Missing discharge dates for CXR imaging patients

Hello,

I'm looking to find censoring times for a survival analysis task using the MIMIC-CXR-JPG dataset. To do this I'm linking with MIMIC IV 2.0.0, patients, and admissions tables. (MIMIC-IV v2.2 has a bunch of patients dropped to form an internal test set which affects the original splits in the CXR JPG data).

Index dates are set by the date of the images which are available.
Dates of death (dod) are available for a subset of the total patients in the patients table.

Calculating censoring times should be possible as the maximum follow-up time for dod is set to 1 year after the last discharge date. However, I can't find discharge dates for 11,221 patients in the admissions table. This accounts for around 7.8% of total images within the CXR dataset.

Am I missing something? Are patients still in the hospital by the time the study ends? What censoring date can I assume for these patients?

Prerequisites

• Put an X between the brackets on this line if you have done all of the following:
  • Checked the online documentation: https://mimic.mit.edu/
  • Checked that your issue isn't already addressed: https://github.com/MIT-LCP/mimic-code/issues?utf8=%E2%9C%93&q=

Description

Description of the issue, including:

• what you have tried
• references to similar issues
• queries demonstrating your question (if applicable)

Start by reviewing the MIMIC-CXR-JPG linkage with the MIMIC-IV patients and admissions tables, focusing on the 11,221 patients without discharge dates. Verify whether the missing dates reflect the dataset boundary or an issue in the linkage, then document the supported censoring-date assumption for these patients.