Antibiotic Infusion Duration in MIMIC Database

Hi,thanks for all the remarkable work that you all do in maintaining and developing the MIMIC database.

In my analysis, I noticed that for antibiotics such as vancomycin, gentamicin, and amikacin, the recorded duration in the 'inputevents' table indicates a fixed duration of 1 minute (ENDTIME = STARTTIME + 1 minute).

However, this duration seems to contradict common medical practices, where these antibiotics are typically administered as continuous infusions rather than bolus doses within such a short duration. I came across additional evidence suggesting that antibiotics are described as "IV Piggyback" in the 'proc_type' field of the 'pharmacy table'. An "IV Piggyback" administration refers to medication diluted in a small volume of IV fluid (e.g.,50–250 ml in a minibag) delivered through the main IV line and administered as a short-duration infusion over 30 minutes.

Similar inquiries from other researchers, such as"vanco administration #372", "Tobramycin administration #691", and "Meropenem infusion duration #1246," raised similar concerns regarding the accurate documentation of antibiotic infusion duration in the MIMIC database.

As you mentioned before, these issues may be able to be resolved in the 'ema' or 'emar_detail' tables. However, I could not locate specific records that provide the infusion rate or actual start and end times for antibiotic administrations. Additionally,The eMAR system operated only from 2011 to 2013.

Given the significance of antibiotic infusion duration in clinical practice and research, I am reaching out to seek your expertise and guidance on this matter. I am particularly interested in exploring if there are other data sources or approaches within the MIMIC database that can shed light on the actual infusion duration of antibiotics. Additionally, if there are plans to incorporate or expand the recording of antibiotic infusion details in future updates of the database.

Thank you for your time and support.

Best regards,

Secnan

Start by reviewing the inputevents, pharmacy, ema, and emar_detail tables, then compare their antibiotic timing fields with the related issues #372, #691, and #1246. Done would require a documented determination of whether another MIMIC source captures actual infusion duration and what limitations apply to the available records.