I am doing an interrupted time series analysis looking at initiations of a medication with the onset of the covid pandemic as the intervention date.
This is population-level data and this medication is frequently discontinued and re-initiated, so individuals can contribute more than one "initiation" during the study period.
We are looking at initiations per 7-day period, and consider a re-initiation to be any initiation after 30 days without being dispensed the medication.
Appreciate the answers to these questions:
1. Is there something I need to do to account for the correlation of data at the participant/individual level?
2. Can I still use the itsa command or is it better to use xtitsa? Or neither are possible and I need to restrict to one entry per individual?
Thanks for any advice!
This is population-level data and this medication is frequently discontinued and re-initiated, so individuals can contribute more than one "initiation" during the study period.
We are looking at initiations per 7-day period, and consider a re-initiation to be any initiation after 30 days without being dispensed the medication.
Appreciate the answers to these questions:
1. Is there something I need to do to account for the correlation of data at the participant/individual level?
2. Can I still use the itsa command or is it better to use xtitsa? Or neither are possible and I need to restrict to one entry per individual?
Thanks for any advice!
