overall mean of clusters

[grateful]

Hi all,

This is a very simple question. 

my data is structured so that each clinician has a set of visits with a dummy variable for imaging. I am able calculate the clinician specific rate (denominator=visits, the numerator=images ordered during visit), but I can’t figure out how to get the summary statistics of the overall per-clinician rates.  

  I am currently using proc means….this code gives me the rate (with the mean statistic), but I want to calculate the overall mean and interquartile ranges across these individual means.  does that make sense?  should I create a new variable?

proc means DATA=out.lbp_NP_spec_flagREVIS;

where np_gen75=1;

class image_rfr_npi;

var x_ray_mri_flag;

output out=all_lvi_rates_NP ; 

run;

thanks for your help!

Monica

[Reeza · Accepted solution]

Check your output data set. Since you don't specify NWAY, there should be two _type_ values. One is overall, and one is for each clinician.

[grateful]

Great--I see that now.  How do I find the interquartile range (min q1 median q2 and max)?  do I need to create a new variable via the out put statement and then run a separate proc means on that?

proc means DATA=out.lbp_NP_spec_flagREVIS mean=rate;

where np_gen75=1;

class image_rfr_npi;

var x_ray_mri_flag;

output out=all_lvi_rates_NP ; 

run;

Thanks!

Monica

[Reeza]

You control the stats on the OUTPUT statement. Add all the statistics you want to that line with an = after the word, ie


output out=.... mean= min= max= Q1= median= / autoname autolabel;