Research vs. Reality: Why Perplexity Health Fails Chronic Patients Where Guava Health Thrives

Written by Stephanie M. Nixon, PhD, CCC-SLP on April 20, 2026. Posted in Apps, Chronic illness, Guava Health, Health care, Navigating health care, Perplexity Health AI. No Comments on Research vs. Reality: Why Perplexity Health Fails Chronic Patients Where Guava Health Thrives

Transparency Disclosure

In a world of sponsored “health tech” content, here is the truth:

I am a paying subscriber of Perplexity Pro ($20/mo) and a paying subscriber of the Guava Health Family Plan.
While I serve on the Guava Health Patient Advisory Panel, I do so pro bono (unpaid).

This review is not a promotion; it is a clinical audit. I’ve spent my own money on these tools because I am a “spoonie” patient, provider, and researcher searching for a way to make a complex life more manageable. (But for the record, I do accept cookies.)

The Dangerous Illusion of “All Your Data”

The biggest risk of Perplexity Health (Beta) is the illusion of completeness. Its marketing promises a seamless “Health Hub,” but my testing revealed a dangerous “blind spot.” Despite being connected to my EMRs, Perplexity ignored my most recent labs and defaulted to a T4 Free result from 2021 as my “current” status.

The Safety Warning: If a patient trusts Perplexity’s dashboard, they might miss key data and believe they are clinically stable when the AI is simply “ignoring” the last five years of their data because it didn’t index in a way the AI could add to your biomarkers.

The Guava Advantage: Guava understands that medical data is messy. Even when a provider doesn’t have a direct API integration, Guava allows you to upload the report. The platform doesn’t just “store” the PDF; it indexes and pulls that data into your biomarkers, ensuring your trends are accurate and complete (or as complete as the data it obtains via APIs and your uploads).

Fig 1. Data Latency Proof. Perplexity remained stuck in 2021 despite current data being physically present in the system. — Data Latency Proof. Perplexity remained stuck in 2021 despite current data being physically present in the system.

The “Spoonie Tax”: Friction as a Barrier

Chronic illness management is often a full-time job. We use technology to save “spoons,” not spend them. Perplexity, however, added to that fatigue through technical friction:

The Quantity and Size File Wall

I could only upload 200 documents. As a person with chronic illness who has been on this earth more than say 40 years, that didn’t begin to cover even half of my documents. You might tell me, ‘Just upload the most important ones’ or ‘Import them from the provider’, well, I tried that.

The connection to my primary care provider’s office linked to her old facility vs. the current one despite having the correct address on file for her.
My dermatologist, hematologist, integrative medicine provider, endocrinologist, cardiologist, gynecologist, primary care provider, and gastroenterologist weren’t accessible via Perplexity. (And these are only a few of the providers whose data couldn’t be accessed by Perplexity.)
- Note. Six of eight of the above aren’t accessible via Guava either, but I can upload those visit notes to Guava Health myself.
To only see 10% of the data can also limit seeing the big picture. So, I grouped the files using Adobe to upload them (after deleting each individual file that I had uploaded).
- BUT if you’ve used AI to try to analyze data in large 150+ page files, you’ll know … it isn’t always easy. (Even Gemini tells me occasionally that my files are “too large”.)

The Administrative Demand

Then I had an interesting ‘conversation’ with Perplexity’s AI when I realized that despite syncing up with the API of one of my facilities with most of the lab data, it was missing specific data points (e.g., all Free T4 data after November 2021 despite that specific lab being drawn between 2-6 times/year.)

So I asked Perplexity Health’s AI where the labs are, and was told “You’re right that [FACILITY X] has them and that you uploaded them; the issue is that they’re not surfacing in the structured “labs/observations” stream I can see”. The AI provided several reasons for why those labs weren’t there:

They may be stored only as a scanned PDF/image of the ER report, not as discrete lab rows, so they don’t appear in the lab feed even though they’re in your chart.
The upload might be filed under “Media,” “Documents,” or an outside-records bucket that isn’t mapped into the lab/observation section my view is pulling from.
The interface that shares data from Facility X’s EMR to this aggregate view may not map ED lab panels from that date or that specific visit type correctly.

But as a patient, provider, or researcher, the issue for those of us using Perplexity Health is that those labs aren’t there, and even uploading them, doesn’t lead to them being added to the biomarker section of the site.

So when I asked Perplexity AI what I should do, it suggested that I call the hospital’s IT department to request they “re-index” my FHIR API mapping.

