ObjectiveHome-time is an emerging patient-centred stroke outcome metric, but it is not well described in the population. We aimed to determine the association between 90-day home-time and global disability after stroke. We hypothesised that longer home-time would be associated with less disability.DesignHospital-based cohort study of patients with ischaemic stroke or intracerebral haemorrhage admitted to an acute care hospital between 1 April 2002 and 31 March 2013.SettingAll regional stroke centres and a simple random sample of patients from all other hospitals across the province of Ontario, Canada.ParticipantsWe included 39 417 adult patients (84% ischaemic, 16% haemorrhage), 53% male, with a median age of 74 years. We excluded non-residents of Ontario, patients without a valid health insurance number, patients discharged against medical advice or those who failed to return from a pass, patients living in a long-term care centre at baseline and stroke events occurring in-hospital.Primary outcome measureAssociation between 90-day home-time, defined as the number of days spent at home in the first 90 days after stroke, obtained using linked administrative data and modified Rankin Scale score at discharge.ResultsCompared with people with no disability, those with minimal disability had less home-time (adjusted rate ratio (aRR) 0.96, 95% CI 0.93 to 0.98) and those with the most severe disability had the least home-time (aRR 0.05, 95% CI 0.04 to 0.05). We found no clinically relevant modification by stroke type, sex or study year. However, for a given level of disability, older patients experienced less home-time compared with younger patients.ConclusionsOur results provide content validity for home-time to be used to monitor stroke outcomes in large populations or to study temporal trends. Older patients experience less home-time for a given level of disability, suggesting the need for stratification by age.