Event ReOccurrence Diagnostics (exit date diagnostic)

[gowthamrao]

What the diagnostic is:

Proportion of subjects in the cohort with:

• 1 event
• 2 events
• 3 events
• 4 events
• 5 events

• 5 events

Time distribution diagnostic: Given an event cohort what is the distribution (mean, sd, median, p1, p5, p10, p25, p50 median, p75, p90, p95, p99) of time in days between event end date of previous event and event start date of next event for the same subject.

• distribution between any consecutive events
• distribution between 1st and 2nd event
• distribution between 2nd and 3rd event
• distribution between 3rd and 4th event
• distribution between 4th and 5th event

Washout impact diagnostic:

• If no two events were allowed to occur within 30 days of each other, what would be the events per person (including proportion of subjects in cohorts with 1/2/3/4/5>5 events
• If no two events were allowed to occur within 90 days of each other, what would be the events per person (including proportion of subjects in cohorts with 1/2/3/4/5>5 events
• If no two events were allowed to occur within 365 days of each other, what would be the events per person (including proportion of subjects in cohorts with 1/2/3/4/5>5 events

What problem does it solve:

• helps inform persistence criteria and cohort era logic and/or washout period choice