The image shows a silhouette of a woman balancing on one leg. She holds the foot of the other leg, which is bent over her back, with both hands. She stands on uneven, dark ground with a lavender, pink and yellow background reminiscent of a sunrise or sunset.

As my diagnostic odyssey plunges ahead through choppy waters, I’ve encountered quite a few eddies. The quest now is to determine the cause of my diffuse bone marrow edema. Despite a recent negative experience with a new pulmonologist (Post 19), I’ve seen several clinicians who’ve filled me with hope, including my:

• Indefatigable PCP, the most patient in the world;

• Sharp-eyed PT, who modified my PT regimen because of bone pain;

• Thoughtful dentist, who’s comped some visits and X-rays as he fixes my 5 cracked teeth;

• Calm counselor, who always recenters me;

• Kindly new rheumatologist, who was very gentle and ordered labs, all of which came back negative. He recommended I see Endocrinology (chronic healthcare is a bit like pinball, bouncing patients among specialties); and

• Proactive Pain infusion NP, who ordered more imaging (which I haven’t done yet).

Despite the bullet points, my writing wasn’t AI-generated.

This list isn’t exhaustive, but I am exhausted by frequent healthcare visits (most of which require out-of-state travel). This Post describes recent visits with 2 new clinicians. In concert with my A-Team listed above, they’ve reminded me that, despite my struggles, healthcare abounds with smart, empathetic folks.

High-Risk Breast

I’ve had pain beneath my mastectomy scar, which is a red flag requiring exploration. My previous clinician left the health system, so I scheduled with an NP named Mandy for an hour-long appointment. A visit this length is a rare luxury!

From the beginning, I had Mandy’s full attention. She was warm and never seemed rushed. I used the eCheck-In’s free-text boxes to highlight my concerns (maybe “lowlight” is a more appropriate term for describing chronic illness). For some reason, Mandy couldn’t access this. I kicked myself for not bringing a printed page with me, but I’ve been rushed and I thought the e-Check-In was enough. I briefly explained my chief complaint. She’d already read my chart. Not everybody does.

Next, we reviewed my med list, medical/surgical history and family cancer history. There were duplicate and even triplicate entries of the same info. I asked why.

“Epic generates multiple entries for multiple visits. It’s ridiculous,” Mandy said, shaking her head as she edited.

After her exam, Mandy ordered my yearly breast MRI 2 months early. I casually mentioned that sarcoidosis can be confused with cancer on imaging.

She nodded, “I’ve diagnosed some patients with sarcoidosis before.”

This statement pinged through my head like a mic drop. Most clinicians know little about sarcoidosis. But Mandy had actually diagnosed it. She got it. Her kindness and knowledge were invaluable. And thankfully, my medical hot streak continued when I saw a new cardiologist the following week.

Cardiology

A medical assistant (MA) called me before this initial visit to ask if the office needed to query notes from any other facilities.

“No,” I responded. Then, “I’m looking forward to seeing Dr. K because the website says he specializes in connective tissue disorders.”

“Oh,” stammered the MA. “The website isn’t always correct. He’s just a general cardiologist.”

My heart sank as I voiced my disappointment, explaining my cutis laxa (CL) diagnosis. Maybe I should cancel?

The MA said, “I get it! I have the rarest form of Ehlers-Danlos. Most providers don’t know about it. But we have good cardiologists here. I think you’ll be OK.”

I reluctantly agreed. I’m really glad I did.

Dr. K was a few minutes late that day. I never care about that because he was likely clicking through my dense chart, piecing my story together like a crocheted quilt. He towered over me when he introduced himself and sat down.

I mentioned my conversation with the MA.

He replied, “Well, connective tissue disorders aren’t my official specialty, but I have Marfan syndrome, so I dabble.”

Marfan syndrome (MFS) is a connective tissue disorder affecting the protein (called fibrillin) that provides a scaffold for elastin fibers. People with MFS are often tall with long, slender limbs and digits. They may have problems with their eyes, heart valves, aorta and skeleton.

I handed Dr. K the paper I’d prepared highlighting my relevant medical history. He scanned it and asked if he could keep it. I nodded.

He unfolded his long limbs as he shifted from the computer to face me.

“When and where were you diagnosed with CL?” he asked.

The resulting conversation made it clear that Dr. K had done his homework. Except for my current dermatologist and the one who’d diagnosed my CL, I’ve never seen a clinician who knew so much about my condition. He was also sarcoidosis-fluent. The heavy pressure that comes from seeing someone new and anticipating another round of gaslighting instantly vaporized. I was so relieved, it felt like I was floating.

He did an exam and then moved to the Beighton Score, which is used to detect joint hypermobility.

“Can you do this?” he grinned, curling his long pinky backwards past 90 degrees. It felt less like an exam and more like a good-natured contest. A former gymnast, yoga enthusiast and rock climber, I’ve always been disturbingly flexible. But my pinky only bends to 90 degrees.

“How about touching your toes?”

Not to be outdone (and admittedly showing off), I folded over, nose to knees, touching the ground several inches behind my feet. My arms don’t extend past 180°, but my knees do (though unfortunately, doing so caused my left knee to repeatedly buckle the next day). I balked at pushing each thumb to touch its respective forearm, as my delicate thumbs have been surgically rewired with arthroplasties.

Dr. K ordered more testing, which I have yet to complete, and asked me to return in 6 months. Most importantly, he understood my rare diseases, took my concerns seriously, and established empathy and rapport by citing his MFS.

Conclusion

Good news: so far, most testing has come back negative, with the exception of a mildly abnormal bone scan that didn’t reveal a cause. More good news: I’m adding Mandy and Dr. K to my A-Team, all of whom have restored my shaken faith in medicine. Good clinicians are out there. Sometimes, it just takes awhile to find them.

Leave a comment describing who is on your A-Team, and why?

One Last Point:

In Post 19, I stated, “I will die young.” Obviously, I don’t know when I’ll die. For instance, Dad’s grandfather, father and brother died before age ~50, so for decades, he paid into term-limited life insurance that expired at age 75. Now he’s almost 79 and kicking himself for it. Everything I’ve read about adult-onset cutis laxa (excepting 1 paper) states it’s associated with poor outcomes. But it’s more accurate (and much less melodramatic) to say that I likely won’t reach my predicted life expectancy.

The image shows an asymmetrical, brown-skinned sweet potato with a bright green background. The sweet potato casts a shadow from overhead lights, as if it were the star of a Broadway show.

I slid my tongue over the roof of my mouth, feeling a painless, convex blob. Suddenly -

POP!

The blob abruptly ruptured, and a thick, briny liquid gushed out. I rushed to the mirror, jaw agape. Tattered pink gum tissue flapped near an angry, reddened area. I’d popped a quarter-sized blister.

I’d just finished roasted sweet potatoes smothered in chipotle vegan mayo (yum!) and must’ve burned my mouth, but it didn’t hurt. I’d also banged the back of my throat while eating (thank you, muscle spasms!), but hadn’t gagged. Intrigued, I jabbed a toothbrush down my throat. No sensation and no gag. Wtf?

My synapses flooded with fear. This was summer of 2020. I’d been dealing with falls and other neuropathy-related issues for 2 years. Was this more neurological fallout?

I messaged my pulmonologist (Dr. E) that I’d lost my gag reflex.

Her response: “I’m so sorry.” She, too, thought this was an ominous sign.

What Was Happening?

I puzzled over these symptoms. Although an isolated cranial nerve impairment was possible (especially with sarcoidosis), it seemed unlikely, and I was pretty spry to have such a severe deficit. The answer hit me suddenly, like a boulder trundled off a 1000-foot (305-m) cliff.

I’d started a steroid inhaler (Qvar) 6 weeks prior for a pesky cough I got after a week in the city for appointments. Wildfires were raging, and the Air Quality Index had soared past 300. The cough made talking and eating difficult, so I’d messaged Dr. E, and she’d prescribed the Qvar.

Online sleuthing revealed steroid inhalers can cause vocal cord paralysis and alter mouth sensations. My pharmacist agreed Qvar was the likely culprit. Since my cough had abated, Dr. E discontinued the inhaler.

The moral of this story could be: It’s not my job to figure this out or, Be wary of medication side effects. Instead, I’m describing what I should’ve mentioned to the new pulmonologist I saw last week. I thought the appointment had gone well, but as I drove away, I realized he’d gaslit me. And in turn, I’d gaslit myself.

The Diagnostic Odyssey Continues

Recently, I’ve had gnawing bone pain. MRIs have revealed bone marrow edema scattered throughout my skeleton, as well as severe degenerative joint disease (DJD). Additionally, my Sports Medicine MD agreed that my rapid osteoporosis and DJD are premature, and sarcoidosis must be ruled out (I’ll omit the extended differential diagnosis list for now). My previous pulmonologist had managed my sarcoidosis, so I scheduled a pulmonology appointment.

The New Pulmonologist

The visit started okay. Dr. R smiled and said he doesn’t treat systemic sarcoidosis – only lungs.

“You don’t want me managing anything else!” he winked.

“Excellent PFTs,” he remarked. Some of my lung volumes were greater than 120% of predicted. Prior values have approached 150%. Like most clinicians, he thought this meant I’m uber-fit.

But that’s not how cutis laxa (CL) rolls. CL occurs when the body breaks down elastin, a protein that allows tissues to snap back after stretching. CL can happen only in the skin (it means loose skin) or throughout the body. My neuropathy and DJD suggest systemic CL, which can age patients by decades over just a few years (see the intriguing I’m 20 But Look 60. Note: lipodystrophy and CL are described here as mutually exclusive, but can co-occur).

Abundant elastin gives the lungs their springiness, so patients with CL are at high risk of lung complications. While CL is likely causing my DJD, it could be playing with sarcoidosis in the sandbox. Things could go south quickly.

When I tried to hand Dr. R the papers collating abnormal labs and sarcoidosis history from several facilities over the previous 8 years, he glanced down and immediately handed them back (so much for bringing paperwork to appointments as a way to be understood!). At this point, I should’ve explained CL better.

Instead, I blurted, “I’m aging faster than my chronological age.”

He burst out laughing.

What happened next? I wish I’d said, “I can’t see a doctor who doesn’t take me seriously,” and left. Or, “Why is it funny that I’ll likely die young?” Or my favorite ploy, to play dumb while leading him to the answer by asking meekly, “Could elastin deficiency be causing abnormally large lung volumes?”

Instead, I felt embarrassed for him, despite having been ridiculed by other clinicians, whom I never saw again and whose dismissals about my rare diseases were ultimately proved wrong.

“Did you know CL has progeroid forms?” I asked. His eyes widened. Progeria is the heartbreaking disorder children who prematurely age and die get. They may be in their teens, but they look like 80-year-olds.

“It’s OK,” I shrugged. “I laugh about it, too.”

Fact. I’m a modern-day, reverse Benjamin Button with grotesque skin-stretching party tricks. Dr. R had likely never heard of CL. But instead of being curious and connecting the dots, he’d reduced me to “excellent” test numbers and gaslit me.

What Does This Have to Do With Sweet Potatoes?

When I mentioned the 2020 inhaler, Dr. R gaslit me again.

“Oh, pulmonologists think we need to prescribe inhalers for the slightest sniffle,” he sniffed.

I didn’t explain why I’d needed the inhaler. Maybe I’d intuited he’d already written me off and what I said wouldn’t matter. Because I’d used portal messaging at a different facility, the reasons for the inhaler weren’t documented. In medicine, this means it had never happened. Dr. R had now also gaslit me because he didn’t believe I could’ve ever needed an inhaler.

Self-Gaslighting

When I left, I thanked Dr. R and said I would give him 5 stars. I don’t know why; maybe because I felt bad for him? After reviewing our conversation, I realized it hadn’t been a good appointment. He hadn’t taken me or my evidence-based concerns about my rare disease seriously. Moreover, I hadn’t listened to the inner voice warning me not to be vulnerable to yet another new clinician. I wrote about medical gaslighting in Post 4 and it’s happened countless times, so you’d think I would’ve been prepared for this. I’d been hopeful that somehow, this visit would be different.

Note to clinicians: patients never forget the sting of ridicule.

I write this on the eve of a visit with a new rheumatologist, a referral I asked for because I realized I have a systemic problem. I won’t bring him any paperwork (though I’ll bring my own prompts). I used Epic’s eCheck-In free-text boxes to bullet-point my sprawling problem list, and I’ll try to be more vocal about my concerns.

I don’t expect every (or really, any) clinician to understand cutis laxa. But I need clinicians who, as Ted Lasso suggests, are curious, not judgmental. Exhaustion and medical trauma, coupled with the grind of knowing I probably won’t be taken seriously yet again, have profoundly altered my view of medicine. I can’t stop clinicians from gaslighting me, but I can stop me from gaslighting myself.

Have you ever gaslit yourself about something important?

References and a recording will be appended once I replete my spoons.

The photo shows a golden sunrise on a mountain with a red-tinged lake, white rocks and green and red grasses in the foreground. Yellow-pink puffy clouds hover above the mountain and evergreen trees dot the talus below it.

Today, I’ll describe some resources that helped make Diagnostic Odyssey a reality. It’s hard to discuss this topic without mentioning financial strain. This isn’t a feeble attempt to garner sympathy or gain donations, but instead gives context to the stark reality of living with chronic illness in the US. A sobering Forbes article calls this a uniquely American problem, noting that up to 2/3 of all bankruptcies result from medical debt and that more than a quarter of Americans forgo medical care because of the cost.

Alas, being sick in the US often leads to poverty, despite support from Social Security Disability Insurance (SSDI), Supplemental Security Income (SSI), Medicare and Medicaid. I don’t receive SSI or Medicaid and Medicare is a sprawling topic, so today I’ll briefly describe SSDI.

SSDI

SSDI benefits rarely cover monthly expenses for disabled folks. We make up some of the difference through Section 8 housing vouchers, food banks and Supplemental Nutrition Assistance Benefits (SNAP) – also known as food stamps or electronic benefit transfer (EBT). We buy grocery store specials and thrift store clothing or “learn” how to sew (see my janky jorts’ Franken-stitching below).

The image shows a close-up of the pantleg hem of a pair of jean shorts with 2 seams crookedly stitched with differently colored blue threads.

Sadly, more and more people need these benefits, at a time when the US government is massively cutting SNAP and Medicaid benefits and increasing barriers for those who need benefits in the years to come.

The average monthly SSDI payment is based on work history and differs in each state but ranges from $1358 to $1694/month for a 1-person household.

Could you live off that amount? Especially if you had out-of-pocket costs for co-pays, out-of-state travel to see specialists and all the items needed to stay healthy and safe that aren’t covered by insurance (e.g., air filters, compression garments and supplements)? Thankfully, folks like me can work part-time and continue receiving SSDI.

Working While on SSDI

Working while on SSDI is allowed provided we stay under certain income and work hour requirements. The 2026 income cap for Substantial Gainful Activity (SGA) is $1690 gross per month for a single-person household. If you overshoot your SGA cap in 9 non-consecutive months, your SSDI payments are suspended.

I started this Substack so I could help others in similar situations and hopefully, to make a few extra bucks along the way. I couldn’t have gotten this far alone. When I started this journey, I reached out to the local Women’s Business Center, and doing so was time well spent.

Women’s Business Center

The name Women’s Business Center (WBC) is a misnomer, since they help anyone starting a small business. The WBC offers free one-on-one coaching sessions and prerecorded online classes, as well as grants and small business loans. Many states and even some cities have their own Business Council or Chamber of Commerce to promote and support small businesses.

I originally wanted to sell greeting cards, prints and T-shirts showcasing photos of my travels (see example above), but didn’t know where to begin. After two 9-week WBC classes, I had viable business and financial plans, including projections for start-up costs and profit/loss. The vestiges of this print-on-demand business venture exist here and at the top of this page. Sales have been unremarkable to non-existent, mostly comprising purchases by sympathetic family members (thank you, cousin Emily!).

The WBC helped me start up and test drive my card business, but their best advice was to refer me to the local Vocational Rehabilitation office.

Vocational Rehabilitation

Vocational Rehabilitation (affectionately called Voc Rehab) provides support for disabled folks looking for employment, assisting with job searches, recruitment, retention and access to care needs. Voc Rehab assigned me a case worker, who coordinates my health- and business-related activities. She’s like a holistically mission-focused cheerleader.

