Incident rate analysis - could we time-to-event visualizations

OHDSI / Atlas

ATLAS is an open source software tool for researchers to conduct scientific analyses on standardized observational data

http://atlas-demo.ohdsi.org/

Apache License 2.0

266 stars 135 forks source link

Incident rate analysis - could we time-to-event visualizations #684

Open gowthamrao opened 6 years ago

gowthamrao commented 6 years ago

Expected behavior

Create life-table estimates and Kaplan Meier-estimates Create time-to-event graph for both life-table and kaplan-meier

Actual behavior

No life tables

More details here

davekern commented 6 years ago

I like this idea. Would also suggest adding confidence bounds around the KM curve. And then it would be great to have that feature extended to the cohort comparison analytics dashboard to show the survival curves of two cohorts on the same graph with the hazard ratio/95% CI/p-value, when we get to that point.

gowthamrao commented 6 years ago

Draft of requirements is here

pbr6cornell commented 5 years ago

I support considering a future feature that descriptively summarizes time-to-event distribution, including a Kaplan-Meier plot. I do not support the idea of allowing for unadjusted comparisons between two cohorts by calculating p-values of differences. Appropriate comparisons should be made through population-level effect estimation analyses, not clinical characterization.