We built a multivariate model in using few stations (3 stations) separated with irregular distances from 26 km to 56 km.
The multivariate model in Table 3 for never-married women shows that the odds of never-married women justifying wife-beating were lower in 2008 and 2013 relative to 2003.
Variables that were not significantly associated with the outcome (α>0.05) were removed from the full multivariate model in succession.
We added variables that were found to be associated with interruption in care or death in univariate analysis (p-value <0.25), to the multivariate model, in order to evaluate multiple factors simultaneously.
Considering a multivariate model in the meta-regression results, case definition was not a significant source of bias in the meta-analysis, while the different mean age of cases and violation of HWE were significant sources of heterogeneity.
We validated our multivariate model in our population (internal validation) [ 39].
Further studies should test the accuracy of our multivariate model in a larger patient cohort.
The most parsimonious multivariate model in our study comprised only performance status and CHOP staining.
Those parameters significant on univariate analysis were entered into the multivariate model in a stepwise fashion.
Table 4 presents a multivariate model in which significant background variables and psychosocial factors were taken into account.
The final multivariate model in the reference group of FTR is presented in the first three columns of Table 2.
