Using Artificial Intelligence in the Medical Field

Full-Length Interview with Dr. Sameer Shaikh: How He Uses Artificial Intelligence in the Medical Field

Dr. Sameer Shaikh is a practicing ER and ICU physician in Toronto, Canada, adept at using artificial intelligence in the medical field! He’s pursuing his Master’s in AI in Healthcare and has kindly shared insights on applying AI in his daily practice and teaching. Without further ado, here’s what he had to say:

1. What are some AI tools you use regularly, and how did you discover them?

ChatGPT

In my daily practice, I leverage several AI tools that significantly enhance both clinical care and medical education. A cornerstone among these is ChatGPT. Its accessibility and widespread adoption make it an essential AI tool in healthcare. The free version of ChatGPT offers substantial functionality, while the premium features, available for a modest $20 per month, provide expanded capabilities. The upcoming GPT store is particularly noteworthy. It allows for the creation of personalised GPTs without any coding knowledge, transforming how we access and retrieve information. We all have trusted open-access resources and personal notes; now, these can be integrated into custom GPTs within minutes. These personalised GPTs can be used at the bedside, offering immediate access to trusted information without the need to scroll through articles. I have created several GPTs myself and have even produced a tutorial video on this process. 

Monic

Another tool I see great utility for is Monic.AI, which is exceptionally beneficial for junior learners and residents. Monic.AI enables the creation of summaries, flashcards and quizzes on any given topic with ease, revolutionising the way we prepare for exams and reinforce learning. Simply upload the document you want to learn from, and this AI-driven platform creates your study materials in an efficient way so that you can spend most of your time learning instead of creating content. 

In the realm of clinical practice, ambient AI is a burgeoning field with numerous applications. 

Nabla

Tools like Nabla offer a practical starting point for physicians looking to integrate AI into their practice. Nabla is an ambient AI software that can automatically produce your consultation note by simply listening to your conversation with a patient. These tools are HIPAA compliant and often available for trial at no cost. For instance, Nabla allows for 30 free consultations per month, providing a hands-on opportunity to evaluate the strengths and limitations of AI in real-time clinical settings.

How to find new uses of AI in medicine

For those seeking to explore a wider range of AI tools, Futurepedia is an invaluable resource. It allows users to search for specific tools, save favourites, and stay updated on the latest developments. Additionally, following thought leaders in healthtech and AI on platforms like LinkedIn is an excellent way to stay informed about new tools and advancements in the field.

In conclusion, the integration of AI tools such as ChatGPT, Monic.AI, and Nabla into medical practice and education not only streamlines various processes but also opens up new avenues for innovative learning and patient care. These tools exemplify the practical application of AI in healthcare, offering tangible benefits to both practitioners and learners.

2. A particularly interesting use of artificial intelligence in the medical field is your use of it to craft teaching cases for your residents. How do you do it?

In crafting teaching cases for residents, I have found AI, particularly ChatGPT, to be an invaluable asset. My process involves several innovative approaches that leverage AI to enhance the learning experience.

Clinical case discussions

Firstly, for clinical case discussions, I use AI to develop learning points tailored to the case and the learner’s level. For instance, in a complex ICU scenario like Metformin-Associated Lactic Acidosis (MALA), I input an anonymized clinical stem into ChatGPT. The AI then generates key learning points appropriate for the learner’s level. This not only aids in reinforcing my understanding of the topic, as I can request a simplified explanation of the pathophysiology, but also provides a quick and efficient way to prepare for bedside teaching. It’s crucial, however, to verify the AI-generated content with trusted sources, especially in areas where I need a refresher myself.

Simulation cases

Another application is in creating simulation cases. As the leader of a simulation program catering to diverse learners, I utilise ChatGPT to design cases around specific topics and objectives. The AI can introduce complexities to the scenarios and suggest how the simulation should flow and create a list of debriefing points. This is particularly useful in simulating real-world clinical situations and preparing residents for a variety of challenges they might face.

My approach with AI is to use it as a starting point. As a practising clinician with my own expertise and insights, I find that ChatGPT excels in organising my thoughts and ideas into coherent and structured teaching materials. It effectively overcomes writer’s block and provides a solid foundation, which I can then refine and adapt. This process saves hours that would otherwise be spent on initial drafting.

A caution regarding the use of AI for teaching and learning

A word of caution is necessary when relying on AI-generated content. It’s important to cross-check the information with reliable sources to avoid potential inaccuracies. A practical solution to this is to provide the AI with specific documents or reference materials from which to base its responses, ensuring the information aligns with trusted and verified sources.

In summary, AI tools like ChatGPT have revolutionised the way I develop teaching cases, offering a blend of efficiency, customization, and adaptability that is invaluable in medical education.

3. Can you provide an example of a teaching case where AI played a pivotal role?

A prime example of AI’s pivotal role in crafting teaching cases can be seen in our ICU simulation sessions. We frequently have residents from various specialties, such as Obstetrics/Gynecology, rotating through the ICU. To tailor the simulation experience to their specific learning needs and objectives, I turned to AI, specifically ChatGPT, for assistance.

On one occasion, I requested ChatGPT to generate three case scenarios suitable for junior Obstetrics residents in an ICU setting. The AI provided three well-constructed cases, from which I selected the one that I thought was most relevant and high-yield for our rotation. 

The chosen case centred on the management of a massive pulmonary embolism (PE) in pregnancy, a critical topic for Obstetrics residents in this setting.

ChatGPT not only created the initial case stem and a brief outline for the simulation but also played a crucial role in the debriefing process. I provided the AI with a document I had prepared on treating pulmonary embolisms, including key clinical pearls I have curated through years of practice. ChatGPT then efficiently summarised key teaching points from this document, which I used in the debriefing session. 

