Background Institutionalisation (entry to nursing home, or other place of residential care) is an important outcome in PD. We investigated frequency of, and baseline predictors of, institutionalisation in Parkinson's disease (PD).

Methods We identified all new cases of PD in Aberdeen, UK, over 4.5 years using multiple, community-based ascertainment methods. Age-sex matched community-based controls were also recruited. Participants were seen annually and place of residence at each follow-up or before death were ascertained, together with date of entry to institution. Kaplan–Meier institution-free survival probabilities were plotted. Potential baseline predictors were investigated using Cox regression.

Results 198 PD patients and 260 controls were recruited. One patient and one control were institutionalised before diagnosis. One patient and four controls were lost to follow-up. 44 (22%) PD patients and 19 (7%) controls were institutionalised during mean 6.1 years follow-up. Hazards of institutionalisation were higher in PD than controls (HR 3.72 [2.16–6.39]). Older age (HR 1.05 [1.01–1.09]), lower MMSE score (HR 0.88 [0.79–0.98]), and baseline dependence (HR 2.48 [1.18–5.25]) were independently associated with institutionalisation.

Conclusions Institutionalisation was higher in PD than controls. Older age, poorer cognition and being dependent at diagnosis led to increased institutionalisation.