Fig. 2 Perplexity delegates technical labor to the patient; Guava automates the extraction through AI parsing.

Even when I provided the connections to the facilities and uploaded the records (labs and visit notes), which took A LOT OF TIME, I learned today from the Perplexity Health AI: “So even though those note sets exist in your health system, what’s flowing into this connection right now is the coded summary, not the full written note body”.

To get Perplexity Health AI to review the visit note (not the summary, the actual note), I had to reupload the visit notes to the AI side of the conversation so it could read those vs. the coded summaries.

With that in mind, Guava’s AI had me copy the line from the visit note that I wanted it to compare with my other visit notes. This is what Guava Health’s AI told me today: “I can review excerpts you paste here, but I can’t directly pull provider notes from your chart unless the app exposes them to me in this chat. If you want, paste the relevant sections”.

The Patient’s Reality

Asking a hospital’s HIM department to fix a third-party AI’s mapping is an exercise in futility. We don’t have the energy to act as unpaid data engineers for a Beta product that is already charging a premium. Also, what do you think the hospital system would say if I asked for this?

Help Desk Emails

I have emailed the help desk for Perplexity Health AI and Guava. Because I did so about the above issue with Perplexity Health AI at the suggestion of the AI after it realized it was missing significant biomarker data points, let’s address the difference.

Perplexity Health: AI Support Agent Sam emailed me back to thank me for my insights and saying that they forwarded my feedback about lab integration limitations, file upload limits, and suggestions to the product team. I sent the email April 14, 2026. I haven’t heard back.

Guava Health: This is one example of an email with the Guava Help Desk. I emailed Guava Health’s Help Desk and asked how to fix an issue when there were two sources for the same lab on August 7, 2025. On August 8, 2025, I received an email from Alex Yau, Founder and President of Guava Health to answer my question. He asked for a screenshot for an example, which I sent and he followed up with additional insight and added that he would forward it to his team to give more thought.

Note. Both Perplexity Health AI and Guava Health have discords. I am not in the Perplexity discord, but I am in the Guava Health Discord.

The Android OS Barrier

While iOS users have a native Hub, Perplexity is “desktop-primary” for Android users. Using a mobile browser to check your health data is clunky and lacks the seamless utility of a native app.

The Repository Advantage: DICOM, Quest, and GI Notes

Guava is a Source of Truth; Perplexity is a search window.

Imaging (DICOM): Guava supports actual X-ray, CT, and MRI image files. You aren’t just storing a “report”; you are carrying your entire imaging library in your pocket. (Note. You need to upload those, but, still.)
Medication Reconciliation: I manage 62 active medications. Guava allows me to merge and deduplicate them across providers. Perplexity was able to retrieve some medication lists from the facilities BUT some of those lists were outdated.

My understanding is that Perplexity Health AI integrates with Apple Health for medication management, but I use Android, so I cannot comment on that.

I can, however, comment on Guava Health’s Medication management:

Data Portability: The “Disability & Tax” Hack

The true power of a health platform is what you can do with the data.

The Guava Hack: I downloaded my encounter history from Guava and used Gemini (in thinking or data analysis mode) to generate a mileage CSV for my taxes and a total encounter count for other paperwork. Guava provided the “raw material” to simplify my legal and financial life.
Perplexity’s Failure: Because Perplexity couldn’t accurately aggregate my history, I question whether it could do so accurately.

I’ll talk more about this in another post.

Setting Health Goals in Perplexity

One of the most revealing disconnects in Perplexity’s “Health Hub” is the Health Goals feature, which feels fundamentally “un-Spoonie.” While the marketing suggests a personalized experience, the available goals are largely aspirational wellness targets—like “Improving sleep” or “Marathon training”—that assume a linear, healthy baseline.

For a patient managing chronic illness, these rigid targets are often inappropriate or even demoralizing, as they ignore the daily fluctuations in energy and capacity that define the “Spoonie” experience. In contrast, the Fitbit Beta with Gemini feels significantly more approachable; instead of assigning a generic wellness category, it begins with a conversation about your specific challenges and health conditions. This “Coach” approach allows the AI to adjust its insights to your actual reality, rather than forcing you into a “fitness enthusiast” mold that your body simply isn’t in today.

Health goals page from Perplexity Health AI — Taken from Perplexity Health AI. It felt like there should be an “other” option.

Auditing for Bias & The Privacy Trap

I decided to test Guava Health and Perplexity AI today on auditing visit notes for biased language from a specific former provider.

