Insurance is a bet against unlikely events
You pay a premium so that if something rare and expensive happens, you're covered. The math works because most people paying in won't need to collect.
Auto insurance is the clean example. You pay your premium, you almost never interact with the insurer, and when you do it's because something genuinely expensive happened. The system kicks in rarely, on things that actually warrant it.
The only issue is that chronic care is not rare and expensive.
Chronic care is not rare
60% of American adults have at least one chronic condition. 51% have two or more. Once you have one, you have about a 70% chance of developing another. Even among 18- to 34-year-olds, 27% already have multiple chronic conditions.
This is just the normal state of American health. Chronic care is recurring and ongoing. It's closer to oil changes and new brakes than it is to a car accident. And unlike a car accident, we can see it coming.
Chronic care isn't shouldn't be expensive
The question isn't whether chronic care is expensive right now. It obviously is. The question is why, and the answer starts with what insurance was actually designed to cover.
Think about the distribution of healthcare events by severity. At one end, a $200,000 hospitalization. At the other, a $35 blood draw. Insurance was built for the expensive end. But routine encounters don't just happen more often than serious ones — they happen exponentially more often, because healthcare severity follows a skewed distribution with a long, sparse high-acuity tail and an enormous low-acuity bulk. The further coverage expands toward routine care, the faster the insured pool grows, and the more the administrative machinery costs relative to the care it's actually delivering.
This model is illustrative, and the dollar figures below are illustrative.
through insurance
system-wide
routed baseline
The industry's response for decades has been whack-a-mole. Wherever the overhead pops up unnecessarily, hire engineers to try to automate it away. Prior auth software, utilization management tools, AI for claims. The irony here is the cost of the engineering teams gets layered into the insurance overhead, and it certainly doesn't go away even if the solution manages to improve efficiency.
The real answer is to stop routing routine care through that system at all. Direct primary care practices do exactly this. They charge a flat monthly fee, take no insurance, and just treat patients. The results are hard to argue with.
DPC is affordable because the overhead is gone.
No claims to file, no prior authorizations to fight, no billing department to staff. The cost of the administrative layer between patient and clinician drops to nearly zero, and prices reflect that.
DPC is effective because doctors have room to actually practice medicine.
Without the pressure of a billing code on every decision, clinicians can follow a patient over time, build on what they already know, and intervene early. Remove per-visit billing and patients show up more: traditional primary care averages 1.5 visits per year; one peer-reviewed study of a DPC practice found patients averaged nearly seven. Continuity research is consistent: regular care with the same clinician is associated with lower costs, reduced acute utilization, and better population-level outcomes.
The problem is that DPC can't get us all the way to scalable personalized care. One clinician can only carry so many patients. You can open more practices, but each one starts over. The model replicates, it doesn't compound. But that's an issue for another post—check out The Art of the State for more on that.
The math was never going to work
Insurance was built for rare, expensive events. Chronic care is common, recurring, and manageable. Routing one through the other doesn't just create overhead—it creates the wrong incentives at every layer of the system. The administrative machinery grows. The distance between patient and clinician grows.
The objection worth addressing is that chronic conditions do spiral into acute episodes, and acute episodes are expensive. That's true, and those are exactly the events insurance was designed for. But the research is consistent: most of those hospitalizations are preventable with appropriate ongoing management. The acute cost is largely a downstream consequence of inadequate chronic care.
Direct care models prove the alternative works. Remove the intermediary, align incentives, give clinicians the time and context to actually practice medicine, and the results follow.
But all of this assumes insurance is functioning as designed. That the bet is being priced correctly, that the risk pool is sound, that the machinery, however inefficient, is at least doing what it promised.
That assumption is getting harder to defend. In the next piece, we look at where the insurance model itself is breaking down, and how the millions of Americans living with chronic conditions are responding.