Introducing an Experimental Whole-Person Health Investigation Using ChatGPT
Millions of people live with chronic health conditions that
continue despite medical appointments, treatments, medications, and
considerable personal effort.
Some have several chronic conditions simultaneously. Others experience persistent symptoms despite medical tests that appear normal.
Many understand their diagnoses but still struggle with
important questions:
Why do I continue to feel unwell?
How might my different health problems be connected?
What might be keeping these problems going?
Are there factors affecting several conditions at once?
What could realistically help me improve?
These questions deserve attention beyond simply identifying
diseases and treating individual symptoms.
Could generative artificial intelligence help us investigate
them more systematically?
I've been developing and testing an experimental AI-assisted
approach that attempts to do exactly that.
It's called:
My Whole-Person Health Investigation
An Experimental AI-Assisted Educational Investigation
Rather than beginning with one symptom and searching for
diagnoses, this approach starts by examining the person as a whole.
It considers physical health, mental and emotional
well-being, psychological processes, daily functioning, medications, social
circumstances, environmental factors, personal history, and other relevant
health determinants.
It then investigates possible connections, maintaining
mechanisms, feedback loops, barriers to improvement, and credible opportunities
for better health.
The objective is not to replace healthcare professionals.
It's to help people become better informed about their health and support more
productive conversations between healthcare visits.
This remains an experimental, unvalidated approach. It is
being shared to encourage critical evaluation, testing, and further
development—not as a clinically proven health-assessment tool.
1. Why Another Approach to Chronic Illness?
Most people are familiar with conventional symptom checkers.
You enter a symptom, answer several questions, and receive
information about possible medical conditions.
These tools can be useful.
But someone living with several chronic conditions may need
something different.
Consider a person who has:
- Diabetes.
- Arthritis.
- Digestive
difficulties.
- Poor
sleep.
- Fatigue.
- Anxiety.
- Reduced
mobility.
- Several
medications.
A symptom checker might investigate fatigue by asking about
anemia, sleep apnea, thyroid disease, and other possible causes.
Those are legitimate medical considerations.
But they don't necessarily explain how the person's existing
diseases, medications, sleep, psychological health, activity limitations, and
daily circumstances might interact.
Nor do they necessarily identify what may be maintaining the
person's overall illness burden.
A more useful question may be: What is happening across
the entire person, and which factors might be contributing to several
difficulties simultaneously?
This is the question behind my experimental approach.
2. The Central Principle: Breadth First, Depth Second
One of the most important lessons from my early testing was
that AI tends to become a symptom checker unless specifically instructed
otherwise.
I initially found that ChatGPT would concentrate on the
first problem mentioned, asking numerous detailed questions before discovering
the person's other difficulties.
That wasn't what I wanted.
The latest version uses a different sequence.
First: Understand the whole person.
Identify the major health conditions, psychological and
emotional difficulties, daily circumstances, treatments, and relevant health
determinants.
Second: Understand the history.
Explore when problems began, what preceded them, how they
developed, and what treatments helped or failed.
Third: Investigate connections.
Examine whether different conditions, systems, or health
determinants may influence one another.
Fourth: Investigate maintaining mechanisms.
Look for credible explanations of what might be keeping
difficulties going.
Fifth: Examine possible feedback loops.
Explore whether certain problems reinforce one another over
time.
Sixth: Identify realistic opportunities for improvement.
Look for potentially modifiable contributors, especially
those that might affect several problems.
Finally: Produce a concise, understandable report.
The report should help the person understand the bigger
picture—not simply repeat their diagnoses.
This sequence is intended to make the investigation more
useful while reducing premature conclusions.
3. Looking Beyond Individual Diagnoses
The human body is made up of interconnected systems.
Changes in one system may affect others.
Physical illness may interact with sleep, activity,
psychological well-being, medications, or environmental conditions.
Likewise, social circumstances, financial pressures,
relationships, and access to healthcare can influence a person's ability to
manage chronic illness.
The investigation considers these possibilities without
assuming that every problem has one underlying cause.
Multiple chronic conditions
Some diseases share recognized biological mechanisms.
For example, certain autoimmune conditions involve related
immune processes and may occur together more frequently than expected by
chance.
However, that doesn't mean every autoimmune condition has
the same cause or can be improved through one treatment.
Other chronic conditions may be entirely independent.
The investigation should examine both possibilities.
Systems that may not be functioning optimally
Concepts such as homeostasis and physiological regulation
are relevant to health.
