While most intervention studies on coronary heart disease have focused on the high-risk person only, the present study used the family as the unit of intervention. In the study 1373 high-risk men, ages 30-54 years, were identified on the basis of high total cholesterol (TC) and/or low relative high-density lipoprotein cholesterol (HDL-C) (HDL-C/TC) following the 1979/1980 survey in Tromsø. The men and their families were randomly allocated to a control or intervention condition. The intervention families were given advice on diet, smoking, and exercise. At rescreening in 1986/1987, significantly lower risk factor levels were found in both the intervention men and their spouses compared with those in the control group. For children, the differences were small and mostly nonsignificant. Men, spouses, and children in the intervention group reported more favorable dietary habits than those in the control group. No differences were found in smoking or leisure time physical activity.