I also have a Voc Rehab financial advisor, who reviews financial plans, projections and SSDI income reporting requirements. She explained how different income levels would affect my SNAP and housing benefits. She has also patiently, repeatedly, explained opaque concepts like SGA.

My financial advisor helped me determine the print-on-demand business would be a bust. Selling hundreds of cards to barely break even wouldn’t work out for me. She assisted me in pivoting to writing this newsletter.

Writing Diagnostic Odyssey

The switch from hawking photos to writing wasn’t as drastic as it seems. Both are creative endeavors. And I’d already started writing a book about my healthcare experiences, producing 61 rambling pages. Yet writing a book felt nearly insurmountable. A book needs structure, a coherent storyline, countless edits, an agent and a publisher. And writing a book while running a separate business and handling my health needs would’ve been impossible.

Starting a Substack made sense, in part because I can publish info in specific quanta whenever I can on whatever topic I want (though most are related to my diagnostic odyssey). Voc Rehab directly funded my Substack start-up costs with a grant, which paid for a much-needed ergonomic keyboard, a subscription to Sarah Fay’s Substack Writers at Work (not a sponsor of Diagnostic Odyssey) and a box of printer paper. They’re subsidizing half of my internet bill for 6 months and have even reimbursed medication fees – $600 over 3 months.

Voc Rehab helped me set up deadlines with the gentle understanding that intervening health issues (like 2 surgeries and a punishing 2-month-long C. difficile infection in 2025 alone) could derail any plan. And they provided the accountability my attention deficit needs. They aren’t magicians, but they’ve removed as many barriers as possible so I can achieve my publishing goals.

Diagnostic Odyssey has profoundly changed my life. It’s given me a sense of purpose and connected me to other folks in the chronic illness community. It’s helped me process my medical trauma in a healing, if at times overwhelming, way. If you’re in a similar situation, I hope your local Business Center and Voc Rehab office can help you. Even if you’re not, perhaps you know somebody who can benefit from these resources. My experience is proof-of-concept that local business councils and Voc Rehab help Spoonies get closer to realizing their dreams.

What resources have helped you achieve your Substack newsletter goals?

The image shows a logo of a blue galley ship that has eyes and a blue caduceus as its mast. Photo credit.

I summarized my previous Posts as I mulled over what to write about this week. Then I realized my readers might appreciate this summary, due to the disjunct nature of Substack newsletters. Reading them thrice monthly reminded me of planning a potluck via text with well-meaning but indecisive friends over a weeklong span. By the end, you’ve forgotten who’s bringing what. All you know is that you’re bringing guac and chips.

That’s really nobody’s fault. Emails, texts, videos, alerts and notifications perpetually intrude upon our lives, imparting a false sense of cognitive fluency and multi-tasking prowess. But experts have correlated task-irrelevant distractions with reduced comprehension and recall.

So, as a sop to our digitally-addled minds, and as an intro for anybody just joining, here’s a refresher.

Summary

I’m a disabled former nurse practitioner with a Ph.D. I started this newsletter in January to describe my wayward quest to receive rare disease diagnoses, with a focus on how medical record errors have stymied the process. I’m still building this plane as it flies, sometimes rattling through unnerving turbulence, but overall, just publishing anything feels like a great success.

Below are some terms I’ve introduced and a bullet-pointed summary of each Post.

Terms (*denotes my own neologism)

• *Creeping narrative – A type of medical record error in which the patient’s physical attributes, disease characteristics or lived experiences are serially altered across multiple notes.

• *Gospel bias – The belief that everything in the medical record is correct, regardless of what the patient says.

• Medical gaslighting – According to Fuss et al., “behaviors inflicted on an individual which invalidate or call into question their ability to judge their own [health-related] lived experience.”

• *Phantom findings – Info documented in the medical record for an exam that wasn’t performed or a symptom or question that wasn’t asked about.

• Spoons – According to Christine Miserandino’s Spoon Theory, spoons represent the daily allotment of energy folks with chronic disease have. When you’re exhausted, you’re out of spoons. The term Spoonie refers to a person who is ill or disabled.

• Zebra – A person with a rare disease.

Posts

• 1: My Doctor’s AI Tool Output Slop

  Describes the historical and current uses of artificial intelligence (AI) in medicine. I list suspected AI-generated note errors I found after I saw a specialist. With an Erratum (Post 9).

• 2: Creeping Narrative

  The diagnostic odyssey began in March, 2018, when my knees swelled overnight. None of my clinicians got my story right.

• 3: We Still Owe Healthcare Workers a Debt of Gratitude

  Musings after a positive initial visit with a surgeon.

• 4: Gaslighting

  I was repeatedly dismissed as a “stressed-out grad student” despite having serious symptoms.

• 5: Rare Diseases

  Overview of zebra facts, with links to relevant websites.

• 6: Symptoms

  Explains the importance of accurate symptom documentation and how clinicians use the OLDCARTS mnemonic to describe an illness.

• 7: A Diagnostic Clue, At Last

  In June, 2018, imaging showed swollen lymph nodes in my lungs – a classic sign of sarcoidosis. Describes the emotions that accompany a life-altering diagnosis.

• 8: What is Sarcoidosis?

  A primer on a disease characterized by granuloma formation and inflammation, with links to patient resources.

• 9: Erratum for Post 1

  When I Zoomed with my specialist to correct note errors, she confirmed she had documented the incorrect diagnosis. All other errors were AI-generated.

• 10: The First Primary Care Provider

  In July of 2018, I got a PCP to coordinate my care. Discusses how the first visit went down and the importance of having a PCP.

• 11: How Do I Amend My Medical Records?

  Describes my process for submitting medical record amendment requests and disagreements.

• 12: More About Medical Record Amendments

  Recounts some of the frustrating issues I’ve had while trying to correct my medical record.

• 13: Professional Patient

  I had ~20 appointments and tests in July of 2018. My identity had begun to morph from being vigorously healthy to being a professional patient.

• 14: Stress Test

  Recounts how the documented test results contrasted with my lived experience and clinical presentation.

• 15: Phantom Findings

  Phantom findings abounded in the note for my first respiratory clinic visit.

• 16: Phantom Findings Redux

  Discusses my personal experiences with phantom findings as an ICU nurse and patient, including factors contributing to this phenomenon.

Sixteen Posts! So far, we’ve covered the first ~3.5 months of my 8-year diagnostic odyssey, with a few (relevant, I hope) tangents on the way. I’ve been meeting a lot of folks on Substack in similar situations. I’m humbled by all the support from everyone. None of this would exist without all of you. See you on the 19^th!

Food Bank Donations

This week, I’m asking if any of you who are able to please donate to Feedingamerica.org. Many of us count on food banks to survive and thrive. Thank you!

References

Alsuwaylim, A. M., Alshehri, O. A., Alzahrani, S. Y., Alsubaie, F. B. S. M., Ali, M. O., & Yacoub, S. A. (2026). The effects of digital multi-tasking and task switching on working memory: A systematic review. Research Journal in Advanced Humanities, 7(2). https:// doi.org/10.58256/r59zkp31

Das, S. (2025). The spoon theory: A journey into the world of chronic illness. Literary and Cultural Disability Studies: A Primer. World View Critical Editions. https://www.researchgate.net/publication/391661264_The_Spoon_Theory_A_Journey_into_the_World_of_Chronic_Illness

Fuss, A., Jagielski, C. H., & Taft, T. (2024). We didn’t start the fire…or did we?-a narrative review of medical gaslighting and introduction to medical invalidation. Translational Gastroenterology and Hepatology, 9, 73. https://doi.org/10.21037/tgh-24-26

I introduced the term phantom findings in Post 6 and mentioned I’ve found > 30 instances of phantom review of systems (ROS) findings in my medical record (MR). Post 15 further explored this phenomenon by recounting my first respiratory clinic visit with Dr. Z, a pulmonologist who documented phantom history, exam and ROS info.

When I mention the phenomenon of phantom findings to my clinicians, they usually fall silent. Which makes me wonder if they agree, but don’t want to say so, or if they simply don’t think this happens.

At a telehealth appointment with a doctor in Turkey last year, I agreed to forward my lab and radiology reports.

“I’d rather not send the clinician notes. My MR has so many errors,” I said gingerly, bracing for an argument.

He frowned and his hand flicked a dismissive gesture as he replied, “I don’t read clinician notes from the US. They’re full of errors copied over and over again.”

I was stunned. He’d just openly validated my experience that phantom findings happen frequently.

More Examples of Phantom Findings

Some phantom findings, like normal respiratory rates, are almost universal (nobody ever looks at me long enough to count anything). Phantom exam findings are less common and usually cover a body system (e.g., a skin or abdominal assessment). Phantom findings for entire or near-entire “normal” exams are rarely documented, but it’s happened a few times:

• In 2020, an ophthalmologist only examined my eyes, calling out his findings while his medical scribe tapped at the computer. I don’t know whether the physician or the scribe copy-pasted the phantom findings for the rest of the exam.

• In 2021, I had 2 visits with a new orthopedic PA who had no stethoscope (but wrote my lung, heart and bowel sounds were normal).

• In December 2025, at an urgent care visit for a nasty Clostroides difficile (C. diff) infection, the PA pressed on my belly to rule out a severe infection – the extent of his exam.

  • I went to the same facility’s ED that night. The physician listened to my lungs and then stepped away for a phone call. He never finished his exam and instead, copy-pasted the PA’s “normal” phantom exam findings from earlier in the day.

Like genetically modified crops drifting onto an organic farm, even small MR errors can blossom into fields of inaccuracy as they cross-pollinate through healthcare. (Maybe not the best metaphor, but you get the idea).

Phantom findings are harmful. So why do they happen? I believe it’s mostly because the healthcare system sets up workers to fail.

Healthcare Enables Phantom Findings

I ran my tail off working in the ICU. I almost never had a sit-down lunch in the break room. Instead, I inhaled my barely palatable protein bars and uninspiring microwaved lunches at the nurses’ station. And after a 4-night stint of 12-hour shifts (more like 13 or 14 hours), my eyelids were heavy as lead.

I’m certain there were instances in which I forgot whether I’d turned a heavily sedated patient every 2 hours or provided oral care on time, yet mindlessly auto-populated confirmation that I’d done so in an exhaustive post-shift haze. Turning patients avoids pressure ulcers, which can become painful and infected and rarely, kill you (this happened to Superman actor Christopher Reeve). Oral care reduces the risk of ventilator-associated pneumonia, which affects up to 40% of ventilated patients.

Both are important preventative tasks. But so is keeping your hemorrhaging patient alive through chest compressions or back-to-back infusions of packed red blood cells.

I’m not proud of skimping on patient care or documenting phantom tasks. Nonetheless, in the stress of such a dynamic environment with constantly changing priorities, something must give. Overwhelmed healthcare workers who are doing the best they can with lack of time and resources must employ workarounds. Other factors contribute:

• Complex and at times, conflicting info must be gathered from multiple sources. Dr. Z had to collate a verbal interview, the paperwork I’d filled out and the MRs that UHS had sent. This is time-intensive.

• Templated electronic MRs. These enable rapid documentation through copy-pasting and auto-population, but may carry over incorrect/outdated info. Clinicians must also shoehorn info into the note-writing template, which can lead to omissions.

• Not taking notes. Medical school is rote-intensive, which can impart overconfidence in memory. But patients are complex and it’s easy to lose track of an individual patient’s details. AI tools can help but they aren’t always accurate (Post 1).

• Confirmation bias. If somebody looks well, it’s easier to predict exam findings will be normal and to skip parts of the exam and document them as normal. Looking normal while feeling shitty is common in patients with sarcoidosis. Having PFTs and stress test results showing I was “physically fit” likely didn’t help my case.

• Burnout. So many factors cause burnout in healthcare workers! An NIH consensus report lists some on the chart below and notes, “… when psychological or psychosocial capacities are exceeded in response to repeated exposure to suffering, death, or social inequities without sufficient structural mechanisms to address them, emotional exhaustion can significantly contribute to burnout.”

  Burnout can affect wellbeing, which affects job performance, which ultimately affects patients.

  

  The image is a chart showing the Work System Factors of the Systems Model of Clinician Burnout and Professional Well-Being. It lists Job Demands as excessive workload, unmanageable work schedules, inadequate staffing; administrative burden; workflow interruptions and distractions; inadequate technology usability; time pressure and encroachment on personal time; moral distress and patient factors. It lists Job Resources as meaning and purpose in work; organizational culture; alignment of values and expectations; job control, flexibility and autonomy; rewards; professional relationships and social support; and work-life integration.

• CYA. Healthcare workers must constantly adapt to complex, ever-changing documentation requirements. One physician friend said he feels his hospital’s administration sees avoiding litigation and billing as the main goals for clinician documentation.

• Billing pressures. A 2021 rule change allows clinicians to bill for medical decision-making time or total time spent on each patient. Before then, addressing more body systems in the ROS and exam sections of a note allowed the facility to bill for more money.

I’m not trying to make abject excuses for why phantom findings occur. Nor can I understate the real, sometimes lasting harm and delays in the diagnostic odyssey that phantom findings can cause. However, I believe recognizing and naming the phenomena that lead to lapses in patient care are the first steps in fixing these systemic problems.

Do you have a story about phantom findings? Leave a comment below.

Glossary

CYA – Covering your ass; documenting to meet mandated standards or to avoid litigation.

Clostroides difficile – Gastrointestinal infection causing severe diarrhea. Previously called Clostridium difficile.

Coding – 1. Short for code blue; occurs when a patient’s heart or breathing (or both) stops, necessitating resuscitation. 2. According to the AACP, “The transformation of healthcare diagnosis, procedures, medical services, and equipment into universal medical alphanumeric codes.”

ED – Emergency department.

ICU – Intensive care unit.

NIH – National Institutes of Health.

PA – Physician’s assistant.

ROS – Symptom list.

Ventilated – Attached to a ventilator that breathes for a patient or assists the patient with breathing.

References

Avsar P, Moore Z, Patton D, O’Connor T, Budri AM, Nugent L. Repositioning for preventing pressure ulcers: a systematic review and meta-analysis. J Wound Care. 2020 Sep 2 [cited 29 Jun 2026];29(9):496-508. doi: 10.12968/jowc.2020.29.9.496. PMID: 32924821. Available from: https://pubmed.ncbi.nlm.nih.gov/32924821/

Drent M, Costabel U, Crouser E et al. Misconceptions regarding symptoms of sarcoidosis. The Lancet Respiratory Medicine, 2021 [cited 29 Jun 2026]; 9, 816-818. https://pubmed.ncbi.nlm.nih.gov/34216548/

Greene, L. CHRISTOPHER REEVE, 1952-2004 // Common pressure wound led to fatal infection. Tampa Bay Times. 2004 Oct 12 [cited 29 Jun 2026]. Available from: https://www.tampabay.com/archive/2004/10/12/christopher-reeve-1952-2004-common-pressure-wound-led-to-fatal-infection/

Kohbodi GNA, Rajasurya V, Noor A. Ventilator-Associated Pneumonia. [Updated 2023 Sep 4] [cited 29 Jun 2026]. In: StatPearls [Internet]. Treasure Island (FL): StatPearls Publishing; 2026 Jan-. Available from: https://www.ncbi.nlm.nih.gov/books/NBK507711/

National Academies of Sciences, Engineering, and Medicine; National Academy of Medicine; Committee on Systems Approaches to Improve Patient Care by Supporting Clinician Well-Being. Taking Action Against Clinician Burnout: A Systems Approach to Professional Well-Being. Washington (DC): National Academies Press (US)[Internet]; 2019 Oct 23 [cited 29 Jun 2026]. 4, Factors Contributing to Clinician Burnout and Professional Well-Being. Available from: https://www.ncbi.nlm.nih.gov/books/NBK552615/

Mada PK, Alam MU. Clostridioides difficile infection. [Updated 2024 Apr 10] [cited 29 Jun 2026]. In: StatPearls [Internet]. Treasure Island (FL): StatPearls Publishing; 2026 Jan-. Available from: https://www.ncbi.nlm.nih.gov/books/NBK431054/

What is Medical Coding? American Academy of Professional Coders [Internet] [cited 29 Jun 2026]. Available from: https://www.aapc.com/resources/what-is-medical-coding

The photo shows a spirit in a black hooded robe. The face and arms can’t be seen. The bottom of the robe splays outward, as if the spirit is floating and turning. The dark, muted background shows a forest of trees with orange leaves, likely at dusk. The tone of the picture is foreboding.

