Table 1. Data quality and performance indicator definitions, mitigation strategies, and references.

	Definition	Mitigation strategies	Relevant studies
Dataquality
Completeness (or, conversely, missingness)	The absence of data points, without reference to data type or plausibility [12]	Automated data extraction; data imputation	[2-6,8,9,24-37,undefined,undefined,undefined,undefined,undefined,undefined,undefined,undefined,undefined,undefined,undefined,undefined,undefined,undefined,undefined,undefined,undefined]
Conformance	The compliance of data with expected formatting, relational, or absolute definitions [12]	Preemptively enforced data format standardization	[2-6,8,14,24-27,29-33,36,38,undefined,undefined,undefined,undefined,undefined,undefined,undefined,undefined,undefined,undefined,undefined]
Plausibility	The possibility that a value is true given the context of other variables or temporal sequences (ie, patient date of birth must precede date of treatment or diagnosis) [12]	Periodic realignment with logic rule sets or objective truth standards; thresholding	[4-6,8,14,25,27,28,30-33,35,37-39,undefined,undefined,undefined,undefined,undefined,undefined,undefined]
Uniqueness	The lack of duplicate data among other patient records [8]	Two-level encounter or visit data structure	[8]
Dataperformance
Correctness or accuracy	Whether patient records are free from errors or inconsistencies when the information provided in them is true [10,13]	Periodic validation against internal and external gold standards	[2,7-9,14,23,24,28,undefined,undefined]
Currency or recency	Whether data were entered into the EHRa within a clinically relevant time frame and is representative of the patient state at a given time of interest [10,13]	Enforcing predetermined hard and soft rule sets for timeline of data entry	[2,4,9,27,32,34,36]
Fairness (or, conversely, bias)	The degree to which data collection, augmentation, and application are free from unwarranted over- or underrepresentation of individual data elements or characteristics	Periodic review against a predetermined internal gold standard or bias criterion	[3,19,22,24,27,35]
Stability (or, conversely, temporal variability)	Whether temporally dependent variables change according to predefined expectations [10,12]	Periodic measurement of data drift against a baseline standard of data distribution	[4,8,19,31]
Shareability	Whether data can be shared directly, easily, and with no information loss [3]	Preemptively enforced data standardization	[2,3]
Robustness	The percent of patient records with tolerable (eg, inaccurate, inconsistent, and outdated information) versus intolerable (eg, missing required information) data quality problems [24]	Timely identification of critical data quality issues	[24]

^aEHR: electronic health record.