Dichotomisation using a distributional approach when the outcome is skewed

摘要: Dichotomisation of continuous outcomes has been rightly criticised by statisticians because the loss information incurred. However to communicate a comparison risks, dichotomised may be necessary. Peacock et al. developed distributional approach dichotomisation normally distributed allowing presentation proportions with measure precision which reflects means. Many common health are skewed so that method for not apply.