2.2. Data Analysis

A database was created using Epi Info (version 7, CDC, Atlanta, GA, USA), and the data that had been collected on hard copy case report forms were transferred therein. Statistical analysis was performed in STATA (version 13, College Station, TX, USA). Categorical variables were detailed in frequency tables, and continuous measures were summarized using means and standard deviations or medians and ranges, as appropriate. After assessing for normality with the Shapiro–Wilk test and plotting histograms and QQ plots, we used the T-test to check for differences in mean values of continuous variables and chi-square tests for binary variables.

To determine the covariates for prevalence of HAND (defined as a positive screen for cognitive impairment on both the MOCA and IHDS), we performed logistic regression modelling. We used backward stepwise selection to identify parameters to fit in the final model. The selected model was run with an interaction term for age and gender. The threshold for statistical significance was set at α = 0.05.