Table 1.

Electronic Health Records Data Types, Examples, Number of Unique Variables, and Number of Grouped Variables

Data TypeExamplesNo. of Unique VariablesNo. of Grouped Variables
Diagnoses*ICD-9 codes (from outpatient visits and problem lists treated separately)12 616189
Medicationsβ-blockers, loop diuretics, etc3952631
LaboratoriesCholesterol, glucose, eGFR, etc2336
Hospitalization§ICD-9 codes associated with admission7071189
Demographics and health behaviorsAge, sex, race/ethnicity, smoking, and alcohol use5
Cardiovascular imaging orders2D echo, transesophageal echo, stress echo, etc18
VitalsPulse, systolic blood pressure, diastolic blood pressure, height, weight, and temperature6
Framingham HF signs and symptomsPositive and negative mentions of acute pulmonary edema, ankle edema, dyspnea on ordinary exertion, paroxysmal nocturnal dyspnea, etc28
  • eGFR indicates estimated glomerular filtration rate; HF, heart failure; and ICD-9, International Classification of Diseases-Ninth Revision.

  • * ICD-9 codes from outpatient visits and problem lists are treated separately.

  • Normalized drug names, ignoring the dosing information, were used.

  • The laboratories and vitals data types contain information about the test order and the numeric result of the test.

  • § Primary and secondary ICD-9 codes associated with the hospital admission were used.

  • The imaging order type, associated ICD-9 codes, and any left ventricular ejection fraction values were used.