Background More than 600,000 Medicare beneficiaries with a diagnosis of dementia are discharged to skilled nursing facilities (SNFs) after hospitalization annually. However, it is unclear how their risks and benefits of a SNF stay compare to beneficiaries without a diagnosis of dementia. Design Retrospective analysis comparing SNF outcomes for Medicare beneficiaries with and without a diagnosis of dementia. Setting One hundred percent sample of Medicare beneficiaries from 2015 to 2016. Participants Dementia was identified using validated diagnosis codes. In beneficiaries who had an acute hospitalization followed by SNF stay, we used propensity score matching to balance demographics, comorbidities, characteristics of the index hospital stay, prior hospital and SNF utilization, and cognitive status on SNF admission. Measurements Outcomes included unplanned hospital readmission, community discharge rate, and mortality during the SNF stay. Multivariate models were adjusted for hospital and SNF characteristics. Results Our sample included 2,418,853 Medicare beneficiaries discharged from hospital to SNF; 830,524 (34.3%) carried a diagnosis of dementia. Overall, 14.7% of the sample had a hospital readmission, 5.0% died, and 61.5% were successfully discharged to the community. In the propensity-matched cohort, beneficiaries with a diagnosis of dementia had a lower odds ratio of mortality (OR 0.87; 95% confidence interval [CI] 0.86-0.89), similar odds of hospital readmission (OR 0.99; 95% CI 0.98-1.00), and reduced odds of discharge to the community (OR 0.92; 95% CI 0.91-0.93). However, these findings varied by the severity of cognitive impairment on SNF admission: in beneficiaries with no impairment, those with a diagnosis of dementia had higher odds of adverse outcomes. In beneficiaries with severe impairment, beneficiaries with a diagnosis of dementia had lower odds of adverse outcomes. Conclusions Cognitive dysfunction on SNF admission is a stronger predictor of outcomes than a diagnosis of dementia, suggesting the need to individualize decisions about the benefits and risks of SNF care in populations with cognitive impairment.