Take Charge of Your Health—Use AI to Analyze Medical Records

Follow a step-by-step demonstration of building a new project in Claude designed to generate actionable health insights and doctor-ready reports. Learn prompting techniques like using capitalization or asterisks for emphasis, and see how to use Google Gemini to refine and improve your project's title and description.

Based on a suggestion for novice users, this brief lecture explains how the course materials are structured for easy reference. Discover how AI is used to generate clear, downloadable step-by-step text guides alongside the video lectures to help you easily replicate the technical processes shown on screen.

Learn how to pin projects for easy access and configure Claude's underlying engine for your tasks. This lesson breaks down the differences between Claude's models—such as Fable, Opus, Sonnet, and Haiku—and explains how adjusting the AI's "effort" level impacts the quality of results and your account's daily and weekly usage limits.

Discover a powerful prompting technique: asking the AI to guide you. This lecture shows how to instruct Claude to ask you clarifying questions so it can generate the best possible project parameters, while Jen shares a real-world example of using AI as a daily coaching tool to monitor activity and pain levels during surgical recovery.

Learn essential keyboard shortcuts for interacting with AI, such as holding down Shift and pressing Enter to create line breaks without accidentally submitting your prompt. Discover how to enable browser notifications so you can multitask while Claude processes information, and understand why providing a precise title, description, and instructions is crucial to focusing the AI's generalized knowledge into a specific expert role.

Add the AI-generated instructions to your Claude project and explore the basics of Markdown formatting, learning how to use asterisks for bolding, hashtags for headings, and dashes for bulleted lists. This lecture briefly introduces the Claude Desktop app (also referred to as Coworker) and demonstrates how to upload a screenshot of your current project details to ask Claude for improved title and description options.

Finalize your project setup by uploading your exported medical records from portals like Stanford, MyChart, and the Mayo Clinic. Learn why uploading PDF files is often preferable to plain text documents, as PDFs preserve the original formatting and can make it easier for the AI to process the clinical data.

Begin querying your health data to generate practical recommendations, ranging from simple lifestyle habits like sleep hygiene to complex insights you can print and review with a doctor. This lesson explores Anthropic's health portal connector but recommends manually uploading PDFs instead to build a continuous, durable timeline of your medical history. You will also learn how to monitor your account's capacity limits and adjust the AI's effort settings based on your daily and weekly usage.

Conclude the first segment of the course as Jen shares her final reflections on navigating complex diagnoses across multiple healthcare networks, including Stanford, UCSF, and Mayo. Discover how using AI to consolidate and instantaneously summarize 95-page medical files empowers you to act as your own case manager, helping you catch crucial details that time-constrained primary care providers might overlook.

Explore the security implications of uploading sensitive personal health data to artificial intelligence platforms. This discussion weighs the theoretical risks of data exposure and weaponization against the substantial personal benefits of gaining deeper medical insights, offering a pragmatic perspective on privacy in the modern digital age.

Prepare for the next phase of the course featuring a new case study with retired professor Dr. Rich Barrett. This segment will reinforce the foundational skills you have already learned by walking through the setup process again, this time utilizing the Claude Desktop application and troubleshooting new technical challenges along the way.

Meet Dr. Rich Barrett and learn how the shift from the "Information Age" to the "Age of Understanding" solves the modern problem of information overload. Discover how AI's ability to synthesize vast amounts of medical data can identify critical oversights—such as contraindicated medications—and learn how dedicated AI projects can focus this computational power to improve healthcare outcomes.

Understand the critical responsibility of being the steward of your own health and recognize that this course provides educational technology training, not medical advice. Learn the "Human in the Loop" (HITL) framework, which mandates that you must always bring the insights and reports generated by AI back to qualified medical professionals for final evaluation and clinical decision-making.

Listen as 82-year-old Dr. Rich Barrett shares his professional background as a Fresno State professor and marriage and family therapist, alongside his extensive personal medical history. Dr. Barrett discusses managing a genetic deficit regarding heart disease, his quadruple bypass in 2000, a prostate cancer diagnosis, and undergoing approximately 20 surgeries over his lifetime, setting the stage for how AI can manage complex, long-term health records.

Learn how AI can help you navigate complex medical concepts like genetic testing and DNA analysis. Listen as Todd emphasizes that he is an educator, not a medical professional, while demonstrating how to use Gemini to find current pricing for DNA testing services like 23andMe and AncestryDNA. Discover how uploading these genetic test files as context to your AI project can uncover new health markers and tests your medical team may not have considered.

