03-01-2017 05:24 PM
I was wondering how best to tackle the following dataset regarding antibiotic utilization and a policy intervention:
Date Abx_Utilization Intervention Time
Feb2013 365 0 1
Jan2017 456 1 48
The antibiotic utlization is a rate, the intervention was implemented on Feb 2016 - i have coded '0' for time before intervention and '1' as time after intervention.
I would like to plot abx_utilization*time and show the trendline without the intervention and the trendline with the intervention. Similar to this example listed previously by a user (Figure 1a https://www.cdc.gov/pcd/issues/2013/12_0268.htm)
Many thanks for the insights!
03-01-2017 05:47 PM - edited 03-01-2017 05:49 PM
Try something like this:
proc sgplot data=intervention;
reg x=date y=abx_utilization / nomarkers group=intervention;
series x=date y=abx_utilization / markers markerattrs=(symbol=circlefilled);
You will probably want to format 0 and 1 to some meaningful text for the legend.
Hope this helps!