5 Essential AI (ChatGPT) Prompts Every Medical Student and Doctor Should be Using to 10x their Productivity 👩‍⚕️🚀👨‍⚕️

In today’s rapidly evolving medical landscape, Artificial Intelligence (AI) is emerging as a significant catalyst for change. For medical students, physicians, and healthcare professionals alike, grasping the integration of AI into medical education and practice has become crucial. Drawing on insights from my conversations with Shiv Gaglani, CEO of Osmosis, and Dr Avneesh Khare, AI expert, this blog post aims to illuminate the transformative potential of AI. We’ll provide specific AI prompts that can enhance your medical education and clinical practice, enabling you to navigate the future of healthcare confidently.

The ripple effect of AI’s influence in medicine extends well beyond the confines of research laboratories and clinical settings. It’s triggering a change in medical education, offering personalized learning experiences and bolstering clinical decision-making capabilities.

“The applications of AI in medicine are just so vast that it’s going to be a requirement for every single health professional to have some basic understanding of AI and data science.” — Shiv Gaglani

Let’s delve into some specific AI prompts that I have created and use at medical school.

AI Prompts💡📚

Think of AI prompts as your key to unlocking the vast reservoir of knowledge housed within AI models. These prompts enable you to guide the AI, specifying the task at hand, and in turn, securing a relevant and precise response. In the following sections, I present carefully crafted AI prompts designed to make your medical education more engaging and clinical practice more effective.

Photo by Levart_Photographer on Unsplash

From mastering complex medical concepts to drafting research abstracts and generating patient cases, these prompts can redefine the way you learn and practice medicine.

Generating Flashcards 🎴

Flashcards are a proven method for facilitating active recall, a key component of effective learning. Many medical students will spend more time making their flashcards than actually learning them!

Ali Abdaal (Youtuber) demonstrating how to make ANKI flashcards for medicine

The introduction of AI can revolutionize your learning experience by creating personalized flashcards, emphasizing critical points, and saving YOU valuable time.

Here is an example of a BAD prompt which is generic and not very specific:

“Generate a set of flashcards on (insert medical topic). Each card should include the disease’s definition, common symptoms, risk factors, diagnostic tests, and current treatment guidelines.”

Here is the response I received having inputed this prompt:

While the generated flashcards provide information on the requested topic, they may not encompass the comprehensive depth of knowledge expected for medical students preparing for their exams. Furthermore, the flashcards may lack engagement and contain excessive information per card, which can hinder effective learning.

Here is an example template of a GOOD prompt which I created for my medical school revision using my intended learning objectives (feel free to copy and paste to try yourself):

“As a final year medical student, you are preparing for your imminent exams, and revision is essential. To aid your studies, generate 50 Anki-style flashcards covering (insert topic). These flashcards will serve as a comprehensive review tool to test your knowledge and understanding. Each flashcard should be designed in a question-and-answer format, focusing on specific areas of medical knowledge.

The flashcards should address the following aspects for each medical topic during your revision:

• Definition and in-depth understanding of the disease or condition
• Key clinical and presenting features characteristic of the condition
• Relevant anatomical and physiological considerations related to the condition
• Important positive and negative aspects of the patient’s history that contribute to the diagnosis
• Abnormal clinical signs and findings associated with the condition
• Understanding the underlying causes and their role in formulating a diagnosis
• Application of risk assessment tools to evaluate the risk of disease development or complications
• Knowledge of relevant basic investigations and the ability to interpret their results
• Identification of the most appropriate further investigations or imaging modalities
• Formulation of a comprehensive management plan, including emergency measures if applicable
• Description of common therapeutic interventions, both pharmacological and non-pharmacological
• Understanding of specific medications, their routes of administration, mechanisms of action, and common side effects
• Awareness of disease prevention and health promotion strategies
• Familiarity with common symptoms, risk factors, diagnostic tests, and current treatment guidelines in the UK

Instructions:

You aim to revise: (insert medical topic).

Craft a question-and-answer flashcard based on the provided components. Each flashcard should detail the disease’s definition and in-depth understanding, key clinical and presenting features, relevant anatomical and physiological considerations, important positive and negative aspects of patient history, abnormal clinical signs and findings, understanding of the underlying causes and diagnostic formulation, relevant risk assessment tools, basic investigations, further investigations or imaging modalities, formulation of a management plan, common therapeutic interventions, specific medications, disease prevention and health promotion strategies, as well as familiarity with symptoms, risk factors, diagnostic tests, and treatment guidelines.”

Now look at the response I received having inputed the GOOD prompt:

This prompt exemplifies the depth and specificity achievable with AI. It’s comprehensive, covering a broad range of patient care components. It demonstrates the AI’s versatility in meeting your unique learning objectives and the depth of knowledge it can encapsulate.

