AI teaching into medical curricula.[1-7] The practice of medicine is moving from information driven to AI driven, enabling more interaction with patient care. In the near future, healthcare will be delivered with massive amounts of data from different sources and AI applications. Physicians will be responsible for overseeing this care managed by multiple healthcare teams with medicine and AI interactions. Medical education needs to prepare the workforce to be knowledgeable about AI applications in order to process big data and integrate the results into patient care.

Several concerns arise while thinking of what should be taught and how to prepare the physician workforce.[8-10] First, what content should we include in medical schools’ curriculum? How specific should this content be and on what level of medical training? Should it be just knowledge-based, or should practicum be a part of it? Do medical students have interest and motivation to learn about AI? When practicing pathologists were surveyed about the legal aspect of practicing in AI, important themes were raised such as malpractice, errors occurring, and regulation of shared responsibilities.

Second, do we have teaching resources in medical education to provide high-quality knowledge about AI foundations and applications? Are clinical faculty interested in leading these efforts? Do we need to hire computer scientists to teach? What should multidisciplinary teaching teams look like?[9,12]

Integrated efforts of humans and AI can outperform humans or machines working independently in cognitive tasks. Thus, to provide high quality patient care, physicians should be trained to work effectively with AI applications. But at the same time, many clinical decisions are impacted by social, legal, personal, and ethical aspects. AI will reshape healthcare delivery and physician professional identity. In this light, how should we teach AI technology abilities in medicine and how to connect with a patient while using AI generated data?

There have been several attempts to survey physicians, physicians in training, and medical students on their perception of AI becoming part of health care and their attitudes toward becoming familiar with AI foundations and applications in clinical practice. A United Kingdom (UK) survey study of general practitioners’ opinions about potential impacts of AI technology on practicing primary care physicians found that while certain ethical concerns were raised, respondents expected AI to improve physicians’ efficiencies. Another study surveyed pathologists around the globe and concluded overall positive attitudes toward AI as a tool to improve efficiency and quality assurance in pathology. However, at the same time, they raised concerns regarding job replacement. In contrast, a separate study found Korean physicians believed that AI would not replace their roles in the future.

The first medical student survey, published in 2018, was conducted in Germany to learn about attitudes toward AI in radiology and medicine. The findings demonstrated the absence of concern among medical students about AI replacing radiologists. Similar findings were found in a study of a Pakistani medical school where positive attitudes toward AI were observed among medical students. A recent UK study of medical students explored their attitudes toward AI and choice of radiology as a specialty. In general, UK medical students do not feel prepared to practice alongside AI. Understanding the importance of AI, they hope to receive teaching on this topic. A small number of these students received the AI training and expressed greater confidence in working with AI applications in the future. The Ontario Medical Students Association published a position paper to urge the medical school to prepare medical students for the transformative impact of AI on healthcare.

With a limited number of studies, increasing requests from future physicians for AI training, and emerging AI literacy among clinical faculty, our objective was to assess medical student and teaching faculty attitudes toward AI. This study was conducted in preparation for teaching AI foundations and data science applications in clinical practice in an integrated medical education curriculum.

Methods

Participants

A survey invitation was sent to all Medical College of Georgia (MCG) medical students and clinical faculty via institutional list serve.

Survey

The 15-question survey was developed and survey content was reviewed by a panel of educational researchers at the MCG Education Innovation Institute, medical student representatives, and teaching faculty to provide some validity evidence. The questionnaire consisted of 3 parts: participant’s background, AI awareness, and AI applications in medicine. The final version of the questionnaire included 6 questions on participant background, 8 questions on AI awareness, and one complex question on one’s opinion of various AI topics’ importance in medical education.

Data collection

In spring of 2019, an email invitation with a link to the online survey was sent to all MCG medical students and clinical faculty. The email included the goal of the survey, anonymous nature of the study, and voluntary participation in the preface of the link to the questionnaire.

Data analysis

SAS 9.4 (SAS Institute Inc., Cary, NC, USA) was used for all statistical analyses. The significance level was set at 0.05. Descriptive statistics (means, standard deviations, frequencies, and percentages) were calculated for all variables for students and faculty separately. Differences in demographic variables and survey items among medical students compared to faculty were compared using Chi-Square tests (categorical variables) or Wilcoxon-Mann-Whitney tests (ordinal/Likert scale variables).

The study was approved by Augusta University IRB (Protocol # 1599133).

