Effects | | | | | Estimate | SE | p-value |
|---|
Intercept | Mean BMI | | | | 15.79883 | 0.54076 | 0.00000 |
Treatment success (yes) | 1.73193 | 0.58372 | 0.00301 |
Lung cavities on X-ray (yes) | 1.47195 | 0.76554 | 0.05451 |
Slope | Time | | | | 0.04917 | 0.09200 | 0.59303 |
Interaction time - Treatment success (yes) | 0.21736 | 0.09380 | 0.02049 |
Interaction time - Lung cavities on X-ray (yes) | −0.18032 | 0.06664 | 0.00682 |
- Data are given as mean (SE); SE standard error; bold p values were < 0.05, which means the corresponding factor was significantly associated with the BMI increase. In the linear mixed model, the repeated measurement of BMI was the dependent variable, and the treatment success and lung cavities on X-ray were the independent variables. Intercept corresponds to the average BMI increase (mean BMI in the table) for patients without lung cavities on X-ray and for the patients who died. Mean BMI was 15.79 kg/m2 at baseline. BMI was higher for cured patients [17.52 kg/m2 (15.79 + 1.73)] and for patients with lung cavities on X-ray [17.26 kg/m2 (15.79 + 1.47)]. Time corresponds to the speed of BMI increase at each visit, which was 0.05 kg/m2 for patients who died and for patients without lung cavities on X-ray. The influence of these factors was evaluated by interaction with the given factor and time. The interaction time and treatment line indicate that the BMI increased faster for cured patients [0.27 (0.05 + 0.22) BMI points per month], and the last line in Table 2 means that the BMI increase was slower for patients who had lung cavities on X-ray [− 0.13 (0.05–0.18) BMI points per month]
