Table 1.
Regression coefficients used to predict the prevalence of overweight/obesity from mean BMI
| Predictor | BMI ≥ 25 kg/m2 (n=1883)b | BMI ≥ 30 kg/m2 (n=1857)b |
|---|---|---|
| Constant |
-24.9 (-27.0, -22.8) |
-33.9 (-36.9, -31.0) |
| Cube of first spline segment (knot at BMI of 21.3) |
-.0000423 (-.000118, .0000334) |
-.0000632 (-.0002191, .0000927) |
| Cube of second spline segment (knot at BMI of 25.1) |
-.00522 (-.00680, -.00365) |
-.00619 (-.00947, -.00291) |
| Cube of third spline segment (knot at BMI of 28.9) |
-.00490 (-.00735, -.00245) |
-.00437 (-.00611, -.00263) |
| Cube of last spline segment |
.00168 (.00038, .00298) |
.00316 (.00222, .00410) |
| Square of mean BMI |
-.0182 (-.0220, -.0144) |
-.0280 (-.0327, -.0233) |
| Mean BMI |
1.46 (1.27, 1.65) |
1.98 (1.71, 2.25) |
| High-income country |
.0077 (-.0287, .0442) |
-.113 (-.181, -.0456) |
| Age (midpoint of age category) |
.00567 (.00499, .00635) |
.00456 (.00335, .00577) |
| Year of survey a |
.00934 (.00492, .0138) |
.0236 (.0157, .0314) |
| Female sex |
.91 (.62, 1.19) |
1.01 (.575, 1.45) |
| Sex * mean BMI |
-.0405 (-.0517, -.0292) |
-.0294 (-.0455, -.0133) |
| County income category * year of survey a |
-.0120 (-.0194, -.0047) |
-.00128 (-.0132, .0106) |
| R2 | 0.97 | 0.92 |
* denotes statistical interaction.
a See methods for further details on how year of survey was used.
b 1884 age-sex groups provided mean and prevalence data with sufficient sample size, but those with prevalence zero were not used in the above regression because the logit(0) is not defined.