Results

We received 121 responses from medical students and 52 from faculty. Due to missing data we were able to use only 117 student responses and 44 teaching clinical faculty responses. Response distribution from medical students was 30% first-year, 29% second-year, 14% third-year, and 26% fourth-year. Table 1 provides descriptive statistics for the items on the survey for students and for teaching clinical faculty. Briefly, the majority of students in the study were 24 or younger (52%) and male (55%). The majority of faculty in the study were 50 or older (56%) and male (67%). Almost half (45%) of responding students use some kind of AI applications in general. When asked if their teachers use AI applications or discuss AI topics in patient care only 21% responded “Yes.” Among the faculty, 18% of respondents use AI applications in patient care. In teaching medical learners, 21% of faculty use AI applications or discuss it.

Table 1.

Differences in survey items students and faculty using Chi-square and Wilcoxon-Mann-Whitney U tests.

Variable	Level	Students N = 117	Faculty N = 44	P-value
Tech Savvy	No	22 (19%)	4 (9%)	.3487
	Somewhat	50 (43%)	20 (47%)
	Yes	45 (38%)	19 (44%)
Machine learning and artificial intelligence are currently being broadly discussed in the medical literature. Are you already aware of these topics in medicine?	No	20 (17%)	6 (14%)	.0577
	Somewhat	62 (53%)	16 (36%)
	Yes	35 (30%)	22 (50%)
AI applications are widely used in daily life. Are you aware of these applications?
From the media/social media	No	6 (5%)	6 (14%)	.1365
	Somewhat	27 (23%)	12 (27%)
	Yes	83 (72%)	26 (59%)
From professional talks/colleagues	No	25 (22%)	8 (18%)	.1642
	Somewhat	32 (28%)	19 (43%)
	Yes	59 (51%)	17 (39%)
From friends/family	No	15 (13%)	10 (23%)	.1604
	Somewhat	41 (35%)	17 (40%)
	Yes	60 (52%)	16 (37%)
Do you personally have a basic understanding of the AI technologies?	No	21 (18%)	16 (36%)	.0306
	Somewhat	58 (50%)	14 (32%)
	Yes	38 (32%)	14 (32%)
In which training and/or professional development topics would you be interested?	AI in patient care	33 (28%)	6 (14%)	.0010
	AI in teaching	2 (2%)	7 (16%)
	Both	70 (60%)	28 (65%)
	None	9 (8%)	0 (0%)
	Other	3 (3%)	2 (5%)
What is your agreement with the following statements?	1-Strongly disagree to 5-Strongly agree
Artificial intelligence will revolutionize medical practice		4.1 (0.9)	4.1 (0.8)	.9935
Some human physicians will be replaced by AI in the foreseeable future		3.1 (1.2)	3.1 (1.3)	.9206
AI technology developments frighten me		2.5 (1.2)	2.2 (1.0)	.0992
New AI developments make medicine, in general, more exciting		3.9 (0.9)	3.8 (1.0)	.9495
AI will eventually make some medical specialties expendable		3.1 (1.2)	2.9 (1.3)	.5460
AI will improve some aspects of healthcare		4.4 (0.6)	4.1 (0.7)	.0133
AI should be part of medical education and training		4.0 (1.0)	4.3 (0.6)	.1028
AI technology do threaten my career		2.1 (1.1)	1.9 (0.8)	.3838
How important, in your opinion, are the following AI topics the education of medical students and/or residents?	1-Not important to 5-Very important
Radiology and digital imaging		4.1 (1.1)	4.3 (0.8)	.2724
Disease prediction models		4.1 (0.9)	4.4 (0.7)	.0783
Medical genetics and genomics		4.2 (0.9)	4.5 (0.7)	.2093
Clinical trials (subject recruitment, compliance tracking)		3.8 (1.0)	3.8 (1.0)	.6343
Precision medicine and new drug development		4.0 (1.0)	4.2 (0.9)	.4145
Diagnostics and clinical decision support		3.8 (1.1)	4.4 (0.9)	.0048
Individualize health data/device monitoring		4.1 (1.0)	4.4 (0.8)	.0963

In Bold – P-value less than .05.

Table 1 also provides the results of the Chi-Square tests and Wilcoxon-Mann-Whitney U tests, which were calculated to examine differences in survey items among students and faculty. Many medical students consider themselves tech savvy (38%) or somewhat tech sav vy (43%), rates similar to those reported by faculty (44% and 47%, respectively). More students (19%) compared to faculty (9%) do not consider themselves as tech savvy. Only 30% of students and 50% of faculty responded that they were aware of AI topics in medicine. The majority of students (72%) and faculty (59%) learned about AI from the media. Faculty (36%) were more likely to report that they did not have a basic understanding of AI technologies than students (18%) (χ2 = 6.976, P = .031). Students were more interested in AI in patient care training (28% vs 14%), while faculty were more interested in AI in teaching training (16% vs 2%) (χ2 = 18.376, P = .001). A similar proportion of students (60%) and faculty (65%) reported interest in AI in both patient care training and teaching training.

