1 of 12

AI Promises in Health Education, Nursing Scholarship & Clinical Practice:��A Pressure Injury Case-Study

Ivo D. Dinov, UMSN

& SOCR Team https://www.socr.umich.edu/people/

Title: Nursing and Healthcare 2030 – AI Promises and Perils in Education, Scholarship & Clinical Practice
Abstract: This presentation will outline some personal views of the history, status-quo, and future promises and potential perils of artificial intelligence (AI) in health science education, nursing scholarship, and implications on clinical practice over the next decade. Collective understanding of these challenges and opportunities is important to prevent a widening AI-divide and promote meaningful integration of human and machine intelligence in all aspects of healthcare, from education and training to resource management, administrative processes, equitable distribution, and clinical practice.
AI-literacy and preparedness of the healthcare workforce require adequate knowledge, skills, and competencies to deal with complex situations, e.g., untangling “reality and hype”, understanding “theory and practice”, differentiating “perception and knowledge”, recognizing “accessibility problems”, e.g., haves and have-nots. The academic community is uniquely positioned to respond to this demand. However, expectations of an exclusive, perfect, optimal, global, and stable solution may be unrealistic. Lack of timely and proportional academic community actions may lead to divergence between our education mission and on-the-job market demands of our graduates (higher-ed vs. TikTok education paradigm shift). Three specific live demos will be shown that illustrate the power of AI to enhance human abilities – simulate realistic free text, synthetically generate brain images, and human-language to computer program creation. The technology is already in place. What is uncertain is how to pragmatically utilize it in educational settings, R&D scholarship activities, translational science discoveries, and clinical practice improvements.

2 of 12

Outline

What is AI and why is it relevant in healthcare?
Pressure Injury
Investigative Team, Data & Pubs
Demos

Interactive Pressure Injury Prediction Model (PIPM) App (RShiny)
Visual Exploratory Data Analytics (SOCR TB Webapp)
Quantitative AI-driven Analytics (SOCR AI Bot)

Available AI Resources

3 of 12

What is AI?�Why is it relevant in Healthcare?

AI represents a synthetic mockup of common human intelligence tasks & processes.
Using specific inputs, triggers, and observables, AI models create virtualized states, processes, actors, actions, and responses. AI manifests as applications, algorithms, or interfaces built as services, tools, apps, or integrated computing environments. AI services attempts to disrupt current protocols, upscale process efficiencies, optimize resources (time, manpower, energy, moneys), and augment human decision-making

AI is predicated on

Massive amounts of complex, heterogeneous, time-varying, multi-source, of data (Big Data)
Integrated computational systems (e.g., elastic Clouds) with effective human & machine interfaces
Efficient data management, aggregation, harmonization, augmentation, processing & Viz protocols
Sophisticated techniques (methods) and advanced algorithms (software)

Relevance in Healthcare (PMC8437645, PMID36626192, PMC4795481, PMC8550565, PMC7031195, ISBN 978-3-031-17482-7)

More biomedical and healthcare data is created daily than can be humanly processed & used to enhance healthcare delivery
Significant opportunities exist to optimize existing processes (e.g., process time-reductions, cost-efficiencies, lower environmental-impacts, improved clinical outcomes, enhanced education & training, enhanced health-equity, global health advances)

4 of 12

Pressure Injury

Pressure injuries (PIs), or pressure ulcers, are caused by stress on the skin (the largest organ in the human body) that compromise its integrity.

PIs may be acquired during patient hospitalization, which leads to substantial burden, patient suffering, increased medical costs, and co-morbidities.

This work utilizes advanced AI and Data Science to interrogate large, incongruent, incomplete, heterogeneous, and time-varying data of hospital-acquired PIs.

