TABLE 2.
Clustersa
| ||||||||||
---|---|---|---|---|---|---|---|---|---|---|
1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | ||
Low Cholesterol |
Healthy | High Blood Pressure |
Low Stress Hormones |
Metabolic Syndrome |
High Blood Sugar |
Low Stress Antagonist |
High Stress Hormones |
High
Pro- Inflammatory Elements |
||
Factor/Componentsb Range (min–max) | ANOVA F-testc | |||||||||
Stress hormones (−3.02 to 3.11) | −0.79d | 0.33 | 0.35 | −0.92 | 0.66 | −0.22 | −.62 | 1.03 | −0.30 | 118.41e |
Metabolic syndrome (−2.81 to 2.90) | −0.55 | −1.08 | −0.40 | 0.16 | 1.22 | 1.00 | −0.74 | 12 | 0.95 | 177.97e |
Pro-inflammatory (−3.03 to 3.08) | −0.41 | −1.19 | −0.71 | 0.29 | 0.99 | 0.57 | −0.27 | 0.12 | 1.08 | 154.72e |
Cholesterol (−4.69 to 2.75) | −1.12 | 0.06 | 0.42 | 0.73 | 0.73 | −0.01 | −0.08 | −0.69 | −0.82 | 93.77 |
Blood sugars (−1.83 to 6.70) | −0.32 | −0.48 | −0.36 | −0.13 | 0.18 | 3.71 | −0.25 | 0.08 | 0.36 | 215.42e |
Stress antagonists (−3.86 to 2.26) | 0.31 | 0.22 | 0.58 | 0.14 | 0.35 | −0.10 | −1.7 | 0.30 | −0.73 | 102.06e |
Blood pressure (−3.91 to 3.17) | −0.60 | −1.10 | 0.94 | −0.06 | 0.47 | 0.26 | 0.15 | 0.21 | −0.52 | 80.78e |
N = 96f | N = 138 | N = 155 | N = 169 | N = 144 | N = 35 | N = 109 | N = 146 | N = 104 |
This nine-cluster solution was obtained using k-means, with standard Euclidian distance measures; convergence criterion was set to zero.
These are the seven factors from Table 1, used to construct the different profiles for the nine clusters. Included below each factor is its min and max score possible, which comes from summing the biomarkers that loaded on it and converting this sum into a z-score.
Unstandardized F scores (ANOVA) demonstrating, for descriptive purposes only, the relative impact the seven factors had in determining cluster membership
The score for each case, for each of the seven factors, was computed (as noted in b above) by summing each case’s scores on the biomarkers for each factor, as shown in Table 1. In turn, these summed factor scores were converted into z-scores to normalize the data.
F test was significant at.000. The factors with the three highest scores are highlighted.
Number of cases in each cluster.