For five years, American hospitals running hospital-at-home programs lived on short extensions. The Medicare waiver that allowed them to be paid for inpatient-level care delivered in a patient’s bedroom was born as a pandemic emergency measure in late 2020, and it kept coming within weeks of expiring. Building a durable service on a waiver that might vanish at the next budget deadline was, to put it mildly, a hard sell to a hospital board.
That changed in February. The Consolidated Appropriations Act, 2026, signed on February 3, extended the Acute Hospital Care at Home waiver through September 30, 2030. For the first time, a US health system can plan a hospital-at-home service on a horizon longer than a fiscal year.
That is good news. It also removes the excuse. For five years, modest programs could be explained by regulatory uncertainty. From here on, a small program is a strategic choice, and the real constraints come into view. Here is the argument: the limit on hospital at home is not regulation or technology. It is who does the unpaid work, and how honestly programs select their patients.
The evidence is real, and narrow
The best-known trial is strong. In 2020, David Levine and colleagues at Brigham and Women’s Hospital published a randomized controlled trial in the Annals of Internal Medicine. Adults admitted through the emergency department with selected acute conditions were randomized to home hospital care or usual hospital care. The home group had lower costs, less sedentary time and far fewer 30-day readmissions: 7 percent versus 23 percent.
Those are striking results. They also came from 91 patients.
That is not a criticism of the trial, which was well designed. It is a caution about generalization. Participants were screened: they had to have a suitable home, specific diagnoses and a level of acuity that fit the model. The trial showed that for the right patient, home is at least as good as a ward. It did not show that most inpatients are the right patient.
In practice, the share of admissions eligible for hospital at home in any given hospital tends to be a small slice of total volume. Programs that report excellent outcomes are often reporting excellent selection. That is fine, as long as nobody mistakes a well-chosen minority for a new model of inpatient care.
There is a subtler point buried in the readmission figure. A home program sees the patient in the place where the patient will live after discharge. The team knows whether the stairs are a problem, whether the fridge has food, whether the pill bottles are organised. Some of the benefit may come less from where the acute care happens than from the fact that the team learns the home. If that is true, the lesson is portable: hospitals could capture part of it with better discharge home visits, without running a full inpatient service in a living room. Programs should want to know which of those two things is doing the work, because one is much cheaper than the other.
The hidden workforce in the spare bedroom
Here is the uncomfortable part. A hospital ward has staff around the clock. A home hospital has scheduled visits, remote monitoring and a phone line. The gaps between visits are filled by someone. Usually that someone is a spouse, an adult child or a neighbour.
Family caregivers help patients to the bathroom, notice confusion at 3 a.m., manage meals and keep track of medicines. That work is real, skilled and unpaid. When a program reports that it delivered care more cheaply than the hospital, part of the saving is labour moved from a paid nurse to an unpaid relative.
It helps to put rough numbers on it. Take a hypothetical four-day admission. Suppose the program provides two in-person visits a day of about an hour each, plus video check-ins. That leaves roughly 22 hours a day in which nobody from the hospital is physically present. Even if a relative is only “on call” for half of those hours, that is around 40 to 45 hours of attention across the stay, which is more than a standard working week. Nobody is sending an invoice for it. If you priced those hours at a home health aide’s wage, the program’s cost advantage would shrink, and in some cases it might disappear. The exact figure will vary by program and patient, and that is the point: almost no one is measuring it.
This does not make hospital at home bad. Many patients and families strongly prefer it, and qualitative research around the Levine trial found patients valued being in their own space. A daughter who sits with her father at home may well consider that time better spent than driving to a hospital car park every evening. But it does mean two things.
First, cost comparisons should say out loud what they exclude. Second, eligibility quietly favours patients who have a capable person at home and a home that can accommodate care. Older adults who live alone, people in crowded or unstable housing and people whose relatives work shifts are less likely to qualify. A model that works best for the well-housed and well-supported can widen gaps in who gets the more comfortable version of care.
The economics hospitals whisper about
The Medicare waiver pays hospital-at-home admissions largely as it pays inpatient admissions. That sounds generous, and for some hospitals it is. But the cost structure of a home program is different. It needs nurses and paramedics who travel, logistics for equipment, courier services for medicines and labs, and a command centre. At low volume, those fixed costs are spread thin.
Think about the command centre alone. A physician or advanced practitioner and a nurse have to be reachable every hour of every day, whether the program has three patients or thirty. At three patients, that staffing is a luxury. At thirty, it starts to look like a ward. Travel is the other hidden variable: a nurse who can see six patients a shift on a hospital floor may see three in a home program once driving, parking and a flight of stairs are counted. Density matters more than technology. An urban program with patients ten minutes apart has a fundamentally different cost curve from a rural one where each visit eats an hour of road.
