Comparing one normal and one skewed group

#1
Dear Forum,
I have 2 groups for which I need to compare the length of stay (LOS) post-op: a control and a treatment group.
My control group has the following descriptive stats: (n=66; Mean = 8.42; Median = 6.00; SD = 7.45; Kurtosis = 17.66; Skewness = 3.81), while my treatment group is as follows: (n=25; Mean = 4.88; Median = 5.00; SD = 2.22; Kurtosis = 0.31; Skewness = -0.78). I would consider the treatment group to be normally distributed, but my control group of patients is definitely not.
My question is: what is the best way to compare these two groups? Should I use a non-parametric test, such as a Mann-Whitney, since one of the groups is not normal?
Many thanks!
Locolindo
 
#2
Check if there are outliers among the control group. It seems strange with such high kurtosis and skewness in the control group (also with the mean and median is higher in the control group).

One can think of all of them as being randomized to either a control group or a treatment group. Should all really be included in the study?

If LOS is length-of-stay is the number of days in the hospital, then maybe a Poissn distribution can be used.