In the last post, The Standardization Tax, Part 2: Why We Need Standardization, I talked about humility as being willing to recognize when following a checklist over your own judgment improves care. Humility has two other dimensions in medicine: one facing the patient and the other facing physicians and clinicians.
Understanding those two additional dimensions helps us see that not every instance of variability in care is a mistake. Sometimes, what seems inconsistent from the outside supports something essential. There are at least two types of these ‘load-bearing walls,’ and each can get destroyed with unintended consequences.
The First Load-Bearing Wall: What the Patient Wants
John Wennberg spent decades studying why healthcare can look so different from place to place. He found variations that shouldn’t exist. But he also found that sometimes, the right answer to a medical question depends on the patient, not just the disease. He called this preference-sensitive care and argued that the variation it creates is justified, not wasteful.
Wennberg explained this with two examples: breast cancer and prostate cancer.
For the first example, imagine you are the doctor and your patient has early-stage breast cancer. You can offer her a lumpectomy, which removes the tumor but keeps the breast, usually followed by radiation. Or you can offer a mastectomy, which removes the whole breast and often does not require radiation afterward. The survival rates are so similar that neither option is clearly better.
What matters most is what she is willing to live with: the follow-up treatments, how she feels about her body, and whether she feels more comfortable removing more tissue or keeping more of it. Two women with the same diagnosis, in the same exam room, minutes apart, might make different but equally good choices. If you walk in already certain about which option she should pick, you have missed the most important part of your job.
For the second example, picture a man with early-stage prostate cancer. Surgery, different types of radiation, or watchful waiting are all reasonable options. Each has its own risks, such as incontinence, impotence, anxiety about living with untreated cancer, or the cost and side effects of treating something that might never have caused problems. There is no single right answer.
If a nurse skips a hand-washing step, there’s no way to defend it. But if two women with the same diagnosis make different equally good choices about their care, neither is wrong. That kind of variation means the system is working as it should. Calling this ‘unnecessary variation’ and trying to standardize it takes away choice. It shifts a decision that should belong to the patient to whoever wrote the protocol.
Wennberg’s solution for preference-sensitive care is shared decision-making: a real conversation in which the doctor explains all the options, even the complicated ones, and the patient’s values guide the choice.
The secret of the care of the patient is in caring for the patient.
— Francis W. Peabody, address to Harvard Medical School students, published in JAMA (1927)
That process takes time. It means spending enough time with a patient to truly understand what they fear and what they are willing to trade for it. This is the opposite of how most doctors are paid and scheduled today. When we do not make time for these conversations, someone ends up paying for the shortcut.
For our pre-med student from the previous posts, I would ask her to notice that the hardest part of protecting this load-bearing wall is not the medicine. Explaining either treatment does not take special skill. The real challenge is being willing to wait for the full answer after you have asked. That does not get easier with more training. In fact, the more you know, the more tempting it is to skip asking.
The Second Load-Bearing Wall: What the Doctor Knows
At the start of this series, I mentioned my years as a rural psychiatrist. The next nearest psychiatric ward was 70 miles away. So when my outpatients needed hospital care, I admitted them myself. I checked on them in the hospital in the mornings and saw my outpatients in the afternoons. When a patient left the hospital, they often came to my office a few days later. I already knew what treatment they got in the hospital — because I was the one who had treated them there.
This setup used to be the rule — not the exception — for doctors. Primary care doctors often checked on their own hospitalized patients before going to their clinics. The doctor who treated you in the hospital was often the same one who saw you afterward.
Then medicine began to remove the variability of this blended practice model. Hospital medicine became its own specialty. Hospitalists now work in shifts, usually a week on and a week off. They do not have clinics, and they may never see a patient they treat again. They never know how patients did on the treatments they were on at discharge. Outpatient doctors also never see their own patients when they need hospital care. Now, rarely does a patient have the same doctor treating them across settings.
This change makes doctors’ schedules more predictable, makes staffing easier, and solves a real problem. When a doctor did both hospital care and clinic, they could not predict how many hospitalized patients they would have on a given day. They might set aside two hours for hospital rounds, but some days, they could end up needing three. Then there was the travel between the hospital and the clinic. All of this made doctors much less efficient in the old setup. If a doctor stayed in the hospital longer than expected, it delayed their clinic. That wasn’t great for the patients who had to wait.
I understand why we moved to the hospitalist model, but we must not ignore what we lost in the change.
These days, I see patient care delivery through a very different lens. I review cases for an insurance company. When I talk to hospitalists about their patients, it’s not uncommon to hear some versions of the following statements:
- “I just came on the service two days ago. I don’t know why they treated this patient with a low dose of this medication for the week before that.”
- “I don’t know why they diagnosed him with neuroleptic malignant syndrome. I wouldn’t have given that diagnosis, because there’s no documentation of fever or muscle rigidity.”
- “I’m not sure what their social situation is. The social worker handles that.”
These doctors aren’t bad doctors. They are doing exactly what the system expects: picking up where the chart ends, managing what is in front of them, and handing off smoothly when their shift is over.
But notice what is happening. The information is not missing because someone was careless. It is missing because it used to live in a doctor’s memory of a patient, built over months or years of inpatient and outpatient care. And now it relies on a tenuous handoff instead.
Vital signs and medication lists make it through a handoff. But the reasoning behind a diagnosis, the decision to start with a low dose, or the patient’s family situation are the kinds of knowledge that quietly disappear when no one person is responsible for carrying them forward.
By far the most frequently used drug in general practice was the doctor himself.
— Michael Balint, The Doctor, His Patient and the Illness (1957)
If the pre-med student I mentioned in the first post in this series, The Standardization Tax: Not Every Tradeoff Is Worth It, hears me explain this to a colleague, she would probably ask, “Why doesn’t someone just write that down?”
The honest answer is that some of it gets written down sometimes, if there is time. The gap between ‘some of it, sometimes’ and ‘all of it, always’ is exactly what this load-bearing wall is made of.
You cannot tell if a wall is load-bearing just by looking at it. You only find out when someone knocks it down, expecting an easy fix, and the ceiling collapses. The knowledge of what was load-bearing was always there, but no one checked the blueprint before someone started tearing it down.
Hospitalist care does not always lead to worse mortality, but mortality is not the only yardstick of quality. Broader research on continuity of care shows that most measured outcomes are better, and costs are usually lower, when the same clinician stays with a patient over time. Fragmented care means more complications and higher costs.
The variation we removed by turning hospital coverage into weekly rotations made the system more efficient, and it deprived the patient of a healer who knew them across clinical settings.
The Patient Pays the Price
Someone pays for what is lost when you take down a wall without knowing it was load-bearing.
Reducing variation by not scheduling enough time for shared decision-making, when preference-sensitive care is the right approach, does not improve quality. Everyone pays a price, but the patient suffers the most.
Reducing variation by choosing consistency over continuity also makes the patient pay the highest price. I believe that this choice has also contributed to burnout among physicians, but that’s for another post.
The One Practice
Pick one of these two questions, whichever fits your work better.
If you make treatment decisions with patients, think of a recent case where more than one reasonable path existed. Did your process allow you to find out what the patient actually wanted, or did it push you toward whichever option was faster and less complicated to land?
If you work across handoffs or rotations, think of a recent handoff. What did the handoff assume the next person would just figure out on their own? How often does that assumption prove wrong?
This post is Part 3 of a four-part series in which I explore where standardization helps in medicine, where it costs us something important that cannot be measured, and how patients and doctors pay the price that I call the Standardization Tax. The next post in the series will be published in one week.
