Here’s a pattern I’ve watched up close: a company builds a genuinely good medical device. They sell the box. The clinic buys the box. The box sits in a corner.

The device worked. The program didn’t.

Because a clinic doesn’t need a box. It needs a screening program that runs, patients identified, images taken, results read, follow-ups booked, every week, without heroics, inside the chaos of a real practice. The device is one ingredient in that. Maybe twenty percent of it. The other eighty percent is workflow: who does what, when, and how you know it happened.

So there’s a flip available, and the smart device companies are starting to see it. Stop selling the device. Sell the outcome. Instead of “here’s a camera,” it’s “here’s a working screening program, and the camera comes with it.” One-time sale becomes ongoing relationship. The customer stops buying equipment and starts buying results.

This is better for everyone. The clinic gets something that actually works instead of an expensive paperweight. The company gets recurring revenue instead of lumpy one-time orders. And patients get screened instead of sitting on a waiting list for a device nobody set up.

The lesson generalizes beyond healthcare, I think. Customers never wanted your product. They wanted the outcome your product was supposed to produce. Sell that.

Diabetes can steal your vision. Slowly, painlessly, and, here’s the tragedy, preventably. A photo of the back of your eye catches it years before you notice anything wrong. The cameras exist. The people trained to read the photos exist. And still, people go blind.

Why? Because the photo never gets taken.

Follow the chain. You’re at your primary-care doctor for diabetes. Someone needs to remember you need an eye photo, find time in a chaotic day, take the photo, send it somewhere, get it read, get the result back, tell you the result, and book the follow-up if something’s wrong. That’s seven handoffs. Every handoff leaks patients. By the end, a fraction of the people who needed screening got it.

Notice what the bottleneck isn’t. It isn’t the camera, that problem was solved years ago. It isn’t the reading, trained eyes and good software both exist. The bottleneck is the boring stuff: who does what, when, and who checks that it happened.

I keep running into this pattern. Most of healthcare’s unsolved problems aren’t medicine problems. They’re logistics problems wearing a medicine costume. We keep inventing better cameras while patients fall through the gaps between the camera and the follow-up.

The next breakthrough in preventive care won’t be a better algorithm. It’ll be someone who makes the seven handoffs happen without heroics, every patient, every Tuesday, no exceptions.

America doesn’t have a specialist shortage. It has a specialist distribution failure.

The scarcest resource in our healthcare system is a specialist’s time. And we hand it out by geography. Live near a big medical center and you might wait days. Live in a small town or farm country and you wait months — for the same question, the same disease, the same human being in need.

Nothing about your illness cares where you live. But everything about your care does.

Here’s the backwards part: we make the patient travel to the knowledge. You drive three hours, take a day off work, arrange childcare — so that a specialist can look at you for fifteen minutes and often just answer a question. A huge share of specialist visits end with advice, not procedures. The knowledge was the product. The travel was pure overhead.

Flip it. An e-consult — that’s just your own doctor asking a specialist a question electronically, doctor to doctor — moves the question instead of the patient. The specialist reads the case, gives guidance, and most of the time no visit is needed at all. Same knowledge, zero windshield time.

We built the whole system around the idea that expertise lives in buildings. It doesn’t. It lives in people’s heads, and heads can answer questions from anywhere.

The zip code lottery isn’t a medical problem. It’s a logistics problem wearing a white coat.

Who the ED calls, and how long the answer takes

In my last two posts I argued that the emergency department’s problem isn’t just missing hands — it’s missing answers. Decision delay, not just doctor shortage. This week I found the data. It says I’m right, and it’s more specific than I expected.

Here’s the setup. When an ED doctor hits a case she can’t resolve alone, she calls a specialist — a consult. Nationally, the CDC’s emergency department survey found a consulting physician was seen in 11.6% of ED visits in 2018, 13.5% in 2019, and 15.7% in 2020 — trending up. Studies that count consults requested (not just completed) put it at 20–40% at teaching hospitals. So somewhere between one in eight and one in three ED visits waits on a specialist’s answer.

