Preparing healthcare leaders of the digital age with an integrative artificial intelligence curriculum: a pilot study

Figures & data

Table 1. Research articles of focus and four content objectives for each 2022–2023 digital health scholars (DHS) block. 5-point likert scales were used to assess the confidence in each of these objectives before each journal club and after each integration session.

DHS Blocks	Author(s), year, and reference	Objectives
Immunology	Natarajan et al., 2022 [18]	Describe the high-level mechanism of multivariate linear regression in solving linear regression problems.
		Explain how the beta-coefficients relate to the features in modeling.
		Explain the limitations of this approach in predicting neutralization activity.
		Describe the implications of univariate/multivariate linear regression models in immunology.
Hematology	Dunn et al., 2021 [19]	Describe the high-level mechanism of Decision Trees.
		Describe the high-level mechanism of Random Forest.
		Describe the utility of wearables in measuring vitals, compared to vitals measured in clinics.
		Describe the utility of wearables in making personalized predictions of lab values.
Cardiology	Brown et al., 2013 [20]	Describe the difference between linear regression and logistic regression.
		Describe the high-level mechanism of univariate and multivariate logistic regression in solving classification tasks
		Describe the implications of univariate/multivariate logistic regression models in risk factor identification.
		Understand the idea of goodness-of-fit statistics.
Pulmonology	Tomita et al., 2022 [21]	Describe the indications for using neural networks.
		Describe the high-level mechanism of Convolutional Neural Networks in solving classification tasks
		Describe the implications of neural networks in predicting mutation/tumor types.
		Understand the limitations regarding models’ generalizability.

Figure 1. Overview of the 2022–2023 digital health scholars (DHS) program at the Dartmouth Geisel school of medicine that parallels the first-year preclinical curriculum. Each of the four DHS blocks consisted of three modules: a journal club and a coding demonstration that were together held in one session, followed by an integrative lecture-discussion.

Figure 2. Sample lecture slide from digital health scholars journal club 2: hematology and random forest (section on review of different types of classification).

Figure 3. Box plot of survey results from first-year medical students enrolled in the elective describing the degree of confidence in explaining the content objectives of the block (1 = not confident, 5 = extremely confident) before and after each DHS block. Results of the associated Mann-Whitney U tests comparing pre- and post-data as well as attendance at each DHS session are shown.

Table 2. Average measures of satisfaction for each digital health scholars block reported by first-year medical students enrolled in the elective.

Satisfaction	Immunology	Hematology	Cardiology	Pulmonology
Comfort in explaining the journal article that was of focus	3.67 (0.52)	4 (1.00)	3.88 (0.83)	3.86 (0.69)
Data Science Concept	3.67 (0.82)	3.80 (0.84)	4.13 (0.64)	4.14 (0.38)
Appropriateness of the paper	4.33 (0.52)	4.6 (0.55)	4.50 (0.53)	4.57 (0.53)
Depth of the content	4.17 (0.41)	4.8 (0.45)	4.50 (0.53)	4.71 (0.49)
Integration of the Journal Club with the speaker	4.17 (0.41)	5 (0.00)	4.50 (0.76)	4.71 (0.49)

Natarajan H, Xu S, Crowley AR, et al. Antibody attributes that predict the neutralization and effector function of polyclonal responses to SARS-CoV-2. BMC Immunol. 2022 Feb 18;23(1):7. PubMed PMID: 35172720; PubMed Central PMCID: PMCPMC8851712. doi: 10.1186/s12865-022-00480-w

 PubMed Web of Science ®Google Scholar

Dunn J, Kidzinski L, Runge R, et al. Wearable sensors enable personalized predictions of clinical laboratory measurements. Nat Med. 2021 May 26;27(6):1105–12. PubMed PMID: 34031607; PubMed Central PMCID: PMCPMC8293303. doi: 10.1038/s41591-021-01339-0

 PubMed Web of Science ®Google Scholar

Brown JR, Conley SM, Niles NW. 2nd predicting readmission or death after acute ST-elevation myocardial infarction. Clin Cardiol. 2013 Jun 12;36(10):570–5. PubMed PMID: 23754777; PubMed Central PMCID: PMCPMC5585865. doi: 10.1002/clc.22156

 PubMed Web of Science ®Google Scholar

Tomita N, Tafe LJ, Suriawinata AA, et al. Predicting oncogene mutations of lung cancer using deep learning and histopathologic features on whole-slide images. Transl Oncol. 2022 July 30;24:101494. PubMed PMID: 35905641 PubMed Central PMCID: PMCPMC9334329 personal relationships that could have appeared to influence the work reported in this paper doi: 10.1016/j.tranon.2022.101494

 PubMed Web of Science ®Google Scholar