The use of biochemical markers in ischaemic heart disease: prefactory remarks

There has been a lack of rigour in many test assessments. The seven well-accepted methodological standards, which are often ignored, are: (1) a suitable spectrum of the investigated population; (2) demarcation of significant clinical sub-groups; (3) avoidance of verification bias; (4) avoidance of review bias; (5) indices of test performance reported with their confidence intervals; (6) adequate handling of indeterminate results; (7) adequate test reproducibility. Additionally, rule-in and rule-out decision thresholds should be reported related to time after infarction. These decision thresholds are dependent on the particular methodology in use and confusion can, and does, result when comparing a value obtained with one analyzer with another. There is an urgent need for the standardization of analyte calibration. There are eight methodological standards for post-infarction prognosis studies. Less than 15% fulfilled these standards in a recent study of 766 such reports.