Author

Data Source

Diagnosis/Treatment Codes and Algorithms

Study Cohort

Validation Methodology

Comments

Quam et al. (1993)

Country: USA

Source: ambulatory physician, hospital, and pharmacy claims from two different medical plans

Years: Jan.1988 - Dec.1989

Codes: Drug codes not specified (anti-hypertensive)

Algorithms: Used three mutually exclusive algorithms

Dx: Has at least one claim indicating a diagnosis of essential hypertension (ICD-9-CM codes) and no prescriptions for the most common antihypertensive medications.

Rx: Has at least one prescription for an antihypertensive medication, and no claims with a diagnosis code for hypertension.

B: Has at least one hypertension diagnosis code, or at least one prescription for an antihypertensive medication.

Between 18 and 65 years of age

Medical charts

Patient survey

Sensitivity, specificity and predictive values are not reported.

“We found that the submission of a diagnosis of hypertension on a single claim form is not a valid indicator of the presence of hypertension.” p.504

Robinson et al. (1997)

Country: Canada

Source: Hospital separations and physician billing claims from Manitoba Health

Years: 1986-1989

Codes: ICD-9 401-405, 642, 362.11, 416.0, 437.2, 796.2

Algorithms: Several algorithms were evaluated that varied in the number of years of administrative data and the number of occurrences of a relevant diagnostic code in one of the 16 diagnostic fields for hospital discharge abstracts and the single diagnostic code in physician claims.

Survey data (Manitoba Heart Health Survey)

Max. Specificity: 0.86

Max. Sensitivity: 0.78

Max. PPV: 0.63

Max. Kappa: 0.59

Muhajarine et al. (1997)

Country: Canada

Source: Physician claims data from Manitoba Health

Years: Oct. 1987- Feb 1990

Codes: ICD-9 401, 402

Algorithm: Any claim filed for services during the 2 years prior to the survey with a diagnostic code of 401 or 402.

Ages 18 to 74 years

Survey data (Manitoba Heart Health Survey)

Max Kappa = 0.65

“The overall proportion agreement between self-reported and physician claims hypertension was 81.7%.” p.714

Borzecki et al. (2004)

Country: USA

Sources: Out-patient clinic (OPC) file (1998 and 1999) and Patient Treatment file (PTF) (1999), from a National Department of Veterans Affairs database

Years: 1998, 1999

Codes: ICD-9-CM 401, 402, or 405

Algorithms: Varied the minimum required number of OPC records with a relevant diagnosis from 1 to 2 and varied the number of years of data from 1 to 2

Outpatients receiving care at 10 different sites across the country

Electronic clinicians’ notes (medical outpatient charts), from the Veterans Health Information System and Technology Architecture (VISTA) for 1999

Specificity, sensitivity, and predictive values are not reported.

Wilchesky et al. (2004)

Country: Canada

Source: Physician billing claims from Quebec

Year: 1995-1996

Codes: ICD-9-CM 401.0-401.9

Algorithm: a single occurrence of the diagnostic code in one year of data

66 years of age and older

Physician charts

Max Sens = 69%

Max Spec = 88%

Rector et al. (2004)

Country: USA

Source: Physician, facility (i.e., hospital), and pharmacy claims.

Years: 1999 and 2000

Codes: ICD-9-CM 401.0, 401.1, 401.9, 402.xx, 403.xx, 404.xx in physician claims in one of up to four diagnosis fields and hospital claims in one of up to nine diagnosis fields.

Current Procedural Terminology (CPT) codes are not specified.

National Drug Codes are not specified.

Algorithms: 38 different algorithms were examined

Age of cohort was not specified.

Survey data

Max Sens = 95%

Max Spec = 96%