MIMIC_Extract_LOS_3: ICU Patient Drug Administration Data for Length-of-Stay Prediction

A clinical dataset derived from the MIMIC-III database, focusing on the PRESCRIPTIONS table for predicting hospital length of stay. The data was created by Alistair Johnson, Tom Pollard, and Roger Mark and is sourced from Kaggle. It includes a detailed key for interpreting the multiclass ROUTE column, which indicates drug administration methods.

Use Cases

• Predicting patient length of stay in the ICU based on drug administration patterns.
• Analyzing the relationship between drug delivery routes and clinical outcomes.
• Developing early-warning systems for patient severity assessment using prescription data.

Strengths

• Based on the widely-used and respected MIMIC-III clinical database.
• Includes a detailed, clinically-relevant key for interpreting over 60 drug administration route abbreviations.

Limitations

• Description metadata is limited; actual data quality requires manual inspection after download.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.

Provenance

Source

MIMIC-III database via Kaggle.

Collection Method

Extracted from the PRESCRIPTIONS table of the MIMIC-III clinical database.

Time Range

null

Freshness

Last update date is unknown; freshness unverified.

Geography

null