Using the monotone logistic aspect of the multiple imputation procedure, we imputed the missing data.
There were no statistically significant differences in marital status or exercise in relation to dementia status.> -wrap-foot> aMissing values imputed using multiple imputations procedure.
It is also worth noting that we applied a multiple imputation procedure to impute the missing values of each covariate.
However, because of missing data, we used a multiple imputation procedure to impute values for the missing data for diabetes and gynecomastia.
We used the multiple imputation procedure to impute missing values of each covariate, assuming that the data are multivariate normally distributed and the missing data are missing at random.
Therefore, missing values were imputed using a multiple imputation procedure following the approach of van Buuren and Groothuis-Oudshoorn, 33 resulting in five imputed data sets.
For these variables, missing values were imputed from a multiple linear regression model (Stata Impute procedure) as described above.
Using this imputing procedure, data sets with different detection limits could be combined, as validated in house.
For those lost to follow-up we imputed (using the "impute" procedure available in STATA) their final frailty status, based on their baseline levels in the following variables: frailty status, sex, marital status, years of school, IADL, cognitive impairment, depression, BMI category, and score on the PF and SF SF-36 subscales.
Intention to treat analyses were performed in comparisons between the intervention and comparison group, and missing data was imputed by the multiple imputation procedure in IBM-SPSS statistical package.
