times the 5 year cumulative incidence of stroke compared to if none of the

# Times the 5 year cumulative incidence of stroke

This preview shows page 3 - 6 out of 6 pages.

times the 5-year cumulative incidence of stroke compared to if none of the exposed had taken aspirin, assuming no residual confounding by sex, no confounding by other variables, no selection bias, and no information bias. Inverse probability weighting L A Y Obs # P(a=1| L) F(A|L) p/f Pseudo 1 1 1 60 3/4 3/4 1 60 1 1 0 90 3/4 3/4 1 90 1 0 1 40 3/4 1/4 3 120 1 0 0 10 3/4 1/4 3 30 0 1 1 75 3/5 3/5 1 75 0 1 0 75 3/5 3/5 1 75 0 0 1 60 3/5 2/5 3/2 90 0 0 0 40 3/5 2/5 3/2 60 Psuedo A=1 A=0 Total Y=1 135 210 345 Y=0 165 90 255 Total 300 300 600 SMR = Pr [ Y a = 1 = 1 A = 1 ] Pr [ Y a = 0 = 1 A = 1 ] = 135 / 300 210 / 300 = 0.643 If all the exposed had taken aspirin, we would have observed 0.64 times the 5-year cumulative incidence of stroke compared to if none of the exposed had taken aspirin, assuming no residual confounding by sex, no confounding by other variables, no selection bias and no information bias. 3. Using standardization, compute the cumulative incidence ratio for the unexposed. Male ASA No ASA Total Female ASA No ASA Total Stroke 60 40 100 Smoke 75 60 135 No stroke 90 10 100 No Smoke 75 40 115 Total 150 50 200 Total 150 100 250 Pr [ Y a = 1 = 1 | A = 0 ] Pr [ Y a = 0 = 1 | A = 0 ] = Pr [ Y = 1 | A = 1, L = 0 ] Pr [ L = 0 A = 0 ] + Pr [ Y = 1 | A = 1, L = 1 ] Pr [ L = 1 | A = 0 ] Pr [ Y = 1 | A = 0, L = 0 ] Pr [ L = 0 A = 0 ] + Pr [ Y = 1 | A = 0, L = 1 ] Pr [ L = 1 | A = 0 ] ¿ 75 150 × 100 150 + 60 150 × 50 150 60 100 × 100 150 + 40 50 × 50 150 = 70 / 150 100 / 150 = 0.7 5 years If all the unexposed had taken aspirin, we would have observed 0.70 times the 5-year cumulative incidence of stroke compared to if none of the unexposed had taken aspirin, assuming no residual confounding by sex, no confounding by other variables, no selection bias and no information bias.  #### You've reached the end of your free preview.

Want to read all 6 pages?

• Summer '14
• FrancisCook
• • •  