The Compression of Morbidity paradigm, introduced in 1980, maintains that if the average age at first infirmity, disability, or other morbidity is postponed and if this postponement is greater than increases in life expectancy, then cumulative lifetime morbidity will decrease-compressed between a later onset and the time of death. The National Long-Term Care Survey, the National Health Interview Survey, and other data now document declining disability trends beginning in 1982 and accelerating more recently. The decline is about 2% per year, contrasted with a decline in mortality rates of about 1% per year, thereby documenting compression of morbidity in the United States at the population level. Longitudinal studies now link good health risk status with long-term reductions in cumulative lifetime disability; persons with few behavioral health risks have only one-fourth the disability of those who have more risk factors, and the onset of disability is postponed from 7 to 12 years, far more than any increases in longevity in the groups. Randomized, controlled trials of health enhancement programs in elderly populations show reduction in health risks, improved health status, and decreased medical care utilization. Health policy initiatives now being undertaken have promise of increasing and consolidating health gains for the elderly.