Always at Capacity
How primary care became the only high-stakes service operating at >100% by design.
When I began as an attending primary care doctor, I knew my practice would be large — around 1,500 patients. But I was surprised to learn that I’d need to fill to 110% of that number, because of expected attrition.
In practice, I reached 160% before I slowly but surely got back “down” to 110% — not through attrition, but through an active effort of redistribution when a new doctor joined our group.
This is the problem with primary care.
There is a reason it’s standard to over-panel doctors. There are perverse incentives to keep every minute of our day filled with patient visits and every appointment slot occupied.
It’s an old adage in systems engineering that every system is perfectly designed to get the result that it does. In primary care, waiting months to get an appointment is not an unfortunate byproduct, but a direct and entirely predictable result of the structures in place.
How much of primary care even needs to be visit-based vs. asynchronous (such as calling or messaging) will be the topic of another post. This one is focused on visits themselves. There is a profound mismatch: The worst-case scenario from an administrator’s perspective is an empty slot. The worst-case scenario from a patient’s perspective is no empty slots.
To satisfy the former, primary care practices run at or above capacity by design.
Fill Every Slot
Traditional primary care in most organizations is reimbursed one way: by the visit. From a health care administrator’s standpoint, the formula for revenue is simple. The more patient visits per day, the more the organization makes. And because primary care services such as prevention visits are generally reimbursed less than procedures or specialist visits, administrators will often make the case that a certain high number of visits per day are needed just to stay afloat.
As such, controlling the primary care purse strings translates into one goal — fill every slot. An empty appointment is money left on the table. This is why primary care runs like an assembly line, our days filled with back-to-back patient visits. Time spent on anything else — reviewing charts, writing notes, answering patient messages, reviewing test results, placing new orders, communicating with other team members, refilling medications, calling insurance companies — is mainly absorbed into doctors’ off hours.
This is also why doctors’ schedules are filled months in advance. To ensure every slot is filled, most primary care practices operate via a straightforward system: Appointment slots open, and they fill on a first-come, first-served basis. The reason for the appointment is largely irrelevant. It goes to whoever calls or clicks the fastest.
The fill every slot approach, where all care is flattened into visits of identical length, scheduled for whenever by whomever, is obviously not ideal primary care. I may see someone for an appointment before they get their blood drawn, only to have them come back to review abnormal results and draw more. I may evaluate a patient’s pain over a video visit and decide they need an in-person evaluation with an ECG and cardiac monitor. In this system, what happens during a visit doesn’t matter. Resolution of issues doesn’t matter. A packed schedule creates the illusion of care without necessarily delivering it.
Good primary care is not a free-for-all of interchangeable visits. Good primary care is getting people to their doctor at the right time.
“Fill every slot” also incentivizes problematic workarounds. My patients often ask me for tricks to get appointments. I wish I had them. So they do what anyone would do in their situation — they send me messages, which spill into my off hours. Some book more than one appointment — say 2, 4 and 5 months in advance — just in case they need it, further blocking access for others. A handful who can afford it leave for concierge or direct primary care.
This system is generally disliked by doctors as much as by patients. After all, we’re the patient-facing ones; we’re the ones who bear the fallout when a sick day becomes a catastrophe because everything was set up four months ago, and cancelling means pushing another few months. But it’s mostly accepted. We absorb endless messages. We develop our own workarounds. We apologize for delays outside our control.
Primary care is a textbook, real-life example of queueing theory: As utilization approaches 100%, average wait times increase dramatically. The most common understanding of long primary care waits is a doctor shortage. But that explanation is very incomplete. Primary care is perfectly manufactured to have impossibly long waits. It is a mathematical certainty that when each doctor runs at or above 100% capacity, there will be waits.
Hiring more doctors into the same system will produce exactly the same outcomes. When the problem isn’t scarcity, but artificial scarcity, any new hire quickly fills and overfills — and simply gets booked months out, too.
No Consequences
Primary care is hardly the only industry financially motivated to squeeze as many through the assembly line as possible. Why don’t other industries face the same problem of impossibly long waits?
Restaurants, airlines, and really any service industry face similar pressures to fill: Fill the tables. Fill the seats. The difference is that there is direct and immediate negative feedback if the wait is beyond reasonable to receive the service. For example, I often stop at Starbucks on my drive to clinic each morning. If my morning coffee arrived the next day, or a 6 PM dinner reservation seated me at midnight, I’d stop going. I wouldn’t be alone. The market would impose its own consequences, quickly and decisively.
Thus, places like restaurants and coffee shops are incentivized to use timely access as a metric critical to their success — as important as the quality of the food and ambience. So they keep some slots open and don’t routinely overbook. Their formula optimizes throughput while also maintaining access.
By contrast, there are no market forces penalizing delays in primary care. Everyone needs a primary doctor. So they’ll wait for one.
And yet, other parts of health care don’t overbook by design. Emergency rooms and hospitals operate with a magic number of around 85%. This is considered a critical threshold; if more patients need care, surge backup comes into play. When I worked on the hospital side, I experienced this regularly. We predicted surges. We had good days and bad days. And when patient volume became too high to manage, the system worked hard to adapt.
There is no equivalent magic number in primary care.