This integration of AI in the preparation process significantly streamlined the creation of educational content, allowing me to develop a comprehensive teaching case in a matter of minutes.

The residents found the simulation experience extremely valuable, highlighting the effectiveness of AI in enhancing medical education. This instance is a testament to the potential of AI tools in augmenting the workflow of busy clinical practices and balancing multiple responsibilities. While AI will not replace physicians, it certainly has the capacity to augment our capabilities in various ways, particularly in the realm of medical education and training.

4. In what ways has AI increased your efficiency when creating cases for teaching your residents? 

AI, particularly ChatGPT, has significantly enhanced my efficiency in crafting teaching cases for residents. Its vast repository of knowledge and ability to quickly synthesise information make it an invaluable starting point in the development of educational content.

When creating teaching cases, I often begin with a set of initial ideas or clinical experiences. Transforming these into structured, concrete cases or teaching points that are tailored to the learner’s stage is where ChatGPT truly shines. While there are tools available for accessing pre-curated simulation cases, ChatGPT’s unique advantage lies in its ability to take my personal clinical experiences and insights and mould them into practical, customised teaching cases. This capability is particularly beneficial for educators who prefer to create content around topics they find intriguing or relevant to their specific clinical environment.

ChatGPT’s efficiency extends to bedside teaching

In the face of unique cases or rare presentations, ChatGPT assists in rapidly organising my thoughts and observations. By feeding the AI trusted documents and resources, it can swiftly generate coherent and relevant teaching material. This is especially useful in real-time clinical settings, where quick access to information and teaching points can greatly enhance the learning experience for residents.

In summary, AI tools like ChatGPT have revolutionised the way I prepare and deliver educational content. They enable me to efficiently convert ideas and experiences into teaching materials that are both relevant and engaging, thereby enriching the educational journey of my residents.

5. Have you run into any limitations of this particular use of artificial intelligence in the medical field that your colleagues should be aware of?

While AI tools like ChatGPT have been transformative in crafting teaching cases and educational content, it’s crucial to acknowledge their limitations, especially in the context of medical education and practice.

Verification is needed

One significant limitation is the need for thorough proofreading and verification by the clinician. 

AI-generated content, while often insightful and comprehensive, is not infallible. These tools have not been specifically designed for medical use cases and should not be equated with dedicated medical resources like UpToDate or DynaMed. 

Anything created using ChatGPT or similar AI resources should be meticulously edited and fact-checked against trusted knowledge and resources. It’s essential to approach AI-generated content with a degree of scepticism and not accept it at face value.

No HIPAA compliance

Another major limitation to consider is that ChatGPT is not HIPAA compliant. Therefore, when using this tool, it is imperative to never share personal patient information or identifying details. This constraint is crucial for maintaining patient confidentiality and adhering to legal and ethical standards in healthcare.

It’s a tool that requires your critical eye

I view ChatGPT as a kind of personal assistant or scribe – a tool to help gather and organise information, facilitate meaningful discussions, and create content. However, it’s vital to maintain a critical eye and not rely solely on its output. This is particularly true for seasoned educators and physicians who, thanks to their training and experience, can easily identify inaccuracies or errors in the content provided by AI.

Concerns with junior learners

My concern with AI tools extends to their use by junior learners, who may not yet have a robust foundation in clinical knowledge. These learners might find it challenging to discern when AI is producing erroneous or misleading information. For them, I recommend using AI to summarise or organise information from already verified and trusted sources, rather than as a primary source of knowledge.

In summary, while AI tools like ChatGPT offer significant advantages in medical education, it’s imperative to use them judiciously and always cross-check their output with established medical knowledge and resources, and to be mindful of their non-compliance with HIPAA when discussing patient-related information.

6. Are there any other thoughts you’d like to share?

AI in healthcare represents a frontier of immense potential, extending far beyond the realm of medical education. As healthcare professionals, we are acutely aware of the myriad challenges inherent in our work and the systems we navigate. Just as critical appraisal of research papers has become a fundamental skill in modern medicine, understanding and utilising AI tools safely and ethically is fast becoming an essential competency for healthcare professionals.

A generational shift is coming…

We are at the cusp of an AI generational shift, a wave of innovation comparable to the advent of the internet in its capacity to revolutionise medicine and knowledge sharing. Andrew Ng likened AI to electricity, underscoring its vast potential and applicability across thousands of use cases. It is incumbent upon us in the healthcare sector to identify and harness these use cases, leveraging AI to enhance our work and the care we provide.

Explore the uses of AI in medicine

I encourage my colleagues to start exploring AI, beginning with applications in education and gradually integrating tools into clinical practice. This exploration should be marked by curiosity and a commitment to safety and ethical considerations. Engaging with AI requires a proactive approach to learning; there is a wealth of online resources and materials available to deepen our understanding of AI and its workings. Following blogs or thought leaders on platforms like LinkedIn can provide valuable insights and keep us abreast of the latest developments.

Stay informed

As we navigate this transformative era, staying informed and adaptable will enable us to not only improve the efficiency and quality of our work but also to redefine the paradigms of healthcare delivery and patient care. The journey into AI in healthcare is not just about adopting new tools; it’s about evolving with the technology to become more effective, insightful, and compassionate caregivers.

References

Educational Resources

• Guilleminot S. AI in Healthcare. LITFL
• ChatGPT Essentials for Clinicians. Medmastery
• The Neural Medwork
• Artificial Intelligence in Healthcare: Master of Science in Applied Computing (MScAC)