However, identifying a particular regulatory disturbance
requires credible evidence.
For example, fatigue alone does not establish mitochondrial
dysfunction, nervous system dysregulation, or hormonal imbalance.
The Master Prompt specifically instructs AI not to invent
physiological abnormalities merely because they fit an appealing explanation.
The goal is to identify credible possibilities, not
manufacture a theory explaining everything.
4. What Might Be Keeping a Condition Chronic?
An important distinction in chronic illness is the
difference between:
What originally contributed to a condition developing.
And:
What may be maintaining the condition or its consequences
today.
These are not necessarily the same.
A past injury might have initiated a problem, while ongoing
structural damage, disrupted sleep, changes in functioning, or another
recognized mechanism may contribute to continuing difficulties.
Similarly, an infection may precede a chronic condition
without necessarily remaining present.
The investigation examines several categories of possible
maintaining factors:
Continuing disease processes: Established conditions
that remain active.
Secondary contributors: Problems that worsen symptoms
or functioning.
Treatment-related factors: Medication effects,
treatment limitations, or relevant gaps in care.
Psychological and behavioral processes: Recognized
mechanisms involving anxiety, avoidance, sleep, stress, or other relevant
factors.
Environmental and social determinants: Circumstances
that may contribute to ongoing health burdens or prevent improvement.
Structural limitations: Permanent damage or other
changes that may not be reversible.
These factors must be evaluated individually.
An AI-generated explanation is not proof that a particular
mechanism exists in a particular person.
5. Chronic Feedback Loops: An Important Possibility
Some chronic difficulties may involve self-reinforcing
cycles.
For example:
Pain → Poor sleep → Increased pain sensitivity → Greater
pain
Another recognized type of cycle can occur with certain
anxiety problems:
Anxiety → Avoidance → Temporary relief → Greater future
avoidance
These examples illustrate possible interactions. They are
not conclusions about every person with pain or anxiety.
An especially important distinction is that a cycle
worsening symptoms does not necessarily maintain the underlying disease itself.
Poor sleep might make spinal pain more difficult to tolerate
without causing the structural spinal condition.
My Master Prompt instructs AI to distinguish these
possibilities and challenge proposed feedback loops rather than accepting them
automatically.
6. What About Mental, Psychological, Emotional, and
Spiritual Health?
A whole-person investigation cannot concentrate exclusively
on physical disease.
Mental health conditions may have important maintaining
mechanisms of their own.
Psychological and behavioral processes can influence
symptoms, functioning, and well-being.
Relevant circumstances may include:
- Anxiety,
depression, PTSD, or other established conditions.
- Adverse
childhood experiences and major life events.
- Long-term
stress.
- Avoidance
or other learned behavioral patterns.
- Grief
and loneliness.
- Emotional
difficulties.
- Relationships
and social support.
- Meaning,
purpose, and spiritual concerns.
However, the investigation should not assume that
psychological distress caused someone's physical illness.
Nor should it automatically classify grief, loneliness, or
spiritual concerns as diseases.
Sensitive subjects should be explored only when appropriate
and welcome.
A genuinely whole-person investigation needs to recognize
both the importance and the limitations of these different perspectives.
7. The Most Important Question: What Might Actually Help?
My initial versions concentrated primarily on identifying
possible maintaining mechanisms.
But understanding a problem isn't the same as improving it.
I believe this is where the approach needs to become
especially practical.
Could addressing one credible factor improve several
problems simultaneously?
For example, an established sleep disorder might contribute
to fatigue, concentration difficulties, and other symptoms.
Appropriately addressing that condition could potentially
improve several areas of functioning.
Likewise, identifying a medication-related issue, improving
management of an established disease, or addressing a significant practical
barrier could have benefits extending beyond one symptom.
These are possibilities to investigate—not automatic
recommendations.
The prompt evaluates potential opportunities according to:
Expected benefit, strength of evidence, safety, cost,
effort, personal goals, and the possibility of improving several difficulties.
It should prioritize credible, manageable opportunities
rather than presenting an overwhelming collection of suggestions.
Medical decisions involving medications, testing, or
specialized treatment remain matters for appropriate professional evaluation.
A Note to People With Chronic Health Conditions
It's understandable that people struggling with chronic
health problems may want to try this prompt themselves. However, it is
experimental and has not been clinically validated. If you choose to use it,
you do so at your own risk. Treat its findings as possibilities to explore,
not medical diagnoses or treatment recommendations. Do not change prescribed
treatments or delay necessary medical care based on its findings. Seek guidance
from qualified healthcare professionals.
