Quantifying alcohol consumption: patterns and problems.

In this paper are summarized the implications of recent population studies in which drinking behavior is used to predict development of heart disease. A key inference is that the association of drinking behavior with mortality may be attributed as much to characteristic lifestyles as to the drinking itself. Increased emphasis on validation studies is needed to improve measurement of drinking behavior for greater precision in predicting mortality. A plan is presented for large-scale longitudinal studies to predict heart disease and other major health problems, calling for the joint efforts of agencies such as the National Center for Health Statistics and the Multiple Risk Factor Intervention Trial.