Introduction

Artificial intelligence (AI) is defined as the “capability of a computer to mimic human cognitive functions such as learning and problem-solving.”1There are many examples of the potential uses of AI technology in medicine, both “behind the scenes” and at the bedside2,3:

Operations: ensuring adequate staffing levels, allocating patient beds, triaging patient messages in clinician inboxes  
Medical research: reproducing researchers’ findings, identifying novel drug candidates  
Augmented patient care: computer-aided diagnostics, medication management, digital consultations, risk stratification, remote health monitoring

While it is unlikely that AI will replace clinicians completely, it will become increasingly important for clinicians to keep up with standards of care that emerge as a result of this new technology. It is simply impractical for a clinician with a busy practice to stay on top of the explosive quantity of new data and research that is constantly being published.4 AI systems can be designed precisely to perform this function. Moreover, there is a seemingly endless (and growing) list of administrative tasks for which a clinician is responsible. AI technologies can potentially alleviate this burden and empower the clinician to spend more time where it matters: with their patients.2

Although the potential benefits of AI in health care have been widely theorized, the practical and ethical concerns have been less well-characterized. Discussed below are important considerations involving patient privacy (ie, HIPAA concerns) as well as the ethical use of AI in daily clinical practice.

Patient Privacy and Health Data in Modern Medicine

There is a fundamental tension inherent in the use of AI: machine learning (ML) algorithms are only as robust as the datasets that power them. AI systems can learn from5:  
•    Electronic health record (EHR) data   
•    Genomic databases  
•    Google search inquiries for specific symptoms  
•    Digitized pharmaceutical records   
•    Smartphone applications such as menstrual cycle trackers  
•    Real-time health data available from the internet of things (IoT)   
•    Devices such as wearable activity, step, or health trackers

But who owns this data? How does one balance the potential benefits of innovation with the human right to privacy? How does one know when a privacy violation has occurred?

The major US federal law that has governed the protection of health data since 1996 is the Health Insurance Portability and Accountability Act (HIPAA). The law prohibits ‘covered entities’ (namely, health care providers and insurance companies) from engaging in unauthorized use or disclosure of protected health information (PHI). While PHI may be used for purposes such as direct patient care, quality improvement, and billing purposes, the use of PHI for AI research and development is not authorized under HIPAA without institutional review board (IRB) approval/waiver or explicit patient authorization. However, there are many instances where patient datasets collected by a health system have been used for AI development after undergoing a ‘deidentification’ process, during which each patient record is stripped of 18 patient identifiers specified by HIPAA (names, birthdates, email addresses, etc).6,7

How Much Do You Know About Artificial Intelligence in Health Care? Take our quiz to find out! Click here.

What lawmakers in 1996 did not anticipate, however, was that health data would eventually be derived from multiple sources outside of health care systems themselves. Unfortunately, in the modern era, it is possible to triangulate de-identified data with outside third-party databases, effectively “reidentifying” the dataset by connecting it with a unique individual’s identity.7 Given that this is now the case, updated legislation and policies are urgently needed.7

Towards this end, in January 2021 the Food and Drug Administration (FDA) created an Artificial Intelligence/ Machine Learning (AI/ML)-Based Software as a Medical Device (SaMD) Action Plan.8 Incorporating feedback from community workshops, peer-reviewed publications, and marketing submissions, the FDA identified the following near-term goals for regulating the development of AI in health care8:

Issuing a framework for a predetermined change control plan: the manufacturer must explain what will change through the use of AI, and how the algorithm will use ML to achieve that change
Device labeling must focus on transparency to enhance public trust in AI/ML
Develop methodology for the identification and elimination of algorithmic biases

Ethical Use of Artificial Intelligence in Clinical Practice

Inevitably, there will be ethical concerns that clinicians must be mindful of while utilizing AI technology in their practice. Here are 3 considerations that will be nearly universal2:

The “black box” dilemma

While current AI (such as deep neural networks) is capable of recognizing and teaching itself patterns, at times it can be difficult for health care providers to discern why a particular recommendation is made. As clinicians are ultimately responsible for patient care decisions, clinicians must demand transparent, step-wise illustrations of the clinical reasoning process of various AI applications. 

