Most definitions of frailty utilize US populations in their development. The concept of frailty has not been well studied in Japan, which has the largest percentage of older patients (per capita) in the world. We created a 5-year prospective cohort study of community-dwelling older Japanese adults. Participants were not frail at baseline, based on our definition adapted from the Canadian Study for Health and Aging Clinical Frailty Scale. Participants underwent a comprehensive geriatric assessment (CGA) at baseline, and final assessments were either in person or via mailed survey. We enrolled 407 individuals (184 men, mean age 78 ± 4 years; 223 women, mean age 77 ± 4 years). Sixty-five participants met criteria for frailty by the end of the study. In univariate analyses, eighteen separate parameters were associated with frailty, some of which included: age, gender, handgrip, timed walk, systolic blood pressure, pulse pressure, cognitive status, living alone, and hearing deficits. In multivariate analyses, the following elements remained associated with frailty: timed walk, pulse pressure, cognition deficits and hearing deficits. We established cut-off points for timed walk (5m/3s) and pulse pressure (60 mmHg). We then created a simple additive score for these four factors (present = 1; absent = 0). A score of 0 had a 93% negative predictive value for frailty while a score of 4 had a 70% positive predictive value. While further study is needed, this work creates an easy-to-administer tool that may be generalizable to other populations.