Trans-thoracic-echo

Project evaluating the use of TTE in septic shock. Data was extracted from MIMIC IV version 2.0 loaded in a postgresql server.

To extract the cohort you need the MIMIC IV derived tables: See Johnson AE, Stone DJ, Celi LA, et al.: The MIMIC Code Repository: enabling reproducibility in critical care research. J Am Med Inform Assoc JAMIA 2018; 25:32–39. These scripts (and instructions) can be found here here: https://github.com/MIT-LCP/mimic-iv/tree/master/concepts.

First run the vasopressor_hourly, fluid_hourly, vasopressor, tte_table and fluid_balance scripts. These scripts will create tables in public/table_name needed for extraction. Note the vasopressor_hourly and fluid_hourly scripts can take a while to run depending on your equipment.

Finally run the cohort_tte scripts to extract the final cohort with all variables. The data is extracted in "long" format with multiple rows per patient, each row corresponding a 4 hour block. Maximum of 30 blocks for an individual patient.