This is an old revision of the document!
Relevant table: http://www.ohdsi.org/web/wiki/doku.php?id=documentation:cdm:drug_exposure If the days_supply and the drug_exposure_end_date fields essentially designate the same thing they are therefore duplicates of the same information and create the danger of of inconsistency. One of them could be dropped.
Alternatively, we could split them into two different meanings:
|drug_exposure_end_date||Yes or No||date||The end date for the current instance of Drug utilization. Depending on different sources, it could be a known or an inferred date and denotes the last day at which the patient was still exposed to drug.|
|days_supply||No||integer||The number of days of supply of the medication as prescribed. This reflects the intention of the provider for the length of exposure.|
|verbatim_end_date||No||date||The known end date of a drug exposure.|
The verbatim_end_dat field is only necessary if we decide to make the drug_exposure_end_date mandatory. In cases where no good data are available to fill that end_date, teh ETL will have to take a guess, and that may not satisfy some algorithms for dose calculation.
All standardized analytics and the drug era constructor will only use the drug_exposure_end_date. The days supply could be used to understand the intent of prescribing versus the reality, and it could be used to calculate MPR.