But why? Do hospitals and ERs just do what’s right for patients regardless of financial incentives? Perhaps, and I sure hope that’s part of it. But it’s also because there is a negative feedback loop here, too.
If the wait to treat a stroke requiring blood thinners within six hours stretched to four days, patients would die or suffer permanent disability. The consequences would be immediately catastrophic, highly visible, and entirely unacceptable.
That primary care runs differently shows a fundamental misunderstanding of the value of primary care. The consequences of delayed primary care are just as damning; they’re simply more distant, less linear, and thus hard to trace back. A young man with slowly worsening anemia waits a month to see his doctor, and a colonoscopy and scans ultimately show stage IV colon cancer. Was the wait to blame, or is this just the devastating reality of cancer? A man messages about a “rash” on his leg and waits patiently to see his doctor, who diagnoses cellulitis and sends him to the emergency room for bacteria that spread into his bloodstream. He develops an acute kidney injury, which persists as chronic kidney disease. Later, kidney disease is a footnote in his chart — treated as an inevitable fact of this man’s medical story, rather than a complication that timely access to his primary doctor could have prevented.
This is how primary care became the only high-stakes service operating at or above capacity by design. It’s precisely because those stakes are underestimated. From a health care administrator’s perspective, there are no consequences. The patients keep coming back. The medical care is the same.
So why fix it?
The Limits of “Advanced Access”
On the surface, there is a solution that fulfills both administrators’ incentives to fill every slot and patients’ need to access them. It’s called “advanced access,” and primary care researchers have been making the case for it since the 1990s.
The idea is simple. Instead of booking out months in advance, keep most appointments open for same-day scheduling. (There are variations of this with different time intervals such as same week visits.)
A JAMA paper described how such a system was implemented at Kaiser and a few other places with remarkable results. Murray and Tantau wrote in the year 2000:
As it turns out, demand is not insatiable, as many practices believe…. When practices enact such a system, several magical things occur:
First, the wait time for a routine appointment is today. No one can beat that.
Second, practices no longer have to hold appointments in anticipation of same-day needs, so they’ve maximized their schedules and gained capacity (or appointment availability) they didn’t have before.
Third, the likelihood that patients will see their own personal physician has increased, which means greater efficiency, a greater sense of control for the physicians and improved satisfaction for everyone.
This approach is directionally right. But in modern primary care, it only partially caught on and therefore hasn’t quite worked out.
The first problem is cannibalization. While my group, for instance, opened a handful of same-day slots, they were the first to go. Because these slots were rare gems and not the norm, they ended up becoming a workaround for desperate doctors to overbook a patient into. I admit I did this routinely. Say I got a scan back showing a recurrence of cancer on a long-time patient. Obviously I wanted to discuss that result, and obviously it needed to be soon. If my schedule had a wide open “same-day” slot next week, I filled it.
The second is continuity. The authors advise that when a patient calls in, they should first be asked “Who’s your doctor?” rather than “What’s your issue?” This doesn’t happen in practice. Same-day slots, when they exist, mostly become an interchangeable stopgap to see other doctors’ patients, even as we all struggle to see our own. Hiring dedicated same-day providers, as my practice did, creates a mini parallel urgent care clinic, but doesn’t advance primary care. Unless these slots are in primary doctors’ own schedules and prioritize their own patients, they exchange access for continuity — when good primary care is both.
The final is the most problematic, and it goes back to the 1,500 (or more) patient cap. Based on the number of slots I had per year, I calculated that my cap allotted two visits per patient per year. But this number underestimates reality. Many of my chronically ill patients require more. As authors Murray and Berwick wrote:
“Advanced access is not sustainable if patient demand for appointments is permanently greater than physician capacity to offer appointments.”
Triaging appointments away from a free-for-all is a bare minimum — a necessary but not sufficient solution. It’s the floor on top of which any other rewiring of doctors’ time can be implemented. If our patients’ need for visits is greater than visits available, no amount of rearranging how they’re filled will reduce waits.
Taking the Leap
It may seem unsavory to speak of care for people in stark economic terms. But unfortunately that is how medicine runs — as a business — meaning a business case must be made that long waits are problematic to change any of the underlying structures driving them. As a primary care doctor, I am rewarded on many metrics: flu vaccinations, mammograms, diabetes control. Waits are not one of them. My patients being able to access me, their doctor, when they need to, is not captured in any metric.
It would be wrongheaded to hold overworked primary care doctors accountable for yet another outcome we cannot control. It would be worse to think up hacks that make the problem worse, such as shortening visit lengths. But until the consequences of delayed primary care are more visible — that incentivize access to one’s own doctor, or dis-incentivize waits — I’m not sure health care administrators will just override existing incentives and do what’s optimal for patients.
This requires a leap of faith: that the solution is not squeezing more in, but less. Even within our current payment model, medical leadership can work with administrators to make the case for leaving slots open and reducing total patients per doctor. In the short-term, they will balk: What if the visits don’t fill? How can we leave money on the table? But they will fill. They’ll always fill. And care will be infinitely better when they fill based on medical need rather than who clicks fastest. Fixing this is not only a matter of queueing theory, but of courage.
More on the “fake math” of primary care: here and the Hard Medicine podcast, Episode 1.