Just a tree

There is a black locust at the edge of one of my fields that I have been meaning to deal with for about nine years. The first year it was a thorny little whip you could have pulled out by hand in about four seconds. The second year it needed a tractor. Somewhere around year four it needed a saw, and by then it had company, because locust spreads by root sucker and the parent tree had been busy. A young locust puts on two to four feet a year, and a grove does not expand in a line, it expands in area. It shades a corner of the field, and dealing with it is no longer a Saturday project. It is a quote. Nobody calls it a problem now. It is just a tree.

Nobody decided to let that happen. Every single year there was something more urgent, and every single year the cost of ignoring it was zero.

Part of the landscape. Much like recurring high cost claimants become part of the landscape in a health plan.

Most employers have one or two, and the six figure cost that was once a thorny problem is now absorbed as a budgeted expense. We are going to walk through some of the reasons that tends to happen.

So here is the theory, stated up front. A large recurring claimant blends into the landscape faster than almost anything else on a plan. The risk gets absorbed, the six figure line becomes a budgeted expense, and the harder questions get deferred to a year when there is more time. Meanwhile the renewal effort goes to employee facing plan design, and plan design stops reaching the money long before the money stops growing. What actually moves the trend is purchasing, and purchasing rarely makes the meeting. Part of the reason is the wild hog from From snout to tail. A renewal under pressure is an angry hog staring at you, and when the hog is staring, plan design is the only stick within reach.

The default answer

Let's break this down by ecosystem, because the same behavior has two very different explanations.

In a fully insured arrangement, downgrading coverage is not one option among several. It is the option, other than changing insurers of course. The carrier prices the risk, the employer sees a renewal percentage, and the only variable the employer actually controls is what the plan pays for. Claims data, depending on size and market and a few other things, is often limited if it is available at all. That leaves nothing visible enough to digest, ponder, or explore. Move the deductible, move the copays, move the tiers, and take the credit against the increase. That is not a failure of imagination. It is the shape of the arrangement.

In a self-funded arrangement, plan design should earn far less airtime, because the entire purpose of self-funding is to control cost where it actually accumulates, which is usually somewhere plan design cannot reach. The employer owns the claims data, the network decision, the pharmacy contract, the stop-loss placement, and the care management approach.

The problem for some plans is that they are afraid of the thorns. What if a change disrupts somebody's care, or lands badly with a member, or creates discomfort in a population that did not ask for any of this? Those are fair worries and they are worth sitting with. For a lot of HR leaders, change means loud, and loud means a week of hallway conversations nobody scheduled. The better question is what happens to the plan that everybody else is also part of, and what is actually in the best interest of the participants as a group. That is the question a plan fiduciary is meant to be asking.

Where the money actually is

Before we talk about the lever, it is worth looking at what the lever is aimed at.

Consider this next part a 101. Some of the solutions named here will be new to many readers, and that is the point. Each will get its own article in time. For now the goal is the lay of the land, introducing the concepts rather than arguing them.

Start on the left, because the left is where most of your people are. Roughly 12 percent of covered members generate no claim billed to the plan at all in a given year, and about 44 percent spend under a thousand dollars. Those two groups together are more than half your population and a rounding error in your spend.

Now go right. About 15 percent of members spend ten thousand dollars or more, and the population thins out fast while the dollars do not. And the far end is growing. Million dollar claims are up 46 percent since 2022.

The concentration argument is not new and I have made it here before, in From snout to tail. What is worth doing this time is putting the plan design lever on the same axis and seeing where it lands.

Now let's look at plan design

Here is the same axis, with your cost sharing drawn on it.

Take the average employer plan at 200 or more workers: a $1,670 deductible, 20 percent coinsurance, and a $6,000 out of pocket maximum. Run a claim through it and the member pays the deductible, then a fifth of everything after, until they hit the maximum. At that point they are done, and every additional dollar belongs to the plan.

That point arrives at a claim of $23,320.

Above $23,320, on that plan, cost sharing has nothing left to do. Not less. Nothing. And you can watch the member's share collapse on the way there: about 47 percent of a $5,000 claim, 33 percent of a $10,000 claim, 21 percent of a C-section at $28,998, 6 percent of a $94,823 orthopedic claim, 2.5 percent of a first year of dialysis at $238,126, and six tenths of one percent of a million dollar claim.

Given that, raising the deductible another five hundred dollars is a change to a number that had already stopped mattering.

A side note on benchmarking, since this is usually where the benchmarking report comes out. Benchmarking tells you whether your plan is competitive against the market you hire from, and whether you are over or under built relative to peers. That is a real question and you should ask it. It does not tell you how anybody controls cost. The moves that actually move a plan's spend rarely appear in one at all, partly because they are contractual rather than design based, and partly because they are not comparable across employers. It is a measuring contest, and nobody enters a measuring contest to learn something.

The dilemma

Put the two charts next to each other and the problem is not really arguable.

At one end sit high volume, low dollar claims where plan design reaches perfectly well and there is almost nothing to collect. A hundred dollar X-ray, a hundred and sixty dollar office visit, a three hundred dollar ultrasound. Shift all of it onto members and your trend does not move.

