Table 5.
Results From Multinomial Logistic Regression With the Final Parsimonious Model Predicting Body Mass Index Categoriesa Within the Bereaved and Nonbereaved Groups
| Predictor | RRR (95% CI) | t | P |
|---|---|---|---|
| Overweight | |||
| Bereavement statusb | 2.61 (1.07–6.38) | 1.19 | .035 |
| Caregiver history of depression before proband death | 1.92 (0.55–6.66) | 1.03 | .303 |
| Bereavement status × caregiver history of depression before proband death | 0.26 (0.05–1.28) | −1.66 | .098 |
| Obese | |||
| Bereavement statusb | 4.78 (2.12–10.80) | 3.77 | .000 |
| Caregiver history of depression before proband death | 3.88 (1.35–11.16) | 2.52 | .012 |
| Bereavement status × caregiver history of depression before proband death | 0.14 (0.04–0.57) | −2.77 | .006 |
a
The outcome variable, body mass index category, was dummy-coded into 2 categories, overweight and obese, with normal weight as the referent condition.
b
Bereavement status served as the independent variable and was introduced as a dummy variable, with nonbereaved offspring as the referent condition.
Abbreviation: RRR = relative risk ratio.