What the research shows, and what it doesn’t.

The most consistent finding is fewer hospital days

In a study of 225 people referred from a Chicago public hospital, those admitted to a respite program used an average of 3.7 hospital days over the following twelve months. Those referred but turned away because no bed was available used 8.3 days, a difference of roughly 4.6 hospital days per person per year.¹

A separate analysis of 735 people discharged from Boston Medical Center found that those discharged into medical respite had adjusted odds of readmission within 90 days of 0.54 compared with people discharged to their own care on the street or in a shelter.²

Both are observational studies. Clinicians, not randomization, decided who went where, so they establish association rather than cause.

The same studies contain results that cut the other way

In the Chicago study, emergency department and outpatient visits did not differ between the groups. The effect was confined to inpatient days.¹

In the Boston study, unadjusted readmission rates did not differ significantly at all. The benefit appears only after statistical adjustment, and the same paper found that a respite discharge was associated with higher total 90-day charges, by about $5,994.²

A systematic review of thirteen studies concluded that medical respite was associated with reduced admissions, inpatient days and readmissions, while results for emergency department use and costs were, in the authors’ words, mixed but promising. They called for future research using adequate comparison groups.³

The randomized evidence is thinner than the field usually suggests

The most cited randomized trial enrolled 405 chronically ill homeless adults and offered the intervention group a post-discharge respite stay plus long-term housing plus intensive case management. After eighteen months it reported relative reductions of 29% in hospitalizations and 29% in hospital days.⁴

Two limits matter. The unadjusted differences in that trial were not statistically significant, and what was tested was a bundle. It is not evidence for recuperative care standing on its own.

On cost, savings are plausible and unproven

The economic analysis accompanying that trial estimated an average annual saving of $6,307 per person against usual care. The result was not statistically significant, and its confidence interval spans everything from large savings to added cost. The authors described their own study as underpowered.⁵

A 2024 federal review of this literature found that although participants show reduced hospital use, there is limited research on whether these programs produce cost savings for hospital systems or communities, and noted that programs do not consistently track or report outcome data.⁶

Why we put this here

It would be easy to write that recuperative care saves money and leave it uncited. Much of this field does. We would rather show the evidence as it stands, including the parts that complicate the case, because the people deciding whether to fund a stay have read it too.