With such a detailed prompt, the AI doesn’t merely provide facts; it curates targeted information, helping you focus on what’s most important. It allows the creation of flashcards that not only centre on disease-specific knowledge but also highlight the full scope of patient care — from disease prevention strategies to management plans, from basic investigations to risk assessment tools.

Such a detailed prompt also serves as a guide for other medical students and doctors aiming to replicate this method. The primary takeaway here is to be as specific as possible with your prompt. Clearly identify your learning objectives, specify the style of flashcards you prefer, and articulate the topics you want the flashcards to cover.

Creating Exam-Style Questions❓

Questions stimulate critical thinking and enhance memory recall. This is due to the phenomenon that when we anticipate the need to teach or explain something to others, we tend to learn more effectively. Question banks like Passmed and Quesmed only have a limited number of questions per topic. I created this prompt so I could have an UNLIMITED question bank without having to pay for more subscription services.

Example of SBA exam style question form medical question bank “passmedicine”

An AI, like ChatGPT, can generate a diverse set of questions, tailored to your learning needs, thereby enhancing your understanding and retention of knowledge. This method not only solidifies your grasp on the material but also ensures you are well-prepared to apply the knowledge in a clinical setting or an examination.

Here is another example of a BAD prompt which is generic and not very specific:

“Generate a set of multiple-choice questions on (insert topic). Include questions on anatomy, physiology, common diseases, diagnostic tests, treatments, and patient care.”

Here is the answer I got after inputting this prompt:

Despite these generated questions still serving as good revision, the questions are very different to the Single Best Answer (SBA) style questions we see in medical school exams.

Here is an example template of a GOOD prompt which I created to have an unlimited SBA style question bank (feel free to copy and paste to try yourself):

“You are developing a question bank for medical exams focusing on the topic of [insert topic]. Generate a high-quality Single Best Answer (SBA) question using the following framework:

Stem:

1. Patient details (gender/age)
2. Presenting complaint (PC)
3. Relevant clinical history (HPC)
4. Physical examination findings
5. Diagnostic test results (optional)

Question stem: [Insert relevant information from the above sections]

Acceptable question style: Ask for the BEST answer, NOT one that is TRUE/FALSE, for example:

Answer options:

1. [Insert plausible answer option]
2. [Insert plausible answer option]
3. [Insert plausible answer option]
4. [Insert plausible answer option]
5. [Insert plausible answer option]

Explanation:

• Identify and explain the correct answer.

• Explain why this is the most appropriate answer based on evidence-based guidelines or expert consensus.

• Briefly explain why the other answer options are less correct or incorrect.

Difficulty level: [Insert difficulty level (e.g., easy, moderate, difficult)]

Using this framework, medical students can generate a variety of high-quality SBA questions for their medical exams by replacing the relevant information in the stem and answer options. Ensure that the questions are focused, relevant, and effectively test the candidate’s knowledge.”

Now look at the response I received having inputed the GOOD prompt:

This prompting framework is designed to generate comprehensive, targeted SBA questions for your medical exams. You can adjust the topic, stem, and answer options according to your specific learning needs. It guides the AI to understand the style and format of the question you want to create. The AI then utilizes this information to generate a question that is focused, relevant, and effective in testing your knowledge.

The significance of creating high-quality questions cannot be overstated. Not only do they push you to remember, understand, and apply what you’ve learned, but they also mimic the style of medical school examinations, preparing you for the real thing. This is an active learning approach, which is proven to be more effective than passive learning methods.

By using this prompt, you are guiding the AI to create a question bank that aligns with your learning objectives, helping you to be more effective and efficient in your studies.

Simplifying Explanations 💡

Medicine is a field replete with complex terminologies and concepts that can sometimes be overwhelming. AI can assist by breaking down these complexities into digestible chunks, making it easier for you to understand and remember.

Concept map example for Hypertension

Many students like to create concept maps to help solidify large chunks of informatio.. I have created this prompt to break down medical topics into clear and concise concept maps based on intended learning objectives for students.

Here is an example of a BAD prompt which is generic and not very specific:

The information generated is not very focused and difficult to digest (usually similar to what is found in medical textbooks 😂)

Here is an example template of a GOOD prompt which I created to help students and doctors create concise concept maps on medical topics based on our intended learning objectives (feel free to copy and paste to try yourself):

“You are creating concept maps for different medical topics to aid in the revision process for medical students.

Based on the below prompt template, please create a concept map for (insert topic):

Concept Map Prompt Template:

Context:

• Provide a brief introduction to the medical condition, including its prevalence, etiology, and impact on patients’ health.