There’s a glossary at the end of this post.

As June of 2018 wound down and summer temps soared, my symptoms worsened. I lamented the 7-week wait until my university health system (UHS) pulmonology appointment. In retrospect, I was privileged to receive a sarcoidosis diagnosis in just a few months. But they were the longest, most tortured months of my life.

After seeing a cardiologist that month, I’d also visited an electrophysiologist (Dr. T) at UHS. Dr. T had ordered my stress test. She’d also supervised the research study I’d participated in (which had revealed my dysrhythmias), so I had her phone number. Alone and overwhelmed, I left her a desperate message, begging for help. I missed her callback that evening, unarousable from one of my prolonged, stuporous slumbers.

Dr. T stammered in her voicemail message, belying her powerlessness to help me. She suggested I visit a local respiratory clinic. This simple gesture – calling a patient outside of business hours, sounding concerned – was the first instance of humanity I’d experienced in my diagnostic odyssey (there would be more throughout the years, restoring my shaken faith in medicine). Moreover, the clinic scheduled an appointment for the following week!

July 10, 2018: Respiratory Clinic Visit With Dr. Z, Pulmonologist

The purpose of this visit was to establish care so Dr. Z could biopsy the swollen lymph nodes in my lungs and confirm the sarcoidosis diagnosis (while ruling out tuberculosis or lymphoma). I completed PFTs that morning, which showed excellent pulmonary function.

Cheerful landscape paintings decorated the waiting room, which was full of coughing, haggard-looking older patients on oxygen. I felt out of place. What was I doing here with all these sick people? I’d figured being unwell was just a blip in my young and otherwise healthy, robust life. I’m ashamed to say this sentiment carried some underlying judgment: I wasn’t a chronically ill patient. I wasn’t one of them.

I filled out forms covering the usual suspects: histories (past medical, surgical, family and social), medications, environmental exposures and my extensive, multi-system ROS. After a few minutes, a petite MA led me to the exam room and took my vital signs before leaving.

Dr. Z entered, his spicy cologne diffusing to my nostrils before he was close enough to shake my hand. He nodded as I outlined my situation, skipping some ROS details I’d written on the forms. He listened to my lungs (the extent of his exam). When he began explaining the next steps in layperson’s terms, I mentioned I’d been a NP. His eyebrows arched in appreciation as he leaned back in his chair.

“So, you’re here for an EBUS. Why didn’t you just say so?” His smile was genuine. His arm swept toward the door as he got up to leave and said, “The scheduling desk is on the way out.”

I’m familiar with this shift in demeanor. It’s partly why I felt out-of-place here as a former NP and now, a reluctant patient. Switching from being the caregiver to the one who’s cared for triggers a cognitive dissonance that’s hard to overcome.

After Dr. Z learned I was part of (what I call) the Clinicians’ Club, he’d become more laid back. I would realize, when I read his note years later, a little too laid back.

ROS. Dr. Z accurately documented the following in the visit note’s HPI and ROS:

• “Fatigue/malaise

• Exercise intolerance

• Palpitations

• Headache

• Presyncope.”

But he omitted the following, which I’d checked or written on the form:

• “Visual changes

• Memory loss and reduced concentration

• Chest tightness

• Color changes on fingers

• Heat intolerance

• Sleep issues

• Urine smells strange (UA –)” (meaning my urinalysis was negative for any infection or abnormalities).

Dr. Z had also documented that I didn’t have some symptoms I’d checked:

• “Normal appetite” (mine was decreased)

• “No vision changes, hoarseness, nasal discharge, chest pain, palpitations, fainting, shortness of breath, cough, abdominal pain, anxiety, depression, muscle weakness, headache, fainting and falls.”

He’d even listed some of these in the HPI, causing discrepancies.

History. Dr. Z’s note claimed:

• “The patient’s past medical … surgical … family … social … and environmental history [were] reviewed, and there are no updates since their last visit.”

These items weren’t reviewed and this was my first visit. He only asked if I was a smoker (which I wasn’t). He’d obviously read neither the forms I’d filled out, nor the forwarded UHS MRs.

Physical Exam. Dr. Z falsely documented a normal exam including:

• “Neck – no masses

• Percussion … [and] palpation of chest

• Rate and rhythm … S1 and S2 … no murmurs

• Pedal pulses: 2+ bilaterally

• Abdomen: Non-tender, no masses … No hepatomegaly or splenomegaly

• Gait and muscle strength/tone

• Skin turgor

• Palpation of lymph nodes in neck … axillae … groin … [and] other areas: No lymphadenopathy.”

Scattered throughout the body, lymph nodes play an important role in immunity. Since sarcoidosis is immune-mediated, a comprehensive lymph node assessment is especially important in patients with this suspected diagnosis. My groin lymph nodes are palpable, which would’ve been documented had he done the exam.

Further, a male physician who performed this exam on a female patient (as well as chest percussion and palpation) would’ve had a female nurse in the room as a chaperone. There was no chaperone documented, because there was no chaperone, because the exams weren’t performed. Instead, these, and most of Dr. Z’s note, comprised phantom findings.

Phantom Findings

After years of chronic illness and countless hours reviewing MRs, I’ve concluded phantom findings are common. But at the time, I was blissfully ignorant that my MRs at UHS and the respiratory clinic were fraught with flawed and missing info. Add the biases based on my stress test, PFT findings and Clinician Club membership, and you get a perfect storm of mostly useless patient info. Perhaps most chilling is the realization that, if phantom findings were in my notes, they were likely in others’ as well.

Have phantom findings popped up in your medical records?

Buy Me a Med Co-Pay

Glossary

Dysrhythmia – Abnormal heart rhythm.

EBUS – Endobronchial ultrasound. Involves snaking a bronchoscope with a camera into the lungs. The scope can also biopsy swollen lymph nodes through the tracheal wall with a fine needle aspiration.

HPI – History of present illness.

Hepatomegaly – Enlarged liver.

Lymphadenopathy – Abnormal swelling of lymph nodes.

MA – Medical assistant.

MD – Doctor of Medicine.

NP – Nurse practitioner.

PFTs – Pulmonary function testing.

Pedal pulses 2+ bilaterally – Both the pulses on the top of my feet felt normal.

ROS – Review of systems. Symptom list.

S1 and S2 – Heart sounds. S1 is caused by the closing of the mitral and tricuspid valves. S2 is caused by the closing of the aortic and pulmonic valves.

Splenomegaly – Enlarged spleen.

The images shows a teal-colored adrenaline molecule, with an IUPAC name of 1-hydroxy-2-(methylamino)ethyl]benzene-1,2-diol. The 6-carbon ring has 2 opposing double bonds and two adjacent OH groups on the left and a chain on the right containing an OH group and an NH group.

In July of 2018, I had a flurry of appointments. I saw a new PCP on the 5^th. This article backtracks to July 2^nd, when I had a stress test. An electrophysiologist (Dr. T), whom I’d seen after a cardiology visit at the end of June, ordered it.

Cardiac Stress Test

This test measures functional aerobic capacity (FAC) and predicts the likelihood of heart issues. It involves monitoring blood pressure (BP) and a continuous electrocardiogram (EKG) as the patient performs staged exercise (the physiologic stressor) on a stationary bicycle or treadmill. If the patient is unable to exercise, medication is used as the stressor.

When this test was ordered, I’d forgotten to mention running on a treadmill would be difficult because of my knee arthritis. Which may seem silly, but I hadn’t yet accepted the moniker of chronically illness, nor had I acknowledge having limitations to my previously healthy life. Oops.

After I checked in, a technician named Ethan led me through a labyrinth of hallways to a tiny, windowless room with a laptop and treadmill. He and a nurse (Michelle) would run the test.

Michelle placed EKG stickers on my chest and Ethan took baseline manual BPs: 120/72 (while I was lying down) and 122/76 (standing). During the test, the speed and incline of the treadmill would be increased every 3 minutes. Ethan would take BPs at each stage and call out the values for Michelle to document on the computer.

The treadmill started rolling at a gentle pace.

“124/72” Ethan said.

Three minutes later, the treadmill groaned as the incline steepened and the pace quickened.

“124/66.”

“122/68.”

As the test continued, I relieved the stabbing pain in my knees by supporting much of my weight on the treadmill’s sturdy arm rails. Was this cheating?

I continued sweating on the treadmill. With my legs pumping faster and faster, body going nowhere, I must’ve looked like Wile E. Coyote.

The cartoon shows a blue roadrunner sticking his tongue out at a brown coyote who is chasing him with outstretched arms. Both the roadrunner’s and coyote’s legs are drawn in a circular way. Copyright Warner Bros. Entertainment. Picture found here.

“116/66.”

Exercise stresses the body, which typically causes BP to increase in a way that’s roughly commensurate with the workload. But my BP values were misbehaving, drifting downward like a soft, gauzy feather cutting half-arcs through the air. Until they suddenly plummeted like a feather dropped on the moon.

Ethan’s brow wrinkled, “I can’t hear anything. Write 120/70.”

My BP had become inaudible and I started feeling nauseous as the room spun. Familiar spots flashed before my eyes. I was about to pass out. Bad idea if you’re “running” (or whatever I was doing) at full tilt on a treadmill! I jumped off.

“Why did you stop?” Michelle demanded.

“I-I felt like I was going to pass out,” I panted.

My post-test BP rebounded to 162/78. My adrenaline had finally kicked in, too late to the party. After the last recovery BP (134/80), the test ended and Michelle left.

I suddenly started shaking violently. Adrenaline increases blood sugar uptake into muscles to prepare them for the jolt of energy needed to battle a vengeful ninja or run from a voracious saber-toothed tiger. This tardy and unruly adrenaline rush had likely dropped my blood sugar too low.

Ethan’s eyes widened.

“J-j-j-juice,” I sputtered.

He ran out and returned with juice. A few minutes after drinking it, I stopped shaking, but still felt weak. Wordlessly, Ethan walked me to the waiting area and I hobbled home on my tortured knees, feeling defeated and strangely dissociative. The stress test report wasn’t populated into my patient portal, so I had to submit a medical records request to read it, which I did last month.

How Were the Test Results Documented?

The fudged BP value was there, as expected. So was my jumping off the treadmill, but my threatening syncope and shaking/needing juice hadn’t been mentioned. The interpreting physician wrote that my BP response to exercise had been blunted.

More like collapsed.

The report also stated I had “excellent” FAC at 150% of predicted for my age, size and sex. This wasn’t a big surprise, because I was fit before my knees had swelled in March and when the pain made most cardio exercise too difficult, I’d switched my workouts to lifting weights and climbing overhanging walls. I also had undiagnosed dysautonomia, which shot up my HR with simple tasks, like showering and reaching up for cereal. It was like running a non-stop, daily marathon I hadn’t signed up for.

My maximum HR (187) was listed, but there was no mention that this value was abnormally high. Unfortunately, the events occurring during my cardiac stress test and the resulting report read like a case study in how not to run and document a cardiac stress test.

What Could the Healthcare Workers Have Done Differently?

• Received better training to prioritize patient safety.

  My BP had become unmeasurable during strenuous exercise – a flaming red flag that means the heart or nervous system isn’t working properly. This test should’ve been immediately terminated for obvious safety reasons, but Michelle and Ethan didn’t recognize the danger in having BP drop during exercise. Worse, at the time I was biking to get around in 95°F (35°C) weather.

• Been curious/thought outside the box.

  Nurses are taught to question everything. Why didn’t Michelle ask Ethan to confirm my BP was inaudible by immediately taking another measurement, instead of agreeing to document a fake “normal” response?

• Documented abnormal responses.

  I wasn’t asked about symptoms or syncope before the test. Would that have changed how my body’s response was perceived? Did Ethan report my shakiness to Michelle? There was a space to enter this info.

• Reported abnormal findings to Dr. T promptly.

  A quick call or portal message would’ve alerted her that something was very wrong with me.

• Avoided tunnel vision.

  I’d previously worn continuous EKG monitors twice, which showed abnormally high HRs. Some were associated with dysrhythmias, but most occurred with “normal” sinus rhythm. The stress test report’s false BP and glaring omissions, coupled with my “excellent” FAC results, tunneled the interpreting physician’s vision enough that he decided I was fine based on just one, contextless number.

What Could I Have Done Differently?

• Ask Michelle and Ethan document accurately.

  Why didn’t I confront them in real time? I couldn’t talk during/for several minutes after the test and I wasn’t thinking clearly, and Michelle had left. I also hadn’t found my voice yet to advocate for myself and sometimes, I still don’t speak out. Having to constantly self-advocate is exhausting and gets you labeled a “difficult” patient.

• Ask Ethan to document my post-test condition.

  I assumed he would do this without prompting.

• Message Dr. T to tell her what had happened.

  This also never crossed my mind. I went home and napped. More importantly, after ~3 months of fairly regular exposure to healthcare, I’d subliminally absorbed one of its core tenets: You can tell the truth, but it doesn’t matter because nobody believes you.

This stress test had been ordered to rule out structural cardiac issues, but had instead revealed severe dysautonomia. Because my abnormal results weren’t properly documented, it would be another month before anybody took my actual results seriously. The diagnostic odyssey continued …

Buy a Co-Pay

The image shows an Excel spreadsheet grid listing appointments and different tests for July, 2018.

Some Housekeeping

Substack promoters recommend publishing on the same day every week. Publishing on most Fridays was too much, so I’ll try publishing on the same days every month instead, on the 9^th, 19^th and 29^th.

There’s a glossary at the end of this post.

On With The Story

My first time writing in a while, and I’ve been avoidant all morning.

A waist-high pile of folders and papers teeters on the floor. The clutter juts into my peripheral vision, distracting me. The apartment feels unkempt. Gotta sweep and mop!

Although disorganization offends my attention deficit, I’m mostly avoidant because I’m writing about experiences that are difficult to relive, even 8 years later. In July of 2018, my suspected diagnosis was sarcoidosis. I’d have to wait 6–7 weeks until my pulmonology appointment, then longer still for the confirmatory lung biopsy.

Meanwhile, my health felt like an untethered wrecking ball careening down a winding road, smashing something to pieces at every turn (Post 6). I started taking public transportation in lieu of driving, for fear I would pass out. Although I’d taken time off from the lab, I was allowed to return because random falling and syncope are scary and I didn’t want to be alone. And the lab was on the same campus as most of my appointments, where I was begging for help into the void.

I asked my then PCP, Dr. O, if I could be admitted to the hospital.

“Nobody’s going to admit you. We already know you have sarcoidosis,” Dr. O rolled her eyes.

Technically, we didn’t know yet. I asked if I should get a MedicAlert button.
“Only old people need those!” she guffawed.

At 46, I wasn’t old enough to be old. I was terrified to be alone and desperate for answers. And somehow, this was funny …

I’ve heard said, “You think you’re being gaslit now? Just wait until you’re older!” Symptoms in older folks can be flippantly ascribed to “old age.” But I was “too young” to have anything serious. Medical gaslighting is an equal-opportunity affliction across all developmental milestones.

Having little to lose, I called a hospital affiliated with a respiratory clinic and explained my symptoms and test results (Post 7). Was that enough to be admitted and diagnosed?

Yes, they confirmed. Finally!

But I was so traumatized from my previous emergency room visit (Post 4) that I couldn’t muster the gumption to visit another. In retrospect, doing so would’ve expedited my diagnostic odyssey. At the time, making coherent decisions was difficult. As Yogi Berra once said, “When you come to a fork in the road, take it!”

It’s impossible to describe the existential confusion and loss of identity I experienced from the sudden barrage of healthcare visits my illness had triggered. That month, I had clinician appointments (PCP x 3, cardiology x 2, pulmonology and electrophysiology) and tests galore (treadmill, brain MRI, pulmonary function, EKG, biopsy, 2-week heart monitor, echocardiogram, 5 blood draws for labs and 3 urinalyses). My identity had morphed; I was now a professional patient.

Professional Patient

I would soon learn any pain or dysfunction could signify sarcoidosis or “worse,” necessitating timely evaluation. Yearly (or 6-monthly at times) surveillance testing became the norm. PCP visits begat referrals to specialists that begat so many more appointments, testing and referrals that their summation rivals the biblical lineage in the Book of 1 Chronicles.

In the ~8 years since my knees swelled, I’ve had > 650 healthcare-related visits, phone calls with clinicians, telehealth and testing (doesn’t include other insurance/disability/healthcare-related calls or portal messages). This includes 5 surgeries. Some months, I’ve had 20 appointments, with half of those being out-of-state. More than a dozen visits/tests were > 1000 miles (1609 km) away. I’ve seen 35 different specialties and endured hundreds of tests. And every 3 months, I see my indefatigable current PCP, Dr. L.

