For five years, a Medicare waiver has let approved hospitals treat genuinely acute patients in their own homes rather than in a ward: daily clinician visits, remote monitoring, infusions and oxygen delivered to the house, and an inpatient-level payment to the hospital. It began as a pandemic measure and has repeatedly been extended a few months at a time, which made it almost impossible for a health system to invest in. In March, Congress extended it through 2030 as part of the Consolidated Appropriations Act.

By then around 360 programs across roughly 140 health systems in 37 to 39 states had been approved. CMS has also started releasing the underlying data for independent analysis. That combination, a durable policy runway and a public dataset, makes this a good moment to look at what the model does and does not demonstrate.

From the Lab to the Ledger

The case for treating acute illness at home is not primarily about comfort. Hospitals carry their own risks that have nothing to do with the admitting diagnosis: hospital-acquired infection, delirium in older patients, deconditioning from days in bed, and sleep disruption. For a patient with pneumonia or a heart failure exacerbation who is sick enough to need daily medical attention but stable enough not to need a monitored bed, those risks can outweigh what the building adds.

What makes it possible now is the monitoring layer. Continuous or frequent vital sign measurement, a communication link to a clinician, scheduled in-person visits by a nurse or paramedic, and a defined escalation route back to the hospital if the patient deteriorates. The waiver requires two in-person visits a day or one plus remote monitoring, which is the structural difference between this and ordinary home health care.

CMS reporting on the program found that patients treated at home had lower 30-day mortality than comparable inpatients, with the gap present across all of the top 25 diagnosis groups and statistically significant in 11 of them. Spending in the 30 days after discharge was lower for home episodes in more than half of those groups. Readmission results were mixed, higher at home for two diagnosis groups and markedly higher for inpatients in three.

Bio-Pipeline Ledger

The Acute Hospital Care at Home waiver: active, extended through September 2030, with roughly 360 approved programs. Operating at scale and now on a policy footing stable enough for health systems to build around.

Remote vital sign monitoring for acute care at home: commercially available and central to how these programs run, with the monitoring itself better validated than the overall care model.

Mortality and spending findings from CMS: favorable to the home model, and explicitly caveated by CMS itself on patient selection. Observational, not randomized.

Randomized trials of hospital at home: a small number of single-centre randomized studies exist and point in the same direction on cost and function. Limited in size and setting, not a substitute for a large trial.

Hospital at home for surgical recovery and complex oncology: being piloted in some systems beyond the medical admissions the model was built for. Early stage, not validated for broad use.

The Clinical Reality Check

The caveat CMS attaches to its own numbers is the one to carry: differences attributable to patient selection and clinical complexity make it difficult to conclude that the program itself lowered spending compared with inpatient care. Programs choose who is suitable, and they choose patients who are stable, who have a safe home environment, and often who have someone with them. Those patients would have done better in any setting. The mortality gap is encouraging and it is not proof of causation.

What is better established is that the model is safe for the people selected for it, which is not a small finding given how many patients have now been treated this way. The escalation routes work, and the dire predictions made when the waiver started have not materialized.

The honest limitation is one of equity rather than clinical safety. The model requires a home with reliable power and connectivity, enough space for equipment, and usually a capable person present. Patients without those things are not eligible, which means the benefit is distributed according to housing and caregiving circumstances rather than medical need. That is a real constraint, and it gets less attention than it deserves.

For anyone who might be offered this, the questions are concrete. What happens at three in the morning if something changes, and how fast does someone arrive. Who is responsible for the parts of care a family member would otherwise never do. What is the cost-sharing compared with an admission, since this is billed as inpatient care. Those answers vary by program, and they determine whether the choice is a good one for a particular household. For most eligible patients, with those answers in hand, this is a legitimate option rather than a downgrade.