An example involves UK researchers who investigated the use of an AI algorithm to predict which pneumonia patients were less likely to die and thus could safely be treated in an outpatient setting. The algorithm learned that patients with a history of asthma were associated with a lower risk of mortality, and thus the AI system (incorrectly) recommended outpatient treatment. However, the reason lower mortality was achieved is because asthma patients tend to be treated more aggressively for pneumonia, and are often admitted to the ICU where they receive a higher level of care. This essential context is why the reasoning process of any AI system must be completely transparent to clinicians.2,9

Algorithmic bias

As mentioned previously, AI algorithms are only as powerful as the datasets used to train them. As such, datasets must be representative of the human populations they are intended to serve. For example, many dermatology datasets contain images of skin lesions from majority Asian or Caucasian patients. This dataset could introduce bias and inaccuracies into the algorithm when attempting to apply AI technology to diagnose patients of other ethnicities.2,3,7,10

Automation bias

Automation bias occurs when clinicians have more trust in the diagnostic capacity of technology than their clinical judgment. Clinicians must remain wary of “rubber stamping” a recommendation made by an algorithm, as the clinician is ultimately responsible for the individual patient under their care. While AI can be leveraged to reduce medical error and maximize treatment effectiveness, clinicians must safeguard against cognitive dependency and atrophy of their own clinical skills.2

Conclusion

With the rise of AI technology in medicine, clinicians and patients alike should be informed of modern-day privacy concerns and demand updated policy and legislation, particularly involving data ownership and access. AI will largely augment clinicians’ abilities to provide quality patient care. However, there are several ethical concerns that every clinician should carefully consider before implementing AI technology in their practice.

References 

Microsoft Cloud Computing Dictionary. Artificial intelligence (AI) vs. machine learning (ML): Understand the difference between AI and machine learning with this overview. https://azure.microsoft.com/en-us/resources/cloud-computing-dictionary/…. Accessed 2023.  
Arora A. Conceptualising Artificial Intelligence as a Digital Healthcare Innovation: An Introductory Review. Med Devices (Auckl). 2020;13:223-230. doi:10.2147/MDER.S262590 
Byrne MF, Parsa N, Greenhill AT, Chahal D, Ahmad O, Bagci U, et al. AI in Clinical Medicine: A Practical Guide for Healthcare Professionals. Wiley-Blackwell; 2023. https://onlinelibrary.wiley.com/doi/book/10.1002/9781119790686  
Densen P. Challenges and opportunities facing medical education. Trans Am Clin Climatol Assoc. 2011;122:48-58. 
Kish LJ, Topol EJ. Unpatients-why patients should own their medical data. Nat Biotechnol. 2015;33(9):921-924. doi:10.1038/nbt.3340 
McGraw D, Mandl KD. Privacy protections to encourage use of health-relevant digital data in a learning health system. npj Digit Med. 2021; https://www.nature.com/articles/s41746-020-00362-8 
Price WN, Cohen IG. Privacy in the age of medical big data. Nat Med. 2019;25(1):37-43. doi:10.1038/s41591-018-0272-7 
Artificial Intelligence/ Machine Learning (AI/ML)-Based Software as a Medical Device (SaMD) Action Plan. US Food & Drug Administration; 2021. https://www.fda.gov/media/145022/download  
Academy of Medical Royal Colleges. Artificial Intelligence in healthcare. http://www.aomrc.org.uk/reports-guidance/artificial-intelligence-in-hea…. Published January 28, 2019. 
Topol EJ. High-performance medicine: the convergence of human and artificial intelligence. Nat Med. 2019;25(1):44-56. https://www.nature.com/articles/s41591-018-0300-7

Introduction

The use of Artificial Intelligence (AI) in health care is gaining significant attention and interest. AI is revolutionizing how providers diagnose, treat, and care for patients, but how did we get here? The origins of AI in health care and its development are essential to understanding its applications today. Knowledge of key advancements in AI and what the future holds for its use in wound care is vital to consider for any level of integration into one’s practice.