While Perplexity was helpful, I found that NotebookLM and Gemini were significantly better at finding subtle linguistic cues.
Guava Health had me copy and paste the lines from the visit note into the AI for analysis against my record, but the AI then gave me an excellent plain language overview of bias in the portion I provided it.

CRITICAL PRIVACY WARNING: Unless you have a signed BAA (Business Associate Agreement)—like Guava—BE AWARE that uploading sensitive records to any non-HIPAA compliant AI (e.g., Gemini via personal Gmail, NotebookLM attached to personal Gmail, ChatGPT, or Perplexity without the Perplexity Health piece) is only as secure as your cloud data/passwords/and more.

Final Verdict: Research vs. Management

Perplexity Health is charging a premium ($20/month) for a Beta product that requires the patient to act as a manual data entry clerk. It is for people who want to research a disease.

Guava Health is for people who have to manage one. Guava Health Premium is only $8 per month. And if your provider has a Guava Health Provider Dashboard and invites you to it, IT IS free for you. Guava Health also offers a free version to patients.

Use Perplexity if: You want a search engine for medical trends. (Maybe? But I still haven’t tested this out as much as I need to. I tend to use Google Scholar.)
Use Guava if: You need to manage medications, prepare questions for medical visits, quickly access provider notes, see overviews of your biomarkers, store your imaging, and have your data work for you in the real world.

COMING SOON:

How to get a copy of your evidence (certificate) of coverage
The Disability & Tax Guide: A step-by-step on using your medical data to navigate these two data heavy tasks.
The Provider Portal: How Guava helps your doctors help you.
Guava Tags: How I use custom tagging to find patterns in flares.

If you have questions about Guava Health, post them here!

Visit the Platforms

Identification Note: Logos used for nominative fair use for critical review. Stephanie Nixon, PhD, CCC-SLP serves on the Guava Patient Advisory Panel pro bono. Review based on the April 2026 Beta of Perplexity Health.

chronic illness, Guava Health, Health apps, Perplexity Health AI, Spoonie life

“Patient’s Log”: Track your Insurance Calls Like a Provider (In 60 seconds)

Written by Stephanie M. Nixon, PhD, CCC-SLP on February 9, 2026. Posted in Apps, Chronic illness, Health care, Health Insurance, Navigating health care. No Comments on “Patient’s Log”: Track your Insurance Calls Like a Provider (In 60 seconds)

Stop scribbling on scrap paper. Here is the exact system I use to hold insurance companies accountable (and keep track of what has been said).

In my last post, I talked about the importance of documenting every single interaction with your insurance company. But let’s be honest: when you are managing a chronic illness, working, or just living life, finding a notebook and a working pen while on hold is just one more hurdle.

My husband and I realized early on that we needed a system that was fast, shared, and impossible to lose.

Our solution? A simple Google Form.

It lives as an icon on our phone home screens. When we get on a call, we tap it, fill in the blanks while we talk, and hit submit. It automatically saves everything into a spreadsheet that we can search later.

Why This Works Better Than a Notebook

It Prompts You: You never forget to ask “Who am I speaking with?” because the form requires you to type it in.
It’s Collaborative: If my husband takes a call, I can see the notes instantly on my computer. No more “Did you call them?” arguments.
It Creates a Timeline: When you need to file a grievance (like I did), you just open the spreadsheet and copy-paste the entire history.

The Fields You Need (Steal My Form)

I created a free Google Form with these specific questions. You can copy this exact structure:

Patient calling about: (Checkbox: … Names of those in the household, etc.)
Date called: (Date picker)
Who contacted? (Checkboxes: Benefits, Care Management, CVS Caremark, HR, etc.)
Method of contact: (Checkboxes: Phone, Email, Secure Message)
Did I record the call? (Yes/No – Check your local laws as many areas require you to ask permission to record.)
- My Script: “I need to record this call so I have a record of what to do next. Do I have your permission to record?” Note. Be sure to note this to any new call participants.
- Note: If they say “No,” I immediately ask: “Since you are recording this for quality assurance, can I request a copy of that recording for my records?” (This usually changes the tone of the conversation!)
Name of representative: (Crucial! Always ask for this first)
Reason for call: (e.g., Prior Authorization, Billing Error, Benefits Question)
Summary of call: (What did they say? What did you say?)
How long were you on the call? (This is important evidence for complaints)
Action items: (What did the representative promise to do? What do you need to do?)
Follow-up date: (When should you check back?)

Settings for the form. Some are personal, but this let’s me and Josh know who entered the data.