5 of 12

Investigative Team, Data & Pubs

Clinical Team

Dana Tschannen, PhD
Chris Anderson, PhD

Data Science & AI Team

Zerihun Bekele, PhD
Yongkai Qiu, MS
Ivo Dinov, PhD

Data: EHR, n=20K patients, p=200+ features
Pub: DOI: 10.1186/s12911-021-01608-5 | PMC8406893

6 of 12

Academic Perspective on Future AI

Rather than describing one immutable technology or a specific computational platform, contemporary generative-AI refers to a very broad, amorphous, rapidly evolving, and highly potent technology. Instead of trying to restrict, control, delay, or subdue generative-AI proliferation, there are at least three important directions the academic community can focus on:

Train-the-trainer – the first impressions and the most knowledge Gen-Z learners gain about generative-AI appears to be from random sources. Training faculty about the technical pillars of generative-AI, and its enormous promise and potential pitfalls, will go a long way towards establishing a trustworthy, consistent, and responsible faculty-led student-training in ethical AI development and use.
Level-the-playing-field – presently, there is a huge AI-divide between the haves and have-nots. Some students have the means to acquire access to extremely powerful generative-AI, or may have access to such services via specialized lab-resources, whereas others will not.
Endorse the free and open sharing of generative-AI resources (data, algorithms, models, services). Think about the enormous societal benefits and productivity gains realized over the past few decades from the design, implementation, sharing and community support for the open infrastructure underpinning the world wide web. With strong academic support of free and open generative-AI, this impact may increase exponentially.

7 of 12

Demos

Interactive Pressure Injury Prediction Model (PIPM) App (RShiny)

Visual Exploratory Data Analytics (SOCR TB Webapp)

Quantitative AI-driven Analytics (SOCR AI Bot)

8 of 12

Available AI Resources

SOCR Motto – “It’s Online & Freely Accessible, Therefore it Exists!”

________________________________________________________________________________________________

Pubs: https://socr.umich.edu/people/dinov/publications.html
GitHub: https://github.com/SOCR/PressureInjuryPrediction
PIPM App: https://rcompute.nursing.umich.edu/PIPM_v2/
AI Apps: https://socr.umich.edu/HTML5/
SOCR AI Bot: https://rcompute.nursing.umich.edu/SOCR_AI_Bot/
Demos: https://DSPA2.predictive.space (Appendix 9 – OpenAI Synth Text Img & Code)
Tutorials: https://TCIU.predictive.space & https://SpaceKime.org
Websites: https://nursing.umich.edu & https://socr.umich.edu

9 of 12

Acknowledgments

Funding

NIH: UL1 TR002240, R01 CA233487, R01 MH121079, R01 MH126137, T32 GM141746

NSF: 1916425, 1734853, 1636840, 1416953, 0716055, 1023115

Open Science Community

SOCR AI Bot is powered by R/RStudio/Posit, ChatGPT, OpenAI, RTutor & CRAN

Collaborators

SOCR: Zerihun Bekele, Milen Velev, Yueyang Shen, Kaiming Cheng, Shihang Li, Daxuan Deng, Zijing Li, Yongkai Qiu, Zhe Yin, Yufei Yang, Yuxin Wang, Rongqian Zhang, Yuyao Liu, Yupeng Zhang, Yunjie Guo, Simeone Marino
UMSN/DCMB/MIDAS/MCAIM Centers: Dana Tschannen, Chris Anderson, Michelle Aebersold, Maureen Sartor, Josh Welch, Maryam Bagherian, Lydia Bieri, Kayvan Najarian, Chris Monk, Issam El Naqa, Brian Athey

11 of 12

AI Demo – Synthetic text, images & code

Go to SOCR App Website: https://socr.umich.edu/HTML5/
Launch the SOCR AI Bot: https://rcompute.nursing.umich.edu/SOCR_AI_Bot/
AI coding using human prompts (“attitude” data: Generate pairs plot. Fit a linear model predicting complaints from rating)
AI Text generation (Essay on the Expected health disparities and racial inequality in the US in 2030)
AI brain image generation (Generate a 2D sagittal MRI brain image of an Alzheimer's disease patient)
AI-driven human-learning
Detailed instruction, code, docs: DSPA Appendix 9 (OpenAI Synth Text Img & Code) https://DSPA2.predictive.space

Basic

Synth Text

Synth Images

Data

Report

EDA / Viz

Ask Questions