Table 1.

Comparison of Electronic Medical Records (EMRs) and Biorepositories at five eMERGE institutions.

Institution	Biorepository Overview	Recruitment Model	Repository Size (Race/Ethnicity)	EMR Summary	Primary Phenotype	Phenotyping Methods*
Group Health (Seattle, WA)	GHC Biobank Alzheimer’s Disease Patient Registry and Adult Changes in Thought Study	Disease specific Cohort	4,000 (>96% Caucasian)	Comprehensive Vendor- based EMR since 2004 20+ years pharmacy data 15+ years ICD9 data	Dementia	Structured data extraction, Free-text searches, Manual chart review
Marshfield Clinic Research Foundation (Marshfield, WI)	Personalized Medicine Research Project. Geographically defined cohort within an integrated regional health care system	Population based	20,000 (98% Caucasian)	Comprehensive internally developed EMR since 1985 75% participants have 20+ years medical history	Cataracts	Structured data extraction, NLP, Intelligent Character Recognition
Mayo Clinic (Rochester, MN)	Vascular Diseases Biorepository. Mayo Clinic Non-Invasive Vascular Laboratory & Exercise Stress Testing Lab	Disease specific cohort	3,500 (>96% Caucasian)	Comprehensive Internally developed EMR since 1995. 40 years history of data extraction	Peripheral Arterial Disease (PAD)	Structured data extraction, NLP
Northwestern University (Chicago, IL)	Nugene Project. Northwestern affiliated hospitals and outpatient clinics	Population based	10,000 (12% AA 8% Hispanic)	Comprehensive Vendor based inpatient (2001) and outpatient (1999)) EMRs 20+ years ICD9 data	Type 2 diabetes	Structured data extraction, Free-text searches
Vanderbilt University (Nashville, TN)	BioVU: Vanderbilt Clinic, diverse outpatient population	Population based, Opt- out consent model	92,000 (11% AA)	Comprehensive internally developed EMR since 2000 35+ years medical history data	Cardiac conduction	Structured data extraction, NLP

^*NLP, Natural Language Processing; Structured data extraction refers to retrieving EMR data that has been stored in a predefined format