8. Try the Experimental Master Prompt
I've made the complete Master Prompt available separately in
Google Docs so readers don't have to scroll through an extremely long article.
Master Prompt V3.4 — My Whole-Person Health Investigation
Open the full Master Prompt:
My
Whole-Person Health Investigation — Master Prompt V3.4
The document contains the instructions for the AI
investigation, including its whole-person approach, evidence safeguards,
feedback-loop analysis, report options, and continuing investigation process.
The prompt is experimental. It has not been clinically
validated or established as safe or effective for personal medical
decision-making.
How to use the prompt
Open the document.
Copy the entire Master Prompt.
Open a new chat in ChatGPT and paste the prompt.
Press Enter (or tap Send) to start the investigation.
ChatGPT should begin by explaining the investigation and
asking how much energy and concentration you have.
You can then describe your major whole person health
problems in your own words.
The investigation is intended to begin automatically.
If it does not, you can type:
Start the investigation now. Ask me the energy-level
question and wait for my answer.
9. Regular ChatGPT or ChatGPT Health?
You can use the Master Prompt with regular ChatGPT.
Regular ChatGPT can help you organize your health history,
explore possible connections, understand medical information, examine credible
research, prepare questions, and develop a clearer picture of your
difficulties.
You can describe your conditions, treatments, symptoms, and
daily circumstances yourself.
ChatGPT Health, when available, can also use relevant
health information that you choose to connect, including supported medical
records and health data.
That may help the investigation consider your history,
medications, other conditions, and changes over time without requiring you to
enter everything manually.
However, connected records can be incomplete, outdated, or
inaccurate. They also may not adequately describe emotional, psychological,
social, environmental, or spiritual circumstances.
The same Master Prompt can be used with either approach.
Both can make mistakes.
Using Regular ChatGPT
Go to:
Sign in or create an account.
Start a new chat.
Paste the complete Master Prompt and submit it.
When prompted, describe all your health concerns.
You can also provide relevant information from medical
records if you wish, after reviewing your privacy settings and removing
unnecessary identifying details.
Using ChatGPT Health
For eligible users with access to Health:
Open ChatGPT and select Health from the sidebar or
More menu.
Follow the instructions to connect supported medical records
or health information if you wish.
Review the permissions and information you are sharing.
Start a new chat in ChatGPT Health, paste the Master Prompt,
and press Enter (or tap Send). The investigation should begin automatically.
Do not assume every medical record or health measurement is
available.
As of November 2026, Health is available to eligible U.S.
users aged 18 or older on supported platforms, with availability dependent on
account eligibility and rollout.
People elsewhere can still use regular ChatGPT for the
educational investigation.
Can I Use Other AI Systems?
This Master Prompt was developed and tested primarily with
ChatGPT, but it may also work with other generative AI systems, such as Claude,
Gemini, or Copilot. However, results may vary because these systems differ in
their capabilities, instructions, memory, and access to health information.
10. Take Your Time — This Isn't a One-Sitting
Questionnaire
People with chronic conditions may have limited energy,
pain, brain fog, anxiety, or difficulty concentrating.
An investigation of complicated health problems may take
hours, days, or weeks – on and off.
There is no hurry.
The prompt is designed to ask one question at a time.
You can answer briefly, skip questions, take breaks, and
return later.
You can also request a preliminary report without completing
the entire investigation.
As more information becomes available, the AI can update the
report.
Choose How Much Detail You Want
The Master Prompt offers three report levels:
Quick Report: The main findings and one worthwhile
next step.
Standard Report: A broader explanation of possible
maintaining mechanisms, connections, and opportunities.
Detailed Report: A more extensive investigation of
systems, evidence, feedback loops, barriers, and uncertainties.
This is especially important for someone who may be too
tired to read a lengthy report.
A Living Investigation
The report is intended to evolve.
You might later remember an important medical event, provide
a new test result, or receive feedback from a healthcare professional.
The AI should reconsider its earlier hypotheses rather than
simply adding more information.
You can ask ChatGPT for a preliminary report at any time,
even if you haven't completed the investigation. You can also take breaks and
return to the same chat to continue where you left off.
11. What Should the Final Report Tell You?
The report is organized around three practical questions.