I’ve had 5 different infections in the past 12 months and in 2026, I’ve had 26 healthcare visits and tests so far. I even canceled a few appointments in January because of a (literally) gut-wrenching, 2-monthlong C. difficile infection. June–August, I’ll have 13 office visits, including my PCP (who will order my yearly EKG and labs), a new geneticist (genetics labs), PT, vision therapy, pain doctor and 2 infusions. I’m not trying to whine or garner pity. I’m just being honest. This is the life of a professional patient (a.k.a., frequent flyer).

I’ve been told, “Must be great to be on disability and not have to work.” Let me be clear: Nobody wants to live like this. I wish I were still running around in the mountains and grading boring nursing papers on weekends. Instead, chronic illness is my job, and it’s enough. Sometimes, it’s too much.

Sometimes Less is More

To reduce the stress of ongoing medical care, Dr. L and I came to an agreement ~4 years ago. I stopped seeing cardiology, pulmonology, neurology and electrophysiology at the respiratory clinic. I didn’t get a new neurologist when my empathetic (if forgetful) neurology PA moved away. Dr. L orders my yearly surveillance testing and refers me to specialists if symptoms worsen or new issues appear. Which happens constantly. Whenever I think I have a handle on my health, something new is always breaking.

To be clear – I’m not suggesting other people should skip vital healthcare! Appropriate and timely healthcare is important and can save your life. But at some point, constant appointments and travel have made me objectively sicker. Sometimes, less is more. I’m fortunate to have a PCP who listens and collaborates. I wish everyone could see her.

I’m Grateful

Sometimes friends seem reticent to talk to me about their health issues. Maybe they think it would upset me, since I deal with a lot. But that’s not how I see it. Illness is not a competition. Everybody has health issues in time that affect their ability to thrive, and I want to be present for that. I don’t just complain about being a professional patient to friends and family; I’ve built an entire Substack around it! The least I can do is afford others the same grace and an empathetic ear.

If you’ve gone through/are going through this, too, I feel for you.

I’m lucky, privileged and grateful it took months, instead of years, to be diagnosed with sarcoidosis. And I’m grateful for the healthcare workers who have helped me. And for friends and family who have lifted me up during my darkest hours. I’m especially grateful for you, dear readers. You’ve given me a purpose as you support and sustain me. If my story can help even one person avoid something awful, all this trouble will have been worth it.

Can you remember the time when you realized you were a professional patient? Or, were you always aware of it? Leave a comment below.

Glossary

• EKG – Electrocardiogram. A tracing of the electrical activity of the heart.

• PA – Physician’s Assistant.

• PCP – Primary Care Provider/Physician. Can be a Medical Doctor, Doctor of Osteopathy, Nurse Practitioner or Physician’s Assistant who coordinates care by ordering tests/referrals, and diagnosing/treating patients.

• PT – Physical therapy, which addresses patients’ mobility problems or pain.

• Syncope – Fainting.

The photo shows Scrabble tiles on a white background. The tiles spell M-E-D-I horizontally, T and C vertically, and then A-T-E horizontally. The puzzle can be read as “MEDICATE,” “MEDITATE” or “MEDICTATE.”

(A glossary is included at the end of this post.)

(Stars denote topics for future posts.)

Just kidding! There’s nothing sexy about medical record (MR) amendments (though Scrabble sure is sexy). Instead, this topic feels like it belongs in The Week magazine’s section entitled, “Boring, But Important.”

This Post explores topics related to Post 11 about medical record (MR) Amendment Requests (ARs) and Disagreements in the United States.

Do Medical Record Errors Really Happen That Often?*

A 2020 study at the Department of Veterans Affairs Health System aimed to answer this question. Unidentified standardized patients presented to physicians for outpatient clinic visits. The visits were audio recorded, and the visit notes were compared to the recordings. The results showed 636 documentation errors in 105 notes, and 90% of notes had at least 1 error.

A few excerpts:

• “Documents asthma patient as a smoker,” but “Patient never smoked.”

• “… patient describes feeling ‘woozy’ with ‘pounding chest’ during syncopal event,” but “symptoms not documented.”

• “Patient ‘stretching’ his Pulmicort medication since loss of job,” but “not documented despite poorly controlled asthma.”

These sobering findings are supported in a study by Bell et al., who found 21% of 22,889 patients found errors in their visit notes. Serious errors were more frequently found by women, more educated patients, sicker or older patients and those who’d read more than 1 of their notes. A few excerpts (the Table of errors is worth a read):

• “The notes reflect exams not done … As a medical profession (sic) I find this alarming.”

• “… for an infusion … the dosage was incorrect by 10-fold.”

• “It was stated … I had lung cancer. I do not and never have had lung cancer. I did have a spontaneous pneumothorax in 1976 …”

• “Reference to my left testicle (I’m female).”

Some argue the presence of errors like this is evidence for the use of AI recording tools. Perhaps, but only if the clinician rereads the note for accuracy. If not, there may be AI-generated note errors, as described in Post 1 and Post 10.

MR errors occur for many reasons,* the biggest likely being lack of time. Clinicians are pressed to see as many patients as possible, which reduces the quality of care. Reviewing and responding to ARs takes even more precious clinician time, but they’re worth it because they can reduce life-threatening errors in care. They also alert clinicians about the kinds of errors they’re making.

I Sent My AR. Now What?

Post 11 describes the submission process in detail. I keep copies of everything I send. If I haven’t heard anything back, I follow up after 60–90 days, usually with the Health Information Management (HIM) department.

I’ve experienced a variety of outcomes after sending an AR. I once sent the AR through certified mail, but it never got delivered. I now keep tracking numbers and confirm receipt of my AR, either through phone calls or email.

Clinicians’ responses to ARs are varied. Some clinicians are confused or unaware that requesting MR amendments are a patient’s right, or they don’t understand the process. This reaction is fairly common. In this instance, I try to coordinate with the facility’s Patient Advocate and HIM.

Some clinicians approach ARs thoughtfully. This doesn’t necessarily mean they make all (or even any) requested amendments; in their clinical judgment or memory of the interaction (or lack thereof), they may disagree. That’s okay; the clinician has the right to deny an AR. I try not to take it personally. If errors remain in the note after a clinician responds to the AR, I submit a Disagreement.

Some clinicians outright deny every AR. This may simply be due to a lack of time or a misunderstanding of the process. Some clinicians seem annoyed at my ARs. Sadly, some may even fire you as a patient for daring to suggest they made mistakes (happened once, after a soul-crushing dermatology visit*)!

What Happens After the MR Has Been Amended?

Technically, ARs & Disagreements become a part of the MR. But they don’t necessarily become part of the specific note that was originally identified as erroneous. Patients must be vigilant to ensure their amendments and Disagreements are seen by clinicians so errors don’t recur.

Just because a note has been amended doesn’t mean the narrative is universally fixed. Often, the charting system (e.g., Epic) retains the error, which can be copy-pasted or auto-populated into the next visit note. This could result in a creeping narrative (See Post 2).

Amended notes may not be uploaded to the electronic MR. The university health system (where I get 85% of my care) didn’t append my Disagreements to the related notes in Epic until I complained to the Patient Advocate. I realized this when I peered over my PCP’s shoulders at an appointment and asked her to show me the Disagreements associated with specific notes.

Even when they’re sent to other facilities, amended notes and Disagreements might still not get seen. New Jersey Health (NJH) sent my voluminous MR to Chronic Cove (CC) Health System (facility names have been invented). In the process, they sent hundreds of pages of original notes as one batch and then sent all the amended notes/Disagreements as a second batch (instead of appending amended notes/Disagreements to each corresponding original note). Every clinician at CC denied seeing amended notes or Disagreements from NJH. This oversight negatively affected my care.*

How Do I Avoid Confusion About Amended Notes?

I bring copies of amended notes with changes highlighted to clinicians who may have seen the original, erroneous note, so we can review the amendments together. I also send amended notes to my insurance if I believe the errors resulted in denials. And I send copies of amended notes to disability lawyers/Social Security. Patient work* in chronic illness is never-ending!

Finally, if a facility/clinician ignores your AR or doesn’t follow the rule of law, you can report them to the Department of Health and Human Services. This last-ditch option isn’t punitive unless flagrant violation of patient rights is a pattern. Instead, complaints direct DHHS to educate clinicians about the MR amendment process.

Healthcare is complex enough without MR errors, which can be harmful or even fatal.* Once errors are identified, timely MR correction improves patients’ health by shortening diagnostic delays and enabling proper treatment. I hope my experiences have been informative.

Please comment below about your experience with MR errors.

Buy Me a Med Co-Pay

The name of the website is “Buy Me a Coffee,” but any donations I receive go toward healthcare. If you already support me, thank you so much!

Glossary

• PCP – Primary Care Provider. A licensed Medical Doctor (MD), Doctor of Osteopathy (DO), Nurse Practitioner (NP) or Physician’s Assistant (PA) who coordinates care by ordering tests/referrals, and diagnosing/treating patients.

• Standardized patient – Person trained to act out patient scenarios with medical students and professionals and assess their clinical skills. Also known as a simulated patient, patient instructor or sample patient.

The image shows a brunette woman in a white lab coat sitting at a white desk. She is clicking on the mouse while reading information on a computer screen.

Amendment requests (ARs) to medical records (MRs) aren’t always welcomed by healthcare workers. And they’re an added burden for patients. But submitting an AR is one of your HIPAA rights that might save your life – or at least make it more tolerable. I’m not an expert on MR ARs, but I’ve submitted many. I’m sharing what works for me.

Here’s my approach:

Call or message the facility and ask how they handle ARs.

At larger institutions, Health Information Management departments (where you get Release of Information forms) handle ARs. The process usually involves filling out a form (example) and uploading it into the patient portal, and/or emailing or snail mailing it. Sometimes, I must do more than one of those (e.g., fill out a form on the portal and snail mail it). Sometimes, I must provide a copy of my driver’s license.

Print the note and highlight the erroneous info.

Some facilities mandate this step. Regardless, I do it for all ARs to organize my thoughts. I send the highlighted note with the AR.

Submit a document listing suggested corrections.

AR forms usually ask, “What do you believe needs to be changed?” But they rarely provide enough space to answer this question, especially if there are many errors (19 is the most I’ve found in one note!). I write “see attached” in that space. Next, I compose a Word document with my name, birthdate, MR number and the facility/clinician’s name and the date of service in the header.

Using an outline format, I number each error as it appears in the note. I write the error in bold and quotations, ensuring it matches the phrases highlighted on the note. Below this, I describe any supporting evidence. I then write, “please amend to … ” and write what I think the MR should say in quotations. There’s an example at the end of this post.

The visit note is a legal document and shouldn’t be changed once it’s been signed. Instead, clinicians should insert or append note corrections on the note. But many clinicians are unaware of this. ARs become a part of the MR (as does a Disagreement). If the clinician thinks the AR is too lengthy, the facility has the option of providing a synopsis instead. That’s never happened to me.

Submit proof.

I submit supporting evidence (e.g., photos, other clinicians’ notes, and lab or radiology reports), numbered as Figures. I highlight my name and date of birth, the visit date and the relevant info on each Figure for easier skimming.

Explain why I want my MR amended.

Some forms ask why I’m submitting an AR. I write “continuity of care” or “I have a complex medical history and accurate medical records are critical to keep me safe.”

Be timely.

It can be traumatic to relive a medical appointment gone awry by rereading the visit note. Despite this, it’s better to submit ARs soon after the appointment, so the details will be fresher for you and the clinician. I discovered writing ARs and even Disagreements is therapeutic because it allows me to be productive and process negative emotions.

Be accurate.

It’s embarrassing to make errors while pointing out others’ errors! I read through my AR materials carefully a few days after writing them before submission.

Be succinct.

I learned the hard way that appending dozens of pages of results from a 2-week heart monitor with an AR is a bad idea. Instead, I append summaries. If necessary, I can always bring supporting documents to my next appointment.

Be respectful.

MR errors have caused me a lot of harm and heartbreak. At times, my frustration and anger about this have boiled over in unfortunate and unprofessional ways. Especially if I was gaslit or the clinician was rude (sadly, many Future Post Alerts). Or when the errors seemed like obvious upcoding grabs to bill for more money (FPA!).

I’m trying to be kinder to myself and others. Live and learn.

When I’m really struggling to be polite, I talk about my medical trauma with my therapist and write a separate document about what angered me and why. I don’t send that to anybody, unless it covers something reportable to the facility’s patient advocate. Maybe it can even be teased into another Substack post?

I also like to write, “Thank you for helping keep my medical record accurate” at the end of every AR.

Be patient.

A clinician has 60 days to respond to the AR. This deadline can be extended by 30 days if a letter is sent stating the clinician needs more time.

The response to your Amendment Request should be a letter stating the clinician has:

Approved,

Partially approved, or

Denied your request.

Wait for a response before agreeing to have the amended note forwarded anywhere.

The AR form usually asks if the erroneous note was sent to anybody else and whether you want the amended note forwarded to those clinicians. I don’t fill this part out because I don’t yet know the outcome. Sometimes the clinician denies the AR. Sometimes the AR is partially approved and I have more to add. Sometimes, the AR is approved, but the “corrections” are still wrong. I submit a Disagreement for each of these scenarios. I can always request that the note and Disagreement be sent to whomever I want later.

I don’t ask that corrected notes get automatically forwarded before I review them because this could cause confusion, as there will be multiple copies of the note – the original, the note with the amendment request and the note with the Disagreement. I think most clinicians would read the first note and believe the next 2 notes are duplicates and ignore them.

Write a Disagreement if needed.

I don’t dismay if the AR is denied. The clinician has every right to do so! Sometimes errors were made on purpose to ensure insurance coverage (see Post 9).

One clinician always listed amendments in 5-point font at the end of the note, which could be easily missed. That note got a Disagreement, too.

I write Disagreements in bold, 14-point, highlighted font. Seems obnoxious, but that makes my points much harder to miss.

Final thoughts until next week …

Humans aren’t perfect, but healthcare is a perfect storm of not enough people, time, resources, education, support and money. Try to give healthcare workers the grace you’d want if you made a mistake.

How do you deal with medical record amendment requests? Leave a comment below.

Please consider supporting this free newsletter with a monthly or yearly donation. Thank you!

Example AR Format (From Post 1)

1. “… heterozygosity for PIGL”

   1. See Genetics Report (Figure 1)

   2. Please amend to: “the gene cited is PYGL, not PIGL.”

2. “Her lactate levels remain low.”

   1. See Lab Reports (Figures 2 and 3)

   2. Please amend to: “Her leptin levels, not lactate levels, remain low.”

3. “Received influenza vaccine this year.”

   1. See Vaccine Verification Forms (Figures 4 and 5)

   2. Please amend to: “Received influenza, Covid-19 and pneumonia vaccine this year.”

The photo shows a brunette woman in blue scrubs and silver rings holding up a pink stethoscope that’s been twisted into the shape of a heart.

One morning, I reached above the stove for a box of Cheerios. My mind suddenly swirled, and the bright yellow box morphed into shiny, prickly stars that quickly faded as my vision pinholed into blackness ...

I came to on the cool linoleum floor, sitting against the kitchen island, shaking violently. I must’ve sat down when I felt dizzy, but I didn’t remember doing so.

Soon, these harrowing head rushes became the norm, especially when I changed positions from sitting or lying down to standing. And regardless of the dizziness, my brain felt hazy, like a mountain range blanketed by thick wildfire smoke.

When I passed out at work one afternoon, my boss recommended I take some time off from the lab. I’d kept him updated about my enervating diagnostic odyssey.

“Why don’t you get a PCP?” he asked.

When I’d moved at the beginning of the year to a new city, I hadn’t bothered to get a PCP. I was healthy, so the thought didn’t cross my mind. Plus, as a graduate student, I was living off of loans. Every appointment cost money I didn’t have. Nobody wants to pay for a copay. Especially if you’re broke.

Despite my nursing background and at a time when I clearly needed a PCP, my thinking had been too fuzzy to comprehend this. Instead, I’d bounced between specialists willy-nilly. At this point, I’d been seen by Orthopedics, Urgent Care, Emergency Medicine and Physical Therapy – a total of 8 providers.