Disease Overview:

• Definition: Provide a concise definition of the medical condition.

• Key clinical and presenting features: List the main clinical symptoms and signs associated with the condition.

• Relevant anatomical and physiological features: Describe the specific anatomical and physiological aspects related to the condition.

Differential Diagnoses:

• List and briefly describe the key differential diagnoses to consider when evaluating a patient with similar clinical features.

Patient History:

• Important positive aspects: Identify relevant positive factors contributing to the diagnosis or understanding of the condition.

• Important negative aspects: Highlight significant negative factors ruling out other conditions or providing important context.

Clinical Signs:

• Abnormal clinical signs: List the notable physical signs or abnormalities commonly observed in patients with the condition.

Investigations:

• Relevant basic investigations and interpretations for each.

Management Plan:

• Stepwise approach to management, including lifestyle modifications, pharmacological interventions, surgical interventions, and other interventions.

Disease Prevention/Health Promotion:

• Strategies for preventing or reducing the risk factors associated with the condition.

Current Guidelines:

• Current treatment guidelines: Refer to the latest guidelines from reputable sources specific to the condition. Summarize the key recommendations for managing the condition.”

Now look at the response I received having inputed the GOOD prompt:

The strength of this prompt lies in its structure and specificity. By asking the AI to break down the topic into defined sections, you guide it to deliver a response that is not only comprehensive but also well-organized, enhancing your understanding and retention of the topic.

Moreover, this prompt acts as a blueprint for tackling other complex medical topics. You can replace “insert topic” with any other medical condition or concept, and the AI will generate a similar structured breakdown, aiding your learning process. This method can be particularly effective when used consistently over time, allowing learners to build a personalized knowledge base in a format that suits their learning style.

In this way, AI doesn’t merely provide information; it functions as a personalized tutor, helping you navigate through complex medical knowledge and grasp difficult concepts with ease. It also tailors its responses to your specific needs, based on the information you provide about your knowledge level and learning objectives. This makes your study sessions more efficient and effective, contributing to better learning outcomes.

Generating Patient Cases 🩺

Patient cases are critical learning tools in medical education. They provide a context in which you can apply theoretical knowledge, enhancing understanding and retention. Moreover, they help you develop clinical reasoning and decision-making skills, preparing you for real-world clinical practice.

I would find it annoying on certain weeks learning rarer conditions and NOT finding a patient on the ward to take a history from. Therefore I created this prompt so that chatGPT could act as a virtual patient presenting with a certain condition to help with my learning.

Here is an example template of this GOOD prompt in action(feel free to copy and paste to try yourself):

“You are an AI model designed to generate detailed patient cases for medical education and clinical practice. Your task is to create a comprehensive patient case simulating real-life scenarios to facilitate clinical reasoning, decision-making, and care planning.

The case should be unidirectional, allowing learners to apply their knowledge in a clinical context and practice their skills in a safe and controlled environment.

Instructions:

• Condition: (insert topic)

• Develop a detailed patient history, including relevant demographic information, medical background, and presenting symptoms.

• Describe the patient’s physical examination findings, highlighting any abnormal clinical signs.

• Consider significant positive and negative aspects of the patient’s history that contribute to formulating a diagnosis.

• Incorporate risk assessment tools and relevant anatomical/physiological features to evaluate the risk of disease development or complications.

• Include a range of appropriate investigations, both basic and advanced, and provide interpretations for each.

• Formulate a management plan based on the findings, including therapeutic interventions, pharmacological and non-pharmacological approaches, and an emergency plan if applicable. • Discuss common medications used in the treatment of the chosen medical condition, including their routes of administration, mechanisms of action, and potential side effects.

• Address disease prevention and health promotion strategies, emphasizing general lifestyle measures and pharmacological control of risk factors.

• Reference current treatment guidelines or best practices specific to the chosen medical condition.

• Ensure the patient case is detailed, realistic, and challenging, promoting critical thinking and clinical decision-making skills.

• Remember, your goal is to provide an invaluable learning experience by generating a comprehensive patient case that allows learners to apply their knowledge in a clinical context.”

This prompt guides the AI to produce a comprehensive patient case, simulating real-life scenarios. It enables you to practice clinical reasoning, decision-making, and care planning in a safe and controlled environment.

This approach doesn’t just offer information, but also an engaging learning experience. By creating a dynamic patient case, the AI encourages active learning, allowing you to think critically, make decisions, and plan patient care in a controlled environment. This is an invaluable tool for honing your clinical skills and preparing for real-world practice.