Now the part nobody talks about: who gets called, and how long each one takes. A 2026 study out of the University of Virginia tracked roughly 35,000 consults over a year and measured the one thing that matters — minutes from “we need you” to an answer the ED can act on. It is, as far as I can tell, the only detailed breakdown of its kind in the US. The CDC doesn’t record consult specialty nationally. Read that again: the country obsessively measures door-to-doctor time in every ED, but nobody nationally measures consult-to-answer time. We clock how fast a doctor arrives. We don’t clock how fast the answer arrives.

Here is what the study found:

Overall median: 108 minutes. Nearly two hours, every time, just waiting for the answer.

A few things jump out.

First, the psychiatry paradox. Behavioral health is only about 7% of consults — yet it has the longest wait of any specialty: 196 minutes, over three hours. And this is the specialty with the worst boarding crisis in the country: roughly 1 in 10 mental-health ED visits among Medicaid-covered youth ends in boarding for 3–7 days. The patients who wait longest for an answer are the ones who then wait longest for a bed. The delay compounds.

Second, the pattern in the slow list. Cardiology, neurology, neurosurgery, ENT, ophthalmology, orthopedics — the longest waits cluster around exactly the specialties that hospitals struggle most to keep on call. Three out of four ED directors say their specialist on-call coverage is inadequate, and the worst gaps are neurosurgery, hand surgery, orthopedics, ENT, ophthalmology, dermatology. These are the 2am specialties: the ones nobody staffs at night because the volume doesn’t justify a full-time body. So the ED calls, and waits.

Third — and this is the one I keep coming back to — one in five of these waits is a number the system chose not to measure. Nationally. The measurement gap is the story. You manage what you measure, and American emergency medicine has spent decades optimizing everything up to the consult order, then shrugging at the 108 minutes after it.

I’m not pretending the data is perfect. This is one academic hospital where every admission requires a consult, which inflates the internal-medicine share. Other studies put orthopedics as high as 25% and psychiatry as high as 15% of consults. But directionally it holds: a large share of ED visits stall on a specialist’s answer, the answer takes hours, and the slowest answers belong to the hardest-to-staff specialties.

Which brings me back to the geometry question from my last post. Every shortage program tries to move the doctor to where the patient is. But the doctor isn’t the bottleneck in this table — the answer is. A cardiology consult that takes 164 minutes doesn’t need a cardiologist relocated to the county. It needs the question to reach a cardiologist faster and the answer to come back sooner. The specialist stays put. The patient stays put. Only the consult moves.

That’s the shape of what I’m building with FabrixMed — e-consult infrastructure that moves answers, not people, into the places and moments where the wait is longest.

I’m still learning this world. If you run an ED, staff one, or study consult patterns — tell me what these numbers get wrong. The map is better when more people draw it.

References

• Layng et al., “Emergency Department Consultation: A Descriptive Study” — JACEP Open, 2026 (University of Virginia; n≈35,000 consults; median 108 min order-to-actionable-answer) — https://www.jacepopen.com/action/showPdf?pii=S2688-1152%2826%2900151-7
• CDC National Hospital Ambulatory Medical Care Survey, 2018 Emergency Department summary tables (consulting physician seen: 11.6% 2018, 13.5% 2019, 15.7% 2020) — https://www.cdc.gov/nchs/data/nhamcs/web_tables/2018-ed-web-tables-508.pdf
• Rao et al., “The Shortage of On-Call Surgical Specialist Coverage: A National Survey of Emergency Department Directors” — Academic Emergency Medicine, December 2010 (442 ED directors; 74% reported inadequate specialist on-call coverage) — https://www.beckershospitalreview.com/hospital-management-administration/study-finds-shortage-of-on-call-specialists-plaguing-ers/
• McConnell et al., “Psychiatric Emergency Department Boarding for Medicaid-Enrolled Youths” — JAMA Health Forum, August 2025 (research letter; doi:10.1001/jamahealthforum.2025.3177; 255,139 ED visits among Medicaid-enrolled youths; 11.9% resulted in 3–7 days of boarding) — https://doi.org/10.1001/jamahealthforum.2025.3177

I spent some time this week on the Bureau of Health Workforce’s website, reading about how the federal government officially labels the places and people that don’t have enough healthcare. I didn’t know this machinery existed. Now I can’t stop thinking about it.

