Attrition table varies between identical runs

[SivamshIndukuri]

I was re-running an analysis and noticed that the attrition table changes between runs. While the subject counts remain consistent, the number of iterations and record counts differ each time. I wanted to check whether this is expected behavior, and it indicates some nondeterminism in the analysis process.

@anthonysena and I also wanted to better understand how to interpret the attrition table, and what each iteration represents to get a better understanding of the overall results and what the package is doing.

I also attached the results of the 2 attrition tables with target_cohort_name and database_id masked.

tp_attrition 2(in).csv
tp_attrition(in).csv

[mvankessel-EMC]

@SivamshIndukuri Could you try the TreatmentPatterns/develop?

remotes::install_github("darwin-eu-dev/TreatmentPatterns@develop")