<aside> 💡
MDI stands for: Medical Data Investigator
</aside>
Table of Contents
MDI is a research and QI tool which can be used to query clinical databases and analyze the data for specific projects. You can access MDI by downloading the Mac or Windows application at ‣ and running it on a secure computer. This tool will store encrypted clinical data on the device where it is installed, and requires a separate device password each time it opens which can not be recovered if lost. No clinical data is ever uploaded to the internet from this computer software.
MDI is designed to work with any relational database (SQL, PostgreSQL, GoogleSQL etc), including the MIMIC-IV Clinical Database as well as databases from hospital partners. Reach out to [email protected] to request a data source adapter to be created for your clinical database.
<aside> 🚨
Note, MDI is a tool which you can use to access clinical databases you separately have obtained permission to access. You should follow all necessary QI or Research IRB approval protocols prior to accessing clinical data through MDI.
</aside>
MDI offers software licenses which include the following features:
🏡 Base License
The base module is your gateway to MDI and allows you to create projects, define cohorts, and download data from data sources which you have approval to access. You can create the following type of report:
3. Data List Report - An exportable table of columns representing a set of features to be extracted from a list of rows each with the same base types (Patient, Surgery, Admission, etc)
📈 Statistics Add-On
The statistics add-on can be added on to enable the viewing and creation of the following type of report:
4. Summary Table Report - An exportable table which summarizes the mean or medians of a set of features across a filterable list of entries. Sub-populations can be defined, matched, and compared using appropriate statistical tests (Chi-squared, ANOVA).
🚧 Prediction Add-On (In Development)
Using this module, you will be able to define outcomes of interest, visualize incidence and analyze time to event, multivariate regression, and decision tree based prediction models. These can be used in conjunction with the Statistic add-on to match patients by propensity-scores