The Patient/Advocate Oncology Passport
Patient Advocate AI Use Case
The Patient/Advocate Oncology Passport
Could AI Help Patients Arrive Better Prepared for Clinical Trials?
This is not medical advice.
This is not a proposal to let AI decide whether a patient qualifies for a clinical trial or replace doctors, coordinators, nurses, principal investigators, pathologists, pharmacists, or trial teams.
The idea is simpler:
Could AI help create a better first draft?
Could it organize the records a patient already carries?
Could it help patients and trial teams begin from a clearer starting point?
Patients Arrive Carrying Too Much
Cancer patients rarely arrive at a clinical trial with one clean story.
We arrive with a range of documents: pathology reports, scans, lab results, medication lists, treatment histories, molecular testing, ctDNA results, side effects, and unanswered questions.
Some information lives in portals, PDFs, clinical notes, pathology departments, or imaging systems. Some lives only in the patient’s memory, which may be strained by surgery, treatment, fatigue, fear, and sleepless nights.
By the time a patient reaches a trial conversation, we may have enough records to qualify as a small, poorly organized medical library.
Patients are often asked to become information fusers. In practice, it can feel like trying to assemble an airplane from a range of medical documents while it is already in the air.
That is difficult on a good day.
It is harder when you are sick.
From Molecular Passport to Patient/Advocate Oncology Passport
A Molecular Passport summarizes the tumor’s biology: genomic findings, biomarkers, target expression, ctDNA, tissue availability, and assay results.
That is useful, but the idea should be wider.
The Molecular Passport describes the tumor.
The Patient/Advocate Oncology Passport describes the person carrying it.
A patient entering a trial is more than a tumor profile. The trial team may also need to understand treatment history, current condition, prior toxicities, safety concerns, communication needs, support systems, open questions, and practical barriers.
The science matters.
The person matters too.
The AI Use Case
The trial team should provide the Passport format first.
Not a generic form. Not whatever the patient or AI invents.
The actual structure the team would find useful.
It might include:
required fields and preferred terminology;
molecular and eligibility-related categories;
prior-treatment history;
source-document requirements;
rules for missing, conflicting, or unconfirmed information;
support and communication needs;
trial-specific screening questions.
AI could then draft a Patient/Advocate Oncology Passport from the patient’s existing records.
The draft would not be final or treated as medical truth. It would go to the patient and appropriate members of the care or trial team for review and validation.
In plain English:
AI drafts.
Humans validate.
Patients understand.
That is the guardrail.
What AI Should Not Do
AI should not:
decide final eligibility;
replace clinical judgment;
recommend treatment;
invent missing information;
hide uncertainty;
make conflicting records appear consistent;
turn hope into a false conclusion.
Clinical judgment must remain with qualified people.
What AI Might Help Do
AI may be useful as an organizer, sorter, translator, and draft builder.
It might help:
create a treatment timeline;
summarize prior therapies;
identify recent imaging and pathology;
extract molecular and ctDNA results;
separate verified records from patient-reported information;
flag missing or conflicting information;
build a plain-language glossary;
draft questions for the study team;
create a “who has the ball?” status list;
place information into the team’s preferred format.
That could reduce the blank-page problem and help everyone begin with a clearer draft.
A Possible Passport Structure
The Passport might have two layers.
Layer One: One-Page Quick View
A page a busy clinician or coordinator can scan quickly:
patient and treating-oncologist information;
diagnosis and current disease status;
major treatment history;
most recent imaging;
known molecular findings;
major safety concerns;
current trial question;
missing items;
next step.
Not the whole story.
Just enough to prevent the first conversation from starting blind.
Layer Two: Supporting Details
This could include:
full treatment timeline;
surgeries and radiation;
prior systemic therapies;
side effects, reactions, and treatment limitations;
pathology and imaging summaries;
genomic and ctDNA results;
biomarkers;
tissue-block location;
medications and allergies;
caregiver and support information;
communication and learning preferences;
practical barriers;
questions for the trial team.
The goal is practical clarity, not medical perfection.
Verified, Patient-Reported, or Missing
Each important item should be labeled clearly:
Verified from a source document
The information comes from an official report, assay, scan, or clinical note.
Patient-reported
The patient supplied it, but the team may still need to confirm it.
Missing or needs confirmation
The information matters but is not yet available or validated.
This protects the science while allowing patients to contribute useful information.
