The development in the United States of America of an automated system for coding mortality data (Automated Classification of Medical Entities--ACME) was undertaken with two major objectives in mind: (i) to introduce consistent and rapid assignment of underlying cause-of-death coding with reduced needs for manpower training; and (ii) to allow better utilization of medical information on death certificates for multiple cause-of-death analyses. The ACME system meets both of these objectives; the National Center for Health Statistics (NCHS) produces all of its underlying cause-of-death statistics for the United States on the basis of this system, and multiple cause-of-death data are routinely available for additional epidemiological study beyond the traditional methods of vital statistics analyses. Enhancements of the automated system, primarily through the software known as MICAR, reduce even further the levels of training necessary for persons doing the basic data entry. MICAR additionally will ease transitions between ICD revisions by reducing the need for coder reorientation and by permitting rapid calculation of comparability ratios when new revisions are introduced.