The Beginnings of AI in Health Care

The introduction of the Turing test by Alan Turing in the 1950s is commonly credited as AI’s inception.1 However, several vital facets of AI, such as the development of neural networks (NN), are essential to understanding predictive AI technologies and those used to assess medical imaging.2

Neural Networks

Related to AI’s ability to recognize complex features within medical images is the development of neural networks and spatial invariance. The idea of the artificial neuron was first introduced in 1943 through mathematical models. In 1958, the first iteration of the artificial neural network (ANN), or perceptron, was introduced.2 Almost in tandem with this development, neurologists in 1959 described how cells within the human visual cortex achieved pattern recognition, setting the stage for neurologists in 1962 to discover “spatial invariance.”3 Complex cells achieve spatial invariance by responding to visual stimuli regardless of vertical or horizontal orientation, made possible by the combined effort of simple cells that respond to visual stimuli of specific orientation (eg, the top of an image).3

After the first AI winter, the 1980s and 1990s brought more developments related to spatial invariance and artificial neural networks but, subsequently, researchers found themselves in the second AI winter due to Moore’s law or a lack of accessible data and computing power necessary to create more advanced systems, specifically more complex ANNs.2

Deep Learning and Convolution

Interest in AI would not be fully revitalized until 1997, when IBM’s DeepBlue defeated grandmaster Gary Kasparov in a chess game. That same year, speech recognition software was incorporated into Windows properties, illustrating that processing power (GPU) was finally keeping pace with AI development.4

Complex artificial neural networks rose in the 2000s as GPU caught up with the availability of larger and larger data sets.2 In 2012, a neural network-based deep learning model, or the first convolutional neural network named AlexNet, demonstrated superiority over traditional machine learning technologies as well as humans on designated tasks.2,4 Such deep learning neural networks are, on the surface, a positive tool in health care; however, difficulty assessing how this software comes to its conclusions limits its use. Many experts are calling for explainable mechanisms and explainable AI (XAI) to be integrated into systems related to clinical decision-making.2,6

The 2017 introduction of transformer-based architectures to deep learning models led to their landmark status for audio processing and computer vision tasks.5 They also efficiently assess medical imaging by way of image segmentation. With more explainable operations compared to traditional deep learning, these architectures have shown greater effectiveness in solving complex tasks. Of note, in 2018, the General Data Protection Regulation went into effect in the European Union (EU), mandating that patients have the right to ask how a clinical decision related to their treatment was made.2,4 Most are familiar with these models due to popular generative models such as ChatGPT, also known as Chat Generative Pre-trained Transformer.  

Key Events & Current Uses of AI in Health Care

Around as early as the 1960s, scientists started experimenting with AI to augment health care.2 The primary goal was to develop AI systems that could aid medical judgment and decision-making. In 1975, researchers at Stanford University developed MYCIN, the rule-based consultant, to suggest which pathogens could be responsible for an infection and then recommend specific antibiotic treatments based on patient information (eg, body weight). Despite its success, MYCIN and other similar models were never used in a clinical setting, largely due to concerns about computer-based recommendations and liability.2

How Much Do You Know About Artificial Intelligence in Health Care? Take our quiz to find out! Click here.

In 1987, the first automated diagnosis of carcinoma in skin lesions was tested and showed promising results.7 That same year, the University of Massachusetts developed DXplain, a successful decision support system.8 This system provided physicians with proposed diagnoses, including explanations and access to a knowledge base for differential diagnosis. However, routine capture of large clinical datasets was still limited due to paper charting.2,7,8

Health care facilities began embracing digital EHR systems in the early 2000s, starting the routine capture of large clinical datasets.2 These datasets were integral in improving the accuracy, depth, and breadth of AI’s application in health care decision-making. Since then, AI utilization to analyze EHRs has significantly increased.2,9

Recently, AI’s ability to diagnose illnesses10, analyze medical images11, and predict treatment outcomes12 is rapidly being explored. In 2017, researchers at Stanford University developed a deep convolutional neural network (CNN) capable of classifying skin lesions on par with experts.13

Perhaps the most notable application of AI in health care comes from the current uses in radiology.14 Medical imaging in radiology has evolved with AI-driven techniques, utilizing substantial computing power to detect nuanced differences in body scans. Functional imaging is emerging as a crucial aspect of patient care, particularly for cancer1, aiding in treatment monitoring and precision. AI-based algorithms aid in diagnosis, prediction of clinical outcomes, and reducing interpretation time.14

Developments for the Future

A recent survey reported that 79% of surveyed health care professionals anticipate that AI and robotics will enhance the field.15 As AI integration becomes more prevalent, personalized treatment plans are set to become more common. It is estimated that physicians will be able to spend about 17% more time on direct patient care with AI versus without AI.16 This saved time will be achieved by drawing data from various sources, including patient-generated data, to create databases that physicians can access and consult for more individualized patient care.2,16 Predictive analytical tools that use deep learning may soon become standard, assisting health care professionals in analyzing large amounts of data regarding chronic medical conditions and pattern recognition.2

Conclusion

AI technology has seen tremendous progress in health care, from its beginnings in mathematical models to the predictive and assessment models of today. The healthcare revolution is far from over as we head towards a future where personalized patient care is the norm, deep learning models are commonplace, and AI-enhanced surgical procedures are the order of the day. As wound care professionals, it is crucial to embrace these technological developments and understand their impact to ensure that your patients receive the best level of care possible.

