You simply treat it as any other phenotype. You can simply say that having a parent with a dx carries one unit of genetic+phenotype risk. Thus a subject with one parent with DX has a genetic risk score as 1, a subject with both parents affected has a genetic risk score of 2. You can also expand this to more distant relatives e.g. assign a genetic risk score of 0.5 to each of the grandparents affected with the illness and so on. Then you can run heritability on this trait. It will explain the proportion of additive genetic variance in the total risk score.