Current rehabilitation care paradigms are not well aligned with the needs of frail older adults, but the resultant impact on rehabilitation outcomes is unclear. Understanding how frailty may impact rehabilitation outcomes, and understanding some of the underlying mechanisms, may help inform payment policy changes. This study was a cross-sectional analysis of data from Round 5 of the National Health and Aging and Trends Study (NHATS). We identified older adults who had completed one or more episodes of rehabilitation care and used a validated 5-item NHATS Fried Frailty scale to categorize patients as frail (3/5 or more) or non-frail (≤2/5). We then evaluated the association between frailty status and three key patient outcomes: (1) achievement of rehabilitation goals, (2) functional improvement during rehabilitation episodes, and (3) discontinuation of therapy after exhausting insurance benefits. Lastly, we used multivariable, survey-weighted logistic regression models to estimate adjusted relationships between frailty and rehabilitation outcomes. An estimated 5.6 million survey-weighted older adults in the United States (95% CI 5.1 to 6.0 million) completed an episode of rehabilitation in the past year, an estimated 1,271,290 (95% CI 921,758 to 1,620,822; weighted: 22.8%) of whom were frail. Frail rehabilitation recipients were generally older, had a greater comorbidity burden, and had a higher prevalence of dementia. In adjusted models, frailty was associated with poorer functional outcomes, a lower probability of meeting rehabilitation goals and a greater likelihood of exhausting rehabilitation insurance benefits. Exercise is a well-supported intervention for the management of frailty, but our results suggest that frail older adults are not getting the volume or intensity of rehabilitation treatment needed to maximally improve outcomes-in part due to limited payer coverage of rehabilitation services in the United States.