Measuring Comorbidity in Cardiovascular Research: A Systematic Review

Table 1

Historical overview of the measurement of comorbidity.

	Kaplan and Feinstein [5]	Charlson et al. [7]	Deyo et al. [8]	D'Hoore et al. [9]	Elixhauser et al. [10]
Time frame	1970s	1980s	1990s	1990s	1990s
Purpose	Classify patients for therapeutic and statistical reasons	Prospectively identify persons at greater risk of death from comorbid diseases	Adapted the CCI for use with administrative datasets	Adapted the CCI for use with administrative datasets	Predict resource use or clinical outcomes
Original population	Diabetics	Medical patients; Female breast cancer patients	Medicare lumbar spinal surgery patients	Hospitalized patients in Quebec, CAN	Acute care patients in CA
Measurement method	Clinician derived from symptom patterns, disease duration, physical exam, and lab tests	Clinician scored from list of weighted diseases; Validated against Kaplan and Feinstein [5]	Used ICD-9-CM codes equivalent to diseases in the CCI	Implemented an algorithm to map the ICD-9 codes to CCI components	Developed a set of 30 comorbidities with their ICD-9-CM codes
Predictors of comorbidity	Clinical (e.g., vascular or nonvascular diseases) variables	Sociodemographic and clinical variables	Sociodemographic variables and clinical variables	Sociodemographic and clinical variables	Sociodemographic and clinical variables
Outcomes assessed	Mortality or vascular complications for those patients who survived	Mortality	Mortality, hospital complications and treatments, discharge destinations	Inpatient mortality	Mortality and fiscal
Surrogate terms for comorbidity	Episodic events, disease, ailment, and chronic condition	Common conditions	Chronic conditions	Complications (if iatrogenic)	Clinical condition, preexisting condition

CCI: Charlson Comorbidity Index, ICD: International Statistical Classification of Diseases.