Assigned 4/21/04
Due 4/28
Returned 5/5
Graded by Adriana Caldwell
[10 points -- point number in brackets]
8.5 All rates are per 100,000 person-years
(A) Risk Ratios
[1] For D1 (lung cancer), RR = 13 / 1 = 13.0
[2] For D2 (heart disease), RR = 100 / 50 = 2.0.
[3] The RR is the most direct measure of the strength of an association. Thus, there is a *stronger* association between
lung cancer and smoking than between heart disease and smoking.
(B) Risk differences (all risk differences are per 100,000 person-years)
[4] For disease D1, RD = 13 - 1 = 12
[5] For disease D2 , RD = 100 - 50 = 50
[6] The RD measures the population-based effect in absolute terms. Thus, smoking
causes more heart disease than lung cancer.
(C) The apparent paradox is explained by the fact that lung cancer is rare compared with heart disease. Although smoking multiplies the rate of lung cancer more than it multiplies the rate of heart disease, far fewer people will be affected with lung cancer than with heart disease.
8.6 Hepatitis B and Liver Cancer
(A) Rates are per 100,000
R1 = 65 / 65,000 � 100,000 = 100
R0 = 5 / 215,000 � 100,000 = 2.33
R = 70 / 280,000 � 100,000 = 25
[7] RR = 100 / 2.33 = 42.9
Interpretation: The rate of liver cancer in the Hepatitis B group is approx. forty-three times
that of the Hepatitis-free group.
[8] (B) RD (per 100,000 person-years) = 100 - 2.33 = 97.67
Interpretation: For every 100,000 person-years of Hepatitis B infection, there
is an excess of 97.67 liver cancer cases.
[9] (C) The RR is a measure of the strength of the association.
[10] (D) The RD measures the effect of the exposure in absolute terms.