FRAMINGHAM ILLUSRATIONS
(June 6, 2011)
The page is one of a number of sub-pages to the Scanlan’s Rule page of jpscanlan.com employing publicly available data to illustrate the patterns described on that page. Similar sub-pages include the NHANES Illustrations and Life Tables Illustrations sub-pages. This sub-page also related to the Subgroup Effects sub-page of the Scanlan’s Rule page, as discussed below.
Table 1 below is based on data from online calculators of heart attack risk developed from the Framingham studies. The table compares heart attack risk of 55-year-old men and women who are non-smokers and not on hypertensive medication. The other risk profile elements are listed. Rows 1 through 10 show for men and women with total cholesterol of 300 and systolic blood pressure of 200 the implications of serial increasing HDL levels without affecting the other indicators,[i] along with the standard measures of differences between the male and female rates of experiencing or avoiding heart attacks. While there is some variation for the odds ratio, the patterns of differences are generally as described in the introduction to the Scanlan’s Rule page. That is, the relative difference in the adverse outcome increases, the relative difference in the favorable outcome decreases, the absolute difference decreases, and the difference measured by the odds ratio increases.
|
Table 1: Illustration of Effects on Differences in Heart Attack Risk
of Men and Women of Modification of Risk Profiles
(Based on Framingham Calculator [b1806 a 1]
|
|
ID
|
TotalChol
|
SBO
|
HDL
|
FHAR
|
MHRR
|
RRAdv
|
RRFav
|
AD
|
OR
|
|
1
|
300
|
200
|
30
|
16.00
|
30.00
|
1.88
|
1.20
|
0.14
|
2.25
|
|
2
|
300
|
200
|
35
|
13.00
|
27.00
|
2.08
|
1.19
|
0.14
|
2.48
|
|
3
|
300
|
200
|
40
|
12.00
|
24.00
|
2.00
|
1.16
|
0.12
|
2.32
|
|
4
|
300
|
200
|
45
|
10.00
|
22.00
|
2.20
|
1.15
|
0.12
|
2.54
|
|
5
|
300
|
200
|
50
|
9.00
|
20.00
|
2.22
|
1.14
|
0.11
|
2.53
|
|
6
|
300
|
200
|
55
|
8.00
|
19.00
|
2.38
|
1.14
|
0.11
|
2.70
|
|
7
|
300
|
200
|
60
|
7.50
|
17.00
|
2.27
|
1.11
|
0.095
|
2.53
|
|
8
|
300
|
200
|
65
|
7.00
|
16.00
|
2.29
|
1.11
|
0.09
|
2.53
|
|
9
|
300
|
200
|
70
|
6.50
|
15.00
|
2.31
|
1.10
|
0.085
|
2.54
|
|
10
|
300
|
200
|
75
|
6.00
|
14.00
|
2.33
|
1.09
|
0.08
|
2.55
|
|
11
|
300
|
180
|
75
|
4.00
|
13.00
|
3.25
|
1.10
|
0.09
|
3.59
|
|
12
|
300
|
160
|
75
|
3.00
|
11.00
|
3.67
|
1.09
|
0.08
|
4.00
|
|
13
|
300
|
140
|
75
|
2.50
|
9.00
|
3.60
|
1.07
|
0.065
|
3.86
|
|
14
|
300
|
120
|
75
|
2.00
|
8.00
|
4.00
|
1.07
|
0.06
|
4.26
|
|
15
|
280
|
120
|
75
|
1.50
|
7.00
|
4.67
|
1.06
|
0.055
|
4.94
|
Looking solely at rows 1 and 10, which reflect the effect of increasing HDL from 35 to 75, one observes that the risk ratio for a heart attack increased from 1.88 to 2.33, the risk ratio for avoiding a heart attack decreased from 1.20 to 1.09, the absolute difference decreased from 14 percentage points to 8 percentage points, and the odds ratio increased from 2.25 to 2.55.
The risk ratio patterns mean that the group with the lower base heart attack risk rate (women) experienced the larger proportionate decrease in the risk of heart attack while men experienced a larger proportionate increase in rates of avoiding heart attacks (in accord with the discussion on the Subgroup Effect page). That is, the female heart attack risk rate was decreased 62.5% (from 16% to 6%) while the male heart attack risk rate was reduced only 53.3% (from 30% to 14%). On the other hand the male rate of avoiding heart attack was increased 24.5% (from 70% to 86%) while the female rate of avoiding heart attack was increased only 11.9% (from 84% to 94%).
Observers who rely on relative differences in adverse outcomes would find a larger effect for women while observers who rely on relative differences in the favorable outcome would find a larger effect for men. The approach described on the Subgroup Effects sub-page of the Scanlan’s Rule page and the Solutions sub-page of the Measuring Health Disparities would find the effects to be virtually identical (.560 standard deviations for women and .556 standard deviations for men).
Rows 11 to 14 then show the effects of serially reducing systolic blood pressure by 20 points. Row 15 then shows the effect of reducing total cholesterol by 20. With minor exception for the odds ratio, the patterns are in accord with those described in the introduction to the Scanlan’s Rule page.
[i] According to the calculator, in some cases change in the risk factor is not reflected in a percentage point change in risk presumably because such change occurred did not round to a full percentage point change. In those cases (e.g., rows 6 and 7), the unchanged rate is reduced by half a percentage point. This adjustment only affects observations between individual rows and not the observations for substantial changes, such as those reflected between rows 1 and 10 that are discussed infra.
|