To investigate the association between treatment modality and functional outcome, logistic regression analysis adjusting for age was applied and 95% confidence intervals were calculated around point estimates. A descriptive analysis of non-responders was performed. To minimize any bias associated with conducting a complete case analysis only, missing values in the four outcome variables (erectile dysfunction, urgency, urinary incontinence and bowel dysfunction), were imputed, both for non-responders to the questionnaire as well as for responders with missing values, both for cases and controls, using multivariate imputation by chained equations following the method described by van Buuren.17 Two alternative methods of imputing missing outcome values were used: imputation based on age only (primary) and imputation based on the variables age, comorbidity, marital status and education, excluding the men without prostate cancer.
Separate analyses were performed restricted to treated prostate cancer patients, adjusting for age, comorbidity status, marital status and education. The statistical analyses were conducted using the R statistical program package, version 2.15.1 (R Foundation for Statistical Computing, Vienna, Austria).
Separate analyses were performed restricted to treated prostate cancer patients, adjusting for age, comorbidity status, marital status and education. The statistical analyses were conducted using the R statistical program package, version 2.15.1 (R Foundation for Statistical Computing, Vienna, Austria).