QUESTION:Do long-term (> 3 months) home or community-based exercise programs improve function, reduce falls and prevent hospital readmissions in older people with cognitive impairment? DESIGN:Systematic review and meta-analysis of randomised, controlled trials. Electronic databases (CINAHL, PubMed, Medline, Embase, AMED) were searched from the earliest date possible until March 2016. PARTICIPANTS:Older adults (≥ 65 years) with cognitive impairment living in the community. INTERVENTION:Supervised home or community-based exercise programs longer than 3 months. OUTCOME MEASURES:The primary outcomes were function (including balance and activities of daily living), falls and hospital readmissions. RESULTS:Of 1011 studies identified, seven trials with 945 participants met the inclusion criteria. Compared with no intervention, long-term exercise programs improved functional independence in basic activities of daily living by a moderate and significant amount (SMD 0.77, 95% CI 0.17 to 1.37, I2=67%), and improved functional independence in instrumental activities of daily living by a small and significant amount (SMD 0.44, 95% CI 0.03 to 0.86, I2=42%). Long-term exercise improved balance (mean difference in functional reach test 5.2cm, 95% CI 0.5 to 9.9, I2=76%). Data from two individual trials suggest that long-term exercise programs also reduce falls in older people with cognitive impairment. However, there was limited reporting of the effect of exercise on hospital readmissions for this group of people. CONCLUSIONS:Long-term home and community-based exercise programs improve function in older adults living in the community with cognitive impairment. Review registration: PROSPERO CRD42015029602. [Lewis M, Peiris CL, Shields N (2016) Long-term home and community-based exercise programs improve function in community-dwelling older people with cognitive impairment: a systematic review.Journal of Physiotherapy63: 23-29].