Background: Limited decision-making capacity (DMC) of older people affects their abilities to communicate about their preferences regarding end-of-life care. In an advance directive (AD) people can write down preferences for (non)treatment or appoint a proxy as a representative in (non)treatment choices in case of limited DMC.The aim is to study limited DMC during the end of life and compare the background, (satisfaction with) care and communication characteristics of people with and without limited DMC. Furthermore, the aim is to describe patient proxies' opinions about experiences with the use of (appointed proxy) ADs. Methods: Using a questionnaire, data were collected from proxies of participants of a representative sample of the Longitudinal Aging Study Amsterdam (n=168) and a purposive sample of the Advance Directive cohort study (n=184). Differences between groups (with and without limited DMC, and/or with and without AD) were tested with chi-square tests, using a level of significance of p<0.05. Results: At a month before death 27% of people had limited DMC; this increased to 67% of people having limited DMC in the last week of life. The care received was in accordance with the patient's preferences for the majority of older people, although less often for people who had limited DMC for more than a week. The majority of the proxies were satisfied with the communication between physician and the patient and them, regardless of DMC of the patient. Of people with an AD, a small majority of relatives indicated that the AD had been of additional value. Finally, no differences were found in the role of the relative and the satisfaction with this role between people with and without a proxy AD. Conclusions: Although relatives have positive experiences with ADs, our study does not provide strong evidence that (proxy) ADs are very influential in the last phase of life. They can best be seen as a tool for advance care planning.