Process treatment_orders with a NULL enddate

[brentlogan]

Rationale

Treatment orders with NULL end dates can be entered into PRIMe.

study.treatmentSchedule.sql queries study.treatment_orders over a time range defined by "dependee" query ehr_looksup.dateRange. It doesn't return results of treatment orders where:

• enddate is NULL, or
• enddate is NULL AND a dateRange date is in the future.

study.treatmentSchedule.sql is used by printable report Clinical Medications and in the EHR animal history, Daily Reports > Clinical Medications Schedule. The animal history reports run for today's date (setting the start date of "dependee" query ehr_looksup.dateRange to today's date with a length of one). The printable report can be run for user-selected dates, including future dates and date ranges greater than one.

study.treatmentScheduleUpdate is used for the TreatmentToDrug ETL, using today's date and range of 1 for the dateRange.

Related Pull Requests

#1111

Changes

• The check for dr.dateOnly (all dates within dateRange) must be <= the treatment order's end date (if it exists).
• The check for date before enddate now returns TRUE if enddate is NULL.

[github-actions]

ERROR: A pull request from 23.11_fb_allowTreatmentNullEnddate is expected to target release23.11-SNAPSHOT, not develop

[jonesgaohsu]

Code meets the objective and is written as was discussed

[brentlogan]

We received feedback that the first commit didn't fully fix the issue. The printed report Clinical Medications failed to produce the desired results for treatment orders having no end date when the report's date was in the future. I reviewed study.treatmentSchedule and realized this results in no overlap between t1.date to t1.enddateCoalesced and the range defined by dr (which is the future date with a range of 1 day), and thus, the treatment order doesn't show in the report.

This latest commit modifies only study.treatmentSchedule and not study.treatmentScheduleUpdate, which might need to be updated.

[brentlogan]

@bbimber please take a look at this pull request. I've determined that allowing treatment orders to have NULL end dates breaks the two queries currently in the pull request. I believe these two proposed changes won't impact the rest of PRIMe, but I'd like your feedback. It might make more sense to require end dates for all treatment orders to avoid this issue.

Should you agree that these changes are fine, I propose making similar changes to the other study.treatmentScheduleX queries that are only slight modifications of treatmentSchedule. I'd also add the second change to study.treatmentScheduleUpdate (which has only one change in this pull request in its current form).

Thanks!

[bbimber]

one minor comment, but i think this will work. I'm approving since the suggestion isnt required for the code to work

[jonesgaohsu]

Reviewed changes with Brent and this is ready for production