A Cochrane review included 4 studies with a total of 1 642 subjects. There were three comparisons: (i) personal assistance versus usual care, (ii) personal assistance versus nursing homes, and (iii) personal assistance versus 'cluster care'. One study was an RCT, three were non-randomised. Personal assistance was defined as individualised support for people living in the community by a paid assistant other than a healthcare professional for at least 20 hours per week.
Overall, participants appeared satisfied with personal assistance, though the models tested were not universally superior to other models of care. In two studies, the personal assistance group experienced fewer unmet needs than the comparison group. Though not a goal of the intervention, most studies reported some data about mortality, which suggest that personal assistance had no comparative impact. Direct measures of functioning suggest that personal assistance does not improve functioning. Broad cost data were available for only one trial, which suggested personal assistance may save a small amount of money compared to treatment as usual.