#648 1994 · Johns Hopkins University (Dr. Bruce Leff, Hospital at Home) · Healthcare / hospital operations
Johns Hopkins treated sick elderly patients in their own beds instead of a hospital ward, and got better outcomes for less
the problem
Hospital care itself was making the sickest elderly patients sicker
background
By the 1990s, geriatricians like Bruce Leff at Johns Hopkins could see a pattern the hospital system was structurally unable to fix: elderly patients admitted for treatable acute conditions — pneumonia, a flare-up of heart failure or COPD — routinely left the hospital worse off than the underlying illness alone would have left them. Confinement to an unfamiliar bed, disrupted sleep and the general disorientation of a ward measurably caused delirium, falls, infection and functional decline in patients over 65, on top of whatever they had been admitted to treat.
Standard practice had no answer beyond more caution inside the same model: closer monitoring, more geriatric-trained staff, stricter fall-prevention protocols — all of it still delivered inside the same room causing part of the harm. Building a gentler hospital was still building a hospital, and the hospital bed itself was the one variable that no amount of better protocol could design out of the risk it created.
what everyone would do
The standard fix, within the existing model, was to make hospital stays for elderly patients safer without leaving the hospital: more geriatric specialists on staff, stricter fall-prevention checklists, closer delirium monitoring — all of it still delivered inside the same room and schedule that was contributing to the harm in the first place.
what they saw
Leff's insight was that 'hospital-level care' and 'a hospital building' were not actually the same thing — what made care hospital-level was the intensity of monitoring and medical attention, not the four walls it happened to be delivered inside. Once he separated the service from the building that had always delivered it, the building itself became optional for patients well enough to be monitored rather than confined.
the move
Leff and colleagues at Johns Hopkins moved the hospital instead of the patient: for patients over 65 admitted with one of four common acute conditions (pneumonia, COPD exacerbation, heart failure exacerbation or cellulitis), a physician and nursing team delivered IV medication, oxygen therapy, vitals monitoring and daily visits directly in the patient's own home, substituting for an inpatient bed rather than supplementing it. The model redefined what a 'hospital admission' physically required — a monitored bed and trained staff on a schedule — rather than accepting that it required a hospital building.
why it works
Removing patients from an unfamiliar ward eliminated the specific mechanisms known to cause hospital-acquired harm in the elderly — disrupted sleep, disorientation, immobility, exposure to hospital-acquired infection — while the physician and nursing visits preserved the actual medical intensity that made the care 'hospital-level' in the first place. Because home care needs no shared-ward overhead (fewer nurses per patient-hour, no bed-day charge), the same clinical intensity cost less to deliver, so the model produced better outcomes and lower costs from the same underlying change rather than trading one for the other.
the payoff
In a 2005 Johns Hopkins-led study across three sites and 455 patients, Hospital at Home patients had a shorter length of stay (3.2 days versus 4.9 for standard inpatient care) and cost 32 percent less per stay ($5,081 versus $7,480), with some evidence of fewer complications and high patient satisfaction — and when offered the choice, most eligible patients chose home care over a hospital bed. Researchers projected roughly $1.4 billion in potential annual U.S. savings if every eligible patient nationally received it.
where it breaks
The model only works for patients sick enough to need hospital-level monitoring but stable enough that their trajectory is predictable — the four qualifying conditions in the original trial were chosen precisely because they rarely deteriorate without warning; it is not a substitute for ICU-level or surgical care requiring equipment a home cannot hold. It also depends on a home environment able to support daily visits (a stable address, a caregiver or a safe solo-living situation) and on a payer willing to reimburse a hospital-level rate for a non-hospital stay, which is why adoption stayed limited for two decades until Medicare created a specific reimbursement waiver during the COVID-19 pandemic.
what came after
The model Leff piloted at Johns Hopkins has since been adopted, with local variations, by Mount Sinai, the Mayo Clinic, the Cleveland Clinic, Geisinger and private operators, and CMS began reimbursing hospital-at-home care nationally under a waiver introduced during the COVID-19 pandemic — three decades after Leff's first patients, the model finally acquired a viable payment pathway to match the clinical case for it.
references
- [1]Home, Sweet HospitalJohns Hopkins Medicine, 2021web.archive.org
- [2]Your Next Hospital Bed Might Be at HomeBioethics.com, 2023bioethics.com