Step back from the technical setup to explore the broader context of human history using Hans Rosling’s Gapminder tool. This video reviews an interactive graph showing the dramatic increase in life expectancy and global wealth since the 1800s, offering a moment of gratitude and perspective for modern medicine while managing personal health challenges.

Discover the instructor's personal rankings of the top three AI chatbots currently available. Learn why Anthropic's Claude is recommended as the primary tool for detailed project work, why Google's Gemini serves as a strong backup for general questions and drafting emails, and how OpenAI's ChatGPT—while highly capable for coding tasks—often presents information in a different style and format.

Address the costs associated with using AI tools effectively. Dr. Barrett asks about pricing, and the instructor explains that while free trials exist, running complex, resource-heavy projects will generally require a paid membership, typically around $20 per month, which unlocks sufficient daily processing limits for most users without needing the $200 professional tier.

Move beyond the web browser and learn how to download and install the dedicated Claude Desktop application (Coworker) on a Windows computer. The video covers locating the downloaded setup file in Chrome, completing the installation process, granting necessary system permissions, and demonstrates how to use Gemini to confirm that a Mac-compatible version is also available.

Follow along as Dr. Barrett creates his own Anthropic account to use with the newly installed Claude Desktop app. This step-by-step guide covers entering a Google email address, opting to skip the Pro upgrade initially, managing data privacy settings by declining to share chats for model training, and verifying the new account via email.

Learn how to configure a new project container specifically for Dr. Barrett's health records using the Claude Desktop application. This lesson demonstrates how to draft a project summary and then use the "Alt+Tab" shortcut to quickly copy and paste that text into Gemini, asking Gemini to refine the intent and generate a polished project title and description to paste back into Claude.

Take a guided tour of the Claude Desktop layout to understand how it organizes your work. Learn the distinction between general "chats" on the home screen and project-specific "tasks," and discover how to "pin" a project so that it remains easily accessible on the left-hand navigation menu.

Embrace the mindset required to successfully navigate rapidly changing technology. In response to Dr. Barrett's question about the availability of the "Projects" feature on Claude's free tier, the instructor encourages adopting an "Indiana Jones" spirit of adventure, urging users to independently explore the software's capabilities and test solutions themselves.

Discover a "meta" approach to generating custom project instructions by using one AI to set up another. Learn how to use the Windows Snipping Tool to capture an image of the blank Claude project screen, upload that screenshot into Gemini using a split-screen setup, and command Gemini as a "CEO" to automatically write the specific role, tone, and safety guidelines needed to focus Claude's analysis.

Take a brief pause with Dr. Barrett as he reflects on the intimidating process of learning new technology. Discover how the course is designed to alleviate this pressure by allowing students to pause, rewind, and re-watch the recorded steps at their own pace to practice the demonstrated techniques.

Understand what "context" means in the realm of AI and why it is critical for accurate project results. This lesson explains the concept of a "context window," comparing it to human working memory, to illustrate how AI models can hold and process millions of tokens—equivalent to multiple books' worth of information—at one time.

Begin adding medical context to Dr. Barrett's project. This video explains the general process of uploading files, noting an important disclaimer about data privacy and ethical AI use. Watch a demonstration of capturing a snippet of a MyChart portal using the Windows Snipping Tool and learn that Claude does not currently support .png image uploads as direct project context.

Troubleshoot the file format limitation discovered in the previous lesson. In response to Dr. Barrett's question about copying images into Word documents, the instructor demonstrates a simple workaround: using the "Print" function to save the MyChart snippet directly as a PDF, which Claude successfully accepts as project context.

Discover the new feature that allows seamless transition between Claude Desktop and Claude's web browser interface. This lesson demonstrates how to query Claude for instructions on connecting to Google Chrome, and walks through the steps of granting Claude access to your Gmail account so it can manage drafts, send emails, and read communications related to your health.

Encounter and resolve a real-world technical roadblock. When the settings for Claude's Chrome connector cannot be found, the instructor discovers that this feature requires a paid subscription. This brief video emphasizes the value of leaving troubleshooting moments in the course to show students how to navigate the actual challenges of setting up complex software.

With the $20-per-month Pro plan activated, follow the steps to successfully link Claude Desktop to Google Chrome. This lecture covers locating the "Claude in Chrome Connector" in settings, downloading the necessary extension from the Chrome Web Store, ensuring it is turned on, and temporarily disabling strict anti-tracking security software to allow the AI to browse sites on your behalf.