What Might Be Contributing to My Continuing Health
Difficulties?
This section explains the most credible maintaining factors,
possible interactions, and important uncertainties.
What Credible Opportunities Might Help Me Improve?
This section identifies potentially modifiable contributors
and evidence-supported opportunities.
It considers whether one opportunity might influence several
difficulties.
What Should I Explore Next?
This section identifies one or two manageable next steps.
Some may involve safe independent learning or observation.
Others may deserve discussion with a healthcare
professional.
The report should not simply reproduce the person's medical
history or invent a hidden explanation.
Its value lies in making the overall picture clearer and
identifying useful questions that may otherwise have been overlooked.
12. Supporting Conversations With Healthcare
Professionals
I believe one of the greatest potential benefits of this
approach is what happens between appointments.
People can use AI to better understand their medical
information, organize observations, prepare questions, and make sense of
professional feedback.
They can also request a concise one-page summary for a
healthcare appointment.
This summary should clearly distinguish documented medical
information from AI-generated hypotheses.
A healthcare professional may confirm an explanation,
challenge it, identify additional evidence, or recommend a different approach.
That feedback can then become part of the continuing
investigation.
The objective is better-informed participation in
healthcare—not competing with healthcare professionals.
13. What Have My Tests Shown So Far?
I've developed several versions of this Master Prompt and
tested how ChatGPT responds to fictional chronic illness scenarios.
The tests revealed important weaknesses.
Earlier versions sometimes concentrated too heavily on the
first symptom mentioned.
Some reports repeated information the person already knew.
Others were too long or failed to identify meaningful
opportunities for improvement.
These observations led to improvements in Version 3.4,
including:
- Breadth-first
questioning.
- Mandatory
whole-system integration.
- Stronger
evaluation of possible maintaining mechanisms.
- Better
safeguards against speculative explanations.
- More
attention to high-leverage opportunities.
- Adjustable
report detail.
- Energy-sensitive
questioning.
- Continuing
investigation and progress summaries.
- Stronger
medical-safety instructions.
These are exploratory development observations.
They do not establish that the prompt reliably identifies
causes, improves health outcomes, or is safe for unsupervised medical
decision-making.
14. An Invitation to Researchers, Developers, and
Healthcare Innovators
I am sharing this experimental framework because I believe
it raises a worthwhile research question:
Can generative AI help people develop a more accurate,
integrated understanding of persistent health problems and identify useful
opportunities for improvement?
I would particularly welcome critical evaluation,
suggestions, and collaboration from:
- Holistic
and whole-person healthcare practitioners who consider physical,
mental, emotional, social, environmental, and spiritual dimensions of
health.
- Functional
medicine practitioners interested in possible underlying contributors,
perpetuating factors, and interactions among chronic conditions.
- Integrative
medicine clinicians who combine conventional medical care with
evidence-informed complementary approaches.
- Lifestyle
medicine professionals specializing in nutrition, physical activity,
sleep, stress management, and other modifiable health factors.
- Systems
medicine and systems biology researchers studying interactions among
physiological systems and chronic diseases.
- Healthcare
professionals experienced in complex chronic illness and multimorbidity,
including primary care clinicians and relevant specialists.
- Mental
health professionals and behavioral medicine researchers studying
psychological, emotional, and behavioral contributors to persistent health
difficulties.
- Rehabilitation
and chronic disease self-management specialists focused on improving
functioning, independence, and quality of life.
- AI
developers, prompt engineers, and digital health innovators developing
patient-facing generative AI tools.
- Researchers
studying AI-supported healthcare and patient self-management,
particularly between medical appointments.
·
Patient advocates, health educators, and patient-safety
researchers who can evaluate the framework's accessibility, potential
benefits, risks, and suitability for future patient use.
- Patient-safety
and clinical research experts who can help assess accuracy, potential
harms, scientific validity, and appropriate safeguards.
My Long-Term Vision
My goal is to help develop a free, accessible,
AI-assisted whole-person health tool that people anywhere in the world can use
to better understand their chronic health problems, explore credible
opportunities for improvement, and support their health over time.
I envision something people can use independently, at
their own pace, whenever they need help—not just during medical appointments.
It should consider the whole person, including
physical, mental, psychological, emotional, social, environmental, and relevant
spiritual dimensions of health. It should explore possible maintaining
mechanisms, interactions among chronic conditions, feedback loops, and
potentially modifiable factors.