Here’s the plain-English version.

The government gives shortage labels to places, groups of people, or buildings that don’t have enough healthcare. The count of geographic shortage areas alone is just over 1,000 — and that’s only one kind of label. There are five, and they stack like this:

1. Not enough doctors here. The official name is a Health Professional Shortage Area. It comes in three flavors: a whole area (a county, a neighborhood), a specific group of people (low-income families, homeless people, migrant farmworkers), or a specific building (a prison, a mental hospital). And it covers three kinds of care: basic primary care, dental, and mental health.

2. Not enough maternity care, specifically. Inside the areas above, some places are short on the people who deliver babies and care for mothers. Those get their own sub-label: Maternity Care Target Areas.

3. Underserved — the area or the people. Medically Underserved Areas are whole geographies (a county, a cluster of counties, city blocks) that lack basic care. Medically Underserved Populations are specific groups inside a geography — people facing money, language, or cultural barriers to care. These labels are what let community health centers get built.

4. Doesn’t fit the boxes, but still can’t get care. The Exceptional Medically Underserved Population — for groups whose situation is unusual enough that the standard definitions miss them.

5. The governor can just say so. A state governor can flag an area as short on providers, and that flag is what lets a Rural Health Clinic get established there.

These labels aren’t just statistics. They unlock real things: loan repayment for clinicians who go work in these places, bonus payments, even visa waivers for foreign-trained doctors. The labels are the system formally admitting: care is missing here.

And here’s the part I’m chewing on. Every program attached to these labels tries to move the doctor to the pin on the map. Loan forgiveness, bonuses, visas — all of them are trying to relocate a human being. But doctors are people with lives, and they mostly don’t relocate on command. That’s not a criticism; it’s just how humans work.

What if the geometry is wrong? What if the doctor doesn’t need to move — only the doctor’s answer needs to travel? A question from a clinic in a shortage county, answered by a specialist two states away, in hours instead of months. The specialist stays where she is. The patient stays where he is. Only the consult moves.

And the answer really is the slow part. A 2026 study of roughly 35,000 emergency department consults at the University of Virginia found the median wait from consult order to an actionable answer was 108 minutes — nearly two hours. For psychiatry, it was 196 minutes. Over three hours. Meanwhile, three out of four ED directors say their specialist on-call coverage is inadequate. The bed isn’t always the bottleneck. Sometimes it’s the answer.

That’s the shape of what I’m building with FabrixMed — digital infrastructure that carries specialist answers into the places the map says are missing them. No relocation required.

I’m still learning how these designations work in practice. If you know this world — a state primary care office, a rural clinic, an FQHC — I’d love to hear what the map gets right and what it misses.

References

• U.S. Health Resources and Services Administration, Bureau of Health Workforce — “What Is Shortage Designation?” — https://bhw.hrsa.gov/workforce-shortage-areas/shortage-designation
• HRSA Data Warehouse — look up shortage areas by address: https://data.hrsa.gov
• “Where Are the Nation’s Primary Care Providers? It’s Not an Easy Answer” — MedBound Times (reports HRSA’s count of just over 1,000 geographic shortage areas and a Health Affairs finding on designation effectiveness) — https://www.medboundtimes.com/medicine/where-are-the-nations-primary-care-providers-its-not-an-easy-answer
• Layng et al., “Emergency Department Consultation: A Descriptive Study” — JACEP Open, 2026 (University of Virginia; ~35,000 consults; median 108 min order-to-actionable-answer, 196 min for psychiatry) — https://www.jacepopen.com/action/showPdf?pii=S2688-1152%2826%2900151-7
• Rao et al., “The Shortage of On-Call Surgical Specialist Coverage: A National Survey of Emergency Department Directors” — Academic Emergency Medicine, December 2010 (survey of 442 ED directors; 74% reported inadequate specialist on-call coverage) — https://www.beckershospitalreview.com/hospital-management-administration/study-finds-shortage-of-on-call-specialists-plaguing-ers/

I read a staffing company’s new report on emergency departments under strain. Their numbers are real and worth sitting with. More than 69,000 emergency doctors in the US. Over 100 rural hospitals closed in a decade. Residency spots growing, yet 140 went unfilled this year, double last year’s number. Patients boarding in hallways for hours.