BONUS: Assisting with Research Papers/Abstracts 📝

Writing research papers, and abstracts in particular, can be challenging due to the requirement for brevity and precise language. Artificial Intelligence (AI), like ChatGPT, can significantly aid this process by generating ideas, editing for grammar and style, suggesting relevant references, and enhancing the structure and flow of your paper.

The AI’s ability to assist in drafting an abstract for a systematic review research paper is particularly noteworthy. Abstracts are crucial as they offer the first impression of your work to readers. A compelling abstract can increase the visibility and impact of your research, summarizing key aspects of your paper and enticing readers to explore the full paper.

Here is again another example of a BAD prompt which is generic and not very specific — the AI just made up an abstract:

Here is an example template of a GOOD prompt which I created to help me structure an abstract based on a systematic review following PRSIMA guidelines (feel free to copy and paste to try yourself):

“AI, your task is to serve as a reviewer and abstract writer for a systematic review on a specific topic. Your job is to analyse the provided systematic review and generate a high-quality abstract based on the PRISMA (Preferred Reporting Items for Systematic reviews and Meta-Analyses) guidelines for Abstracts.

Systematic Review Title: [Provide the title of the systematic review]

Systematic Review Background: [Provide a brief overview of the background and context of the systematic review]

Systematic Review Methods: [Provide a summary of the methods used in the systematic review, including eligibility criteria, information sources, risk of bias assessment, and other relevant details]

Systematic Review Results: [Summarize the main findings and results of the systematic review, including the number and type of included studies, synthesis of results, description of effect sizes, and any other important outcomes]

Systematic Review Discussion: [Discuss the strengths and limitations of the evidence, provide interpretations of the results, and highlight important implications]

Systematic Review Funding: [Specify the primary source of funding for the systematic review]

Systematic Review Registration: [Provide the registration number and registry name, if applicable]

Based on this information, generate a high-quality abstract for the systematic review following the PRISMA for Abstracts guidelines. The abstract should be structured, concise, and accurately reflect the key aspects of the systematic review.

PRISMA for Abstracts Template:

Please use the following structure for your abstract and ensure that all necessary items from the PRISMA for Abstracts checklist are included:

Section 1: TITLE Item 1: Title.

Section 2: BACKGROUND Item 2: Objectives.

Section 3: METHODS Item 3: Eligibility criteria. Item 4: Information sources. Item 5: Risk of bias assessment.

Section 4: RESULTS Item 6: Included studies. Item 7: Synthesis of results. Item 8: Description of effect.

Section 5: DISCUSSION Item 9: Strengths and limitations of evidence. Item 10: Interpretation. Section 6: OTHER Item 11: Funding. Item 12: Registration.

Ensure that the abstract accurately reflects the key aspects of the systematic review and includes all necessary items from the PRISMA for Abstracts checklist.

Note for prompter: Please attach the full systematic review document for analysis and abstract generation.”

This prompt positions the AI as a valuable collaborative tool in crafting a concise, informative, and engaging abstract. The AI helps articulate your ideas, ensures language accuracy, and guides you towards relevant literature, enhancing the quality of your work significantly. By following this template and adjusting the systematic review details, you can generate abstracts for various research papers, thereby streamlining your research writing process.

DISCLAIMER: USING CHATGPT TO PASS OFF WORK AS YOUR OWN IS PLAGARIASM! YOU SHOULD USE THIS PROMPT TO ONLY HELP YOU STRUCTURE AND CHANEL YOUR CREATIVITY/ THOUGHT PROCESS

Facing the Challenges and Welcoming the Future with Open Arms 🌄🤖

While AI carries the promise of transforming personalized learning and streamlining diagnosis processes, it’s equally vital to recognize and address the potential hurdles that accompany its application. Ethical considerations, data privacy concerns, and the risk of excessive reliance on technology are notable issues that must be carefully navigated as we further weave AI into the fabric of medicine.

“AI is going to be, and already is, an incredibly useful tool. But just like any tool, it has to be used correctly.” — Shiv Gaglani

Concluding Thoughts 🎁💡

The landscape of medical education and practice is vast, ranging from budding medical students to seasoned physicians. Everyone in this spectrum stands to gain immensely from incorporating AI into their workflow and learning protocols. By comprehending and effectively harnessing AI, we position ourselves at the forefront of medical advancements, ensuring our ability to deliver the highest standard of care remains steadfast.

As we stand on the brink of an exciting new era in medical education, let’s wholeheartedly embrace the transformative power of AI, revolutionizing healthcare one AI prompt at a time. Remain committed to continuous learning and growth and stay engaged for further insights on the role of AI in medical education. The future awaits, and it is ours to shape. 🚀🎓

P.S. If you found this article useful, I would really appreciate if you share to other medical students and doctors.

THANK YOU! THIS ARTICLE TOOK ME AGES!