Hospital-at-Home Comes of Age: The Clinical Evidence Base, the Medicare Waiver Settlement, and What Comes Next

The Acute Hospital Care at Home programme was established by the Centers for Medicare and Medicaid Services in November 2020 as a COVID-19 emergency response measure. The original intent was to expand acute care capacity by allowing Medicare-certified hospitals to deliver inpatient-level care, including physician visits, IV medications, respiratory therapy, and remote patient monitoring, in a patient's own home, reimbursed at the standard inpatient DRG rate.

From the outset, the programme was designed as temporary. It was tied to the COVID-19 public health emergency and was assumed to sunset when the emergency ended. What followed instead was a five-year sequence of short-term congressional extensions that created a stop-start investment environment. Extensions lasted between 78 days and two years, making it commercially impossible for health systems to justify the capital outlay required to build meaningful hospital-at-home programmes. The programme even lapsed entirely during the 43-day government shutdown in September 2025, leaving hospitals in a temporary capacity void that required emergency patient transition back to inpatient beds.

The commercial significance of this distinction is material. Health systems, technology vendors, and payers can now plan multi-year programmes with a stable reimbursement horizon. Cleveland Clinic, which has admitted approximately 3,000 patients to its hospital-at-home programme since April 2023, averaging 24 patients per day, has built a proprietary care coordination infrastructure called CIViC — the Clinically Integrated Virtual Command Centre — that would not have been economically rational under a 90-day extension cycle. Penn Medicine launched across three hospitals in early 2026, building on a nine-month pilot and designing the programme for full-system expansion. NewYork-Presbyterian launched Hospital at Home across Columbia University Irving Medical Center and Weill Cornell Medical Center in November 2025, before the waiver extension was signed, signalling confidence in eventual congressional action.

Hospital-at-home sceptics have historically raised three objections: that outcomes are not equivalent to inpatient care, that the patient population is selected and therefore not generalisable, and that the cost savings do not account for the infrastructure required to deliver the service. The evidence accumulated between 2020 and 2026 addresses each of these directly.

The most cited US evidence remains Levine et al. (Annals of Internal Medicine, 2020), which found a 30-day readmission rate of 7% for hospital-at-home patients versus 23% for those who received traditional inpatient care, alongside a 38% reduction in cost of care. That finding has since been replicated and extended.

A February 2026 propensity-matched cohort study published in Frontiers in Digital Health, covering 2,905 episodes of hospital-at-home care across NHS England, found an overall odds ratio for 30-day readmissions of 0.55 (95% CI 0.42 to 0.70, p < 0.001) and a mortality odds ratio of 0.43 (95% CI 0.35 to 0.53, p < 0.001). Time in hospital over 90 days was 2.64 days fewer than matched inpatient comparators (p < 0.001). The West Hertfordshire Teaching Hospitals NHS Trust, where the study was conducted, reduced mean surge bed occupancy from 65 to 20 between July 2022 and August 2025 as hospital-at-home capacity expanded, a direct operational benefit of programme scale.

The patient selection objection is partially valid and partially obsolete. Earlier hospital-at-home programmes did concentrate on lower-acuity patients, partly because the logistics and technology required to safely manage high-acuity patients at home were not yet available. A 2024 study published in BMC Health Services Research covering 2,905 episodes reported an in-episode mortality rate of 0.32% and a 30-day mortality rate of 4.35%, both comparable to conventional care in the selected population. The authors noted substantial variation in inclusion criteria across programmes, which is the primary methodological limitation of the existing evidence base.

Cleveland Clinic's programme, which averages 24 patients per day and has admitted approximately 3,000 patients since 2023, has expanded its eligible condition list to include heart failure exacerbations, community-acquired pneumonia, cellulitis, COPD exacerbations, and deep vein thrombosis — conditions that would previously have required inpatient admission. The technological infrastructure that enables this expansion is the combined capability of continuous vital sign monitoring, 24/7 physician-led command centres, and pre-positioned rapid response teams who can reach enrolled patients within a defined time window.

Cost analysis of hospital-at-home is more complex than it appears. The 38% cost reduction cited in Levine et al. was calculated for a specific programme operating in a specific health system context. That reduction does not account for the capital cost of building the command centre infrastructure, the logistics cost of deploying paramedics and nurses for twice-daily in-person visits, or the technology cost of equipping each patient with vital sign monitoring hardware. Programmes that do not achieve sufficient patient volume to amortise these fixed costs can end up more expensive per episode than the inpatient alternative.

The evidence suggests that volume is the critical variable. Cleveland Clinic averaging 24 patients per day represents a programme that has reached a scale at which cost per episode is materially lower than inpatient. Penn Medicine's two-programme structure, combining hospital-at-home with a lower-acuity pathway called PATH (Practical Alternatives to Hospitalisation), addresses the patient selection problem that trips up programmes operating at insufficient scale: without a lower-acuity adjacent pathway, hospital-at-home programmes struggle to find enough appropriate patients to sustain their cost structure.