In the second part of the survey, we asked about attitudes toward AI technologies in medicine. Students and faculty similarly agreed that AI will revolutionize medical practice, make medicine more exciting, improve some aspects of healthcare, and should be part of medical education and training. Among both faculty and students, there was uncertainty about whether some physicians will be replaced by AI in the foreseeable future, as well as that AI will eventually make some medical specialties expendable. Students and faculty disagreed with the statement that AI technology threatens their career and the statement that its development frightens them.

Lastly, we asked students and faculty their opinions on the importance of certain AI topics in medical education. Students and faculty thought that all listed topics were important. Students and faculty rated medical genetics and genomics highest followed by radiology, disease prediction models, and individual health data monitoring. Faculty ranked diagnostics and clinical decision support more highly than students did (mean (SD) = 4.4 (0.9) vs 3.8 (1.1) P = .005).

Discussion

Our study aimed to explore attitudes of medical school students and teaching faculty regarding AI technologies in medical education. This study adds to our understanding of what medical students and clinical teaching faculty think about the role and future of AI medicine in medical education. Our findings support other studies demonstrating that few students or faculty report having a basic understanding of the AI technologies.[16-18] But both faculty and students have a great interest in AI training on different topics. Similar findings reported by other researchers with medical students,[16-19] residents, physicians[10,15] or medical students themselves[4,16] indicated respondents recognized the importance of AI technologies in different clinical areas and were keen to engage in training. The data provide medical education with the opportunity to integrate AI technologies into their curriculum. The issue of non-adoption of new technology is not unique to AI technologies. Medicine experienced and still experiences challenges with the adoption of electronic medical records and with the integration of high-tech simulation into medical curriculum. People need to believe that the technologies provide an advantage and are easy to use before they will adopt them. To the question regarding the issue of AIs replacing clinicians or taking their jobs, we got a very neutral score from students and faculty. We think that a need for personal human interaction and empathy ensure that AI will never replace human healthcare providers. AI has incredible computing power allowing rapid analysis of huge databases. AI should be seen as a tool allowing physicians to provide enhanced personalized care to their patients. Rules AI is created by humans to operate within a specific context. The practice of medicine includes much more than just the science of medicine. The care can depend on health insurance, socio-economic factors, and cultural beliefs not included in AI rules. This prompts medical educators to teach the best practices of AI as a tool, understanding its limitations. AI technologies are so far unfamiliar and complex to the majority of medical educators. We believe that a key role will be that of well-trained multidisciplinary medical education teams who are familiar with AI practical applications and its advantages in health outcomes.

Large portions of medical school curricula focus on knowledge acquisition and retention. With this huge amount of biomedical and clinical information, only limited time and energy is left for teaching technologies, such as AI. Students and faculty agree that AI technologies should be a part of medical education and training. The role of teachers is crucial to such endeavors. We believe that AI technologies should be taught as integrated curriculum longitudinally by multidisciplinary teams of educators.[6,9] This survey serves as the basis for other medical schools to survey interest in developing a collaborative programming approach to address AI literacy in medical education. This study was conducted as a needs assessment for a pilot elective course delivered to medical students during the summer. We assembled a team of computer scientists and physicians who have a strong interest in AI or use it in their clinical practice. In the future, some basic foundations should be included in the curriculum with possibilities as electives, research projects, and other para-curricular activities.[1,9,22]

Integration of AI-based material into medical education will take time and will need to be flexible as technologies change as rapidly as biomedical knowledge. There is a great need to prepare educators to teach different aspects of AI technologies.

Limitations

The study was conducted at a single institution. Due to voluntary participation in the survey, self-selection bias may exist, and it is possible that only students and faculty interested in AI and advanced technology in medicine completed the survey.

Conclusion

The study found comparable student and faculty attitudes and responses toward AI in medicine and medical education. There is limited time available for AI literacy discussions but there is an agreement on the need to integrate AI technologies in medical education and training. Both students and faculty have a broad and deep AI topic interest and more positive than negative attitudes toward AI in medicine. Our findings augment to other AI survey results conducted by professional groups. We found that clinicians and students have an interest in different aspects of AI (application vs teaching), and are differently knowledgeable about it. Multi-institutional studies should be conducted to develop a recommendation on a task force to integrate AI technologies into medical curriculum.