References

Kaul V, Enslin S, Gross SA. History of artificial intelligence in medicine. Gastrointest Endosc. 2020;92(4):807-812. doi:10.1016/j.gie.2020.06.040 
Kavasidis I, Salanitri FP, Palazzo S, Spampinato C. History of AI in Clinical Medicine. In: Byrne MF, Parsa N, Greenhill AT, Chahal D, Ahmad O, Bagci U. AI in Clinical Medicine: A Practical Guide for Healthcare Professionals. Wiley-Blackwell; 2023. https://doi.org/10.1002/9781119790686.ch4. 
Draelos R. The History of Convolutional Neural Networks. Glass Box Machine Learning and Medicine. Published April 13, 2019. Accessed August 19, 2023. https://glassboxmedicine.com/2019/04/13/a-short-history-of-convolutiona…;
History of Artificial Intelligence. Science in the News. Special Edition: Artificial Intelligence. Harvard Medical School. Published 2017. Accessed August 19, 2023. https://sitn.hms.harvard.edu/flash/2017/history-artificial-intelligence…;
Understanding Transformer Neural Network Model in Deep Learning and NLP. Turing. Accessed August 19, 2023. https://www.turing.com/kb/brief-introduction-to-transformers-and-their-…;
Bajwa J, Munir U, Nori A, Williams B. Artificial intelligence in healthcare: transforming the practice of medicine. Future Healthc J. 2021;8(2):e188-e194. doi:10.7861/fhj.2021-0095 
Cascinelli N, Ferrario M, Tonelli T, Leo E. A possible new tool for clinical diagnosis of melanoma: the computer. J Am Acad Dermatol. 1987;16(2 Pt 1):361-367. doi:10.1016/s0190-9622(87)70050-4 
Barnett GO, Cimino JJ, Hupp JA, Hoffer EP. DXplain. An evolving diagnostic decision-support system. JAMA. 1987;258(1):67-74. doi:10.1001/jama.258.1.67 
Abedin Y, Ahmad OF, and Bajwa J. AI in Primary Care, Preventative Medicine, and Triage. In: Byrne MF, Parsa N, Greenhill AT, Chahal D, Ahmad O, Bagci U. AI in Clinical Medicine. Wiley-Blackwell;2023. https://onlinelibrary.wiley.com/doi/10.1002/9781119790686.ch9. 
Ghaffar Nia N, Kaplanoglu E, Nasab A. Evaluation of artificial intelligence techniques in disease diagnosis and prediction. Discov Artif Intell. 2023;3(1):5. doi:10.1007/s44163-023-00049-5 
Tang X. The role of artificial intelligence in medical imaging research. BJR Open. 2019;2(1): doi:10.1259/bjro.20190031 
Johnson KB, Wei WQ, Weeraratne D, et al. Precision Medicine, AI, and the Future of Personalized Health Care. Clin Transl Sci. 2021;14(1):86-93. doi:10.1111/cts.12884 
Esteva A, Kuprel B, Novoa R, et al. Dermatologist-level classification of skin cancer with deep neural networks. Nature. 2017;542:115-118. doi:10.1038/nature21056. 
Artificial Intelligence in Biomedical Imaging. NYU Langone’s Department of Radiology. https://med.nyu.edu/departments-institutes/radiology/research/ai-biomed…. Accessed August 19, 2023. 
Essential AI in Healthcare Statistics in 2023. https://zipdo.co/statistics/ai-in-healthcare/. Zipdo. Published June 30, 2023.  
Time Saved by AI in Healthcare Administration in Europe as of 2020. Statista. https://www.statista.com/statistics/1202254/time-ai-could-save-in-healt…. Accessed July 19, 2023. 

