Dispatches from Discharge Hell: A 25-Part Series on the Patterns Nobody Warns You About
Every person in the discharge ecosystem is doing their job correctly.
The insurance company denies coverage when their reviewer decides meaningful recovery has plateaued — because the utilization model says the patient should be done. The SNF refuses the complex patient — because Medicare reimbursement won't cover the nursing hours. The DME vendor works within a 6-week custom fabrication timeline — because ATP evaluations, fabrication, and fittings can't be compressed out of existence. The hospital pushes for faster discharge — because the bed generates revenue only when it turns over.
Everyone is responding rationally to their incentives.
The patient is the only one whose incentives aren't in the model.
By my tenth month in catastrophic care, I stopped being surprised by individual failures and started seeing the architecture. Not a conspiracy. Something worse. A system where every actor's rational behavior produces irrational outcomes for the person the system ostensibly exists to serve.
Health economists call this the principal-agent problem: when the person making decisions (the agent) has different incentives than the person affected by those decisions (the principal). In healthcare, the principal-agent chains are so long and so tangled that nobody can trace who's actually working for the patient.
The hospital is the patient's advocate, except when census pressure makes the patient's extended stay a financial liability. The insurance company manages the patient's benefits, except when those benefits are managed by denying the care the patient needs. The SNF provides post-acute recovery, except when the patient's complexity makes them unprofitable.
A 2018 study in Health Affairs documented that post-acute care spending accounts for 73% of the variation in Medicare spending across regions, and that this variation is driven primarily by institutional incentives, not patient needs. The patients in high-spending regions aren't sicker. The systems around them just extract more.
One Patient's Discharge Chain
Let me walk you through one patient's discharge and show you every incentive at play.
The hospital admitted a patient with a severe TBI, trach, g-tube, and behavioral dysregulation. The admission was clinically appropriate. The CMI goes up. The reputation is reinforced. But 25 days in, the patient can't discharge because no facility will accept the complexity. Every additional day is unfunded. The hospital now has an incentive to discharge the patient to anywhere rather than the right place.
The insurance company authorized 21 days based on a standard length-of-stay model. The patient made progress but not enough to meet discharge criteria. The payer denies continued stay. The P2P happens. The P2P is not overturned. An appeal exists, but at a program like TIRR, that appeal belongs to the family. They were told the complexity at admission, they are informed of the denial, and the decision to file is theirs. The appeal does not keep the patient inpatient. A pending family appeal is not a clinical justification for continued stay, and the hospital carries real liability exposure for every unfunded inpatient day it cannot clinically defend. The only exception is when the medical picture requires it: when the patient's needs are complex enough to demand hospital-level monitoring and the team has no safe alternative. Then the clinical reality overrides the financial one, and the hospital holds the patient because there is no other option.
The SNF receives the referral packet. Complex patient. High nursing needs. Medicare rate. The SNF's incentive is to accept patients who require standard care at standard reimbursement. This patient would require 1:1 staffing that the reimbursement doesn't cover. The SNF declines. Rationally.
The DME vendor has the wheelchair order. Custom seating takes 6 weeks — not because the vendor is dragging their feet, but because the process requires an ATP evaluation, custom fabrication, and a fitting before delivery. Vendors who work with a high-volume catastrophic care program generally stay aligned with discharge timelines. They don't want to lose the referral stream. The system's timeline is the constraint, not the vendor's motivation. When delays happen, they're usually measured in days, not weeks.
The home health agency gets the referral. Trach patient. Behavioral episodes. Remote zip code. The agency's incentive is to accept patients who can be efficiently served in geographic clusters. This patient is a 45-minute drive each direction for a 45-minute visit. The agency declines.
Every single actor made a rational decision. The patient has nowhere to go.
Misaligned Incentives
The case manager and social worker are the only people in this chain whose incentives are aligned with the patient, because their job is literally to get the patient safely to the next setting. But they have no authority over any of the actors whose cooperation is required.
We can't make the SNF accept. We can't make the vendor expedite. We can't make the payer authorize. We can't make the home health agency staff the case. We can only call, document, follow up, escalate, and absorb the frustration of a system that works exactly as designed. Just not for the patient.
The economist would say the incentives need realignment. The administrator would say the system is working within constraints. The policymaker would say reform is complex.
The case manager says: it's 4 PM, the patient has nowhere to go, and administration is asking what the plan is.
There's a wrinkle in catastrophic care that complicates the whole picture.
In a routine IRF case, system inefficiency is pure friction. Every delay hurts the patient. But in catastrophic care — DOC patients, high cervical SCIs, severe TBIs with no viable discharge destination — the family is rarely the one demanding speed. They want more time. More therapy. More days before a nursing home becomes unavoidable.
The SNF that declines the referral? That's another week at the IRF. Every facility that says no is another day the patient stays, not because anyone authorized it, but because there is nowhere safe to send them and the medical complexity leaves the team no choice. In those cases, the family that wanted more time gets it. Not by design. Not through an appeal. Because the system ran out of options and the clinical picture made discharge impossible.
But that's only one version. There's another: the family came in expecting the system to place the patient. They didn't know what was coming. Most families don't. They arrive with high expectations and real hope, and nobody has told them yet what "complex discharge" actually means. The trach isn't the problem. Programs like TIRR train families for home ventilator management. That part is solvable. What isn't solvable is when every potential receiving facility looks at the clinical picture and declines — too medically complex for the SNF rate, too behaviorally unpredictable for a standard post-acute unit, too resource-intensive for anyone's reimbursement model. The family isn't the obstacle. The patient has simply exceeded what the downstream system is built to absorb.
The system's dysfunction is sometimes a lifeline. Sometimes it's just a wall. The case manager navigates both: one family desperate to stay, and another desperate to get home with nowhere left to send them.
Same misaligned incentives. Completely different consequences.
I used to think individual bad actors caused bad outcomes. Now I know better. The system doesn't need bad actors. It just needs misaligned incentives and enough distance between the decision-maker and the consequence to make the math feel abstract.
The insurance reviewer who denies coverage will never meet the patient. The SNF administrator who declines the referral will never see the family's face. The home health agency that turns down the referral will never watch the family absorb what that decision means.
The case manager sees all of it. Every face. Every consequence. Every gap between what the system promises and what it delivers.
That's not a job description. That's a front-row seat to a system that works perfectly for everyone except the person it's supposed to serve.
Next in the series: Part 11, "It Depends on Who Shows Up." When the difference between a safe discharge and a crisis is whether the family member answers the phone.