The last step (Step 3) was the multivariate prediction model construction.
Two data transformations including the Johnson system of distribution and Box-Cox method were used in the construction of the multivariate model for the Jiangsu cohesive soils.
For the construction of the multivariate model, we used forced simultaneous entry all candidate variables remained in the model regardless of statistical significance.
The multivariate model 2 is more appropriate for construction of a predictive model for protein expression, as it considers the combined effect of all miRNAs and thus weighs the contributions of the individual miRNAs to obtain optimal prediction.
Furthermore, the large number of women included allowed a stratified analysis to be performed and construction of a multivariate model exclusively for them in which the set of variables found to be statistically significant differs from that of men.
Cox proportional hazards model for time to event data was used to evaluate the hazard of every important variable, as well as for the construction of a multivariate model, explaining the relative significant influence of the covariates on outcomes (unconditional backwards procedure).
Another strong point of the current study was the construction of a comprehensive multivariate model, including several sociodemographic, clinical and treatment variables in its adjusted analysis.
For the multivariate model, a significant association was observed between malaria positives and ethnicity, forest cover, altitude, floor construction, household density and treatment preference from the malaria control program.
For each zone, the final multivariate model was also estimated for men and women separately, in order to understand the cultural construction of gender.
The final multivariate model was adjusted for ethnicity, household income, father's or mother's education, type and year of construction of residence, and time between the reference date and interview.
