We then constructed a second Cox proportional hazards regression model to test if preventability belief was related to FH and a third model to test if belief could be viewed as a mediator of the FH-CHD relation.