Validity of Administrative Data for Identifying Patients Who Have Had a Stroke or Transient Ischemic Attack Using EMRALD as a Reference Standard

Manual review of electronic medical records as a reference standard for case definition development: a validation study

“…They reported insufficient validity, with sensitivity ranging from 0.05% to 0.49%. The EMRALD database in Ontario, based on a single primary care EMR system, has been used as a reference standard for several studies validating case definitions; [7][8][9][10] however, these studies tend to validate case definitions using administrative data with EMR data as the reference standard.…”

Regional differences in use of endoscopic ultrasonography in Ontario: a population-based retrospective cohort study

“…19 The Ontario Health Insurance Plan (OHIP) database contains information as of 1991 on claims billed for physician and laboratory services, including endoscopic procedures. OHIP database diagnostic codes have been shown to have a high positive predictive rate (> 90%) [20][21][22][23] and high specificity (> 85%) [24][25][26][27] but somewhat limited sensitivity (35%-85%). [24][25][26][27] The Registered Persons Database is a roster of OHIP beneficiaries and maintains information on age, sex, postal code and vital statistics for all Ontarians with a valid OHIP number.…”

“…However, OHIP billing data have been shown to be highly accurate for other procedures. [20][21][22][23][24][25][26][27] We attempted to maximize the specificity of the endoscopic ultrasonography codes by excluding procedures not performed by physicians in related health care specialties; however, future validation studies are needed. Second, it is possible that endoscopic ultrasound procedures performed in the earlier years of this cohort were not captured by OHIP billing codes; however, as the study period was designed to begin 2 years after codes for endoscopic ultrasonography were introduced in Ontario, we feel that this is unlikely.…”

Use and Utility of Administrative Health Data for Stroke Research and Surveillance

“…Most studies validating ICD codes select patients who have records with the diagnostic code of interest. [24][25][26] The charts are then reviewed for accuracy. Fundamentally, this approach is a comparison of the agreement between coders and reabstractors of health records and has shown moderate to high sensitivity and specificity.…”