Editor’s Note: In this interview, Nancy Morgan, RN, BSN, MBA, WOC, reviews several artificial intelligence (AI) modalities that can augment wound and health care, assuage provider’s fears of adopting these new technologies, and some considerations for integrating these technologies in one’s practice.

How Can AI Augment Clinical Judgement? from HMP on Vimeo.

Transcript

How can AI help or augment clinical decision-making in a wound care practice?

AI speeds up the decision-making process by rapidly processing data and providing relevant insights. Clinicians can focus on patient care and less on manual data analysis. By using an AI, it’s going to help give us kind of an early warning system. The AI can actually identify potential complications or risks, allowing clinicians to intervene sooner.

Consistency and standardization: AI technology can ensure consistency in wound assessment and documentation, reducing the variability among different clinicians, and promoting standard of care.

For providers who are hesitant to incorporate AI systems into their practice, what advice would you give?

So, my advice that I would give other providers, clinicians, in regards to AI, well, I would say just start slowly, okay, baby steps. I want you comfortable, but this is a direction that we all have to go. If you want to have a career and you still got several years ahead of you to work, time to jump in the game, guys, don’t be afraid. Remember when the computer came out years ago? Some of you probably don’t.

Well, we all had to get with the program then and start working on computers, and now look what it has done to us. It’s just changed our lives. So, think of AI the same way. So, start slowly. I want you to educate yourself. Take time to understand how different AI systems work.

Involve your patients. Some of your patients are pretty techy savvy as well. Why not? If they’re interested in that, that kind of brings them into the whole care plan of things as well. I would talk to them about data privacy. Assure patients that their personal and medical data will be handled with the utmost care and adherence to privacy regulations. Address any concerns that they might have about data security.

Another thing that I would tell clinicians is gradually integrate certain AI technologies, you know. Learn an AI tool, get comfortable with you and your team, get some feedback from your team and your patient, and then, hey, maybe it’s time to add another tool. And just kind of do some gradual waves of integration, so it’s not overwhelming to you and your office.

And also, I would always try to keep a feedback loop just for my staff so that way they can provide any kind of feedback or any certain experiences that we can enhance. Maybe we can do a better job on. So that way, it’s open door for your employees. And if the patients want to say something to it, we should actually document that as well.

How do you ensure patient trust in AI?

Transparent communication. Be upfront with your patients on the role of AI in their care. Explain to them that this is a tool that assists you, it’s not to replace you, or their doctor, it’s just to help them with their individual needs and preferences. That’s why we use these tools, to help customize the care for that patient.

Also, another thing that we should do is, let’s demonstrate AI’s value not only to our patients but also to our colleagues that maybe are scared. Maybe they’re just like, oh, they don’t want to learn something new. Oh, they’re going to be retiring soon. “I don’t want to get into something new.” This is easy. If I can do it, you can do it.

How can clinicians ensure the AI technology they incorporate into their practice does not put them at risk?

Make sure that you do a thorough evaluation before adopting any AI technology. Conduct a thorough assessment. Consider its purpose, its benefits, its potential risk. Is it compatible with your practice workflow? Does it make sense with the nature of your business?

How Much Do You Know About Artificial Intelligence in Health Care? Take our quiz to find out! Click here.

Vendor due diligence. Research the AI technology provider extensively. Check their reputation, experience, client reviews, and whether they adhere to medical regulations.

Clinical validation. Ensure that AI technology has been clinically validated and tested in real-world scenarios relevant to your specialty. Published research and peer-reviewed studies can provide valuable insights. Integration with the EHR. Ensure that the AI system that you’re using can integrate with your electronic health record. Make sure it’s going to facilitate efficient information exchange and reduce the risk of errors.

Training and education. Well, invest in proper training for yourself and your team to effectively use AI technology. This training helps prevent misuse or misinterpretation of its outputs. Regularly assess the AI system performance and outcomes. If you notice any inconsistencies or errors, address them promptly and document your actions. Collaboration. Involve your IT department, tech experts, and hospital administration that can offer insights into implementation, security, and risk management.

About the Speaker 

Nancy Morgan, RN, BSN, MBA, WOC is an experienced clinician, successful business leader, and accomplished educator in the field of wound management. She is the co-founder of the Wound Care Education Institute, (WCEI®) and Wild on Wounds Productions. Nancy is a distinguished wound care educator, delivering nearly 1200 lectures, conference keynote addresses, seminars, webinars, and bedside consultations for more than a quarter-century.