My insurance allowed me to see specialists without a referral. But I was coordinating my own care at a time when my brain and body felt as if I were moving through molasses. I lacked the insight and ability to take this on.

No wonder I felt so alone.

I selected a PCP from the internal medicine website for the university health system, where I was also a student and had been seen for the knee swelling described in Post 2.

I figured a younger physician, and specifically, a woman, would listen better and be more supportive. I scheduled an appointment for Dr. O, who looked like she was in her 30’s.

July 5, Initial PCP Visit

I’d been robust and healthy my entire life in the Before Times. When I worked in the ICU, I felt thankful after every shift that I was going home while the patients had to stay. And I never thought I would be one of those people with a mile-long medical history and matching medication list. At least, not until I was much older.

I also lacked access to my previous medical records. The pediatricians who’d cared for me had long since retired, at a time when medical records were exclusively written on paper. My parents hadn’t collected my medical records as a child; I had just a few immunization cards. I remembered neither the date of my appendectomy nor the doctor who’d done it 20 years prior.

At this point, my documented medical history comprised migraines, depression, my recent knee arthritis, an appendectomy, hysterectomy and removal of my right ovary and both fallopian tubes, and endometriosis. The eCheck-In didn’t solicit my medical history.

Thankfully, Dr. O took a thorough health history. Her specific questions (mostly) jogged my addled memory.

“Any childhood illnesses?”

“Chicken pox. And mononucleosis at age 19.”

“Any other illnesses as an adult?”

Yep – 2 short episodes of vertigo. Shingles, in the stressful month when I was studying for my nurse practitioner boards. That was pure hell.

I wasn’t asked about, and didn’t remember, that I’d had at least 3 childhood concussions (likely many more). I also forgot to report (mostly because these seemed minor/had been diagnosed years prior):

• Dupuytren’s contractures

• Traumatic kissing spine/Baastrup syndrome, from a botched backflip in high school

• Patellofemoral syndrome from biking a decade prior

• Hip bursitis from biking 5 years prior.

(Spoiler Alert: These musculoskeletal issues, plus some baseline clumsiness and my disturbing flexibility as a gymnast, point to a connective tissue disorder. But it would be a while before I put those pieces together … ).

“Any surgeries, including oral surgery?”

I hadn’t thought about my oral surgeries in decades. I reported 4: impacted tooth removal, 2 wisdom tooth surgeries and drainage of an infected socket in my jaw.

The facility where I’d had my hysterectomy was affiliated with the university health system. Thus, Dr. O could access the related pathology report, but not the surgical note.

“The path report says you don’t have endometriosis,” Dr. O challenged.

At the time, I didn’t understand that she couldn’t see the surgical note. I was mostly focused on staying upright and following the conversation through the ever-present roar in my ears.

“OK, but I have endometriosis …” I stammered.

“No, you don’t!” Dr. O snapped.

The world was swimming around me, and at this point, I was too exhausted to respond. I had no evidence on hand, and I lacked the will to protest. I sat, light-headed and sluggish, as Dr. O rewrote my medical history.

My heart sank. Back to feeling dismissed and a little desperate.

Although Dr. O wasn’t as good a listener as I’d hoped, seeing a PCP was an important step toward finding answers in my diagnostic odyssey. In the ensuing years, I would learn a lot about navigating healthcare and choosing the right clinicians to make up my A-Team.

I’m happy to report: I have an awesome PCP now. I see her every 3 months because of everything I have going on. I wonder how my diagnostic odyssey would’ve been different had I started with a PCP (and especially, her) instead of going to Orthopedics.

If You’re Healthy, Do You Really Need a PCP?

Yes, because PCPs do the following for you:

• Provide a healthcare home base

• Coordinate your care

• Diagnose and treat common illnesses

• Guide you through age-appropriate wellness visits, preventative care and screening

• Fill out forms (for example, for disability, FMLA or absentee notes)

• Fight insurance denials on your behalf

• Refer you to specialists (and they may know some good ones, too!)

• Provide holistic care, because specialists rarely look at the whole picture; instead, they tend to focus on their specific field

• If you get sick, you’ll be able to see a clinician sooner if you’ve already established care with a PCP.

What Could I Have Done Differently?

• Visited a PCP earlier, before it became urgent

• Collected notes from all my healthcare visits

• Called my parents to fill the holes in my memory of my childhood illnesses

• Written down my comprehensive health history before my appointment.

What Could Clinicians Have Done Differently?

• Suggested I get a PCP. None of the clinicians I’d seen in a 3-month period did so. This was a missed opportunity to vastly improve healthcare delivery

• Been more curious. Instead of shutting down the conversation, Dr. O could’ve asked me to explain why I said I had endometriosis.

What have your experiences with PCPs been like? Please comment below!

Diagnostic Odyssey is a free newsletter. If you’re able, please consider signing up for a monthly or yearly donation. $5 (monthly) is the co-pay for my fludrocortisone. $50 (yearly) covers one box of lidocaine patches. Or, just buy me a (decaf) coffee (honestly, the $ will be used for healthcare costs, not coffee … but whatever!). Thank you!

Subscribe now

Buy Me a Coffee

Glossary

• Bursitis – Inflammation of the fluid-filled sac surrounding joints.

• Dupuytren’s disease – Thickening of the connective tissue (fascia) on the palms that can cause fingers to curl (especially the pinky and ring fingers). It’s unusual for a woman in her 30’s to be diagnosed with it, like I was. And I wasn’t a heavy alcohol drinker or smoker.

• Endometriosis – Inappropriate growth of uterine tissue outside of the uterus.

• FMLA – Family and Medical Leave Act.

• Kissing spine/Baastrup syndrome – Happens when the spinous processes of vertebrae touch each other.

• PCP – Primary care provider or primary care physician who coordinates care. Can be a nurse practitioner or physician’s assistant.

• Patellofemoral syndrome/Runner’s knee – Pain near the patella/front of the knee, often from overuse injuries.

The image shows black typed text on white paper with a bolded phrase stating “Intractable chronic migraine with aura and without status migrainosus.” The bolded phrase is circled in orange ink and the phrase “Error — not AI-generated” is written in cursive beneath it.

In Post 1 for Diagnostic Odyssey, I explored how Dr. F’s artificial intelligence (AI) tool appeared to have transcribed errors in my visit note. A few months after this visit, I submitted a medical record (MR) amendment request to have the errors fixed.

Dr. F recently said she’d intentionally made one of the errors I’d flagged as possibly AI-driven.

This post describes our interaction and highlights a recent article on how AI may amplify latent MR errors.

About Dr. F: I adore her. Like my PCP, she’s dedicated to her patients. She’s even called me on weekends. It can take years to find the right clinicians, but once you do, you’d fly to the moon to keep them.

That’s why I wasn’t surprised when Dr. F called me a few weeks ago, asking if we could Zoom.

“I want to get these corrections right,” she said in her soft, kind voice.

I’ve submitted dozens of amendment requests (far too many!), but nobody has ever called and offered to do this. Like I said, Dr. F is special. Minutes later, a pixilated image of her smiled from my dusty laptop as we discussed the amendment request.

Our Conversation

We plodded through the corrections and without hesitation, Dr. F agreed to amend all 11 errors. I’ll write more about the process and pitfalls of amending MRs at another time (Future Post Alert!).

I told her I was concerned about what I’d thought had been an AI hallucinated diagnosis:

“Intractable chronic migraine with aura and without status migrainosus.”

My migraines are treatable with meds and don’t last for days on end, so they aren’t intractable. And they don’t start with an aura (a neurologic phenomenon occurring before a migraine that can include confusion, seeing spots or flashing zigzags, or experiencing other sensory disturbances).

“No, that wasn’t AI. I needed to pick a diagnosis for billing purposes, so I picked that one, based off your med list,” Dr. F explained.

The wrong migraine diagnosis wasn’t a fluke. It was an intentional error.

Dr. F had probably picked it, in part, so my migraine meds would be covered by insurance. Unfortunately, clinicians in the US must often document intentional errors so insurance will cover their patients’ needs or to avoid liability (Future Post Alert!).

But the diagnosis was also selected for billing purposes. This frustrates me, since I’ve been harmed before by spurious diagnostic codes (Future Post Alert!).

Before ending our Zoom, I said, “Dr. F, I’m worried about AI because it made mistakes in my note and is likely making mistakes in other patients’ notes.”

Her response surprised me.

She shrugged, “There are always mistakes in notes.”

Not wanting to monopolize Dr. F’s time, I thanked her profusely as we logged off. Now I needed to correct my Substack Post. And not just correct, but correct and reflect.

Reflections on My Error

One of my assumptions about the AI output in my MR was wrong. Though I included a caveat of uncertainty in Post 1, I wrote it without verifying all the info. I didn’t think I’d get a chance to discuss these errors until I saw Dr. F again in October. She’s an insanely busy researcher, teacher and clinician.

When I wrote Post 1, I felt the crux – the fact that clinical AI tools can cause errors – was more important and timelier than slowing down and fact-checking. I’m ashamed to say this, but working this way reminds me of the tech bro motto, “Move fast and break things.”

But now I see how impatience can affect the veracity of my writing, so I’ll rethink how I challenge my own assumptions and biases. I’m learning, like all of us. I hope you can all forgive me for being a little too kneejerk and alarmist.

Speaking of alarmist, I thought my warning to Dr. F about my concerns with AI-based tools would resonate with her. Instead, it fell flat.

Reflections on Dr. F’s Response About AI-generated Errors

I thought Dr. F’s response to my concerns about AI was a bit … cavalier.

I’m thankful she realizes MRs contain mistakes, because many healthcare workers seem to view anything in the MR as sacrosanct and 100% correct by default. I call this belief Gospel bias (Future Post Alert!).

I was disappointed Dr. F didn’t seem willing to do anything to avoid, recognize or fix the specific MR errors she knows are happening.

I wanted to insist, “You need to read every note that AI tool transcribes before you sign it!”

But I know she doesn’t have time to reread notes. That’s why she’s using the AI tool in the first place.

I don’t know how to solve this problem. Admittedly, I’ve felt very alone in my quest to highlight MR errors. Just hearing she’s not an adherent to the Gospel bias was a huge relief. But it’s not enough.

So, what can be done?

Final Thoughts on AI and MR Errors

Thankfully, stakeholders are finally starting to realize some of the inherent dangers in using AI in an already flawed healthcare documentation system.

In a 2025 article, Morreim states, “Unfortunately, evidence suggests that the [electronic MRs] serving as the main dataset for therapeutics-focused AI are considerably more error-laden than we commonly realize” [italics added by me].

Worse, Morreim states MR errors may enable AI systems to produce inaccurate algorithms and note errors that are “difficult to detect and even more difficult to correct.”

Dr. F’s willingness to correct errors is admirable, but her statement about errors always occurring belies a fundamental truth about MRs. They’re not always reliable and in some instances, dangerously so. And superimposing AI models, which are prone to hallucination, on top of our current system compounds the potential risks for patients.

The impetus for writing Diagnostic Odyssey is to educate others about the serious impact of MR errors and specifically, how these errors delay diagnosis of rare diseases.

This important mission isn’t possible without all of you.

Please help get the word out by sharing these newsletters with anybody you know who is a patient, family member/caregiver, healthcare worker and especially, a student in healthcare. Thank you!

The image shows a logo of a blue galley ship that has eyes and a blue caduceus as its mast.

Sarcoidosis (pronunciation here) is a non-communicable, immune-mediated disease, which differs from autoimmune diseases in that the body isn’t attacking itself. Instead, it’s mounting a massive immune response to something foreign that causes cells in the immune system to go haywire. However, if you have sarcoidosis, your chance of getting an autoimmune disease increase.

Sarcoidosis causes T-lymphocyte cells known as T-helper (CD4) cells to become overactive. These are the same cells that are suppressed in HIV/AIDS. The CD4 cells send out chemical messengers known as cytokines to summon other types of immune cells to the battle. All the cells clump together, forming a granuloma. Granulomas can calcify and turn into tough scar tissue.

This is the most adorable cartoon of a granuloma I’ve ever seen. It looks like an alien dance circle! The picture shows immunologic cells forming a granuloma, with an epithelioid cell in the center surrounded by a layer of neutrophils, and then a layer of giant cells, dendritic cells, T-lymphocytes and B-lymphocytes. The outermost layer is made of fibrocytes. Photo credit.

Sarcoidosis is considered a multisystem granulomatous disease because granulomas can wreak havoc anywhere in the body. > 90% of people with sarcoidosis have granulomas in their lungs. Granulomas can be cutaneous (on the skin) or they may be found in any organ.

Granulomas can appear in the heart, preferentially in its conduction system. Seemingly healthy people may experience sudden cardiac death, only to have cardiac sarcoidosis discovered on autopsy. Granulomas in the nervous system (neurosarcoidosis) can be particularly devastating.

Granulomas can produce active vitamin D, which increases blood calcium levels. The resulting hypercalcemia can be fatal. Patients with sarcoidosis may have low vitamin D levels, which makes vitamin supplementation a challenge.

The incidence of sarcoidosis clusters around certain occupations. Miners, construction workers and the people who cleaned up the World Trade Center debris pile after 9-11 have a higher risk of sarcoidosis than the general population. But jobs involving contact with lot of other people impart a higher risk of sarcoidosis, too, including hairdressers, nurses and teachers.

Regardless of the means of exposure, anybody who has sarcoidosis can be seriously affected by it.

What Causes Sarcoidosis?

Nobody quite knows. Because pulmonary involvement is common in sarcoidosis, it’s believed the trigger is inhaled. Viruses, fungi and bacteria are possible triggers. So are environmental pollutants, including silica, dust, mold, metalworking fumes and even cat litter.

Certain genetic mutations may increase the risk of having sarcoidosis. Recent increases in wildfire smoke across the U.S. could lead to a higher prevalence of sarcoidosis. Certain demographics also seem to be more susceptible.

Who Gets Sarcoidosis?

When chest X-rays were used as a mass screening tool during WWII, up to 15% of apparently healthy prospective GIs had mediastinal lymphadenopathy. So, at least a small percentage of people may have pulmonary lymphadenopathy and not even know it, until perhaps they’re exposed to a trigger and then have sarcoidosis symptoms.

Oddly, smoking cigarettes hasn’t been strongly linked to sarcoidosis, though research on this is of low quality. Some studies have shown smokers are less likely to get sarcoidosis than non-smokers, unless the smokers have a certain gene allele (HLA-DRB1*03).

In the US, more women have sarcoidosis than men. African-Americans in the Southeast have a higher rate of sarcoidosis (it’s important to study disease demographics because patients from marginalized communities may take longer to get diagnosed and may experience a more severe disease course). There’s also a high sarcoidosis prevalence in people of Danish and Irish descent.

People of Scandinavian descent have a higher risk of Löfgren’s syndrome, which typically presents with lymphadenopathy, a migratory arthritis that wanders from joint to joint and erythema nodosum (painful red bumps on the shins – see pic here).

Sarcoidosis is rare in Japan, and cardiac sarcoidosis is also relatively uncommon, but 25–70% of Japanese patients with sarcoidosis have it in their heart – and it’s often undiscovered until autopsy.

How Does Sarcoidosis Affect Patients?

Sarcoidosis has a varied presentation and also produces varied effects on different patients. Up to 60% of patients can have remission within 2 – 5 years. Other patients have a chronic and progressive disease course. Sarcoidosis is fatal in ~7% of cases.

Sarcoidosis causes harm and chaos beyond the pain and dysfunction of granulomas. Inflammation is common, and it leaves patients with crippling fatigue, arthritis, joint swelling, muscle pain and weakness, weight loss, headaches, cough, shortness of breath and brain fog.

I’ve felt all of these symptoms (and more) and at my worst, I was sleeping 20 hours a day. And although both my knees swelled at the start of my illness (Post 2), it’s more common for both ankles to swell in sarcoidosis.

Systemic inflammation may also cause small fiber neuropathy, which leads to burning pain, tingling and numbness especially in the feet, temperature dysregulation and dysautonomia, leading to a high heart rate and low blood pressure.

Perhaps most frustratingly, individuals with sarcoidosis suffer from a “… wide range of … general symptoms … but … can often appear completely healthy.” This can lead to gaslighting and delayed diagnosis, which makes us sicker. And having a life-altering disease can lead to psychological issues, like depression and anxiety. Being gaslit doesn’t help!

How is Sarcoidosis Diagnosed?