It also prevents AI-generated language from sounding more certain than the records justify.
The Trial Team’s Format Comes First
If every patient creates a different Passport, the study team may receive ten different versions of confusion.
The team should define the structure first. AI should organize the patient’s information into that structure.
That makes the Passport a bridge, not another burden.
A Passport would not replace expert judgment. Like a good checklist, it would help expert judgment begin with the right information in view.
Fitting Within Existing Trial Infrastructure
Centers such as UC San Diego Moores Cancer Center already have serious clinical-trial infrastructure, including research support, study-design resources, monitoring, patient-care follow-up, and translational research services.
The Passport would not replace that work.
It would serve as a patient-side companion: a cleaner first draft in the team’s preferred format, ready for human review.
No one is asking AI to drive the bus.
At most, it helps gather the luggage, label the bags, and makes sure the map is not upside down.
Why This Might Help Patients
The Passport could help patients:
understand their own story;
see what is known and what is missing;
ask better questions;
arrive less scattered and better prepared.
That may not mean less afraid.
But perhaps less lost.
Sometimes less lost is a very big thing.
The Passport is not meant to make the patient sound like a doctor. It is meant to keep the patient from arriving like a mystery wrapped in a records request.
Why This Might Help Trial Teams
For study teams, the Passport could:
reduce repeated information gathering;
make preliminary review cleaner;
identify missing records sooner;
show what information is already available;
bring patient needs into the process more clearly.
Trial participation is not only about whether a tumor matches a protocol.
It is also about whether the person can safely, realistically, and knowingly walk the trial road.
My Own Lens
I am not proposing this as an AI expert or medical researcher.
I am proposing it as a cancer patient who has had to learn too many terms too quickly.
I also bring experience as a former principal investigator on a federally funded research-and-development program, a degree in Human Communications, and personal experience with dyslexia.
That does not make me an expert in clinical-trial operations.
But it does make me sensitive to one problem:
The right information can exist and still not be usable when someone needs it.
That is a design problem.
The Passport is one possible response.
A Human-Centered Workflow
Step 1: The trial team defines the template
The team decides what information is useful for screening, communication, education, and trial entry.
Step 2: The patient supplies or authorizes access to records
These may include pathology, imaging, molecular assays, treatment history, medications, and patient notes.
Step 3: AI creates a draft
The information is organized into the team’s format and labeled as verified, patient-reported, missing, conflicting, or awaiting confirmation.
Step 4: The patient reviews it
The patient corrects obvious errors and adds questions, support needs, and practical concerns.
Step 5: The study team validates it
Appropriate trial personnel confirm what is usable and what still needs correction or documentation.
Step 6: The Passport becomes a living document
It can be updated after scans, biopsies, assays, eligibility decisions, or changes in the patient’s condition.
The goal is not automated eligibility.
The goal is better-prepared human review.
Questions Worth Asking Before Building It
Before developing this AI use case, I would ask:
What information do coordinators need first?
What slows preliminary review?
Which records are most useful early in the process?
Which fields require formal validation?
How should missing or conflicting information appear?
Could the Passport reduce duplicate record requests?
Could it help patients understand who has the ball?
How could it work better for dyslexic, fatigued, chemo-brained, or overwhelmed patients?
How could it remain useful without creating more work for trial teams?
Those are design questions.
Good trials deserve good design questions.
AI Experiment
I designed a simple AI prompt to review all of my medical data and format a table of my imaging for the past two years. The table is searchable, and can be sorted by who ordered the imaging, what type they ordered (PET CT, CT, etc), Contrast (Y/N) and other data. What I found was fascinating. surgeon was watching my cancer with PET scans, until he turned my care over to an Oncologist who used CT scans. What was caught? What was Missed. Is this following protocols? I did find the tabular format very useful for asking questions, and for some overview and discovery of my care.
Closing Thought
The Patient/Advocate Oncology Passport would not replace the medical record, doctor, coordinator, principal investigator, trial team, or patient.
Its purpose would be practical:
Help organize the patient’s story in the team’s preferred format.
Show what is known, missing, or unconfirmed.
Help the patient arrive better prepared.
Help the study team begin with a cleaner draft.
Help everyone spend less time hunting for the map and more time deciding what it means.
Precision medicine should not stop at the tumor.
It should include the patient.
With the right guardrails, this patient-advocate AI use case may help build the first draft of that bridge.
—Ty