The hospital-at-home model has four non-negotiable infrastructure requirements that determine whether a programme can operate safely at the acuity level required for Medicare DRG reimbursement. Each represents a commercial opportunity for medtech, digital health, and home care service providers — and each represents a capital commitment that health systems cannot make under a 90-day waiver extension.

The hospital-at-home market in 2026 has two structural tiers: health systems that own and operate their own programmes, and specialised operators that partner with health systems to provide the logistics, technology, and clinical support infrastructure on an outsourced basis. The two tiers are not mutually exclusive — health systems like Cleveland Clinic operate proprietary programmes while simultaneously engaging technology vendors for monitoring hardware and platform software.

The consolidation of DispatchHealth and Medically Home is the defining M&A event in this market to date. The merged entity has the geographic footprint, the logistics infrastructure, and the health system relationships to function as the primary outsourced hospital-at-home operating partner for health systems that cannot or choose not to build proprietary programmes. Its direct competitor is the do-it-yourself model employed by Cleveland Clinic, Penn Medicine, and similar large academic medical centres with the capital and clinical infrastructure to build in-house.

For smaller and mid-size health systems, the outsourced operator model is likely the only viable path. Building a proprietary CIViC-style command centre requires a minimum patient volume to justify the fixed cost, and most regional health systems cannot reach that threshold without an operator partner providing shared infrastructure across multiple client organisations.

The five-year extension creates a credible investment horizon for health systems that have been deferring the capital decision. The Penn Medicine two-programme model — combining hospital-at-home with an adjacent lower-acuity pathway — is the most commercially rational structure for health systems entering the market now. Programmes that attempt to operate hospital-at-home in isolation, without a feeder pathway for patients who need more than outpatient care but do not meet the full AHCAH acuity threshold, consistently report patient volume shortfalls that undermine the cost structure of the programme. Health systems should size their initial catchment radius against the twice-daily in-person visit logistics requirement before committing to command centre infrastructure investment.

Hospital-at-home creates a durable demand signal for continuous vital sign monitoring hardware, cellular-enabled transmission platforms, and integrated EHR connectivity. The RPM market at USD 67.3 Billion in 2026 is growing at 8.3% CAGR to USD 117.9 Billion by 2033, with wearable biosensors as the fastest-growing segment at 21.1% CAGR. The primary commercial opportunity for device vendors is not the hospital-at-home command centre itself — which large operators like Cleveland Clinic are building with proprietary technology — but the health system programmes that require standardised, off-the-shelf monitoring solutions that integrate with Epic or Oracle Health. The February 2026 Biotronik partnership with MiCare Path for remote cardiac monitoring and the Philips-Medtronic collaboration announced in June 2025 are both indicative of the integration-platform strategy that device vendors are pursuing.

The command centre model requires telehealth platform capability for daily physician check-ins, clinical decision support, and patient-facing communication. NewYork-Presbyterian's programme combines twice-daily in-person visits with daily physician telehealth check-ins — the telehealth layer reduces the clinical labour cost of the programme without reducing the visit frequency that CMS requires. Digital health platforms that can demonstrate deep EHR integration, HIPAA-compliant data transmission, and clinical workflow embedding in a hospital-at-home context are the primary commercial beneficiaries of the five-year extension, because the investment horizon now justifies multi-year platform contracts.

Hospital-at-home was a constrained investment opportunity under the short-term waiver extension regime: the exit risk associated with programme discontinuation if the waiver lapsed was not compatible with standard fund timelines. The five-year extension materially reduces that risk, making hospital-at-home technology and operator companies more investable. Chilmark Research's projection of the market reaching USD 300 Billion by 2028 reflects both the scale of the addressable acute care market and the pace of health system adoption following the waiver extension. The evidence study that CMS is required to deliver to Congress by September 2028 will be the primary determinant of whether the programme becomes permanent or faces another extension decision — and that timeline is directly relevant to investment exit horizon planning.

The five-year extension provides a runway, not a guarantee. Making hospital-at-home a permanent feature of the Medicare programme requires three things that the existing evidence base has not yet conclusively established: standardised quality measurement, demonstrated equitable access across rural and low-income populations, and a sustainable payment model that does not require a special waiver.

MedPAC's June 2024 Report to Congress identified these three policy questions explicitly. The payment differential — hospital-at-home is currently reimbursed at the same DRG rate as traditional inpatient care — has attracted scrutiny because the cost structure of a home-based episode is materially different from an inpatient episode. If the CMS evidence study confirms that hospital-at-home costs significantly less per episode than traditional inpatient care, Congress may adjust the payment rate downward, which would reduce the economic viability of the model for programmes operating at the margin.

The equity question is more acute. As of January 2026, 373 approved hospitals are concentrated in 37 states, with the majority of programmes located in urban and suburban catchment areas. Rural health systems face a different logistics problem: the twice-daily in-person visit requirement that is manageable within a 30-minute urban catchment becomes economically unviable across rural geographies where drive times are 45 to 90 minutes per visit. Saint Francis Health System in Tulsa's explicit rural expansion strategy — nearly 60% of its patient base comes from outside Tulsa County — represents one approach to this problem, but the logistics model required to serve genuinely rural patients at AHCAH DRG reimbursement rates remains unresolved.