Sarcoidosis can take months or years to diagnose, partly because of the non-specific nature of its symptoms. Fatigue and headaches can occur in many diseases (Post 6). I was privileged to have access to a university health system and a nursing background when I got sick. Still, I needed to fight hard to been heard and diagnosed.

Sarcoidosis is typically diagnosed by biopsy. In cutaneous sarcoidosis, skin granulomas are usually easy to access. If the granulomas are the lungs, patients must be sedated for a bronchoscopy to complete the biopsy.

How is It Treated?

The Foundation for Sarcoidosis Research (FSR) has a primer on sarcoidosis-specific treatments here. Taking, inhaling or applying steroids is first-line. Methotrexate and hydroxychloroquine are other treatments. The goal is to suppress the massive immune response.

These medications have serious side effects. Taking steroids can lead to weight gain, hypertension, insomnia, diabetes and mood swings. Methotrexate may decrease blood cell production or damage the liver. Monoclonal antibody medications can cause cancer, autoimmune diseases and neuropathy. With sarcoidosis, you must pick your poison!

Other medications may be prescribed to treat sarcoidosis symptoms. Fludrocortisone (Florinef) causes my kidneys to retain sodium and water, thus driving my low blood pressure to normal levels. ADHD medications like modafinil (Provigil) are used off-label for my fatigue. I previously received intravenous immunoglobulin (IVIg) infusions to slow neuropathy progression and treat pain. I’ll spare you the rest of my (way too long) med list for now!

Resources

Patients with sarcoidosis must become experts of their own disease. Below are some resources for patients and caregivers:

• Bernie Mac Foundation

• Foundation for Sarcoidosis Research

• Sarcoidosis News

• Sarcoidosis Research Institute

• Say Sarcoidosis

• World Association for Sarcoidosis and Other Granulomatous Diseases (WASOG)

• WASOG Recognized Sarcoidosis Clinics and Centers of Excellence

  If you enjoy Diagnostic Odyssey, please consider supporting me by buying me a (decaf) coffee or donating a monthly or yearly subscription. Thank you for being here!

Buy Me a Coffee

DISCLAIMER: This post is for entertainment purposes only and isn’t intended to diagnose or treat an illness. Always ask your healthcare provider if you have any questions.

Glossary

• ADHD – Attention deficit/hyperactivity disorder.

• AIDS – Acquired immune deficiency syndrome.

• Communicable – Infectious; able to be passed from person to person.

• GI – Military term for an enlisted soldier, mostly used during WWII. Can stand for General Issue, Government Issue, or Galvanized Iron.

• HIV – Human immunodeficiency virus.

• Lymphadenopathy – Enlarged lymph nodes.

• Mediastinal – Central compartment of the chest.

• Sarcoidosis – A systemic, granulomatous disease involving massive immune system activation and dysfunction.

Photo credit

The image shows a black-and-white chest X-ray showing the heart, esophagus, diaphragm, ribs and clavicles in white and the lungs mostly in black. There are arrows pointing to the middle section of the lungs. The arrows show thick, white areas caused by disease.

Getting a new diagnosis – or even a probable diagnosis – can be a loaded emotional adventure.

For me, there was the surprise, of finally find out what I had.

And fear – How will this illness change my life?

Guilt – Did eating too many peanut butter cups and Pringles, or partying too much in my 20’s, cause this?

Stigma – I was always so healthy and strong. What will everybody think of me now?

Relief and validation – See?! I really am sick.

Triumph – I knew I had [disease X]!

Anger – Why is this happening now??

Sadness – Will I still be able to hike and climb?

And if it’s a rare diseases, it can impart a little bit of wtf is that?!

Receiving a diagnosis is like leaping across a giant chasm, but then feeling the ground crumble like a graham cracker beneath your feet as you land on the edge. It can alleviate previous anxiety while fostering new-found concerns. It’s at once grounding and disorienting. It defines your existence, even as you lose yourself in it. Knowledge is gained as vitality is lost.

It’s easy to think being diagnosed is the end of the diagnostic odyssey. Like reading the last page in a yellowed, dog-eared paperback (remember those?). But for most people, diagnosis – or even the specter of one – is merely the end of one journey and the beginning of another.

I got my first rare diagnosis more quickly than many other patients do. I was lucky; working at a university gave me access to privileges other patients don’t have. When my symptoms spiraled, I called a cardiologist affiliated with our lab (Dr. M) and begged him for help. He ordered a cardiac MRI for the next day, June 25th. I’d already scheduled an appointment with him for the 27^th.

I lived and worked on campus, so I could walk to my healthcare appointments. Being a Ph.D. student made my schedule flexible. And I had pretty decent insurance (or so I thought). More privilege.

Cardiac MRI

“Stay as still as you can,” the resident directed over the intercom, before the MRI tube sucked me in. “OK, I’ll try, but my body shakes when I have PVCs,” I warned.

He seemed flummoxed and replied, “Don’t do that; you have to stay still!”

As if I somehow had control over how my body was acting.

The resident’s bizarre attitude made the experience unpleasant. I was getting this test because I was having PVCs and a racing heart. Now I was being told the thing they were testing to figure out would mess up the test itself, and that it would somehow be my fault.

I’m not claustrophobic and MRIs don’t make me anxious, but I felt my heart race as the MRI groaned and pinged. As soon as it stopped, my heart rate slowed (As an aside, my frequent anecdotal experiences in MRIs have led me to believe the anxiety people feel inside of MRI machines is due at least in part to physiologic changes directly induced by the machine).

The results were posted on the patient portal that evening:

“Hilar and mediastinal lymphadenopathy of uncertain etiology … Findings could be compatible with sarcoidosis although other supportive features are not seen” (The linked imaging in this paragraph is pretty … well, rad).

Translation: The lymph nodes in the middle of the chest are enlarged. Not sure what’s causing it. Could be sarcoidosis, but she doesn’t have other things we’d expect to see in somebody with sarcoidosis.

I blinked. Emotions swirled like cream in a teacup. Sarcoidosis? Hadn’t been on my radar. Years prior, I had considered sarcoidosis as a topic to teach my nursing students. There was only one paragraph on sarcoidosis in the textbook and I’d never treated anybody with it, so I flipped past it with a shrug. The students already have enough to read, I thought. Looking back, I feel ashamed about that.

Thankfully, Dr. M saw the results and immediately ordered the recommended chest CT to be done, and I got it 2 days later. The CT scan findings were pretty much the same: “compatible with sarcoidosis.”

Cardiology Visit

Dr. M and I had met weeks before in the lab where I worked. We discussed how he’d participated in one of the lab Director’s studies as a medical student and that it had fostered his passion for research. He was young, bright and easy to talk to.

When I met with Dr. M this time, he wasn’t so cheerful. He lowered his eyes and said, “It’s likely sarcoidosis.”

I didn’t know what to say. Neither did he. I nodded.

The CT scan results were another clue, but didn’t confirm my diagnosis. I needed a lung biopsy to rule out other possible diagnoses. The differential diagnoses for my scan results were pretty scary and included:

• Autoimmune diseases, like lupus, Sjögren’s syndrome or rheumatoid arthritis

• Cancer

• COPD

• Cystic fibrosis

• Cysts

• Fungal infections, like coccidioidomycosis or histoplasmosis

• Leukemia

• Lymphoma

• Sarcoidosis

• Tuberculosis

More stuff to stress about!

Dr. M referred me to a pulmonologist. But the next available appointment was in 2 months. My world was falling apart and I had to wait. At least it wasn’t a 6-month wait. Or a year. Or longer. I realize in retrospect that, despite my frustration, this accelerated schedule is an example of yet more privilege I experienced in my diagnostic odyssey.

How Did Getting Closer to a Diagnosis Feel?

Everything about this experience was humbling. But after months of patronizing gaslighting (Post 4), just knowing there was a possible diagnosis gave me hope.

I was thankful to be in a position to access quality healthcare quickly. And grateful that my nursing and research background would help me navigate these uncharted waters. But I dreaded having an illness I knew little about and I was overwhelmed at the thought of what would come next.

Leave a comment below: What were your first thoughts and feelings when you or a loved one received a life-altering diagnosis?

If you enjoy this work and would like to support me, please consider buying me a (decaf) coffee or giving a donation as a paid monthly or yearly subscriber. Every little bit helps. Thank you!

Buy Me a Coffee

Subscribe now

DISCLAIMER: This post is for entertainment purposes only and is not intended to be used to diagnose or treat any illnesses.

Glossary

• Autoimmune – A systemic, immune-mediated disease in which the body’s immune system, which fights off infections and cancer, attacks the person’s own body (e.g., lupus or rheumatoid arthritis).

• CT scan – Computed tomography, a type of radiologic imaging. Also called a cat scan.

• Differential diagnosis – The list of the possible illnesses that could explain a patient’s chief complaint.

• MRI – Magnetic resonance imaging.

• PVC – Premature ventricular contraction. Abnormal heartbeat in which the lower chambers (ventricles) beat out of sync. Also called premature ventricular complex.

Illustration credit — well worth viewing more macabre pics from twisted German folklore!

The cartoon shows a young girl in a green dress with bloomers and red shoes who is on fire. She also has a red belt and red ribbons in her hair. Smoke billows behind her and she stands on one foot. The girl is waving her hands in the air in fear. Two gray tabby cats stand on either side of her, their mouths agape and their paws in the air. There is a tree branch on the right side of the picture. The tone of the image is alarm.

See glossary at the end of this post for medical terminology.

In spring of 2018, my knees acutely became painful, swollen and stiff. I visited several orthopedic clinicians for this. Soon, I felt fatigue, palpitations and dyspnea and visited the ER, Student Health and urgent care in the span of ~week. Despite my obviously (to me) severe change in health, I was gaslit (Post 4) and told I was just a stressed-out grad student.

I believe I was gaslit because:

• I’d been socialized as a child to keep a stiff upper lip and act happy and confident, even when I wasn’t, which masked my illness

• I looked healthy

• I was a woman in my mid-40’s, so menopause was likely blamed

  • The pejorative term is whiny gyny

• My chart’s medical history was an erroneous creeping narrative (Post 2)

• Clinicians are trained “when you hear hoofbeats, think horses, not zebras,” so they viewed my experiences as psychological by default (Post 5)

• I lacked a PCP to consider the big picture and coordinate my care

• My symptoms were largely ignored.

More symptoms followed. Headaches, chest pain, chills, difficulty concentrating and dizziness. One day, my heart thudded so quickly and violently that my body shook with each beat.

I became emotionally labile. When something funny happened, I laughed too hard in embarrassing outbursts that felt beyond my control. A horror movie afficionado, I was shocked when my legs felt shaky and weak after watching (and thoroughly enjoying) Get Out.

Most of all, I felt a total loss of control and a sense of impending doom. As described on my About page, it felt as if I were being tumbled and crushed within a relentless, churning wave. This made the gaslighting all the more frustrating. Why didn’t these clinicians see something was terribly wrong? That, it turns out, was part of the problem. Symptoms are felt, and often not seen.

Signs vs. Symptoms

Signs comprise a patient’s abnormalities, as observed during the physical assessment. Think heart rate, blood pressure, skin temperature, joint swelling, etc. Signs are often measurable and are documented as objective data.

Symptoms are the abnormal sensations patients have. Think pain, feeling hot or cold, fatigue, nausea, or dizziness. Symptoms are documented as subjective data described by the patient.

Note that patients may show signs of their symptoms. A patient with the symptom of severe pain from appendicitis may have signs like tachycardia, tachypnea, perspiring profusely, leaning forward, clutching her stomach and moaning.

Clinicians document your HPI symptoms using the mnemonic OLD CARTS:

• Onset

• Location

• Duration

• Characteristics

• Associated symptoms

• Relieving/aggravating factors

• Timing

• Severity.

Many clinicians favor concrete, objective data over symptoms – think white blood cell counts that confirm an infection or CT scans revealing a mass at the source of pain. Symptoms are squishy. This squishiness can lead to having them downplayed in favor of objective data.

Unfortunately, many diagnoses are symptom-based. And many symptoms are non-specific, meaning they’re found across multiple disorders, so their presence doesn’t strongly suggest a certain diagnosis. Think fatigue, unintentional or unexplained weight loss, headache, loss of appetite, night sweats, malaise and dizziness.

Patients with systemic diseases may have wide-ranging symptoms. OLD CARTS is typically used to describe one symptom. But what if you have many?

Review of Systems

Symptoms are documented in the review of systems (ROS), a list categorized by body system (example). The ROS can be filled out with paper and pen at your appointment or online through the eCheck-In process, or the clinician or nurse can ask you about it.

After hundreds of clinician visits in the past 8 years, I’ve encountered issues with how the ROS is used:

• Checking boxes on prefilled forms or online templates is convenient, but limiting. What if your symptom isn’t listed?

• Specialists often only ask about symptoms related to their specialty

• Clinicians sometimes fudge part or all of the ROS, by indicating body systems are “normal” or “negative,” or that the patient “denies” symptoms, even if those symptoms haven’t been asked about

  • I call this documenting phantom findings

  • Auto-population, copy-pasting or computer templates are often to blame

  • It’s happened > 30 times, and I haven’t even reviewed all my notes yet!

• The ROS is no longer a documentation requirement, meaning the clinician can skip it altogether.

Inaccurate or omitted ROS documentation can hurt patients because:

• Those with complex, multi-system (often rare) disorders will face even longer wait times until diagnosis

• Insurance can more easily deny treatment coverage

• Disability claims are easier to deny

  • When I applied for disability, my symptom list included every body system

  • In his denial letter, the judge cited notes from every clinician who’d documented a phantom ROS as evidence I didn’t have these symptoms

• Patient agency and credibility are undermined when our lived experiences aren’t accurately documented.

How can we patients ensure our symptoms are taken seriously?

• Keep a symptom journal

  • Clinicians won’t read your 40-page journal. Summarize patterns and triggers into a sentence or 2.

• Type it onto a paper you bring to your appointment

  • Describe the symptom that bothers you the most using OLD CARTS, as well as your chief complaint

  • Include your ROS

    • May not be appropriate for urgent care or the ER

  • Be brief: I write everything on one side of a page

  • Ask that it’s scanned into your chart

    • Every clinician won’t agree to do so, but it’s a way to ensure your concerns are documented

• Fill out eCheck-Ins thoroughly

• If there isn’t an ROS on the eCheck-in, ask your PCP to add one

  • You may have to ask a patient advocate to have this done

  • Ask for a free-text box so you can document less common symptoms

• If your clinician isn’t taking your symptoms seriously, find somebody who does

  • Especially if it’s your PCP

  • I know; easier said than done. But having an A-Team that you trust makes a huge difference in your care.

• Read your notes!

  • Find errors and ask for them to be amended (this will be covered in a future post. Stay tuned!).

Remember that humans are complex and our illnesses aren’t always obvious. Clinicians are time-crunched and doing the best they can. Part of our frustration as patients stems from poor communication. Being direct and concise about your symptoms and how they affect you can help.

Comment below about how symptom report (or lack thereof) affected your diagnostic odyssey.

Diagnostic Odyssey is a free newsletter supported by readers like you. If you’re able, please consider donating a monthly or yearly subscription, or buying me a (decaf) coffee. Thank you so much for being here!

Buy Me a Coffee

Glossary

• Chief complaint — The main reason why you’re seeking medical care.

• Clinician — A catchall term for a Doctor of Medicine (MD), Doctor of Osteopathy (DO), nurse practitioner (NP) or physician’s assistant (PA) who bills as an independent provider and diagnoses and treats patients.

• CT scan — Computed tomography, a type of radiologic imaging. Also known as a cat scan.

• Dyspnea — Shortness of breath.

• ER — Emergency room. Also known as the emergency department (ED).

• HPI — History of present illness.

• Palpitations –— Abnormal heart sensations, like fluttering, pounding or rapid heart rate.

• PCP — Primary care provider. A licensed MD, DO, NP or PA who coordinates care by ordering tests/referrals and diagnosing/treating patients.

• Specialist — A MD or DO who specializes in a specific field of medicine (e.g., pulmonology, neurology and cardiology).

• Tachycardia — Abnormally fast heart rate (> 100 beats per minute).

• Tachypnea — Abnormally fast respiratory rate (> 20 breaths per minute).

Photo credit

The image shows a zebra standing sideways on a brown prairie. The zebra looks directly at the camera. The background has a light-blue sky and a blurry landscape.

