My father-in-law died the day he came home from the hospital. The hospice intake nurse was scheduled to arrive that afternoon.
Hospice is the one part of American healthcare built to own a whole situation rather than a piece of it, and the organization designed to hold his whole situation never held any of it.
This post is about why hospice can do that at all, when nothing in the two or three years leading up to it can. I went in believing the answer was the way hospice gets paid. Payment is part of it and less of it than I thought. Coordination produces nothing you can put on an invoice, which makes it very hard to buy per item, and every Medicare hospice is also under a federal requirement to coordinate. Payment and requirement cannot be separated on the evidence, and the one place they come apart points to the requirement.
What hospice actually has
A hospice patient gets a team, by regulation. A physician, a nurse, a counselor, someone doing the social and practical work. A registered nurse from that group is required to be designated to coordinate the care and keep assessing what the patient and the family need. Somebody answers the phone at three in the morning. And after the patient dies, the hospice is required to keep bereavement services available to the family for up to a year.
None of that exists in the two or three years of decline that usually come first. Same person, same family, same handoffs coming apart in the same places, and no team, no after-hours number, and nobody whose job is the whole of it.
Two things are different about hospice and I cannot separate them. Hospice gets paid a flat amount per day to handle the situation rather than per item. Hospice is also under a federal mandate to have that team. Every Medicare hospice has both at once, and there is no hospice anywhere paid per diem without the mandate, or under the mandate without the per diem. So a post that credits the payment is choosing between two candidates on no evidence, and I am going to come back to that.
Why that changes what is possible
Alchian and Demsetz named this problem precisely in 1972, though they did not answer it the way I want. Working on why firms exist at all, they defined a kind of production where you cannot separate out who contributed what. The output is joint. No portion of it belongs to any one input.
Coordination is the extreme version. Think of a caterer whose invoice runs to eleven line items, none of which corresponds to the four days somebody spent in March finding out that the venue's kitchen ran on a single forty-amp circuit and that two of the warmers on the equipment list would have taken it down together. She swapped equipment before anyone knew there was a question. What she produced was a dinner that happened, which is the same thing the cooks produced and the servers produced. There is no slice of that dinner that is hers.
A fee schedule prices items. Every line on the caterer's invoice is an item you can count and point at. Coordination is not an item, so under a pay-per-item arrangement it gets buried in the price of something else or comes out of somebody's evening.
Here is where I have to concede something to the paper I just cited. Alchian and Demsetz did not conclude that joint production makes coordination unbuyable. They concluded the opposite: somebody specializes in watching the whole thing, and you pay him by handing him whatever is left over after everyone else is paid. Non-separable output is the problem their residual claim was invented to solve. So the strong version of my claim is wrong. You cannot price coordination as an item, and you can pay for it, by the residual or by the hour, since hours are countable even when output is not. What a fee schedule cannot do is make it appear on its own.
The fact that complicates my own argument
Here is where I have to hold the story looser than I would like.
Every hospice is required to run an organized bereavement program and to make those services available to the family for up to a year after the death. That is a condition of participation and it is not optional. Note the verb the regulation uses, which is make available rather than provide; the hospice has to have the program and offer it, not deliver a year of counseling to everyone.
And the statute says, in so many words, that no payment may be made for bereavement counseling.
So the clearest case in American healthcare of somebody being required to own the whole arc, past the patient's death and into the family's year afterward, is also a case of the payer mandating a service and declining to fund it. The cost gets absorbed into a daily rate calculated without it.
That is worse for my argument than a qualification. Bereavement is the one place where payment and requirement come apart. Bereavement has a mandate and zero payment, and it happens anyway. If payment were the operative variable, it would not.
Which leaves me holding a thinner claim than the one I started with. Paying for the situation rather than the item removes an obstacle to coordination. It does not by itself produce it, and the case I know best suggests the requirement is doing more work than the money.
The evidence can run both ways
The natural experiment I leaned on may run the other way. Bereavement is unpaid, but it sits inside an organization already funded per diem and already carrying an interdisciplinary team it was required to assemble for other reasons. Adding a bereavement program to a staffed team is cheap in a way that standing one up from nothing is not. The case may show that the per diem funds the capacity that makes the mandate affordable to satisfy.
That restores payment as the operative variable and takes back what I just conceded. Treating payment and mandate as separable also creates a problem. A condition of participation is enforced by survey and by the loss of Medicare certification, which is a financial instrument wearing a regulatory label. The requirement is a payment consequence attached to a different trigger. I have one observation with which to tell the two apart.
The trade nobody escapes
Paying for the situation solves the measurement problem and creates its own. A daily rate rewards a provider for spending less on a patient exactly as cleanly as a fee schedule rewards one for doing more. Neither form is neutral. Moving between them is a choice about which failure you would rather have to watch for.
I said in a book last year that coordination shows up when somebody is paying for coordination, and stays missing when the only way to fund it is out of an employee's evening. I would put it differently now. Payment form sets what is possible. A requirement is what makes it happen. Hospice has both, and bereavement, which carries the requirement and no payment, happens anyway.
One aside, for anyone reading this in the middle of it
Electing hospice does not mean giving up curative treatment, and it does not mean giving up Medicare. I have watched families believe both and decline hospice over it. The waiver reaches services related to the terminal condition or equivalent to hospice care. Unrelated care stays covered under Parts A and B. And the election can be revoked at any time, by the patient or their representative, during any election period.
Something to look at
Take any invoice your organization sends, and find the work that has no line on it.
Then find out who is doing that work, and when.