At the other end sits roughly half your spend, above the point where cost sharing has already run out. Dialysis, specialty infusion, the million dollar claim, the six figure orthopedic case.

Stacked in the middle is the part that gets the most attention and deserves the least optimism: maternity at about $20,416 an episode, newborn care, emergency admissions. High dollar, reasonably common, and almost entirely unschedulable. Nobody shops a delivery.

In other words, the lever works precisely where the money is not, and the money sits precisely where the lever does not reach.

The other deductible

Here is the part that usually goes unsaid.

If you are self-funded, you have two deductibles. We have spent this whole piece on the one that maxes out around six thousand dollars. The other one is your specific stop-loss attachment point, and for a group of 200 to 499 employees it is most commonly $100,000, with about 59 percent of that band sitting somewhere between $76,000 and $150,000.

Most buyers I talk to react to a higher attachment point the same way: "I don't want that much risk." That is a completely fair instinct and I am not going to argue with it. But it is an answer to a question nobody asked. The question is not whether you want the risk. The question is what somebody else is charging you to take it, and whether that price is any good.

Here's why that matters. The premium difference between a $100,000 and a $150,000 attachment is a quoted number you can have in an afternoon, and the expected claims in that corridor is something your own history has an opinion about. If the premium you save exceeds what you would expect to pay in the layer, you were renting protection you could have carried yourself.

And the number moves on you even when you leave it alone. As claims inflate, a fixed attachment point catches more claims every year, which is why stop-loss trend runs ahead of medical trend. An attachment set in 2019 is functionally a lower attachment in 2026 without anyone deciding to lower it. This is not a fringe view, at least not explicitly. Sun Life's own 2026 report tells employers to periodically evaluate and consider increasing their deductibles to offset premium impacts, and names deductible strategy and risk forecasting as critical. When the carrier's report is telling you to move the number, the number has probably been sitting still for a while.

To be fair, a well performing plan raising its attachment is genuinely taking on volatility, and one bad year can erase several years of premium savings. The point is not that higher is always better. It is that this should be an annual decision priced against a quote, and for most plans it is not a decision at all.

So every year the room moves the deductible that stops working at twenty three thousand dollars, and leaves the one set at a hundred thousand exactly where it was.

Back to the tree

The reason deductibles, copays, and out of pocket maximums keep getting raised is not that anyone in the room believes it works. It is that it is the move that fits.

It fits in the renewal timeline, which is measured in weeks. It fits in a spreadsheet cell, because you can quantify it before the meeting ends. It fits the authority of the people in the room. Renegotiating a pharmacy contract, moving infusions out of a hospital outpatient department, or repricing your stop-loss layer are all bigger than a renewal cycle and involve people who are not in that meeting.

So the small move gets made, the big ones get deferred, and next year the same thing happens. That is how a weed becomes a tree. Not neglect exactly. Deferral, on an annual schedule, for a decade.

For what it is worth, the moves that do reach are not mysterious. A total knee replacement runs about $21,800 in an ambulatory surgery center, $27,600 in a hospital outpatient department, and $38,300 inpatient, on commercial allowed amounts. That is not a projection, it is what already happened: commercial knee replacements went from 81 percent inpatient in 2019 to 8 percent in 2023. Most of that volume landed in the hospital outpatient department rather than the ambulatory center, so roughly seventy percent of commercial knee replacements are still not at the cheapest site. Infusions in a hospital outpatient department cost the plan about 42 percent more than the same infusion in an ambulatory center, a home, or a physician's office. That is price and not volume: same therapy, matched patients, measured on the day of the infusion. What changes is the facility fee. There was no measured difference in adverse events, or in adherence a year out. Over the following week the hospital outpatient patients were admitted somewhat more often, which the authors could not attribute to the setting. On the day of the infusion, when a complication from the infusion itself would show, admissions did not differ at all. Commercial plans paid $238,126 for a first year of dialysis against Medicare's $80,509 for the same care.

None of those are plan design. All of them are purchasing.

A note on pharmacy

Many of the recurring, year over year, high cost claimants on a plan trace back to pharmacy. It would be easy to spend the back half of this piece there, which is exactly why I am not going to. This writeup stays on the fundamentals.

Pharmacy is getting its own article soon, in depth, with the same kind of interactive tools you have been using here. Until then, Your mileage may vary works through one drug class and has a calculator attached.

What to do

The CFO wants lower costs and HR wants better benefits. Running the tools above against your own plan tends to help both of them sharpen what they are actually asking for.

Four things, all of which can be started this quarter.

1. Run your own ceiling. Put your actual deductible, coinsurance, and out of pocket maximum into the second chart above and find the claim size where your cost sharing stops. Then ask what share of your spend sits above that number.
2. Ask where your joint replacements, infusions, and imaging are being performed, and what the same service costs at an alternative site in your network. Not what it is billed at. What your plan paid.
3. Look at what your own ecosystem actually allows. Fully insured, level funded, self-funded, captive. Which of the levers above are even available to you today, and what would it take to get to the ones that are not?
4. Pull your stop-loss premium at two or three attachment points and set the difference against your own claims history in that corridor.

Fringe Theory is independent and unaffiliated. Views expressed are my own and do not represent those of my employer. Nothing here is legal, tax, medical, or investment advice.