A hospital gets paid roughly the same for an admission whether the patient stays four days or nine. Under DRG payment, the reimbursement is fixed. So every day a medically-ready patient waits in a bed for a nursing-facility slot, a prior authorization, or a home-health start date, the hospital absorbs the cost of that bed with no revenue against it.
Commonly cited cost to a hospital of one avoidable inpatient bed-day. At scale, one state hospital association's analysis of 52 hospitals counted roughly 60,000 delayed discharge days in a single quarter, about $169 million.
The delay days are not evenly spread. They concentrate in patients headed to post-acute care, and the numbers there have been moving the wrong way. American Hospital Association data showed average length of stay up 19.2% between 2019 and 2022, and up around 24% for patients being discharged to post-acute settings. Referral data from that period showed skilled nursing facilities rejecting the majority of first-pass referrals, often near nine in ten, which means case managers fax and phone their way down a list while the patient waits upstairs.
Little's Law says the number of patients in your hospital equals the arrival rate times the average time each one spends there. You mostly can't control arrivals. So census, and everything downstream of census, is a length-of-stay problem, and the cheapest length-of-stay days to remove are the ones where care is already done.
Removing them matters beyond the bed itself, because the inpatient unit is the buffer for the whole building. When beds don't empty, admitted patients board in the emergency department. ED boarding is routinely diagnosed as an ED problem and handed to the ED to fix. It is usually an inventory problem three floors up. That's the classic misdiagnosis an operations lens catches: an hour saved anywhere other than the binding constraint is worth approximately nothing.
Walk the discharge path and label each step as clinical or administrative. Most of the delay is administrative: assembling and sending referral packets, waiting for facility responses, chasing payer authorization, coordinating transport. Those steps have three properties worth noticing. They are measurable from data the hospital already has. They run on documents and phone calls, which modern automation handles well. And nobody owns their end-to-end clock.
That last one is the real finding in most flow work. The delay isn't one department's fault, so it's no department's metric. The hospitals that have attacked it, some with capacity command centers, some with dedicated placement teams and better tooling, started the same way: measure the queue, price the delay per day, and put one owner on the number.
If you run a hospital, three questions tell you whether this is your money on the table. How many bed-days last quarter were patients who were medically ready to leave? What was the average time from discharge order to actual departure for post-acute-bound patients? And who, by name, owns reducing it? If the first two take more than a week to answer, that delay is itself the finding.
Figures: DRG payment structure and avoidable-bed-day cost ranges are standard in hospital finance literature; the 52-hospital, 60,000-day, $169M quarter comes from a state hospital association analysis published in 2023; length-of-stay increases are American Hospital Association reporting for 2019–2022; skilled-nursing first-pass rejection rates come from national discharge-referral platform data for early 2022. Details and sources for any figure on request: ask.