Since the release of ChatGPT, I've gradually become my family's "personal doctor," and I've discovered that many people have significant blind spots in their disease/health management.
Most people believe that regular checkups are enough to avoid major health risks.
They think that going to the best hospital and finding the most professional doctor is sufficient.
In reality, my personal experience has revealed problems including, but not limited to:
1. Even high-end checkup packages costing 3000+ RMB, or even 6000+ RMB, from high-end medical institutions can (in fact) produce results that are vastly different from those from top-tier hospitals.
2. Even in top-tier hospitals in places like Shanghai, with their most prestigious special needs/international departments and leading doctors, carelessness/lack of responsibility and even a considerable probability of misdiagnosis can seriously delay your or your family's treatment. A doctor sees dozens of patients a day; for you, it's 100%, but for them, you might only be 2% of their day, and even the same doctor's condition varies from day to day. The various medical achievements listed in a doctor's profile do not necessarily correlate directly with the value of the services they provide.
Previously, our concept was that entrusting the bottom line of managing our own and our family's health and illnesses to doctors was sufficient.
In the AI era, it should be: we are the primary responsible parties for our own and our family's health, and we should treat doctors as those who review and implement treatments.
This is how we take control of our quality of life related to health risks, rather than leaving it to luck.
How to do this: Regarding data, learn to download electronic versions (PDF/JPG/even DICOM image files) of your and your family's accumulated medical examination reports, hospital medical records/examination reports. Organize the file names by date and category and archive them locally. Use Dropbox or other personal data file storage tools to perform cloud backups, ensuring everything is easily accessible and readily available on your phone or computer.
Most hospitals' WeChat official accounts/mini-programs offer data export functions. In more complex cases, such as in many hospitals in Shanghai, medical record export is outsourced to a dedicated "Medical Record Assistant" mini-program. These WeChat official account web systems even allow you to link your family members' identities to your own account, so you can operate these functions on their behalf within your WeChat account, since it's difficult to expect most elders to know how to do these things themselves.
In some cases, you only have paper documents, so you'll need a printer/scanner that supports continuous automatic document feeders. For home use, consider the HP 323sdnw; for faster business scanning, the HP 4104dw is an option, both supporting Wi-Fi connectivity. Scanning one document at a time with a glass plate is too cumbersome in the long run. For scanning software, the free and open-source NAPS2 is highly recommended (http:/naps2.com).
Regarding AI analysis, in ChatGPT, you can create "projects" by name. Under each project, you can maintain different conversations to address different issues. The "sources" for a project currently support a maximum of 40 files (so the above data may require periodic PDF merging). NotebookLM doesn't have a clear limit on the number of source files; after importing, you can reference these sources in Gemini.
For a more advanced approach, you can create a dedicated directory on a local machine, such as Claude Code or Lobster Workspace, to synchronize the aforementioned data for long-term maintenance and analysis.
If the above is done correctly, the AI-generated medical opinions will allow you to "know not only what" but also "why," covering 80% or even far exceeding the opinions of outpatient doctors. Based on this foundation, during follow-up appointments, you can boldly ask additional questions about the doctor's opinions, such as "Do I need to do any more XXX tests?" "Is there any other possibility of XXX problems?" "Would the XXX treatment plan be more suitable?" As mentioned above, "treat your doctor as a reviewer," as a collaborator, and don't be shy. A few doctors may sense that you've brought AI knowledge to the clinic; if their attitude is dismissive or impatient, it precisely indicates that such a doctor shouldn't be your sole reliance, and you should consider seeking a second opinion from another hospital/doctor. Some forward-thinking and responsible doctors now even encourage patients to use AI themselves.
For problems with significant potential impact, it's not advisable to go to only one hospital and see only one doctor; it's best to seek second opinions from different hospitals and different doctors. For the same symptom, Doctor A might tell you it's inflammation, while Doctor B might suggest further tests to rule out worse possibilities.
For abnormal values in a physical exam, it's best to go to another hospital for a second test to rule out errors before giving the doctor a diagnosis. Otherwise, if the exam data is incorrect (potentially due to limitations in the lab facilities or even logistics; expensive tests don't necessarily equate to reliable ones), even the best doctor can only make an incorrect diagnosis.
The AI era is one where all institutions and individuals vie for various contexts.
The competition among AI companies is essentially a war over user context. You must have control over your own and your family's health and life context.
Then, regardless of how platforms, tools, and medical methods change,
you'll have the initiative to seamlessly upgrade and migrate at any time.