I have the email addresses required by default and the same for questions. (You can set some as not required.)

How to Set It Up

Go to forms.google.com and click “Blank Form.”
Add the questions listed above.
Click “Send,” copy the link, and email it to yourself and your spouse/caregiver.
Pro Tip: Open the link on your phone, tap “Share” (iOS) or the menu dots (Android), and select “Add to Home Screen.” Now it looks and acts just like an app.

The Result

When my pharmacy billing nightmare happened, I didn’t have to rely on my foggy memory. I opened my spreadsheet and saw exactly who disconnected on me on January 29th, and exactly what “Curtis” told me on February 6th.

That data wasn’t just notes; it was evidence. And evidence is the only thing that wins insurance appeals.

Disclaimer: I am a Speech-Language Pathologist and person with chronic illness, not an insurance broker or attorney. This post shares my personal experience and is not intended as legal or financial advice.

February 9, 2026

Access and advocacy, communication log, health insurance

Simple Kanban Task Tracker! Free!

Written by Stephanie M. Nixon, PhD, CCC-SLP on April 23, 2025. Posted in Apps, Conversations, Discord, Education, organization. No Comments on Simple Kanban Task Tracker! Free!

Organize your tasks visually with this easy-to-use Kanban board!

This tool helps you manage your workflow by moving tasks through “To Do”, “In Progress”, and “Done” columns. It runs directly in your web browser and saves your data locally, making it a simple, private way to stay organized.

Key Features:

Visual Workflow: Classic Kanban layout with “To Do”, “In Progress”, and “Done” columns.
Drag & Drop: Easily move tasks between columns using your mouse or touchscreen.

Custom Categories: Add your own categories (e.g., “Project X”, “Follow Up”, “Home”) to better organize tasks. Predefined categories include “Work”, “Personal”, and “Urgent”.

Color Coding: Tasks automatically change color based on their column (“To Do”, “In Progress”, “Done”) for quick visual status checks. Category badges also have distinct colors.
Confetti Fun!: Get a burst of confetti whenever you move a task to the “Done” column. 🎉

Local Storage: Your tasks and custom categories are saved directly in your browser, so they’ll be there when you reopen the app on the same device and browser.
Clear Completed Tasks: Easily clear all tasks from the “Done” column with a dedicated button.

How to Download and Use:

Download:
- Click the download link provided
- The license terms (GPLv3) are detailed below and available on the blog post/GNU website.

Nixon Speech and Language LLC KanBan Task Tracker Download

Open: Double-click the downloaded .html file. It will open in your default web browser. No internet connection is needed after opening.
Add Tasks: Type a task description, select a category from the dropdown, and click “Add Task” or press Enter.

Image of adding a text and selecting a category

Add Categories: Type a new category name in the “Add New Category” section and click “Add Category”. It will now appear in the dropdown list for tasks.

Move Tasks: Click and drag (or tap and drag) tasks between the columns.
Clear Done: Click the “Clear Done” button in the header of the “Done” column to remove all completed tasks.

Important Notes:

Local Data Storage: Task and category data is saved only in the browser you are currently using on this specific device. It will not sync automatically across different computers, tablets, phones, or even different web browsers (like Chrome vs. Firefox) on the same device.
Updates: For the latest version of this tool, please check the Nixon Speech and Language Blog or our Discord Community. Follow our blog or social media channels (linked below) for update announcements.

License:

This program is free software distributed under the terms of the GNU General Public License Version 3 (GPLv3). Essentially, this license guarantees you the freedom to use, study, share, and improve the software. You can redistribute it and/or modify it under the terms of this license.

Key points to understand:

Freedom: You are free to use, modify, and share this software.
Sharing Changes: If you modify the software and distribute your version, you must also license your modified version under the GPLv3 and make the source code available. You cannot make a modified version proprietary (closed-source).
Commercial Use: You can charge a fee for distributing copies or offering support/warranty for the software, provided you follow all GPLv3 terms (like providing the source code and keeping it under the GPL).
NO WARRANTY: This software is provided “AS IS” without any warranty. Nixon Speech and Language, LLC is not liable for any issues or damages arising from the use or modification of this software by others, as detailed in the full license text.
Brand Protection: The GPL license applies to the code. The Nixon Speech and Language name and logo are trademarks and are not automatically licensed for use by the GPL.

The full terms can be viewed on the GNU GPL website. Please refer to the full text for complete details.

Kanban Task Tracker, Organizer, Planner