SIG: Real-World Evidence (RWE) Collaborative

Dear all, 

I am seeking advice regarding using post-initation information (i.e. using EHR data points which occur after the start of the treatment assignment when defining eligibility, treatment strategies, or covariates) in TTE studies using EHR data. 

I'm well aware that it is a "no-go" to do so.

However, this restriction seems to me a bit rigit given that a lot of information in EHR record contains registration date which is not the same as a start date. For example: you would like to include cancer as one of the covariates to adjust for in a TTE study. In the EHR you will have cancer diagnoses with a registration date. This registration date does not mean that cancer started on this date. The cancer was most likely there before the registration date. So if we say "do not use post-initation data" we would not use use cancer diagnoses registred after initation of treatment (time zero). To me this seems a bit odd, given that the patient very likely had cancer at time zero. 

I would very much like to hear RWE community members opinion on this and how to deal with such information.

Thank you in advance,

Joanna Klopotowska

Amsterdam UM, The Netherlands

------------------------------
Joanna Klopotowska PharmD, PhD
Assistant Professor
Amsterdam University Medical Centers
Amsterdam
+31 614393672
------------------------------

EMR registration date of a patient will not be used as a total time for cancer progression time. In that case, disease progression duration will be counted from history.

------------------------------
MAHADIE JAHAD MPH, BPharm
International institute of Health Sciences, Bangladesh Open University
Dhaka
------------------------------

Original Message:
Sent: 05-07-2025 11:11
From: Joanna Klopotowska
Subject: Post-initiation information in TTE using EHR data

Dear all, 

I am seeking advice regarding using post-initation information (i.e. using EHR data points which occur after the start of the treatment assignment when defining eligibility, treatment strategies, or covariates) in TTE studies using EHR data. 

I'm well aware that it is a "no-go" to do so.

However, this restriction seems to me a bit rigit given that a lot of information in EHR record contains registration date which is not the same as a start date. For example: you would like to include cancer as one of the covariates to adjust for in a TTE study. In the EHR you will have cancer diagnoses with a registration date. This registration date does not mean that cancer started on this date. The cancer was most likely there before the registration date. So if we say "do not use post-initation data" we would not use use cancer diagnoses registred after initation of treatment (time zero). To me this seems a bit odd, given that the patient very likely had cancer at time zero. 

I would very much like to hear RWE community members opinion on this and how to deal with such information.

Thank you in advance,

Joanna Klopotowska

Amsterdam UM, The Netherlands

------------------------------
Joanna Klopotowska PharmD, PhD
Assistant Professor
Amsterdam University Medical Centers
Amsterdam
+31 614393672
------------------------------