Leaving this potential confounding variable to go undetected in subsequent

These data sources allow the patient-level linkage of health resource utilization data to demographic and vital statistics data. When studying clinic-based populations, patients with severe HF are likely to be overrepresented, but administrative databases provide a means for identifying risk factors for HF, and quantifying the effects of treatment in unselected populations. However, administrative databases are only useful for HF research if the diagnostic codes contained within are valid; that is, if they can be used to distinguish those who actually have HF from those who do not. Their validity can be assessed by comparing the administrative database diagnosis to an accepted ��gold standard�� reference diagnosis. This diagnosis is typically obtained through more resource-intensive processes such as patient self-report, retrospective chart review, or prospective clinical examination. Principal measures of validity include sensitivity and specificity. Unfortunately, there is some uncertainty surrounding the validity of diagnoses recorded in administrative databases since most databases are not established for research purposes. Validity is of particular concern when studying HF patients, as they tend to have high comorbidity burdens and be hospitalized for other cardiovascular and respiratory conditions. While HF may have contributed to the need for these hospitalizations, this diagnosis may not be entered on the discharge record, leaving this potential confounding variable to go undetected in subsequent Cetrimonium Bromide epidemiologic investigations. Although several assessments of the validity of HF codes in administrative databases have been published, there is considerable heterogeneity amongst them with regards to the clinical settings and reference standards used. Of note, many of these assessments were limited to specific Bumetanide populations ) so may not be generalizable to the HF diagnoses recorded for other individuals. As a part of a Canadian Rheumatology Network for establishing best practices in the use of administrative data for health research and surveillance, we have conducted a systematic review of studies reporting on the validity of diagnostic codes for identifying cardiovascular diseases in administrative data.