Most importantly, it should help people identify
realistic, evidence-supported opportunities to improve their health and quality
of life—not simply produce another list of symptoms or diagnoses.
The aim is to empower people to do more for themselves,
while recognizing when professional healthcare is important and using it when
needed and available. It should never encourage someone to avoid necessary
medical care.
I want this approach to be useful to people
everywhere—not only those with reliable healthcare access, but also those in
underserved communities, countries where professional healthcare is limited or
unaffordable, and people whose health insurance does not adequately cover the
care they need.
Even in countries with established healthcare systems,
insurance coverage may be limited or unavailable for certain functional
medicine, integrative medicine, holistic healthcare, and whole-person
approaches. Coverage varies by country, insurer, provider, and specific
service. As a result, people may struggle to obtain comprehensive care that
considers the many interconnected factors affecting their health.
My vision is to help make whole-person health education
and practical self-management support freely accessible worldwide, regardless
of income, location, healthcare availability, or insurance coverage. This
would complement necessary professional healthcare, not replace it.
The current Master Prompt is an experimental starting
point, not a clinically validated health tool. Further testing, independent
evaluation, and improvements are needed to establish its reliability, safety,
accessibility, and effectiveness.
I welcome collaboration with researchers, healthcare
professionals, holistic and integrative practitioners, AI developers, patient
advocates, and others who share this vision.
Ultimately, I hope to help create something that people
worldwide can use to better understand their health, make informed choices,
improve where possible, and sustain those improvements over time.
Relevant questions include:
Can AI reliably investigate the whole person rather than
becoming a symptom checker?
Can it distinguish credible maintaining mechanisms from
speculative theories?
Can it identify important interactions among multiple
chronic conditions?
Can it recognize when several problems are unrelated?
Can it identify worthwhile opportunities for improvement
without producing unsafe treatment recommendations?
Can it support people between healthcare visits without
replacing professional care?
What would proper validation require?
The Master Prompt is offered as an experimental starting
point for discussion, testing, and improvement—not as a finished clinical
system.
15. A Final Thought
Living with chronic illness often involves far more than
managing individual diagnoses.
It may involve multiple medical conditions, treatments,
emotional challenges, practical limitations, and unanswered questions.
Healthcare professionals provide essential expertise, but
people must still live with their conditions between appointments.
Generative AI may eventually help bridge part of that gap.
My hope is that we can develop tools that help people
understand their health more comprehensively, identify credible opportunities
for improvement, and become more effective participants in their healthcare.
The ultimate goal isn't simply to produce a longer health
report.
It's to help people answer three questions:
What might be keeping me unwell?
What might realistically help?
What should I investigate or discuss next?
That would be a meaningful step toward more useful
AI-assisted whole-person health support.
Disclaimer
I am not a healthcare professional. This article is provided
for educational, informational, and research purposes only.
The Master Prompt described here is experimental and has not
been clinically validated or established as safe or effective for diagnosing,
treating, curing, or managing medical conditions.
Generative AI can produce inaccurate, incomplete,
misleading, or fabricated information. It may overlook serious conditions,
propose unsupported causes, or suggest inappropriate interventions.
AI-generated findings and hypotheses should not be treated
as established medical conclusions or personalized treatment recommendations.
Do not use this article or Master Prompt to diagnose
yourself, independently change prescribed medications or treatments, undertake
unsafe interventions, or delay necessary medical or mental health care.
Important individualized medical decisions should be
discussed with appropriately qualified healthcare professionals.
No particular health improvement, recovery, or cure is
promised or guaranteed.
Published by TG Ideas LLC
https://sites.google.com/site/tgideas/
Additional Disclaimers
https://sites.google.com/site/tgideas/ideas-for-products-or-services/disclaimer
My LinkedIn Newsletter — Thrive: Mind, Body & Tech
https://www.linkedin.com/newsletters/thrive-mind-body-tech-7178508146028556289/
My Amazon Author Page
https://www.amazon.com/author/tomgarz
My Custom GPT
Try my free Custom GPT, Chronic Illness: Why Am I Still
Sick? — a whole-person tool for people with persistent symptoms, multiple
chronic conditions, or normal tests who still don't feel well.
https://chatgpt.com/g/g-69fa4cd970448191ace058c5d4ca15f2-chronic-illness-why-am-i-still-sick
Connect on Facebook
https://www.facebook.com/pages/Tom-Garz-Writer-and-Inventor/422266054499307
