About the GEMINI Data Dictionary

The GEMINI Data Dictionary is an essential resource for anyone using GEMINI data. This document describes the metadata of all the data tables within GEMINI. Individuals will be granted access to a subset of tables needed to investigate their research question. For any questions or feedback, please contact GEMINI.data@unityhealth.to.

About GEMINI Data

GEMINI holds granular hospitalization-level data on 31 hospitals from 20 healthcare organizations in Ontario. GEMINI data include all patients admitted to the department of medicine or intensive care unit and capture >60% of all adult medical and intensive care unit beds in Ontario. GEMINI data are comprised of administrative data (CIHI DAD, NACRS) linked to clinical data extracted from hospital electronic health records. Clinical data include but are not limited to laboratory measurements, radiology, vital signs, and blood transfusion. As of September 14, 2022, GEMINI data include 1,295,574 unique hospitalizations discharged from April 1, 2015.

File Versions History

• Version: v2.1.2
  • Hospital and Institution names are now anonymized which results in the following changes:
    • hospital_id column is removed from every table and replaced with hospital_num (3 digit integer)
    • lookup_hospital is replaced with lookup_hospital_num where the hospital_num is mapped to the institition_id to preserve the relationship between network hospitals and individual sites
  • Note: Metadata and ER diagram in this file are yet to reflect the changes

• Version: v2.1.1
  • GEMINI internal formatting fix to Vitals table

• Version: v2.1.0
  • Removal of non-standardized variables from the following tables:
    • admdad, echo, ipscu, lab, pharmacy, physicians, radiology, roomtransfer, and transfusion
• Version: v2.0.0
  • Addition of all available CIHI and clinical data from 29 GEMINI sites from April 1, 2015 to December 31, 2021
  • Removal of data from April 1, 2010 to March 31, 2015 for 8 original sites to align on April 2015 onward with Hospital Cohort Study
  • Addition of two new tables: clinical notes and edconsults
  • Addition of supplemental variables to admdad, er, ipdiagnosis, erintervention, radiology, and transfusion
  • Addition of de-identified clinical notes and radiology text result data
  • Update of metadata descriptions for derived scores and lookup_statcan variables
  • Update of recalculated derived variable fields.
  • Update of clinical derived variables based on mapped OMOP codes
  • Addition of lookup table lookup_transfers: used to calculate readmission
  • Update of GIM physician definition in physician table:
    • Physician who attends GIM wards is now defined by ‘y’, instead of being separated by ‘GP-GIM’ and ‘Geriatics’
• Version: v1.0.0
  • GEMINI data for 8 GEMINI sites from April 1, 2010 to December 31, 2020 (availability varies by site)

About GEMINI

GEMINI is a hospital and analytics study that collects, formats, standardizes and analyzes hospital data with the aim of improving how healthcare is delivered. GEMINI data are used for clinical research and to improve the quality of hospital care.