The variance analysis of the model coefficients shows that the integrated AR/ARCH model can get more concentrated distribution.
However, after adjusting for need variables (model 3), the coefficients showed a reduction relative to the coefficients in model 1, ranging from 82% for hospital admission to 41% for institutional care.
The model's coefficients show that the duration of ART was a significant predictor of the changes in health utility scores of HIV/AIDS patients, after adjusting for study characteristics.
In terms of prediction performance, the prediction model constructed from the synthesized coefficients showed slight improvements over the conventional approach, particularly in BS.
Regarding the regression coefficients of the interaction term CSE × Age in these models, only two out of nine coefficients showed statistical significance, as their confidence intervals did not contain zero (see Table 5).
We also used backwards elimination (with a 15% significance level 14), which excluded the variable sex from the model (coefficients not shown).
Each model for a domain of HRQoL included all demographic and clinical variables and the model coefficients are shown by variable in figures 1 and 2. Demographic and clinical variables which were significant at the 10% level were extracted and are presented in table 3 by HRQoL domain.
