Abstract: INFO16
A Curated Data Set for Clinical Research in Continuous Renal Replacement Therapy
Session Information
- Informational Posters
November 04, 2021 | Location: On-Demand, Virtual Only
Abstract Time: 10:00 AM - 12:00 PM
Category: Dialysis
- No subcategory defined
Authors
- Gottlieb, Eric Raphael, Brigham and Women's Hospital, Boston, Massachusetts, United States
- Kumar, Bhawesh, Harvard University T H Chan School of Public Health, Boston, Massachusetts, United States
- Bonventre, Joseph V., Brigham and Women's Hospital, Boston, Massachusetts, United States
- Celi, Leo Anthony, Beth Israel Deaconess Medical Center, Boston, Massachusetts, United States
Description
Innovations in data science are transforming clinical research, including in some areas of nephrology. However, progress is tempered by a limited number of clinical nephrologists with the requisite data science skills and interest to participate in this type of research. Management of continuous renal replacement therapy (CRRT), given its clinical complexity, is one area of research that could benefit both from clinician involvement and application of large databases such as the Medical Information Mart for Intensive Care (MIMIC), which was developed by the Laboratory for Computational Physiology (LCP) at the Massachusetts Institute of Technology (MIT, Cambridge, MA) and the Beth Israel Deaconess Medical Center (BIDMC, Boston, MA). To advance research in this area, we have developed an easily accessible curated CRRT-focused data set derived from MIMIC-IV.
We used the R Shiny platform to develop a tabbed web interface to facilitate access and exploration of these data. Data are extracted from the MIMIC-IV database with real-time queries. This data set can be customized to include CRRT periods of a minimum length and a specified patient age range. A subset of data categories relevant to CRRT research are accessible from this interface. These include CRRT parameters, basic demographic and admission data, Elixhauser comorbidities, fluid intake/output, vasopressor and inotrope infusions, laboratory values, illness severity scores, and medications. Additionally, customized database queries (in Structured Query Language, SQL) and R code for data wrangling are provided to support interaction with the original data source when needed. A customizable scatter plot data visualization tool allows for exploratory data analysis. Depending on parameters chosen, a final data set will include as many as 7,442 CRRT sessions over 1,872 hospitalizations.
This interface will make large data CRRT research accessible to a wider community of researchers. It may be used to answer questions such as how to optimize CRRT prescriptions, manage vasopressors in patients on CRRT, and predict outcomes. It will also serve as a resource for easy access to data for use in time-constrained “Datathons.” Further development and outside partnerships will streamline this system and may expand it to include comparable CRRT data from other institutions.
Funding
- Research reported in this publication was supported by NIH grants T32DK007527 (Gottlieb); NIDDK/NCATS UH3TR002155, NIDDK 2R01DK072381, NIDDK R37DK039773 (Bonventre); NIBIB R01EB017205 (Celi).