Table 2.

Datasets used in clinical natural language processing community challenges.

Dataset	Provenance	Documents	Sizea	Annotations	Studies
2010 i2b2/VA [123]	PHCb, BIDMCc, UPMCd	Discharge summaries, progress reports	871	Medical problems, treatments, tests, and relations among them	[20,49,64,67,96,104]
2011 i2b2/VA [124]	PHC, BIDMC, UPMC, Mayoe	Discharge summaries, progress reports, radiology reports, pathology reports, other reports	978+164	Coreference chains for the problem, person, test, result, treatment, anatomical site, disease or syndrome, sign or symptom, etc	[63]
2012 i2b2 [125]	PHC, BIDMC	Discharge summaries	310	Clinical events, temporal expressions, temporal relations	[64]
2013 ShARe/CLEF eHealth [126]	BIDMC	Discharge summaries, electrocardiogram reports, echocardiogram reports, radiology reports	300	Disorders, acronyms, and abbreviations	[54,57,88,98,114]
2014 i2b2/UTHealth [127,128]	PHC	Longitudinal medical records	1304	Protected health information; risk factors for heart disease	[18,21,26,52,62,64,80,82,91,107,108]
2015 SemEval/THYME [129]	Mayo	Clinical notes, pathology reports	600	Times, events, and temporal relations among them	[60]
2016 CEGS N-GRID [130,131]	PHC	Psychiatric intake records	1000	Protected health information; symptom severity	[23,27,42,53,58,65,78,83,92]

^aSize is expressed as the number of documents.

^bPartners Health Care (PHC) is a nonprofit hospital and physician network that includes Brigham and Women’s Hospital and Massachusetts General Hospital.

^cBeth Israel Deaconess Medical Center (BIDMC) is a teaching hospital of Harvard Medical School. Both organizations are based in Boston, Massachusetts, United States.

^dThe University of Pittsburgh Medical Center (UPMC) is a global nonprofit health enterprise that integrates over 35 hospitals, 600 clinical locations, and a health insurance division.

^eThe Mayo Clinic is a nonprofit academic medical center based in Rochester, Minnesota, which focuses on integrated clinical practice, education, and research. The clinic specializes in treating difficult cases through tertiary care.