This is the question bothering me long time.

If we want to look at effects of some dichotomous variables on a continous variable in a whole model, should we use these dichotomous variables as independent variables and the continous variable as dependent variable in a multiple linear regression model? These dichotomous variables can be sex, events (1 represents yes and 0 represents no) and are very common used in clinical research.
I have seen some papers used it and was published. for example:
http://www.sciencedirect.com/science?_ob=ArticleURL&_udi=B6T4S-4JRM0FH-3&_coverDate=05%2F15%2F2006&_alid=399398589&_rdoc=1&_fmt=&_orig=search&_qd=1&_cdi=4982&_sort=d&view=c&_acct=C000004218&_version=1&_urlVersion=0&_userid=37161&md5=1264e4a2b93c3cd37dac3d4266bc2dc3
But this is confusing because for dichotomous variables it is not making sense to put them into linear model.

Please correct me if I am wrong. Or is there any better solutions to deal with this situation?

Thank you.

ANOVA is more suited to a situation with categorical independents and continuous dependents, but you can use either ANOVA or multiple regression - they're basically the same thing. ANOVA is a special case of regression.