The strongest business case is often not the per-case margin at all. It is capacity. A hospital that is chronically full, boarding patients in emergency department corridors, can treat home beds as extra beds it did not have to build. Each patient moved home frees a physical bed for a surgical case or a sicker admission, and the revenue from that displaced case is where the money really is. For a hospital with empty wards, the same program may simply move patients from a paid-for bed to a more complicated one.
That is why hospital at home has grown fastest in systems with capacity pressure, and why it may never make sense for others.
The strongest case against my argument
A fair critic would say this: every model of care leans on families. Patients discharged after a standard admission also go home to relatives who manage their medicines and watch for trouble. Hospital at home does not invent unpaid care; it simply makes it visible earlier. And selection is not a scandal. Every service line selects. Nobody accuses a day-surgery unit of cherry-picking because it does not do heart transplants.
That critique partly holds. The comparison with ordinary discharge is fair, and it would be wrong to hold hospital at home to a standard of zero family involvement that no part of medicine meets. Selection, too, is a feature of good clinical design, not a flaw.
Where it stops holding is in the claims. Day-surgery units do not claim to be a replacement for the hospital. Hospital at home is often marketed that way, and its savings are often quoted as if they would scale to the whole inpatient population. The difference between “a good option for a defined group” and “the future of the hospital” matters for policy, because Congress is going to decide in 2030 which one it is paying for.
Second-order effects nobody is planning for
If programs grow, several knock-on effects follow. Home health agencies and hospital-at-home programs will compete for the same nurses, and in many regions home health is already short-staffed. A hospital that recruits a community nurse to its home program may be pulling her from the post-acute service that its own discharged patients rely on.
Employers will feel it too. If more acute care happens at home, more working-age adults will take unplanned leave to cover the nights. That cost lands on payrolls and personal finances rather than hospital budgets, which is exactly why it will be ignored unless someone insists on counting it.
And the home itself becomes clinical infrastructure. Broadband, a reliable power supply and enough space for equipment become eligibility criteria. That quietly turns housing quality into a determinant of access to acute care, in a way few health policies have done before.
What the five-year window should be used for
With the waiver secure until 2030, the useful work is no longer lobbying for extensions. It is building the evidence that Congress and CMS will need to decide what comes after, and each group has a part to play.
Hospital chief executives should stop asking whether hospital at home “pays” in the abstract and ask a narrower question: are we capacity-constrained, and how many beds’ worth of patients could we safely move? If the honest answer is “not many” or “we have empty beds,” a small or no program may be the right decision, and there is no shame in it.
Program leaders should report eligibility, not just outcomes. What share of admissions were screened, what share were eligible, and why were the others excluded? Without that denominator, outcome data cannot be interpreted. They should also measure caregiver burden directly by asking families what the admission cost them in time, missed work and stress. If the model depends on them, their experience is an outcome.
Payers, including Medicare Advantage plans and commercial insurers watching from the sidelines, should ask for the same denominators before building their own contracts. A plan that pays a home admission at inpatient rates without knowing the selection criteria is buying a result it cannot interpret.
Researchers and funders have the clearest opening. Programs can deliberately design support for patients who live alone, such as extra visits or paid overnight aides, and test whether outcomes hold. That is the experiment that tells us whether the model can reach beyond the well-supported, and it is cheap relative to what is at stake. They should also be honest about the counterfactual: savings measured against a full hospital are different from savings measured against an empty one.
Clinicians, finally, are the people who make selection decisions at the bedside. They are well placed to document why a patient was declined, and that record is the raw material for any honest equity analysis.
What to watch before 2028
The next two years will show whether the extension produces growth or merely stability. Watch whether large systems announce expansions beyond their flagship hospitals, which would signal that the capacity case is winning. Watch whether any program publishes eligibility rates and caregiver-burden data alongside outcomes, because the first credible one will set the reporting standard for everyone else. And watch CMS: any study or reporting requirement it attaches to the waiver will tell you what questions the 2030 decision will turn on.
Watch, too, for the first serious adverse event that makes the news. Every model of care has them, but a home program has fewer witnesses and less institutional cover. How programs handle that moment, and how transparent they are, will shape public trust more than any trial.
The takeaway
Hospital at home is one of the few genuine care-delivery innovations of the past decade with randomized evidence behind it. The extension to 2030 gives it room to become permanent. Whether it deserves to depends on whether programs can show that it works beyond the carefully chosen few, and that it does not simply shift the hardest hours of hospital care onto families who never signed up for a nursing job.
If your system runs a hospital-at-home program, what share of patients actually qualify, and who covers the night? Tell us in the comments.