Learn the importance of dedicating 5 to 10 minutes to review and optimize Claude's general settings. The instructor walks through adjusting interface appearance, enabling AI-powered artifacts, granting network egress for package managers (dependency handling), and giving Claude system-level permissions to control the computer's mouse and keyboard to maximize its capabilities, while acknowledging the inherent security trade-offs of granting broad access.

Understand how to leverage the strengths of different AI models in combination. When Dr. Barrett asks about handling photographs, the instructor demonstrates how he used Claude—the superior text generator—to write a highly detailed image prompt, which he then fed into Google Gemini Images—the superior image generator—to create a custom LinkedIn banner for a professional project.

Learn advanced prompting techniques for instructing Claude to independently navigate Google Chrome and extract MyChart data. This lecture introduces three key prompt strategies: asking the AI for clarifying questions, requesting suggestions for improvement before execution, and asking the AI to recommend the most cost-effective "model" and "effort" level (such as Sonnet vs. Fable) to minimize unnecessary token usage.

Watch the Chrome extension in action as Claude begins to autonomously navigate the MyChart portal. This video demonstrates how to verify when Claude has finished generating a response by checking for the static logo, explains how to copy or print Claude's output, and walks through resolving permissions roadblocks to finally grant Claude continuous browsing access to your health records.

In response to Dr. Barrett's feedback about needing written reference materials, this quick update explains how the course provides AI-generated step-by-step documentation. The instructor outlines the process of exporting video transcripts into Adobe Premiere and using AI to consolidate them into a clear, downloadable guide available in the course resources.

Returning to the project after a ten-day break, the instructor uses the natural confusion of "where were we?" as a teaching moment. Discussing changes to the studio's lighting and acoustics, this video emphasizes that struggling, adapting, and iterating are inherently human processes—and that maintaining a grounded, human perspective is essential when dealing with highly capable AI systems.

Reorient yourself with the Claude Desktop interface after a break. Review how to check your account usage limits, verify project instructions, and select the appropriate AI model (like Opus or Fable) for your next task. This lesson also highlights the difference between casual "chats" and action-oriented "tasks," comparing AI agents to secret agents running missions on your behalf.

Pause for a brief, personal reflection with Dr. Barrett. He shares a powerful message about the necessity of maintaining hope—currently operating at an "85% full" tank—when facing severe medical challenges, emphasizing that hope provides the vital energy needed to continue pursuing treatment and maintaining a sense of self-efficacy.

Encounter a major technical issue: discovering that Claude failed to save the MyChart data it extracted during a previous browsing session. Learn how to troubleshoot this problem by directly questioning the AI about its missing context, searching past conversation logs for the data, and instructing Claude to update its project instructions to prevent future data silos.

Continue troubleshooting the siloed data issue. Using the Opus model at maximum effort, the instructor forcefully commands Claude to update the core project instructions to ensure that crucial information—like the 2008 medical records extracted in the previous session—is stored centrally across the entire project rather than isolated in individual task threads.

When automated AI browsing proves unreliable, pivot to a manual solution. This lesson demonstrates how to navigate the Epic MyChart "Sharing Hub" to request a full, machine-readable export of all medical visits. When direct printing is blocked, learn how to copy and paste the raw text into a Notepad file, save it, and manually upload it directly to Claude to ensure the context is permanently saved.

Supplement the full medical history upload with additional context. Following the same manual extraction process, this brief video shows how to locate the "Short Health Summary" within the Epic MyChart Sharing Hub, copy the text into a new Notepad document, and append it to the ongoing Claude task to ensure the AI has the most current and concise snapshot of Dr. Barrett's health.

Discover the true power of AI agents by automating a daily research task. Inspired by a Joe Rogan podcast featuring David Sinclair's age-reversal research, the instructor demonstrates how to copy a YouTube video transcript, paste it into an Opus-level Claude task, and instruct the AI to run a daily web search for related human medical trials, notifying Dr. Barrett if any relevant opportunities arise.

Wrap up the course's practical setup by reviewing your active AI tasks and executing the final health analysis. This lesson explains how Anthropic now stores Claude project files in the cloud for cross-device access, and demonstrates how to answer the AI's clarification questions to refine an automated daily search for anti-aging trials. Finally, achieve the ultimate goal of the course by prompting Claude to synthesize the complete set of uploaded medical records—dating back to 2001—to act as an artificial intelligence health advisor and generate personalized recommendations for improving Dr. Barrett's health.