February 28 is Rare Disease Day. This post is dedicated to all you zebras out there. There are more of us than most people realize. Stay strong!

Horses, Not Zebras

Clinicians are taught the simplest explanation for a patient’s symptoms is likely the correct diagnosis, an example of Occam’s Razor (a.k.a. the Law of Parsimony). In medicine, Occam’s Razor is typified by the adage, “when you hear hoofbeats, think horses, not zebras.” This means common diseases are considered first among the list of possible diagnoses. Unfortunately, this approach often overlooks zebras.

People with rare diseases have embraced the term “zebra” to describe our existence. The Ehlers-Danlos Society has even adopted the zebra as its logo.

But what exactly makes us zebras?

What is a Rare Disease?

In the U.S., a rare disease is one affecting < 200,000 people. In Europe, it’s defined as a disease affecting ≤ 50 per 100,000 people. These definitions typically exclude infections, poisonings and rare cancers.

Some rare disease facts:

• ~3–6% of people have a rare disease, meaning at least 1 out of 20 patients could be a zebra

• Up to 30 million adults in the U.S. have a rare disease

• > 300 million people worldwide have a rare disease

• 70% of rare diseases present in childhood

• 7000 rare diseases have been identified, and

• Only ~500 of those have treatments

• Rare diseases are often difficult to diagnose. One study showed diagnosis takes an average of 4.7 years, with 25% of zebras receiving a diagnosis 8 years after symptoms start.

Why is Diagnosing Zebras Difficult?

It can take months, years or even decades to diagnose rare diseases, and some are never diagnosed. This is partly because:

• Rare diseases are often complex, involving multiple body systems

• They may have a heterogeneous presentation (different people present differently)

• Few relevant research studies may exist

• Existing studies have small sample sizes, which may skew results

• Many are hereditary, and genetic testing is costly

• Clinicians lack the time and knowledge to investigate them thoroughly

• Patients may see multiple specialists before a unifying diagnosis is found

• U. S. healthcare is fractured

• Gaslighting occurs, like I described in Post 4

• Travel for diagnosis and care is expensive.

Why are Zebras Difficult to Treat?

Treatment is difficult for many of the same reasons listed above, and also because:

• Treatment for a complex disease is often complicated

• It’s expensive for Big Pharma to develop meds that will be used by only a few people

• Big Pharma passes its expenses on to patients with rare diseases

• Insurance companies may deny coverage, especially for off-label treatments

• Patients and caregivers may struggle to maintain employment and insurance coverage

• Patients and caregivers must become experts on their disease and advocate for themselves or their family.

Rare Disease Self-Advocacy

Zebras can learn more about rare diseases and connect with other zebras through:

• Patient support groups

• Government websites

• Clinical centers of excellence

• Research databases

• Clinical trials

• Rare disease foundations.

Zebras, Not Horses

Being a zebra presents multiple challenges beyond the physical and psychological effects of the disease. Financial barriers, stress, patient advocacy, patient work, disability, discrimination and social isolation are all factors that can affect zebras and their families. The following resources can help.

Rare Disease Resources:

• Genetic and Rare Diseases Information Center (GARD)

  • National Institutes of Health (NIH) program

  • Helps the rare disease community access info, get a diagnosis and find resources

• National Organization for Rare Disorders (NORD)

  • Database with info sheets, undiagnosed disease support, patient organization list and more

• Orphanet

  • “Rare diseases are rare, but rare disease patients are numerous.”

  • Encyclopedia of rare diseases, inventory of orphan drugs (meds for rare diseases), directories of patient organizations and info on institutions and clinical trials

• Rare & Undiagnosed Network (RUN)

  • Run by impassioned mother of 3 children with an undiagnosed disease

  • Resources and support for parents of children with rare or undiagnosed diseases

• Undiagnosed Disease Network

  • NIH program

  • “… combines basic and clinical research services to understand health conditions for individuals and their families who have sought a clinical diagnosis without success.”

Thank you for reading about rare diseases! If you’re able, please consider supporting NORD, Orphanet, RUN or other rare disease organizations with a donation.

Have something to share about being a zebra? Did I forget something about rare diseases you think is important? Leave a comment!

Leave a comment

Instead of posting a newsletter next week, I’ll record previously posted newsletters and send out some Notes. Be well, and have a Happy Rare Disease Day! :]

Buy Me a Coffee

This trailer is so dramatic! Totally worth watching.

The image is a screen grab of a trailer for a black-and-white movie. There’s a young woman with her hand at her face, looking concerned. A man stands behind her, leaning over and furrowing his brows. The woman has curly hair and wears a lacy Victorian blouse and a dark jacket. The man wears a dark suit with a tie. The mood of the people in the image is concerned and pensive.

See glossary below for medical terms.

I wrote in Post 2 how my medical history was distorted across successive notes into a creeping narrative. During that time, life as a Ph.D. student was stressful. I’d been tasked with writing my comprehensive exam (comps) paper and my first 3 dissertation chapters over a 7-week period.

I kept a grueling schedule and ultimately wrote 85 pages, reminding myself grad school is all about suffering. During this time, my knees had swelled up painfully, but I lacked the time to address this properly. Somehow, at the end of May, I finished my papers. I expected to feel better once the yoke of this madness had been lifted.

Something Was Still Wrong

Instead, I felt drained and mentally fuzzy. I started having palpitations. Sometimes it felt like an angry gnome was trapped inside me, intermittently giving a sharp kick to my sternum. Other times, I felt a fluttering sensation in my chest, leaving me woozy. When I showered, my heart rate raced up to a dizzying 150 bpm.

Loathe to spend money on healthcare, I joined a research study in which I wore a heart rate monitor for 2 weeks. The results showed 2% of my heartbeats were PVCs, runs of SVT and a couplet. There was also a lot of sinus tachycardia (ST) — heart rates that were too fast, but otherwise normal. The results were unexpected in a fit, healthy 46-year-old woman. Notably, the ST wasn’t mentioned in the final report.

I scheduled a cardiology appointment at the end of June, the earliest availability. I figured I’d just deal with my symptoms. Until I couldn’t anymore.

Student Health Visit

One Friday in June, I felt dyspnea while cycling ½ mile (0.8 km) home, so I visited Ms. K, a Student Health NP. Labs were drawn and an EKG showed a single PVC. Ms. K listened with compassion, but had no answers. She suggested I visit the emergency room (ER) if things got worse.

Soon, the little gnome was kicking my chest a few times per minute. I cycled back to the lab, panting, and hooked myself up a continuous EKG monitor. I saw PVCs on every 5^th beat. I’d seen this before, usually in older patients with heart problems. Something was wrong, so despite my hesitancy, I went to the ER.

ER Visit

When I told the ER resident I was having PVCs every 5^th beat and dyspnea, he ordered an EKG that was normal. My (usually normal) blood pressure was high at 150/90. Still feeling PVCs every 5^th beat, I asked to be attached to a continuous EKG monitor.

“I’m not going to do that,” he said with a shrug. “You have a cardiology appointment in 2 weeks.”

I was shuffled off to the PA and discharged after lab results came back as normal. Something was wrong and almost every clinician I’d seen so far had been dismissive. In retrospect, their notes held some clues.

Student Health and ER Notes

Years later, I read that Ms. K had accurately documented my reported dyspnea. But the ER team wrote I’d denied dyspnea, even though it was one of the main reasons I’d gone to see them. Being invalidated like this made me feel so alone. Especially since I’d buried myself in my work and hadn’t had time to make friends.

Thankfully, somebody else realized I wasn’t well.

External Validation

The Director of the lab where I was doing my research called me the night before my oral comps exam. He was out of town at the time.

After I greeted him, he asked, “What’s wrong?”

I hedged, “I’ve been tired.”

He responded, “Something’s wrong with you. Let’s postpone your comps. If you don’t take yourself to the ER today, I’ll drive 6 hours and take you myself.”

Finally, somebody else had noticed what I’d been feeling for months! I was grateful my Director was looking out for me. I told him I had no intention of returning to the ER because of the way they’d treated me (and because of the cost). We agreed I should visit Urgent Care.

Urgent Care

At Urgent Care, another EKG was done and I watched closely. Normal, again, as were my labs. I was sent back into the waiting room and after a few minutes, the physician approached me. She took my arm as she sat beside me.

Her diagnosis?

“Graduate school is hard.”

When I read her visit note years later, I realized the MA who’d taken my vital signs had invented my RR (she wrote it was 28 breaths per minute). I think she did this because she decided I was anxious because I was watching the EKG screen. If I’d been breathing that quickly, I would’ve felt short of breath just sitting there. Which I didn’t.

When the physician saw a RR of 28, she likely concluded I was having a panic attack. Which I wasn’t.

Months earlier, when I thought my swollen knees could be immune-mediated, orthopedics clinicians disagreed. Other clinicians dismissively thought stress was my only problem. Everyone was missing what seemed so obvious to me. I was being gaslit.

Gaslighting

Dismissing a patient’s health concerns is a form of gaslighting. The term is based on the movie Gaslight, in which a husband fiddles with the gas lights in the house, in an attempt to convince his wife she’s insane so he can steal her fortune.

Medical gaslighting is also known as medical invalidation, or a questioning of patients’ authority to judge their own lived experiences. The word has become so popular, it was christened Merriam-Webster’s 2022 Word of the Year. And of course, it has its own Reddit forum.

Why Was I Gaslit?

An article in JAMA attributes gaslighting to:

• Clinicians’ lack of knowledge

• Implicit bias – Believing negative stereotypes toward certain groups of patients, based on demographics like race, ethnicity, gender, age, economic status, geographic location, sexual orientation or patients who’ve had many appointments (a.k.a., frequent flyers)

• Paternalism – Believing that clinicians alone know what’s best for patients.

Gaslighting more frequently happens to people from marginalized communities. I was gaslit because I was a middle-aged woman, known pejoratively in medicine as a “whiny gyny.” I was seen as a graduate student just going through menopause and making a big deal out of it. My experience also shows that inaccurate documentation contributes to gaslighting.

How Did Gaslighting Affect Me?

I was privileged to have been trained as an NP, so I knew my abnormal findings were a drastic change from my baseline. Yet, I felt unseen and unheard at almost every healthcare visit. With the exception of the Student Health NP (and my lab’s Director), it seemed like nobody cared. This made me feel frustrated and demoralized. I began to question my lived experience, and with that, my sanity. The stress of gaslighting makes patients sicker and delays proper diagnosis.

Yes, grad school is stressful. My symptoms were likely exacerbated by stress, but psych issues weren’t the main cause. If clinicians had taken my report of dyspnea seriously and taken a chest X-ray, they would’ve seen telltale signs explaining why I was having these symptoms. Instead, I was gaslit.

Buy Me a Coffee

Glossary

• BPM – beats per minute

• Couplet – 2 premature ventricular contractions in a row

• Dyspnea – Shortness of breath

• Dysrhythmia – Abnormal heart rhythm

• EKG – Electrocardiogram. Also written as ECG

• MA – Medical assistant

• NP – Nurse practitioner; A mid-level provider who can diagnose and treat patients and prescribe medications. Physician supervision is dependent upon the state where the NP practices.

• PA – Physician’s assistant; A mid-level provider who can diagnose and treat patients and prescribe medications under the supervision of a physician.

• Premature ventricular contraction (PVC) – Abnormal heartbeat in which the lower chambers (ventricles) beat out of sync. Also called premature ventricular complex.

• RR – Respiratory rate

• Resident – Medical doctor (MD) or Doctor of Osteopathy (DO), graduated from medical school and is in post-graduate training and is supervised by a board-certified MD.

• Supraventricular tachycardia (SVT) – Rapid heart rate originating from the upper chambers (atria).

The image shows a blue sky at dawn with white and gray speckled clouds. The bottom of the picture has black treetops, while a flock of black birds is flying in front of the clouds. The image is peaceful and inviting. Photo credit.

Click below for an audio version of this post, which includes hyperlinks to a YouTube video of people cheering and banging pots and pans for essential workers and a YouTube video about active listening.

This week, I had an initial visit with a surgeon. My stomach was knotted in apprehension as I sat in the waiting room, thoughts pinging through my head. Will she listen? Will I be ridiculed for thinking my issue is serious enough to warrant surgery? Will she brush off the articles I brought about my rare disease?

I ask these questions because sometimes, unfortunate interactions with providers occur. Patients who’ve experienced medical trauma or gaslighting can become cynical and mistrustful. Sometimes, we project our previous experiences onto the healthcare workers in front of us now, wrongly seeing them as barriers to our care. When you’re sick or in pain, it’s easy to forget that the people who are caring for you are doing so because they want to help you get better.

At times, I’ve let anger and frustration at the healthcare system get the better of me and acted in ways I regret. As a nurse and nurse practitioner, I’ve also been on the receiving end of patient and family anger – sometimes for making mistakes, but also for doing nothing wrong at all. It’s easy to list a litany of grievances against the monolith that is healthcare, but let’s step back and remember why healthcare workers do what they do.

Service in Healthcare Involves Sacrifice

Healthcare workers are driven by a selfless dedication to serve others by bettering our lives, offering comfort and hope and ameliorating illness and pain. It’s not an easy career choice. Healthcare workers often go deep into debt for their schooling. They work long hours, sacrificing coveted personal and family time in service of their patients. The best example of their selfless devotion to their patients occurred during the Covid-19 pandemic.

I didn’t work in healthcare during the pandemic, but I know a lot of people who did. Many worked long hours with few resources, at times even without protective equipment. But they didn’t just experience the physical travails of working long shifts. They also endured the psychological trauma of watching so many patients die while feeling helpless to do anything about it.

We should commend those who stepped up to care for the vulnerable when we needed them. Remember when we used to thank essential workers who risked their own lives each day for us? We banged pots and pans and cheered for them in gratitude. Now that the pandemic has receded, do we still value the people who live to make the world a better place than how they found it?

Healthcare Workers Are Frustrated About Healthcare, Too

On hold for 45 minutes, a patient snaps at a clinic receptionist. A pharmacist is chastised for a delayed refill. A doctor and her staff repeatedly fight behind the scenes to get a treatment approved. In a nod to civility, a clinic’s on-hold recording reminds patients their office is staffed with intelligent and kind employees. A nurse practitioner routinely stays 2 hours past the end of her shift to carefully write notes for all her patients. These are symptoms of a healthcare system strained beyond its means.

There aren’t enough people to do the work. Workers face unrealistic and ever-increasing documentation pressures to get insurance approvals, meet government requirements and avoid lawsuits. These pressures affect patients and healthcare workers alike, making it easy for healthcare to devolve into an us-versus-them mentality. But sometimes, all it takes is a visit with a compassionate provider to remind us we should value all healthcare workers because they value us.

The surgeon I visited let me know she cared about me. She and her staff put on masks when they saw I was wearing one. She demonstrated active listening. She was curious, asked questions and didn’t make assumptions. She seemed delighted when I gave her articles on my rare disease. She said I was eligible for surgery, but empowered me to make the decision when I feel ready.

Most of all, she reminded me that everyone on both sides of the stethoscope must give each other grace while navigating healthcare. And that, under pandemic conditions or not, we patients owe all healthcare workers a debt of gratitude. Their burdens are many; the thanks they receive, few. So, the next time you encounter a healthcare worker, take a moment to see them as someone who cares and thank them for all they do. Maybe just skip the pots and pans …

Buy Me a Coffee

Photo credit. The image is a fuzzy white snowball sitting on snow that covers the ground. The snowball casts a shadow. The background is dark as if this is nighttime and there are circles of bright light, colored yellow and blue.

See glossary below for definitions of medical terms.

From the beginning, nobody got my medical history right.

For years, I didn’t even know.

Like clues at a crime scene, a patient’s history of their present illness (HPI) gives crucial evidence for proper diagnosis. But it’s really hard for clinicians to efficiently and accurately integrate the many details of the HPI. Research suggests we humans are unaware of how poorly we listen in our busy environment. And patients may omit or forget important info, emphasize the wrong info, or mislead altogether.

My meandering diagnostic odyssey shows how small mistakes in the HPI can snowball across multiple healthcare notes and become an unrecognizable creeping narrative.

How it All Started

Before becoming ill, I worked hard and played harder. I hiked through fog in Hokkaido so dense, I couldn’t see my own hands; careened down a zip line in Costa Rica onto a rickety wooden deck and soared above the Catskills on a tandem hang glider. Every new park or peak, from Bolivia to Banff, defined my vigorous existence.