Boarding means this: a patient who has been admitted still sits in the emergency department, waiting for a bed upstairs. The whole department slows down behind them.

The report’s answer is mostly about staffing. More doctors, more flexible coverage, better retention. That makes sense coming from a staffing company, and I am not dismissing it. Short-staffed departments are real.

But I keep coming back to something their lens cannot quite see. At 2am in a small-town ED, the binding constraint is often not another pair of hands. It is a decision that cannot get made.

The ER doctor has a patient with a tricky heart tracing, or a wound that might need a hand surgeon. She knows what she needs: a specialist’s eyes on the case. So she pages, waits, pages again. The patient waits. The bed waits. The whole department waits behind them. That is boarding too, except no staffing plan fixes it, because the specialist she needs does not work in that building and never will.

You cannot staff your way out of that math. There are not enough specialists to put one in every ED at 2am, and there never will be.

What I am learning, building FabrixMed, is that this is an infrastructure problem, not a headcount problem. The question is not how to move more doctors into more buildings. It is how to move the decision to the doctor faster. Get the right eyes on the case in minutes, and the ER doctor can decide: admit, discharge, transfer. The patient moves. The department breathes.

Lay down the rails. Run whatever trains.

Reference: Medicus Healthcare Solutions, “A System Under Strain: Rising Demands in Emergency Medicine”.

Your doctor says, “I’ll refer you to a specialist.” It sounds like a plan. Whether it becomes one depends on something nobody mentions: are both doctors inside the same fortress?

Let me explain the fortresses. A health system is a company that owns hospitals and employs doctors — Baylor Scott & White in Texas is one. An IDN, an integrated delivery network, goes one step further: it owns the hospitals, employs the doctors, and owns the insurance plan. Kaiser Permanente is the classic example — they collect your premium and employ your doctor, so the whole chain is one company. The insurance side has a name too: the payer, the one who foots the bill. In an IDN, the payer and the provider are the same company.

Now the referral. If your primary-care doctor — your PCP, the doctor who owns your overall picture — sends you to a specialist inside the same company, that’s in-network. Outside it, out-of-network. The boundary between those two words is where referrals go to die.

Inside the fortress, things mostly work. One EHR — the electronic health record, usually Epic — so the referral lives in one chart, not a fax. The system gets paid for the specialist visit, so it wants you to complete it. The industry even has words for this: keepage when you stay inside, leakage when you don’t. And there are humans whose whole job is chasing referrals.

But “mostly” is doing heavy lifting. The system can see the open order. It can’t make you schedule. You face a three-month wait and give up. Prior auth — the insurer’s permission slip, required before many visits — gums up the works. And the final step fails everywhere: the specialist writes a note, your PCP gets a twelve-page document in an inbox, nobody reads it, your plan doesn’t change. “Visit happened” is not the same as “loop closed.”

And here’s the thing: most of America doesn’t live in the fortress. Community clinics — FQHCs, federally qualified health centers, the clinics that serve everyone regardless of ability to pay — rural practices, independent doctors: they refer out into the void every day. No shared chart. No keepage incentive. No coordinators. Just a fax and a hope.

So the honest version of the claim: the referral doesn’t die everywhere. It dies at the boundary — and American healthcare is mostly boundaries.

From first principles, then: the unit of value was never the referral. It’s the closed loop — the moment the specialist’s knowledge actually changes your PCP’s plan. Everything before that is potential energy. Build for the loop closing, not the order being sent. Measure completions, not sendings. And if you’re building for the frontier rather than the fortress, you’d better build the plumbing the fortress got for free.