An Evolutionary Computation Approach for Optimizing Multilevel Data to Predict Patient Outcomes

Table 1

Common multilevel predictor data available in electronic health records.


Multilevel predictors	Description	Examples

Reasons for visit	Descriptors of the reason for the healthcare system encounter	Ambulatory care chief complaints; inpatient admission diagnoses
Diagnoses	Descriptors of patients’ differential or final diagnosis departing the healthcare system	International classification of disease codes (e.g., ICD-10); read codes
Medical history	Descriptors of previous medical history and chronic conditions	EHR problem lists (e.g., diabetes, previous coronary artery bypass graft (CABG), hypertension)
Diagnostic and therapeutic procedures	Descriptors of diagnostic and therapeutic courses of action taken	Procedure coding system (ICD-10-PCS), surgical procedures, rehabilitation
Diagnostic exams	Descriptors of medical tests conducted	Laboratory exams, imaging exams, physical exams
Medication	Descriptors of medications administered	US Food and Drug Administration Drug Class (e.g., opioids and hydrocodone)
Administrative	Descriptor of the administrative status of patients	Inpatient, outpatient, observation