My troubles began in 2018, after I moved to a new city for grad school. I lifted weights and climbed at a local gym to stay in shape. After weeks of training, my knees became a little stiff. This was odd but painless, so I ignored it. Two weeks later, after climbing hard on March 22nd, my knees swelled overnight, becoming very stiff – I could barely bend them – and very painful. What was going on?

March 23: Saw Dr. E, a Board-Certified Family Care MD at Urgent Care

As I drove to urgent care the next morning, I winced at the stabbing pain. Swallowing tears, I told Dr. E I’d felt 2 weeks of slight stiffness and then overnight, my knees had swelled, becoming severely stiff and painful. She seemed bored and took no notes.

Dr. E said the tangerine-sized balloons on the back of my knees were Baker’s cysts. She refused to drain them, saying they would only swell again. Instead, she injected steroid and lidocaine into my (more painful) left knee, flooding it with short-lived relief.

Years later, I found mistakes in Dr. E’s note. She’d written my pain was “moderate,” although I’d called it “severe,” an 8/10 on the 1–10 scale. She wrote my pain and swelling had happened “gradually over 2 weeks,” instead of overnight. And she didn’t mention anything about my right knee.

March 26: Saw Ms. S, an Orthopedic PA

This and all visits described below were at the same University Health System.

The pain and swelling remained, so I needed further evaluation. I was nervous because I knew sudden, severe, bilateral joint pain and swelling can happen in autoimmune diseases. Before this visit, I asked the urgent care facility to forward my visit note to Ms. S. I didn’t think to request a copy for myself.

As I described my HPI, Ms. S nodded, hands resting in the pockets of her crisp, white coat. Like Dr. E, she took no notes. X-rays showed mild bilateral arthritis.

Ms. S then drained both cysts. I felt a painless pop! as she guided the needle into my left knee and aspirated 50 mL of clear, yellow fluid, which she sent to the lab. But as she aspirated the right cyst, a burning lightning bolt shot down my leg. Ouch! The cyst had likely collapsed onto a nerve. Overall, these procedures provided immense relief for several days, until my knees filled up again.

Ms. S wrote that my knees had become painful and swollen “gradually,” over a 2-week period. She likely copied the inaccurate HPI from Dr. E’s note. And she mentioned aspirating fluid from only the left knee.

April 9: Follow-Up with Ms. S

The pain and swelling remained and Ms. S ordered a left knee MRI.

I asked, “Could this be autoimmune?”

“No, your lab results were negative,” she said.

The note for this visit stated my knees had swelled and become painful “over the past month.” It had been 2 ½ weeks. The narrative continued to creep …

April 23: Saw Dr. A, Board-Certified Orthopedic Surgeon and Dr. K, Orthopedic Resident

Ms. S referred me to Dr. A because I wasn’t improving. Dr. A and Dr. K said the MRI showed “fairly severe” left knee arthritis and they surmised the right knee had similar issues. They said I was too young (age 46) for knee replacement.

Again, I asked, “Is this autoimmune?”

“No.”

I was referred to PT. The surgeons wrote my swelling and pain had been occurring for “several months,” though it had been exactly 1 month. More creep …

May 9: Saw J, DPT

J asked about my HPI. Just like before, I said I’d had 2 weeks of mild stiffness and then my knees had abruptly swelled overnight. I never said swelling had occurred “gradually,” but that word appeared in her note, too. She wrote the swelling had occurred over “several” months, and also described its onset as “insidious.”

June 18: Saw Dr. A and Dr. T, Orthopedic Resident

I saw Dr. A again with a different resident. The visit note had few HPI details, stated I had “mild” arthritis and cited none of my imagining.

Creeping Narrative Synopsis

Across several notes at 2 facilities:

• The inaccurate qualifier “gradually” was added

• The onset and duration time frames were exaggerated

• Problems with, and a procedure in, the right knee were omitted

• Pain severity was downplayed.

These errors cast strong doubts that my diagnosis could be immune-mediated, so no referrals to a rheumatologist were made. These interactions with clinicians left me feeling confused and frustrated that I wasn’t being heard.

What Could I Have Done Differently?

• Visit an ED

  • I went to urgent care instead of the ED to save money, but also because my issue didn’t feel life-threatening. I would’ve gotten a more thorough work-up at an ED.

• Emphasize my chief complaint

  • When recounting my HPI, I first mentioned “2 weeks of slight stiffness.” But stating “my knees swelled overnight and became very painful” first would’ve emphasized the onset and severity.

• Read medical records

  • I didn’t for years, so I missed errors and the timely ability to correct them.

• Find a PCP

  • PCPs are essential for quality healthcare.

• Self-advocacy

  • I didn’t understand patients must advocate for appropriate care.

What Could Clinicians Have Done Differently?

Take notes

• This reduces the chance of errors. The use of AI recording tools can help with this, but my first post shows how it can also lead to errors.

• Consider other diagnoses

  • My knee workup was limited and no differential diagnoses were documented. Knee swelling can be caused by aneurysms and blood clots, which shouldn’t be punctured by needles.

• Be mindful

  • Clinicians didn’t just copy incorrect info from note to note. They embellished or omitted details and seemed unaware they were doing this.

• Beware of Gospel bias

  • In my experience, the belief that documented medical info is accurate, often despite what the patient says, seems common. I call this Gospel bias.

• Training

  • A negative result for immune-mediated blood or fluid markers doesn’t automatically rule out an immune-mediated diagnosis. Patients with these disorders often visit orthopedic clinics first, so orthopedic clinicians deserve better training on this topic.

• Connect patients to appropriate services

  • If a patient lacks a PCP, specialists or their staff should suggest they get one, explain why it’s important and help them obtain one (See Post 10).

If you enjoy my work, please consider buying me a cup of (decaf) coffee or tea.

Buy me a coffee

Glossary

• Aneurysm – A bulge in an artery or vein.

• Autoimmune – A systemic, immune-mediated disease in which the body’s immune system, which fights off infections and cancer, attacks the person’s own body (e.g., lupus or rheumatoid arthritis).

• Baker’s cyst – A fluid-filled cyst on the back of the knee, typically associated with injury or arthritis.

• Bilateral – occurring on both sides of the body (e.g., both knees).

• Chief complaint – The main reason for a patient’s visit.

• Clinician – A catchall term for a Medical Doctor (MD),Doctor of Osteopathy (DO), nurse practitioner (NP) or PA who bills as an independent provider and diagnoses and treats patients.

• Differential diagnoses – The list of the possible illnesses that could explain a patient’s chief complaint.

• Doctor of physical therapy (DPT) – A licensed healthcare professional who treats patients’ mobility problems or pain.

• Emergency Department (ED) – A healthcare facility providing care for patients with acute and life-threatening issues.

• Immune-mediated – A systemic disease involving the immune system. Includes autoimmune diseases and also diseases in which an attacking antibody hasn’t been identified.

• Insidious – Not acute; occurring slowly, over a long period of time.

• Physical therapy (PT) – Field of medicine in which a DPT works.

• Physician’s assistant (PA) – A licensed, mid-level provider who can diagnose and treat patients and prescribe medications under the supervision of a physician.

• Primary care provider (PCP) – A licensed MD, DO, NP or PA who coordinates care by ordering tests/referrals, and diagnosing/treating patients.

• Resident – MD or DO, graduated from medical school and is in post-graduate training and is supervised by a board-certified physician.

• Specialist – A MD or DO who specializes in a specific field of medicine (e.g., pulmonology, neurology and cardiology).

Photo credit. The image is a coiled red stethoscope, draped over a black computer keyboard.

See the glossary at the bottom of this post for definitions of (mostly medical) terminology.

The purpose of Diagnostic Odyssey is to describe barriers to diagnosis and treatment of rare diseases, with a focus on the impact of documentation errors. I expected to post chronologically about my illness journey. But after recently seeing a specialist whom we’ll call Dr. F, I found some errors in the visit note, which was generated by an artificial intelligence (AI) tool that recorded our conversation. So instead of starting at the beginning of my story, let’s jump into the present with a timely concern: the growing use of AI in medicine and how it can generate inaccurate slop that could affect patients.

What is AI?

AI involves the use of computers to autonomously learn and complete tasks approximating complex human thinking skills, like problem solving, pattern recognition and information synthesis. AI tools are everywhere; think Google Gemini, Open AI’s ChatGPT, Canva AI and DeepSeek.

When you search in Google, an AI response pops up. Have you noticed these can be wrong? When an AI tool outputs inaccurate or nonsensical info, it’s called a hallucination. There are plenty of comical examples of hallucinations, including a recommendation to eat rocks daily and a spaghetti recipe that uses garlic sauteed in gasoline. Less comical is the potentially dangerous impact of using AI in medicine, which historically has had serious bugs.

AI in Medicine – Then

In 2018, STAT News detailed concerns with IBM’s Watson for Oncology AI tool. Using training data taken from Memorial Sloan Kettering Cancer Center (MSKCC), Watson botched cancer care recommendations for the physicians and health systems that bought it worldwide. Apparent failures included the following:

• The model was trained on a small number of cases

• It was sometimes trained on synthetic cases based on groups of patients

• Some MSKCC treatment protocols conflicted with evidence-based cancer care guidelines

• Suggested medications were unavailable in some countries

• The accuracy of Watson’s performance was exaggerated

• Product flaws weren’t corrected in real-time.

One oncologist stated bluntly, “This product is a piece of shit.”

Watson for Oncology provides a case study on how not to use AI in medicine. IBM ultimately sold its AI division in 2021. You may protest that Watson was rolled out almost 10 years ago. Surely AI has improved since then, especially in medicine!

AI in Medicine – Now

A 2024 Cureus article touts the use of AI in medicine as “revolutionary” but also addresses its disadvantages.

Possible uses of AI in medicine include:

• Radiologic imaging interpretation

• Prognosis prediction

• Diagnosis and clinical decision-making

• Analysis of large datasets

• Medication development

• Driving health-related chatbots

• Monitoring patients remotely

• Providing treatment options

• Improving efficiency in administrative tasks.

Disadvantages include:

• Difficulties maintaining health information privacy and system security

• Poor transparency on how AI tools complete tasks

• Reinforcement or amplification of pre-existing biases and discrimination

• Higher costs

• Difficulties integrating AI with current systems

• Accessibility issues

• Errors

• Clinician overconfidence in AI performance.

Let’s see the last 2 bullet points in action in Dr. F’s note.

Note Errors and Analysis

I found 11 note errors/omissions, and for brevity, I’ve highlighted the 6 most likely AI-generated examples. I haven’t spoken about these errors to Dr. F, whom I see out-of-state once a year and with whom I can’t do telehealth. I also don’t know which AI tool Dr. F used. I can’t prove what info was written by Dr. F and what was AI-generated, but I have a pretty good idea:

1. The PYGL gene was written correctly once and as “PIGL” twice

2. I mentioned I’d seen a Dr. Jaylander (not his name, but it sounds pretty close to this), but his name was written as “Dr. Goldan”

3. Under labs, “leptin” was written correctly once and as “lactate” twice

At my appointment, I gave Dr. F a paper listing these issues and we discussed them. Dr. F also knows Dr. J. I think the AI software had problems transcribing our conversation to text because I wear a mask that muffles my voice (I’m immunocompromised). To be fair, human scribes or transcription software without AI could’ve made the same errors, as they make more mistakes than most people realize.

4. “She has not undergone a mammogram due to a previous mastectomy.”

After an abnormal mammogram in 2018, I had a right mastectomy and I get a mammogram on my left breast yearly.

The point of discussing mammograms with Dr. F was to tell her about substandard care I’d received from a rheumatologist (Dr. V), whom I’d seen at another health system. At that appointment, Dr. V didn’t ask about my labs or radiology reports, but he could’ve accessed this info through a software feature known as Care Everywhere. Despite this, Dr. V’s note lacked important abnormal testing results, including several mammograms. These tests were listed as “Never done.”

I was perplexed and frustrated with Dr. V’s conclusion that I’d never had a mammogram. Notes from other physicians at this facility also had errors and omissions that could’ve hindered my care. Because Dr. F had previously asked for these notes, I brought Dr. V’s note to defend my decision not to forward them to her.

I’ve written 2 paragraphs to describe this complicated point, so it’s possible Dr. F misunderstood me and dictated this inaccurate entry. But it’s also possible it was AI-generated. Note that I was advised to disable Care Everywhere after my appointment with Dr. V, so Dr. F’s AI tool presumably couldn’t access testing data from other facilities.

5. “Two-point discrimination is mostly sharp dull.”

This sentence mixes 2 statements I made about previous neurological exam findings. I said two-point discrimination on my feet is wide at > 1.5 cm. I also said I’ve lost sharp-dull discrimination in my feet.

Dr. F would know these are 2 different exams and likely didn’t dictate this.

6. Diagnosis of “Intractable chronic migraine with aura and without status migrainosus”

I have a history of migraines without aura and my migraines aren’t intractable.

Dr. F and I didn’t discuss my migraines, and they weren’t mentioned in the document I gave her. I take 2 migraine meds, so perhaps my medication list prompted the AI tool to hallucinate this diagnosis.

[Erratum 17 April, 2026: This diagnosis was not hallucinated by AI. Dr. F said she selected this diagnosis on purpose. See Post 9 for details].

Taken together, these errors paint a vastly different clinical picture than my real medical issues. AI-generated errors like these could lead to serious consequences for patients including missed diagnoses, misdiagnoses and inappropriate testing or treatments. As Futurism.com notes, we’re currently being force-fed imperfect AI models as if they’ll magically improve most human tasks. I believe this hype fosters overconfidence in AI’s ability to efficiently and safely improve healthcare delivery.

Clinician Overconfidence in AI Performance

I know Dr. F trusted the accuracy of the AI tool and didn’t read my note before signing it because at least some of the errors would’ve been corrected if she’d done so. Many clinicians don’t have enough time to read notes before signing them. Which isn’t surprising, since healthcare workers are under immense stress and pressure to see as many patients as possible, necessitating workarounds. I can empathize because, when I worked in healthcare, I often barely had time to document at all, let alone review what I’d written. Of course healthcare workers find AI tools appealing! But clearly, we must be more objective about AI.

AI Tools in Medicine: Approach with Caution

The New York Times recently published “Stop Worrying, and Let AI Help Save Your Life” by Dr. Robert Wachter. He argues for the use of AI, including AI scribes, in medicine, based on clinical experience. He is also clear-eyed about the need to review AI output and he admits AI-generated responses are sometimes “just plain wrong.”

I didn’t experience negative consequences from these AI-generated note errors, so my story is a cautionary tale. Mistakes can compromise healthcare by giving clinicians, patients and insurance companies inaccurate data. And if these errors are happening in my note, they’re likely happening in other patients’ notes, too. Mistakes can also undermine any AI-driven gains in workflow efficiency, as time must be spent correcting them.

The biggest consequence for me is the time it takes to submit an amendment request to have these errors corrected. But I urge all stakeholders to avoid blindly trusting the accuracy of AI-generated notes, and to carefully review them and promptly amend errors. Because AI slop has no place in medicine – although apparently, it’s already here.

If you’re able, please consider helping to support me by buying a medication co-pay.

Buy a Med Co-Pay

Glossary

• AI Slop – Low-quality output made with artificial intelligence. Usually refers to mass-produced content.

• Aura – Sensory disturbance and/or neurological deficits occurring before a migraine, including feeling confused and having difficulty speaking, numbness, tingling or blurred/altered vision like blind spots or seeing colors, flashing lights, dots or zigzagging patterns.

• Clinicians – Catchall phrase for physicians and mid-level healthcare workers (nurse practitioners and physicians assistants) who can diagnose illnesses and prescribe medications.

• Intractable – Difficult to alleviate or treat.

• Migraine – A complex neurologic condition that can cause a severe, throbbing headache, tiredness, nausea or vomiting and sensitivity to light and sounds.

• Sharp-dull discrimination – A neurologic exam in which patients are gently poked with either a sharp object (like a pin) or a dull object (like a cotton swab) and asked which sensation they perceive.

• Specialist – A physician who provides healthcare in a specific field. For example, neurology, infectious disease, cardiology or gynecology.

• Status migrainosus – A migraine that lasts longer than 72 hours.

• Two-point discrimination – A neurologic exam in which patients must distinguish whether 2 pins touching the skin are not one point of contact, but 2 distinct ones. This exam has somewhat fallen out of favor because it isn’t considered to be very reliable.