St. Louis University (SLU) and Community Health-in-Partnership Services (CHIPS) are collaborating to develop programs to reduce frailty as identified by physical weakness and instability of gait and balance) in community-dwelling African-American seniors living in a 3.25 square mile area of north St. Louis city. The project described herein is one of the initial steps toward meeting this goal. The specific aims of the project are to define (1) the distribution of frailty measures in elderly subjects living in this community; (2) factors associated with progression in frailty measures; and (3) the effect of frailty on deterioration in self care functioning. The effect of frailty on injurious falls, restricted activity days, use of home health services, hospitalization, long term care institutionalization, and mortality will also be examined. Methods involve the acquisition of one group of 400 subjects selected at random from the catchment area using a sample frame developed from HCFA's Medicare Eligibility Enrollment File recruited during the first grant year and a second group of 400 to 500 subjects recruited from SLU-CHIPS seniors, programs recruited during the second year. Both groups will be followed every two months for the duration of the project. Measured variables include sociodemographic factors; reported physical, psychological, and social functioning, including activities of daily living; social support; socioeconomic status; medical diagnoses; medications; cognitive impairment; depression; timed physical performance (hand, shoulder, lower extremity, and walking); stability of gait and balance; health services utilization; and mortality. Extensive quality control of data collection will be employed. Distribution of frailty measures will be estimated using the randomly selected sample. Risk factors for frailty progression will be determined using both samples and multiple linear regression (MLR). MLR on multiple linear regression will be used to evaluate the impact of frailty or decline in self care functioning (depending on the measure).